Add files using upload-large-folder tool

CIGS/0_data_process.ipynb

@@ -0,0 +1,1113 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "bead31c0",
+   "metadata": {},
+   "source": [
+    "# 先去下载好数据： https://cigs.iomicscloud.com/"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "0db2da68",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "02cf9b84",
+   "metadata": {},
+   "source": [
+    "## 转成 parquet 格式更方便处理"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "006f7147",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import os\n",
+    "import glob\n",
+    "\n",
+    "# 定义要处理的根目录\n",
+    "base_path = \"/data/boom/CIGS/\"\n",
+    "\n",
+    "# 使用 glob 找到所有以 _Counts.xlsx 结尾的文件\n",
+    "count_files = glob.glob(os.path.join(base_path, \"*_Counts.xlsx\"))\n",
+    "\n",
+    "if not count_files:\n",
+    "    print(f\"在路径 '{base_path}' 下没有找到任何以 '_Counts.xlsx' 结尾的文件。\")\n",
+    "else:\n",
+    "    print(f\"找到 {len(count_files)} 个匹配的文件，开始批量处理...\")\n",
+    "\n",
+    "# 循环处理每个文件\n",
+    "for count_path in count_files:\n",
+    "    try:\n",
+    "        print(f\"\\n--- 正在处理文件: {count_path} ---\")\n",
+    "\n",
+    "        # 1. 处理 Counts 文件\n",
+    "        counts = pd.read_excel(count_path, engine=\"openpyxl\", header=1)\n",
+    "        # 清理空行/列\n",
+    "        counts = counts.dropna(axis=1, how=\"all\").dropna(axis=0, how=\"all\")\n",
+    "\n",
+    "        # 规范化首列\n",
+    "        id_col = counts.columns[0]\n",
+    "        if str(id_col).lower() != \"sample_unique_id\":\n",
+    "            counts = counts.rename(columns={id_col: \"Sample_unique_id\"})\n",
+    "        counts[\"Sample_unique_id\"] = counts[\"Sample_unique_id\"].astype(str)\n",
+    "\n",
+    "        # 构建 Parquet 文件名\n",
+    "        parquet_path = count_path.replace(\"_Counts.xlsx\", \"_Counts.parquet\")\n",
+    "        counts.to_parquet(parquet_path, index=False)\n",
+    "        print(f\"成功处理并保存 Counts 文件到: {parquet_path}\")\n",
+    "\n",
+    "        # 2. 处理对应的 MetaData 文件\n",
+    "        meta_path = count_path.replace(\"_Counts.xlsx\", \"_MetaData.xlsx\")\n",
+    "        if os.path.exists(meta_path):\n",
+    "            meta = pd.read_excel(meta_path, engine=\"openpyxl\", header=1)\n",
+    "            meta = meta.dropna(axis=1, how=\"all\").dropna(axis=0, how=\"all\")\n",
+    "\n",
+    "            parquet_meta_path = meta_path.replace(\"_MetaData.xlsx\", \"_MetaData.parquet\")\n",
+    "            meta.to_parquet(parquet_meta_path, index=False)\n",
+    "            print(f\"成功处理并保存 MetaData 文件到: {parquet_meta_path}\")\n",
+    "        else:\n",
+    "            print(f\"警告: 未找到对应的 MetaData 文件: {meta_path}，跳过处理。\")\n",
+    "\n",
+    "    except Exception as e:\n",
+    "        print(f\"错误: 处理文件 '{count_path}' 时出错: {e}\")\n",
+    "        print(\"跳过此文件，继续处理下一个。\")\n",
+    "\n",
+    "print(\"\\n所有文件处理完毕。\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "2514cecc",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "合并后的数据集形状: (13221, 7)\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Compound name</th>\n",
+       "      <th>Catalog Number</th>\n",
+       "      <th>CAS Number</th>\n",
+       "      <th>MOA</th>\n",
+       "      <th>Clinical Information</th>\n",
+       "      <th>Approved Type</th>\n",
+       "      <th>Catalog Number.1</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Pirozadil</td>\n",
+       "      <td>HY_100144</td>\n",
+       "      <td>54110-25-7</td>\n",
+       "      <td>Others</td>\n",
+       "      <td>No Development Reported</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>NKL 22</td>\n",
+       "      <td>HY_100384</td>\n",
+       "      <td>537034-15-4</td>\n",
+       "      <td>HDAC</td>\n",
+       "      <td>No Development Reported</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Toll-like receptor modulator</td>\n",
+       "      <td>HY_10018</td>\n",
+       "      <td>926927-42-6</td>\n",
+       "      <td>Toll-like Receptor (TLR)</td>\n",
+       "      <td>No Development Reported</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Lu AF21934</td>\n",
+       "      <td>HY_100366</td>\n",
+       "      <td>1445605-23-1</td>\n",
+       "      <td>mGluR</td>\n",
+       "      <td>No Development Reported</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Vonoprazan</td>\n",
+       "      <td>HY_100007</td>\n",
+       "      <td>881681-00-1</td>\n",
+       "      <td>Proton Pump</td>\n",
+       "      <td>Launched</td>\n",
+       "      <td>FDA; Other Countries</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13216</th>\n",
+       "      <td>Methoxyeugeno l\\n4-O-rhamnosyl(\\n1→2)glucoside</td>\n",
+       "      <td>Cpd1991</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>903519-86-8</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13217</th>\n",
+       "      <td>(2S)-1-O-β-D-\\nglucopyranosyl-\\n2-O-acetyl-3-O...</td>\n",
+       "      <td>Cpd1992</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13218</th>\n",
+       "      <td>Manninotriose</td>\n",
+       "      <td>Cpd1998</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>13382-86-0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13219</th>\n",
+       "      <td>1-Methylhydant\\noin</td>\n",
+       "      <td>Cpd1999</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>616-04-6</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13220</th>\n",
+       "      <td>Agrimol B</td>\n",
+       "      <td>Cpd2000</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>55576-66-4</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>13221 rows × 7 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                           Compound name Catalog Number  \\\n",
+       "0                                              Pirozadil      HY_100144   \n",
+       "1                                                 NKL 22      HY_100384   \n",
+       "2                           Toll-like receptor modulator       HY_10018   \n",
+       "3                                             Lu AF21934      HY_100366   \n",
+       "4                                             Vonoprazan      HY_100007   \n",
+       "...                                                  ...            ...   \n",
+       "13216     Methoxyeugeno l\\n4-O-rhamnosyl(\\n1→2)glucoside        Cpd1991   \n",
+       "13217  (2S)-1-O-β-D-\\nglucopyranosyl-\\n2-O-acetyl-3-O...        Cpd1992   \n",
+       "13218                                      Manninotriose        Cpd1998   \n",
+       "13219                                1-Methylhydant\\noin        Cpd1999   \n",
+       "13220                                          Agrimol B        Cpd2000   \n",
+       "\n",
+       "         CAS Number                       MOA     Clinical Information  \\\n",
+       "0        54110-25-7                    Others  No Development Reported   \n",
+       "1       537034-15-4                      HDAC  No Development Reported   \n",
+       "2       926927-42-6  Toll-like Receptor (TLR)  No Development Reported   \n",
+       "3      1445605-23-1                     mGluR  No Development Reported   \n",
+       "4       881681-00-1               Proton Pump                 Launched   \n",
+       "...             ...                       ...                      ...   \n",
+       "13216           NaN                       NaN                      NaN   \n",
+       "13217           NaN                       NaN                      NaN   \n",
+       "13218           NaN                       NaN                      NaN   \n",
+       "13219           NaN                       NaN                      NaN   \n",
+       "13220           NaN                       NaN                      NaN   \n",
+       "\n",
+       "              Approved Type Catalog Number.1  \n",
+       "0                       NaN              NaN  \n",
+       "1                       NaN              NaN  \n",
+       "2                       NaN              NaN  \n",
+       "3                       NaN              NaN  \n",
+       "4      FDA; Other Countries              NaN  \n",
+       "...                     ...              ...  \n",
+       "13216                   NaN      903519-86-8  \n",
+       "13217                   NaN              NaN  \n",
+       "13218                   NaN       13382-86-0  \n",
+       "13219                   NaN         616-04-6  \n",
+       "13220                   NaN       55576-66-4  \n",
+       "\n",
+       "[13221 rows x 7 columns]"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import pandas as pd\n",
+    "\n",
+    "cas_csv = \"/data/boom/CIGS/Drug_infoM.xlsx\"\n",
+    "\n",
+    "# 读取两个sheet，假设第二个sheet名为\"Supplementary Table 4\"\n",
+    "sheet1 = pd.read_excel(cas_csv, sheet_name=\"Supplementary Table 3\", header=1)\n",
+    "sheet2 = pd.read_excel(cas_csv, sheet_name=\"Supplementary Table 4\", header=1)  # 替换为实际的第二个sheet名\n",
+    "\n",
+    "# 合并两个DataFrame\n",
+    "merged_df = pd.concat([sheet1, sheet2], ignore_index=True)\n",
+    "\n",
+    "# 查看合并结果\n",
+    "print(\"合并后的数据集形状:\", merged_df.shape)\n",
+    "merged_df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "05c9cbb1",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "合并后的数据集形状: (13221, 7)\n",
+      "转换完成，已保存为 merged_with_smiles.csv\n"
+     ]
+    }
+   ],
+   "source": [
+    "import requests\n",
+    "import pandas as pd\n",
+    "import time\n",
+    "\n",
+    "def cas_to_smiles_single(cas, sleep_time=0.2):\n",
+    "    \"\"\"\n",
+    "    输入: 单个CAS号\n",
+    "    输出: (cid, smiles)\n",
+    "    \"\"\"\n",
+    "    try:\n",
+    "        # 通过 PubChem 查询 CID\n",
+    "        url_cid = f\"https://pubchem.ncbi.nlm.nih.gov/rest/pug/substance/name/{cas}/cids/TXT\"\n",
+    "        r_cid = requests.get(url_cid)\n",
+    "        if r_cid.status_code != 200 or not r_cid.text.strip():\n",
+    "            return None, None\n",
+    "        cid = r_cid.text.strip().split(\"\\n\")[0]\n",
+    "\n",
+    "        # 通过 CID 获取 canonical SMILES\n",
+    "        url_smiles = f\"https://pubchem.ncbi.nlm.nih.gov/rest/pug/compound/cid/{cid}/property/CanonicalSMILES/TXT\"\n",
+    "        r_smiles = requests.get(url_smiles)\n",
+    "        if r_smiles.status_code != 200 or not r_smiles.text.strip():\n",
+    "            return cid, None\n",
+    "        smiles = r_smiles.text.strip()\n",
+    "\n",
+    "        time.sleep(sleep_time)  # 防止被限流\n",
+    "        return cid, smiles\n",
+    "    except:\n",
+    "        return None, None\n",
+    "\n",
+    "\n",
+    "# ================= 主程序 =================\n",
+    "if __name__ == \"__main__\":\n",
+    "    cas_csv = \"/data/boom/CIGS/Drug_infoM.xlsx\"\n",
+    "\n",
+    "    # 读取两个sheet\n",
+    "    sheet1 = pd.read_excel(cas_csv, sheet_name=\"Supplementary Table 3\", header=1)\n",
+    "    sheet2 = pd.read_excel(cas_csv, sheet_name=\"Supplementary Table 4\", header=1)\n",
+    "\n",
+    "    # 合并\n",
+    "    merged_df = pd.concat([sheet1, sheet2], ignore_index=True)\n",
+    "    print(\"合并后的数据集形状:\", merged_df.shape)\n",
+    "\n",
+    "    # 遍历 CAS 列，生成新列\n",
+    "    smiles_list, cid_list = [], []\n",
+    "    for cas in merged_df[\"CAS Number\"]:\n",
+    "        cid, smiles = cas_to_smiles_single(str(cas))\n",
+    "        cid_list.append(cid)\n",
+    "        smiles_list.append(smiles)\n",
+    "\n",
+    "    merged_df[\"PubChem_CID\"] = cid_list\n",
+    "    merged_df[\"SMILES\"] = smiles_list\n",
+    "\n",
+    "    # 保存结果\n",
+    "    merged_df.to_csv(\"merged_with_smiles.csv\", index=False)\n",
+    "    print(\"转换完成，已保存为 merged_with_smiles.csv\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "032a4d8c",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(11122, 10725)"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(merged_df[\"CAS Number\"].unique()), len(merged_df['SMILES'].unique())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "5e891d70",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(True, 2)"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "merged_df['SMILES'].isna().any(), merged_df['SMILES'].isna().sum()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "522c3c6b",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Compound name</th>\n",
+       "      <th>Catalog Number</th>\n",
+       "      <th>CAS Number</th>\n",
+       "      <th>MOA</th>\n",
+       "      <th>Clinical Information</th>\n",
+       "      <th>Approved Type</th>\n",
+       "      <th>Catalog Number.1</th>\n",
+       "      <th>PubChem_CID</th>\n",
+       "      <th>SMILES</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>5991</th>\n",
+       "      <td>Atuveciclib S-Enantiomer</td>\n",
+       "      <td>HY_12871C</td>\n",
+       "      <td>250279-81-1</td>\n",
+       "      <td>CDK</td>\n",
+       "      <td>No Development Reported</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>None</td>\n",
+       "      <td>None</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10734</th>\n",
+       "      <td>FITC-Dextran (MW 10000)</td>\n",
+       "      <td>HY_128868</td>\n",
+       "      <td>60842-46-8</td>\n",
+       "      <td>Others</td>\n",
+       "      <td>No Development Reported</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>None</td>\n",
+       "      <td>None</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                  Compound name Catalog Number   CAS Number     MOA  \\\n",
+       "5991   Atuveciclib S-Enantiomer      HY_12871C  250279-81-1     CDK   \n",
+       "10734   FITC-Dextran (MW 10000)      HY_128868   60842-46-8  Others   \n",
+       "\n",
+       "          Clinical Information Approved Type Catalog Number.1 PubChem_CID  \\\n",
+       "5991   No Development Reported           NaN              NaN        None   \n",
+       "10734  No Development Reported           NaN              NaN        None   \n",
+       "\n",
+       "      SMILES  \n",
+       "5991    None  \n",
+       "10734   None  "
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "nan_rows = merged_df[merged_df['SMILES'].isna()]\n",
+    "nan_rows"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4380d330",
+   "metadata": {},
+   "source": [
+    "# 通过手动搜索来补全"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "977eb4bb",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Int64Index([5991], dtype='int64')\n",
+      "Int64Index([10734], dtype='int64')\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 250279-81-1\t： N=[S@@](CC1=CC(NC2=NC(C3=CC=C(F)C=C3OC)=NC=N2)=CC=C1)(C)=O\n",
+    "# 60842-46-8： O=C1C2=CC=CC=C2C(=O)C1C3=CC=C(N=C=S)C=C3O  只能把“FITC 核心 + 连接臂”这部分拆开表示\n",
+    "\n",
+    "# 先确认行号\n",
+    "print(merged_df.loc[merged_df['CAS Number'] == '250279-81-1'].index)   # 假设返回 5991\n",
+    "print(merged_df.loc[merged_df['CAS Number'] == '60842-46-8'].index)    # 假设返回 10734\n",
+    "\n",
+    "# 一次性写进去\n",
+    "merged_df.loc[5991, 'SMILES'] = 'N=[S@@](CC1=CC(NC2=NC(C3=CC=C(F)C=C3OC)=NC=N2)=CC=C1)(C)=O'\n",
+    "merged_df.loc[10734, 'SMILES'] = 'O=C1C2=CC=CC=C2C(=O)C1C3=CC=C(N=C=S)C=C3O'   # FITC 骨架"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "e3bffa9d",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(False, 0)"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "merged_df['SMILES'].isna().any(), merged_df['SMILES'].isna().sum()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "d9f6a788",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "merged_df.to_csv(\"CIGS_with_smiles.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "014e0723",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "merged_df['SMILES'].to_csv('CIGS_smiles_list.csv', index=False, header=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bcbfa041",
+   "metadata": {},
+   "source": [
+    "# 检查这些SMILES是否是规范化的\n",
+    "\n",
+    "## 上面的才是规范化的，用这个，下面这个如果原 SMILES 本身错/含非法字符，MolFromSmiles 直接返回 None，会丢行"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "70cc22cb",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "规范化后的 SMILES 数组:\n",
+      "13221 13221\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "[12:50:19] WARNING: not removing hydrogen atom without neighbors\n",
+      "[12:50:19] WARNING: not removing hydrogen atom without neighbors\n",
+      "[12:50:19] WARNING: not removing hydrogen atom without neighbors\n"
+     ]
+    }
+   ],
+   "source": [
+    "# import numpy as np\n",
+    "# from rdkit import Chem\n",
+    "\n",
+    "# # 假设你的 results_CP['smiles'] 数组如你所示\n",
+    "# smiles_array = np.array(merged_df['SMILES'])\n",
+    "\n",
+    "# # 存储规范化后的 SMILES\n",
+    "# canonical_smiles_list = []\n",
+    "\n",
+    "# # 遍历数组中的每一个 SMILES\n",
+    "# for byte_smiles in smiles_array:\n",
+    "#     # Chem.MolFromSmiles() 会处理非规范的 SMILES，并返回一个分子对象\n",
+    "#     mol = Chem.MolFromSmiles(byte_smiles)\n",
+    "\n",
+    "#     # 3. 如果解析成功，生成规范 SMILES\n",
+    "#     if mol is not None:\n",
+    "#         # Chem.MolToSmiles() 默认生成规范 SMILES\n",
+    "#         canonical_smiles = Chem.MolToSmiles(mol)\n",
+    "#         canonical_smiles_list.append(canonical_smiles)\n",
+    "#     else:\n",
+    "#         # 如果 SMILES 无效，可以添加 None 或原始字符串\n",
+    "#         print(f\"警告: 无法解析 SMILES: {byte_smiles}\")\n",
+    "#         canonical_smiles_list.append(None) # 或者 smiles_str\n",
+    "\n",
+    "# # 将列表转换回 NumPy 数组\n",
+    "# # dtype='O' 可以存储不同长度的字符串\n",
+    "# canonical_smiles_array = np.array(canonical_smiles_list, dtype='O')\n",
+    "\n",
+    "# print(\"规范化后的 SMILES 数组:\")\n",
+    "# print(len(smiles_array), len(canonical_smiles_array))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7014f8af",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ea42c2fb",
+   "metadata": {},
+   "source": [
+    "# 中药没有给CAS，所以漏了SMIELS，先用名字生成CAS，然后再生成SMIELS"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "id": "dacc1145",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Compound name</th>\n",
+       "      <th>Catalog Number</th>\n",
+       "      <th>CAS Number</th>\n",
+       "      <th>MOA</th>\n",
+       "      <th>Clinical Information</th>\n",
+       "      <th>Approved Type</th>\n",
+       "      <th>Catalog Number.1</th>\n",
+       "      <th>PubChem_CID</th>\n",
+       "      <th>SMILES</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>11356</th>\n",
+       "      <td>Acteoside</td>\n",
+       "      <td>Cpd0001</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>61276-17-3</td>\n",
+       "      <td>445063</td>\n",
+       "      <td>CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11357</th>\n",
+       "      <td>Asiaticoside B</td>\n",
+       "      <td>Cpd0002</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>125265-68-1</td>\n",
+       "      <td>445063</td>\n",
+       "      <td>CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11358</th>\n",
+       "      <td>Brandioside</td>\n",
+       "      <td>Cpd0003</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>133393-81-4</td>\n",
+       "      <td>445063</td>\n",
+       "      <td>CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11359</th>\n",
+       "      <td>Clematichinenos\\nide AR</td>\n",
+       "      <td>Cpd0004</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>761425-93-8</td>\n",
+       "      <td>445063</td>\n",
+       "      <td>CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11360</th>\n",
+       "      <td>Saikosaponin\\nB1</td>\n",
+       "      <td>Cpd0005</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>58558-08-0</td>\n",
+       "      <td>445063</td>\n",
+       "      <td>CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                 Compound name Catalog Number CAS Number  MOA  \\\n",
+       "11356                Acteoside        Cpd0001        NaN  NaN   \n",
+       "11357           Asiaticoside B        Cpd0002        NaN  NaN   \n",
+       "11358              Brandioside        Cpd0003        NaN  NaN   \n",
+       "11359  Clematichinenos\\nide AR        Cpd0004        NaN  NaN   \n",
+       "11360         Saikosaponin\\nB1        Cpd0005        NaN  NaN   \n",
+       "\n",
+       "      Clinical Information Approved Type Catalog Number.1 PubChem_CID  \\\n",
+       "11356                  NaN           NaN       61276-17-3      445063   \n",
+       "11357                  NaN           NaN      125265-68-1      445063   \n",
+       "11358                  NaN           NaN      133393-81-4      445063   \n",
+       "11359                  NaN           NaN      761425-93-8      445063   \n",
+       "11360                  NaN           NaN       58558-08-0      445063   \n",
+       "\n",
+       "                                        SMILES  \n",
+       "11356  CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O  \n",
+       "11357  CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O  \n",
+       "11358  CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O  \n",
+       "11359  CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O  \n",
+       "11360  CC(=O)NC1C(CC(OC1C(C(CO)O)O)(C(=O)O)O)O  "
+      ]
+     },
+     "execution_count": 29,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# merged_df.head()\n",
+    "\n",
+    "# 把merged_df中的Catalog Number列中以“Cpd”开头的行提取出来\n",
+    "merged_df_cpd = merged_df[merged_df['Catalog Number'].str.startswith('Cpd')]\n",
+    "merged_df_cpd.head()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "485f1864",
+   "metadata": {},
+   "source": [
+    "## name to SMILES"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "26725140",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import re, requests, pandas as pd, time\n",
+    "\n",
+    "# 去掉换行、多余空格、括号备注\n",
+    "def clean_name(n):\n",
+    "    return re.sub(r'\\s+', ' ', n.split('(')[0]).strip()\n",
+    "\n",
+    "name_list = [clean_name(n) for n in merged_df_cpd['Compound name'].tolist()]\n",
+    "name_list = [n for n in name_list if 2 <= len(n) <= 200]   # 去掉极端长/短\n",
+    "\n",
+    "BATCH = 50\n",
+    "URL = \"https://pubchem.ncbi.nlm.nih.gov/rest/pug/compound/name/property/CanonicalSMILES/TXT\"\n",
+    "\n",
+    "def names2smiles_post(names):\n",
+    "    smi_map = {}\n",
+    "    for i in range(0, len(names), BATCH):\n",
+    "        chunk = names[i:i+BATCH]\n",
+    "        payload = '\\n'.join(chunk)          # POST 正文\n",
+    "        r = requests.post(URL, data=payload, headers={'Content-Type': 'text/plain'}, timeout=60)\n",
+    "        if r.status_code == 200:\n",
+    "            for line in r.text.strip().split('\\n'):\n",
+    "                if '\\t' in line:\n",
+    "                    n_query, smi = line.split('\\t')\n",
+    "                    if smi != 'N/A':\n",
+    "                        smi_map[n_query] = smi\n",
+    "        else:\n",
+    "            print(f\"批次 {i//BATCH + 1} 失败：{r.status_code}\")\n",
+    "        time.sleep(0.5)\n",
+    "    return smi_map\n",
+    "\n",
+    "smiles_dict = names2smiles_post(name_list)\n",
+    "print('命中率：', len(smiles_dict)/len(name_list))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "43405512",
+   "metadata": {},
+   "source": [
+    "# 失败，放弃中药"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e1dcdb41",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "c3bf93e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import os\n",
+    "import numpy as np\n",
+    "import pyarrow as pa\n",
+    "import pyarrow.parquet as pq\n",
+    "\n",
+    "file_path = \"/data/boom/CIGS/MDA-MB-231-JQ1-RNA-seq-RawCounts.xlsx\"\n",
+    "\n",
+    "meta = pd.read_excel(file_path, sheet_name=0, engine='openpyxl',\n",
+    "                   header=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "7e226bfa",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Ensembl</th>\n",
+       "      <th>DMSO_1</th>\n",
+       "      <th>DMSO_2</th>\n",
+       "      <th>DMSO_3</th>\n",
+       "      <th>JQ1_1</th>\n",
+       "      <th>JQ1_2</th>\n",
+       "      <th>JQ1_3</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>ENSG00000000003.15</td>\n",
+       "      <td>335</td>\n",
+       "      <td>334</td>\n",
+       "      <td>428</td>\n",
+       "      <td>285</td>\n",
+       "      <td>261</td>\n",
+       "      <td>330</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>ENSG00000000005.6</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>ENSG00000000419.13</td>\n",
+       "      <td>1624</td>\n",
+       "      <td>1615</td>\n",
+       "      <td>1781</td>\n",
+       "      <td>1512</td>\n",
+       "      <td>1471</td>\n",
+       "      <td>1763</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>ENSG00000000457.14</td>\n",
+       "      <td>69</td>\n",
+       "      <td>100</td>\n",
+       "      <td>126</td>\n",
+       "      <td>90</td>\n",
+       "      <td>107</td>\n",
+       "      <td>152</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>ENSG00000000460.17</td>\n",
+       "      <td>314</td>\n",
+       "      <td>301</td>\n",
+       "      <td>338</td>\n",
+       "      <td>74</td>\n",
+       "      <td>63</td>\n",
+       "      <td>89</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>60646</th>\n",
+       "      <td>ENSG00000288695.1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>60647</th>\n",
+       "      <td>ENSG00000288696.1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>60648</th>\n",
+       "      <td>ENSG00000288697.1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>60649</th>\n",
+       "      <td>ENSG00000288698.1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>1</td>\n",
+       "      <td>1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>60650</th>\n",
+       "      <td>ENSG00000288699.1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>60651 rows × 7 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                  Ensembl  DMSO_1  DMSO_2  DMSO_3  JQ1_1  JQ1_2  JQ1_3\n",
+       "0      ENSG00000000003.15     335     334     428    285    261    330\n",
+       "1       ENSG00000000005.6       0       0       0      0      0      0\n",
+       "2      ENSG00000000419.13    1624    1615    1781   1512   1471   1763\n",
+       "3      ENSG00000000457.14      69     100     126     90    107    152\n",
+       "4      ENSG00000000460.17     314     301     338     74     63     89\n",
+       "...                   ...     ...     ...     ...    ...    ...    ...\n",
+       "60646   ENSG00000288695.1       0       0       0      0      0      0\n",
+       "60647   ENSG00000288696.1       0       0       0      0      0      0\n",
+       "60648   ENSG00000288697.1       0       0       0      0      0      0\n",
+       "60649   ENSG00000288698.1       0       1       1      0      0      0\n",
+       "60650   ENSG00000288699.1       0       0       0      0      0      0\n",
+       "\n",
+       "[60651 rows x 7 columns]"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "meta"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2b4d7a01",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

CIGS/CIGS_Gene_names.csv

@@ -0,0 +1,3408 @@
+Gene_name
+A2M
+A4GNT
+AARS
+ABCA1
+ABCB1
+ABCB6
+ABCC5
+ABCC8
+ABCF1
+ABCF3
+ABCG1
+ABCG2
+ABHD3
+ABHD4
+ABHD6
+ABL1
+ABL2
+ACAA1
+ACAA2
+ACACA
+ACACB
+ACAD10
+ACAD11
+ACAD9
+ACADL
+ACADM
+ACADS
+ACADSB
+ACADVL
+ACAT1
+ACAT2
+ACBD3
+ACD
+ACE
+ACKR3
+ACLY
+ACO1
+ACO2
+ACOT12
+ACOT6
+ACOT7
+ACOT8
+ACOT9
+ACOX1
+ACOX2
+ACOX3
+ACSBG1
+ACSBG2
+ACSL1
+ACSL3
+ACSL4
+ACSL5
+ACSL6
+ACSM3
+ACSM4
+ACSM5
+ACTA2
+ACTB
+ACTC1
+ACTG1
+ACTN1
+ACTN2
+ACTN3
+ACTN4
+ACTR2
+ACTR3
+ACVR1
+ACVR2A
+ACVRL1
+ADA
+ADAM10
+ADAM17
+ADAM23
+ADAM33
+ADAMTS1
+ADAMTS13
+ADAMTS8
+ADAR
+ADARB1
+ADAT1
+ADCK3
+ADCY1
+ADCY2
+ADCY3
+ADCY5
+ADCYAP1
+ADCYAP1R1
+ADH1C
+ADH5
+ADI1
+ADIG
+ADIPOQ
+ADIPOR1
+ADIPOR2
+ADM
+ADO
+ADORA2A
+ADRA2B
+ADRB1
+ADRB2
+ADRB3
+AEBP1
+AES
+AFF1
+AGA
+AGER
+AGFG1
+AGGF1
+AGL
+AGR2
+AGRP
+AGT
+AGTR1
+AGTR2
+AGTRAP
+AHNAK
+AHR
+AHSG
+AICDA
+AIFM1
+AIM2
+AIMP1
+AIMP2
+AIPL1
+AJAP1
+AJUBA
+AK2
+AK4
+AKAP8
+AKAP8L
+AKR1B1
+AKR1C2
+AKR1D1
+AKR7A2
+AKT1
+AKT1S1
+AKT2
+AKT3
+ALAS1
+ALDH2
+ALDH3A2
+ALDH7A1
+ALDH8A1
+ALDOA
+ALDOB
+ALDOC
+ALG13
+ALK
+ALOX5
+AMBRA1
+AMDHD2
+AMH
+AMHR2
+AMOT
+AMOTL1
+AMOTL2
+AMPD3
+ANAPC1
+ANAPC2
+ANG
+ANGPT1
+ANGPT2
+ANGPTL1
+ANGPTL3
+ANGPTL4
+ANKRA2
+ANKRD10
+ANO10
+ANPEP
+ANTXR1
+ANXA1
+ANXA4
+ANXA5
+ANXA7
+AP1B1
+APAF1
+APBA1
+APBB2
+APC
+APCS
+APEX1
+APEX2
+APOA1
+APOA2
+APOA4
+APOA5
+APOB
+APOC3
+APOD
+APOE
+APOF
+APOL1
+APOL2
+APOL3
+APOL4
+APOL5
+APP
+APPBP2
+AQP1
+AQP2
+AQP4
+AR
+ARAF
+AREG
+ARF6
+ARFIP2
+ARG1
+ARHGAP1
+ARHGAP5
+ARHGDIA
+ARHGDIB
+ARHGEF12
+ARHGEF2
+ARHGEF7
+ARID4B
+ARID5B
+ARL4C
+ARNT
+ARNT2
+ARNTL
+ARPP19
+ARRB1
+ARRB2
+ARRDC3
+ARVCF
+ASAH1
+ASAP1
+ASB13
+ASB2
+ASCC3
+ASH1L
+ASPH
+ATF1
+ATF2
+ATF3
+ATF4
+ATF5
+ATF6
+ATF6B
+ATG10
+ATG12
+ATG16L1
+ATG16L2
+ATG3
+ATG4A
+ATG4B
+ATG4C
+ATG4D
+ATG5
+ATG7
+ATG9A
+ATG9B
+ATM
+ATMIN
+ATP11B
+ATP1B1
+ATP2B1
+ATP2C1
+ATP5A1
+ATP5C1
+ATP5G2
+ATP5S
+ATP6V0B
+ATP6V1A
+ATP6V1D
+ATP6V1F
+ATP6V1G2
+ATR
+ATRIP
+ATRN
+ATRX
+ATXN3
+AURKA
+AURKB
+AXIN1
+AXIN2
+AXL
+B3GALTL
+B3GNT1
+B3GNT2
+B3GNT3
+B3GNT4
+B3GNT8
+B4GALT1
+B4GALT2
+B4GALT3
+B4GALT5
+BAALC
+BACE2
+BACH1
+BAD
+BAG1
+BAG3
+BAG4
+BAI1
+BAIAP2
+BAK1
+BAMBI
+BARD1
+BAX
+BBC3
+BCAR1
+BCAT1
+BCCIP
+BCL10
+BCL2
+BCL2A1
+BCL2L1
+BCL2L10
+BCL2L11
+BCL2L2
+BCL3
+BCL6
+BCL7B
+BCL9
+BCR
+BDH1
+BDH2
+BDNF
+BECN1
+BFAR
+BGLAP
+BHLHE40
+BID
+BIK
+BIRC2
+BIRC3
+BIRC5
+BIRC6
+BLCAP
+BLK
+BLM
+BLMH
+BLNK
+BLVRA
+BMF
+BMI1
+BMP1
+BMP2
+BMP3
+BMP4
+BMP5
+BMP6
+BMP7
+BMPER
+BMPR1A
+BMPR1B
+BMPR2
+BNIP2
+BNIP3
+BNIP3L
+BOD1
+BPGM
+BPHL
+BRAF
+BRCA1
+BRCA2
+BRD2
+BRE
+BRIP1
+BRMS1
+BRP44
+BRS3
+BSG
+BTG1
+BTG2
+BTG3
+BTK
+BTLA
+BTNL10
+BTRC
+BUB1B
+BZRAP1
+BZW2
+C17ORF85
+C19ORF24
+C1D
+C1GALT1
+C1GALT1C1
+C1ORF159
+C1QA
+C2CD2
+C2CD2L
+C3
+C3AR1
+C5
+C5AR1
+C7ORF50
+C9
+CA4
+CA9
+CAB39
+CAB39L
+CACNA1F
+CACNA2D3
+CAD
+CADM1
+CALCA
+CALCR
+CALCRL
+CALD1
+CALM1
+CALM3
+CALR
+CALU
+CAMK2N1
+CAMKK1
+CAMSAP1
+CAMSAP2
+CANT1
+CAP1
+CAPN1
+CAPN2
+CARD11
+CARD18
+CARD6
+CARTPT
+CASC3
+CASK
+CASP1
+CASP10
+CASP14
+CASP2
+CASP3
+CASP4
+CASP5
+CASP6
+CASP7
+CASP8
+CASP9
+CASR
+CAST
+CAT
+CAV1
+CAV2
+CAV3
+CBL
+CBLB
+CBLL1
+CBR1
+CBR3
+CBX3
+CCDC103
+CCDC85B
+CCDC86
+CCDC90A
+CCDC92
+CCK
+CCKAR
+CCL1
+CCL11
+CCL13
+CCL14
+CCL15
+CCL16
+CCL18
+CCL19
+CCL2
+CCL20
+CCL21
+CCL22
+CCL23
+CCL24
+CCL25
+CCL26
+CCL27
+CCL28
+CCL3
+CCL3L1
+CCL4
+CCL5
+CCL7
+CCL8
+CCNA1
+CCNA2
+CCNB1
+CCNB2
+CCNC
+CCND1
+CCND2
+CCND3
+CCNE1
+CCNE2
+CCNF
+CCNG1
+CCNG2
+CCNH
+CCNO
+CCNT1
+CCP110
+CCR1
+CCR10
+CCR2
+CCR3
+CCR4
+CCR5
+CCR6
+CCR7
+CCR8
+CCR9
+CCRL2
+CCT4
+CCT5
+CD14
+CD180
+CD1A
+CD1B
+CD1C
+CD1D
+CD2
+CD209
+CD22
+CD27
+CD274
+CD276
+CD28
+CD320
+CD34
+CD36
+CD3D
+CD3E
+CD3G
+CD4
+CD40
+CD40LG
+CD44
+CD47
+CD5
+CD55
+CD58
+CD59
+CD69
+CD7
+CD70
+CD74
+CD79A
+CD80
+CD81
+CD82
+CD83
+CD86
+CD8A
+CD8B
+CD97
+CD99
+CDC123
+CDC16
+CDC20
+CDC25A
+CDC25B
+CDC25C
+CDC27
+CDC34
+CDC42
+CDC42EP3
+CDC45
+CDC6
+CDCA4
+CDH1
+CDH11
+CDH13
+CDH2
+CDH3
+CDH4
+CDH5
+CDH6
+CDK1
+CDK19
+CDK2
+CDK4
+CDK5R1
+CDK5RAP1
+CDK6
+CDK7
+CDK8
+CDKN1A
+CDKN1B
+CDKN1C
+CDKN2A
+CDKN2B
+CDKN2C
+CDKN2D
+CDKN3
+CDON
+CDR2
+CDSN
+CEACAM1
+CEACAM3
+CEACAM5
+CEACAM6
+CEBPA
+CEBPB
+CEBPD
+CEBPZ
+CEL
+CELA3A
+CELA3B
+CENPE
+CEP57
+CERK
+CES1
+CETN3
+CETP
+CFB
+CFD
+CFL1
+CFLAR
+CFTR
+CGN
+CGRRF1
+CHAC1
+CHD4
+CHD7
+CHD9
+CHEK1
+CHEK2
+CHERP
+CHGA
+CHI3L1
+CHIA
+CHIC2
+CHMP2A
+CHMP4A
+CHMP6
+CHN1
+CHP
+CHRD
+CHSY1
+CHUK
+CIAPIN1
+CIB1
+CIDEA
+CIDEB
+CIITA
+CIRBP
+CISD1
+CKLF
+CKS1B
+CKS2
+CLC
+CLCA1
+CLCA2
+CLDN1
+CLDN10
+CLDN11
+CLDN12
+CLDN14
+CLDN15
+CLDN16
+CLDN17
+CLDN18
+CLDN19
+CLDN2
+CLDN3
+CLDN4
+CLDN5
+CLDN6
+CLDN7
+CLDN8
+CLDN9
+CLEC3B
+CLEC4C
+CLEC4E
+CLEC7A
+CLIC4
+CLIC5
+CLK1
+CLN3
+CLPS
+CLPX
+CLSTN1
+CLTB
+CLTC
+CLU
+CMA1
+CMKLR1
+CMTM1
+CMTM2
+CMTM3
+CMTM4
+CNBP
+CNDP2
+CNN2
+CNOT2
+CNOT4
+CNPY3
+CNR1
+CNTF
+CNTFR
+CNTN1
+COASY
+COG2
+COG4
+COG7
+COL11A1
+COL12A1
+COL14A1
+COL15A1
+COL16A1
+COL18A1
+COL1A1
+COL1A2
+COL3A1
+COL4A1
+COL4A2
+COL4A3
+COL5A1
+COL5A2
+COL6A1
+COL6A2
+COL7A1
+COL8A1
+COLEC12
+COMMD4
+COPB2
+COPS7A
+CORO1A
+COX5A
+COX6C
+CP
+CPA3
+CPD
+CPE
+CPNE3
+CPSF4
+CPT1A
+CPT1B
+CPT1C
+CPT2
+CR1
+CRADD
+CRAT
+CRB1
+CRB2
+CRB3
+CREB1
+CREB3
+CREB3L4
+CREBBP
+CREG1
+CRELD2
+CRHR1
+CRHR2
+CRIM1
+CRK
+CRKL
+CRLF2
+CROT
+CRP
+CRTAP
+CRY1
+CRYAB
+CRYZ
+CS
+CSDE1
+CSF1
+CSF1R
+CSF2
+CSF2RB
+CSF3
+CSF3R
+CSK
+CSNK1A1
+CSNK1D
+CSNK1E
+CSNK2A1
+CSNK2A2
+CSNK2B
+CSRP1
+CST6
+CST7
+CSTB
+CTBP1
+CTGF
+CTLA4
+CTNNA1
+CTNNA2
+CTNNA3
+CTNNAL1
+CTNNB1
+CTNNBIP1
+CTNNBL1
+CTNND1
+CTNND2
+CTSB
+CTSC
+CTSD
+CTSE
+CTSK
+CTSL1
+CTSS
+CTTN
+CUBN
+CUEDC2
+CUL1
+CUL2
+CUL3
+CX3CL1
+CX3CR1
+CXADR
+CXCL1
+CXCL11
+CXCL12
+CXCL13
+CXCL14
+CXCL16
+CXCL2
+CXCL3
+CXCL5
+CXCL6
+CXCL9
+CXCR1
+CXCR2
+CXCR3
+CXCR4
+CXCR5
+CXCR6
+CXXC4
+CYB561
+CYB5R3
+CYCS
+CYLD
+CYP11A1
+CYP19A1
+CYP1B1
+CYP27A1
+CYP2E1
+CYP2S1
+CYP39A1
+CYP3A5
+CYP46A1
+CYP4A11
+CYP4V2
+CYP51A1
+CYP7A1
+CYP7B1
+CYSLTR1
+CYTH1
+DAAM1
+DAB2
+DAB2IP
+DAD1
+DAG1
+DAPK1
+DARC
+DAXX
+DBN1
+DCHS1
+DCHS2
+DCK
+DCN
+DCTD
+DCUN1D4
+DDB1
+DDB2
+DDIT3
+DDIT4
+DDIT4L
+DDR1
+DDR2
+DDX10
+DDX11
+DDX39B
+DDX42
+DDX58
+DECR1
+DECR2
+DEFB1
+DEGS1
+DENND2D
+DENND4A
+DENR
+DEPTOR
+DERA
+DESI1
+DFFA
+DFFB
+DGAT1
+DGAT2
+DGKA
+DGKZ
+DHCR24
+DHCR7
+DHDDS
+DHRS7
+DHX29
+DIABLO
+DIAPH1
+DIAPH2
+DIO2
+DIRAS2
+DIXDC1
+DKC1
+DKK1
+DKK3
+DLAT
+DLC1
+DLD
+DLG1
+DLG5
+DLK1
+DLL1
+DLST
+DLX2
+DMC1
+DMTF1
+DNAJA1
+DNAJA3
+DNAJB1
+DNAJB2
+DNAJB6
+DNAJC15
+DNAJC3
+DNM1
+DNM1L
+DNM2
+DNMT1
+DNMT3A
+DNTT
+DNTTIP2
+DOCK1
+DOCK2
+DOCK4
+DOK1
+DOK2
+DOK3
+DOLK
+DPH2
+DPP10
+DPP4
+DPYSL4
+DRAM1
+DRAM2
+DRAP1
+DRD1
+DRD2
+DSC1
+DSC2
+DSC3
+DSG1
+DSG2
+DSG3
+DSG4
+DSP
+DST
+DUSP1
+DUSP11
+DUSP14
+DUSP22
+DUSP3
+DUSP4
+DUSP6
+DVL1
+DVL2
+DYNLT3
+DYRK3
+E2F1
+E2F2
+E2F3
+E2F4
+EAPP
+EBI3
+EBNA1BP2
+EBP
+ECD
+ECH1
+ECHS1
+ECI2
+ECM1
+ECSIT
+ECT2
+EDA2R
+EDEM1
+EDEM2
+EDEM3
+EDIL3
+EDN1
+EDNRB
+EED
+EFNA1
+EFNB1
+EFNB2
+EGF
+EGFR
+EGR1
+EGR2
+EGR3
+EHD3
+EHHADH
+EI24
+EIF2AK2
+EIF2AK3
+EIF2B1
+EIF3M
+EIF4A1
+EIF4B
+EIF4E
+EIF4EBP1
+EIF4EBP2
+EIF4G1
+EIF4G2
+EIF5
+EIF5B
+ELAC2
+ELAVL1
+ELK1
+ELK4
+ELMO1
+ELN
+ELOVL6
+EML3
+EMP1
+EMR1
+ENAH
+ENG
+ENO1
+ENO2
+ENO3
+ENOPH1
+ENOSF1
+ENPP1
+EOMES
+EP300
+EPB41
+EPB41L2
+EPHA1
+EPHA2
+EPHA3
+EPHA4
+EPHA5
+EPHA7
+EPHA8
+EPHB1
+EPHB2
+EPHB3
+EPHB4
+EPHB6
+EPN2
+EPO
+EPOR
+EPRS
+EPX
+ERBB2
+ERBB3
+ERBB4
+ERCC1
+ERCC2
+ERCC3
+ERCC4
+ERCC5
+ERCC6
+ERCC8
+EREG
+ERG
+ERO1L
+ERRFI1
+ESAM
+ESR1
+ESR2
+ETFB
+ETFDH
+ETS1
+ETS2
+ETV1
+ETV4
+EVL
+EWSR1
+EXO1
+EXOC2
+EXOSC4
+EXOSC8
+EXT1
+EZH2
+EZR
+F11R
+F2
+F2R
+F3
+F8
+FABP1
+FABP2
+FABP3
+FABP4
+FABP5
+FABP6
+FABP7
+FADD
+FADS2
+FAH
+FAIM
+FAM20B
+FAM32A
+FAM57A
+FAM63A
+FAM69A
+FANCA
+FANCD2
+FANCG
+FAP
+FARP2
+FAS
+FASLG
+FASN
+FASTKD5
+FAT1
+FAT2
+FAT3
+FAT4
+FBP1
+FBP2
+FBXL12
+FBXO11
+FBXO21
+FBXO32
+FBXO38
+FBXO7
+FBXO9
+FBXW11
+FBXW4
+FCER1A
+FCER1G
+FCER2
+FCGR1A
+FCGR2A
+FCGR2B
+FCHO1
+FDFT1
+FDPS
+FEN1
+FER
+FES
+FEZ1
+FEZ2
+FGF1
+FGF10
+FGF13
+FGF2
+FGF20
+FGF4
+FGF7
+FGF9
+FGFBP1
+FGFR1
+FGFR2
+FGFR3
+FGFR4
+FGR
+FH
+FHIT
+FHL2
+FIGF
+FIS1
+FJX1
+FKBP14
+FKBP1A
+FKBP4
+FKBP5
+FKBP8
+FLNA
+FLNB
+FLT1
+FLT3
+FLT3LG
+FLT4
+FN1
+FOS
+FOSL1
+FOSL2
+FOXA1
+FOXA2
+FOXC2
+FOXD3
+FOXG1
+FOXI1
+FOXJ3
+FOXN1
+FOXO1
+FOXO3
+FOXO4
+FOXP1
+FOXP2
+FOXP3
+FPGS
+FPR1
+FRAT1
+FRK
+FRS2
+FRS3
+FRZB
+FSD1
+FST
+FSTL3
+FTH1
+FTSJ1
+FUCA1
+FUCA2
+FURIN
+FUT1
+FUT11
+FUT8
+FXC1
+FXYD5
+FYN
+FZD1
+FZD2
+FZD3
+FZD4
+FZD5
+FZD6
+FZD7
+FZD8
+FZD9
+G3BP1
+G6PC
+G6PC3
+G6PD
+GAA
+GAB1
+GABARAP
+GABARAPL1
+GABARAPL2
+GABPB1
+GADD45A
+GADD45B
+GADD45G
+GAL
+GALE
+GALM
+GALNT1
+GALNT10
+GALNT11
+GALNT12
+GALNT13
+GALNT14
+GALNT16
+GALNT2
+GALNT3
+GALNT5
+GALNT6
+GALNT7
+GALNT8
+GALNT9
+GALNTL5
+GALNTL6
+GALR1
+GALR2
+GANAB
+GAPDH
+GAS2L3
+GATA1
+GATA2
+GATA3
+GATA4
+GBE1
+GBP1
+GCA
+GCDH
+GCG
+GCGR
+GCK
+GCLC
+GCLM
+GCNT1
+GCNT3
+GCNT4
+GDF2
+GDF3
+GDF5
+GDF6
+GDF7
+GDNF
+GDPD1
+GDPD5
+GEMIN2
+GEMIN5
+GFI1
+GFOD1
+GFPT1
+GH1
+GH2
+GHR
+GHRHR
+GHRL
+GHSR
+GJA1
+GJA3
+GJA4
+GJA5
+GJA8
+GJA9
+GJB1
+GJB2
+GJB3
+GJB4
+GJB5
+GJB6
+GJB7
+GJC2
+GJC3
+GJD2
+GK
+GK2
+GLB1
+GLI1
+GLI2
+GLOD4
+GLP1R
+GLRX
+GLT8D1
+GLUL
+GML
+GMNN
+GNA11
+GNA15
+GNAI1
+GNAI2
+GNAQ
+GNAS
+GNB5
+GNG11
+GNL3
+GNPDA1
+GNPTAB
+GNPTG
+GNRH1
+GOLGB1
+GOLT1B
+GOT1
+GPATCH1
+GPATCH8
+GPC1
+GPC5
+GPD1
+GPD2
+GPER
+GPI
+GPM6A
+GPNMB
+GPR17
+GPR18
+GPR56
+GPS1
+GPX3
+GRB10
+GRB2
+GRB7
+GREM1
+GRM1
+GRM2
+GRM4
+GRM5
+GRM7
+GRN
+GRP
+GRPR
+GRWD1
+GSC
+GSK3A
+GSK3B
+GSR
+GSTA2
+GSTA3
+GSTM2
+GSTP1
+GSTZ1
+GTF2A2
+GTF2E2
+GTF2I
+GTPBP8
+GTSE1
+GUCY1A2
+GUCY1A3
+GUCY1B3
+GULP1
+GYPA
+GYS1
+GYS2
+GZMA
+GZMB
+H2AFV
+H2AFX
+H6PD
+HADH
+HADHA
+HAL
+HAS1
+HAS2
+HAT1
+HAVCR2
+HCK
+HCLS1
+HCRT
+HCRTR1
+HDAC1
+HDAC2
+HDAC6
+HDAC9
+HDGFRP3
+HDLBP
+HEATR1
+HEBP1
+HELZ2
+HERC6
+HERPUD1
+HES1
+HEXA
+HEXB
+HEY1
+HGF
+HGS
+HHIP
+HIC1
+HIF1A
+HIPK2
+HIST1H2BK
+HIST2H2BE
+HK1
+HK2
+HK3
+HLA-A
+HLA-B
+HLA-C
+HLA-DMA
+HLA-DPA1
+HLA-DRA
+HLA-E
+HMCN1
+HMG20B
+HMGA2
+HMGB1
+HMGCL
+HMGCR
+HMGCS1
+HMGCS2
+HMMR
+HMOX1
+HN1L
+HNF1B
+HNF4A
+HNRNPA0
+HNRNPA1
+HNRNPA2B1
+HNRNPUL1
+HOMER2
+HOOK2
+HOPX
+HOXA10
+HOXA3
+HOXA5
+HP1BP3
+HPRT1
+HPSE
+HPX
+HRAS
+HRG
+HRH1
+HRK
+HS2ST1
+HSD17B10
+HSD17B11
+HSP90AA1
+HSP90AB1
+HSP90B1
+HSPA12A
+HSPA1A
+HSPA1B
+HSPA4
+HSPA4L
+HSPA5
+HSPA8
+HSPB1
+HSPBAP1
+HSPD1
+HTATIP2
+HTATSF1
+HTR2A
+HTR2C
+HTRA1
+HTT
+HUS1
+HYOU1
+IAPP
+IARS2
+ICAM1
+ICAM2
+ICAM3
+ICMT
+ICOS
+ICOSLG
+ID1
+ID2
+ID3
+ID4
+IDE
+IDH1
+IDH2
+IDH3A
+IDH3B
+IDH3G
+IDI1
+IDI2
+IDO1
+IER3
+IFI6
+IFITM2
+IFNA1
+IFNA2
+IFNAR1
+IFNB1
+IFNG
+IFNGR1
+IFNGR2
+IFRD1
+IFRD2
+IGF1
+IGF1R
+IGF2
+IGF2BP2
+IGF2R
+IGFBP1
+IGFBP3
+IGFBP4
+IGFBP5
+IGFBP7
+IGHMBP2
+IGJ
+IGSF5
+IGSF6
+IKBKAP
+IKBKB
+IKBKE
+IKBKG
+IKZF1
+IKZF2
+IKZF3
+IL10
+IL10RA
+IL10RB
+IL11
+IL12A
+IL12B
+IL12RB1
+IL12RB2
+IL13
+IL13RA1
+IL13RA2
+IL15
+IL16
+IL17A
+IL17C
+IL17D
+IL17F
+IL17RA
+IL17RB
+IL17RC
+IL17RE
+IL18
+IL18R1
+IL18RAP
+IL1A
+IL1B
+IL1R1
+IL1R2
+IL1RAP
+IL1RL1
+IL1RN
+IL2
+IL20
+IL21
+IL22
+IL23A
+IL23R
+IL24
+IL25
+IL27
+IL27RA
+IL2RA
+IL2RB
+IL2RG
+IL3
+IL31
+IL32
+IL33
+IL3RA
+IL4
+IL4R
+IL5
+IL5RA
+IL6
+IL6R
+IL6ST
+IL7
+IL7R
+IL8
+IL9
+IL9R
+ILF2
+ILK
+INADL
+ING4
+INHA
+INHBA
+INHBB
+INPP1
+INPP4B
+INPPL1
+INS
+INSIG1
+INSIG2
+INSL3
+INSR
+INSRR
+INTS3
+IP6K2
+IPO13
+IQGAP1
+IRAK1
+IRAK2
+IRAK4
+IRF1
+IRF2
+IRF3
+IRF4
+IRF7
+IRF8
+IRF9
+IRGM
+IRS1
+IRS2
+IRS4
+ISG15
+ISG20
+ISOC1
+ITCH
+ITFG1
+ITGA1
+ITGA11
+ITGA2
+ITGA2B
+ITGA3
+ITGA4
+ITGA5
+ITGA6
+ITGA7
+ITGA8
+ITGA9
+ITGAE
+ITGAL
+ITGAM
+ITGAV
+ITGAX
+ITGB1
+ITGB1BP1
+ITGB2
+ITGB3
+ITGB4
+ITGB5
+ITGB6
+ITK
+ITPKB
+ITPR1
+ITPR2
+JAG1
+JAK1
+JAK2
+JAK3
+JAM2
+JAM3
+JMJD6
+JPH3
+JUN
+JUNB
+JUND
+JUP
+KAL1
+KANK4
+KAT2B
+KAT6A
+KAT6B
+KCNIP1
+KCNK1
+KCNQ1
+KCTD5
+KDELR2
+KDM3A
+KDM5A
+KDM5B
+KDR
+KEAP1
+KIAA0100
+KIAA0196
+KIAA0355
+KIAA0494
+KIAA0528
+KIAA0753
+KIAA0907
+KIAA0947
+KIAA1033
+KIAA1279
+KIF14
+KIF20A
+KIF2C
+KIF5B
+KIF5C
+KISS1
+KISS1R
+KIT
+KITLG
+KLF1
+KLF10
+KLF13
+KLF15
+KLF2
+KLF3
+KLF4
+KLF5
+KLF9
+KLHDC2
+KLHL13
+KLHL21
+KLHL24
+KLHL9
+KLK3
+KNG1
+KNTC1
+KPNA2
+KRAS
+KREMEN1
+KRT14
+KRT18
+KRT19
+KRT5
+KRT7
+KRT8
+KSR1
+KTN1
+LAG3
+LAGE3
+LAMA1
+LAMA2
+LAMA3
+LAMB1
+LAMB3
+LAMC1
+LAMP1
+LAMP2
+LAMTOR3
+LAP3
+LAT
+LAT2
+LATS1
+LATS2
+LBR
+LCAT
+LCK
+LDHA
+LDLR
+LDLRAP1
+LECT1
+LEF1
+LEFTY1
+LEP
+LEPR
+LGALS3
+LGALS4
+LGALS8
+LGMN
+LGSN
+LHCGR
+LIF
+LIFR
+LIG1
+LIG3
+LIG4
+LIMD1
+LIMK1
+LIPA
+LIPE
+LIX1L
+LLGL1
+LLGL2
+LMNA
+LMNB1
+LMNB2
+LMO1
+LMO2
+LMO7
+LOX
+LOXL1
+LPAR1
+LPAR2
+LPGAT1
+LPIN1
+LPL
+LPP
+LRG1
+LRMP
+LRP1
+LRP10
+LRP12
+LRP1B
+LRP5
+LRP6
+LRPAP1
+LRRC16A
+LRRC32
+LRRC41
+LSM5
+LSM6
+LSR
+LTA
+LTA4H
+LTB
+LTB4R
+LTBP1
+LTBP2
+LTBR
+LTK
+LY86
+LY96
+LYL1
+LYN
+LYPLA1
+LYRM1
+LYZ
+LZTR1
+MACF1
+MAD2L1
+MAD2L2
+MADD
+MAG
+MAGI1
+MAGI2
+MALT1
+MAMLD1
+MAN1A1
+MAN1A2
+MAN1B1
+MAN1C1
+MAN2A1
+MAN2A2
+MAN2B1
+MANBA
+MAP1B
+MAP1LC3A
+MAP1LC3B
+MAP2K1
+MAP2K2
+MAP2K3
+MAP2K4
+MAP2K5
+MAP2K6
+MAP2K7
+MAP3K1
+MAP3K2
+MAP3K3
+MAP3K4
+MAP3K7
+MAP4K1
+MAP4K4
+MAP7
+MAPK1
+MAPK10
+MAPK11
+MAPK12
+MAPK13
+MAPK14
+MAPK1IP1L
+MAPK3
+MAPK6
+MAPK7
+MAPK8
+MAPK8IP2
+MAPK8IP3
+MAPK9
+MAPKAP1
+MAPKAPK2
+MAPKAPK3
+MAPKAPK5
+MAPRE1
+MAPRE2
+MARCO
+MARK2
+MAST2
+MAT2A
+MATK
+MAX
+MAZ
+MBD1
+MBD4
+MBL2
+MBNL1
+MBNL2
+MBOAT7
+MBTPS1
+MC3R
+MCAM
+MCEE
+MCHR1
+MCL1
+MCM2
+MCM3
+MCM4
+MCM5
+MCOLN1
+MCOLN3
+MCPH1
+MDC1
+MDH1
+MDH1B
+MDH2
+MDK
+MDM2
+MDM4
+ME1
+ME2
+MECOM
+MED1
+MEF2A
+MEF2C
+MEFV
+MEIS1
+MELK
+MEN1
+MERTK
+MEST
+MET
+METAP2
+METRN
+MFGE8
+MFSD10
+MFSD11
+MGAT1
+MGAT2
+MGAT3
+MGAT4A
+MGAT4B
+MGAT4C
+MGAT5
+MGAT5B
+MGMT
+MGST1
+MICA
+MICALL1
+MICB
+MIF
+MIS12
+MKI67
+MKNK1
+MLEC
+MLH1
+MLH3
+MLLT11
+MLLT4
+MLST8
+MLXIPL
+MLYCD
+MMD
+MME
+MMP1
+MMP10
+MMP11
+MMP12
+MMP13
+MMP14
+MMP15
+MMP16
+MMP2
+MMP3
+MMP7
+MMP8
+MMP9
+MMS19
+MN1
+MNAT1
+MOB1A
+MOB1B
+MOGS
+MOK
+MOS
+MPDZ
+MPG
+MPL
+MPO
+MPP5
+MPP6
+MPZL1
+MRC1
+MRCL3
+MRE11A
+MRPL12
+MRPL19
+MRPL57
+MRPS16
+MRPS2
+MS4A1
+MS4A2
+MSH2
+MSH3
+MSH4
+MSH5
+MSH6
+MSN
+MSRA
+MST1
+MST1R
+MSTN
+MSX1
+MSX2
+MT1A
+MT1E
+MT2A
+MTA1
+MTCP1
+MTDH
+MTERFD1
+MTF2
+MTFP1
+MTFR1
+MTHFD1
+MTHFD2
+MTHFR
+MTO1
+MTOR
+MTSS1
+MUC1
+MUSK
+MUT
+MUTYH
+MVD
+MVK
+MVP
+MX1
+MYB
+MYBL1
+MYBL2
+MYC
+MYCBP
+MYCBP2
+MYCL
+MYCN
+MYD88
+MYH10
+MYH3
+MYH9
+MYL9
+MYLK
+MYO10
+MYO1C
+MYOD1
+NAGPA
+NAIP
+NAMPT
+NANOG
+NAP1L1
+NARFL
+NAV2
+NBN
+NCAM1
+NCAPD2
+NCK1
+NCK2
+NCL
+NCOA2
+NCOA3
+NCOA6
+NCOR2
+NDRG3
+NDUFB6
+NEIL1
+NEIL2
+NEIL3
+NENF
+NET1
+NEU1
+NEU2
+NEU3
+NEU4
+NEUROD1
+NF1
+NF2
+NFAT5
+NFATC1
+NFATC2
+NFATC3
+NFATC4
+NFE2L2
+NFIB
+NFIL3
+NFKB1
+NFKB2
+NFKBIA
+NFKBIB
+NFKBIE
+NFKBIL1
+NFRKB
+NGFR
+NGRN
+NHLH2
+NIPSNAP1
+NISCH
+NIT1
+NKD1
+NKX2-1
+NKX3-1
+NLK
+NLRC4
+NLRC5
+NLRP1
+NLRP12
+NLRP3
+NLRP4
+NLRP5
+NLRP6
+NLRP9
+NLRX1
+NMB
+NMBR
+NME1
+NME4
+NMT1
+NMU
+NMUR1
+NNT
+NOD1
+NOD2
+NODAL
+NOG
+NOL3
+NOLC1
+NOS2
+NOS3
+NOSIP
+NOTCH1
+NOTCH2
+NOTCH3
+NOTCH4
+NOV
+NPC1
+NPC1L1
+NPDC1
+NPEPL1
+NPHP4
+NPM1
+NPPB
+NPR1
+NPRL2
+NPY
+NPY1R
+NQO1
+NR0B2
+NR1H2
+NR1H3
+NR1H4
+NR2C2
+NR2F6
+NR3C1
+NR4A1
+NR4A2
+NR4A3
+NR5A2
+NRAS
+NRCAM
+NRF1
+NRIP1
+NRP1
+NRP2
+NSDHL
+NSF
+NT5DC2
+NTHL1
+NTRK1
+NTRK2
+NTRK3
+NTS
+NTSR1
+NUCB2
+NUDCD3
+NUDT13
+NUDT9
+NUP133
+NUP62
+NUP85
+NUP88
+NUP93
+NUSAP1
+NVL
+NXT2
+OAS1
+OCLN
+ODC1
+OGDH
+OGG1
+OGT
+OLR1
+OPCML
+OPRD1
+OPRK1
+OPRM1
+OR10J3
+ORC1
+OSBPL1A
+OSBPL5
+OSM
+OSMR
+OXA1L
+OXCT1
+OXCT2
+OXSR1
+P2RX6
+P2RX7
+P4HA2
+P4HTM
+PA2G4
+PABPC1
+PACSIN3
+PAF1
+PAFAH1B1
+PAFAH1B3
+PAG1
+PAICS
+PAK1
+PAK2
+PAK3
+PAK4
+PAK6
+PAN2
+PANX1
+PANX2
+PANX3
+PAPD7
+PARD3
+PARD6A
+PARD6B
+PARD6G
+PARP1
+PARP2
+PARP3
+PARVA
+PARVB
+PARVG
+PAWR
+PAX5
+PAX8
+PC
+PCBD1
+PCCB
+PCGF2
+PCK1
+PCM1
+PCMT1
+PCNA
+PCSK9
+PDCD1
+PDCD7
+PDE3B
+PDGFA
+PDGFB
+PDGFC
+PDGFD
+PDGFRA
+PDGFRB
+PDHA1
+PDHB
+PDHX
+PDIA5
+PDK1
+PDK2
+PDK3
+PDK4
+PDLIM1
+PDLIM4
+PDLIM7
+PDP1
+PDP2
+PDPK1
+PDPR
+PDS5A
+PDX1
+PEA15
+PECAM1
+PECR
+PELI1
+PER1
+PER2
+PERP
+PES1
+PEX11A
+PF4
+PF4V1
+PFKL
+PFN1
+PGAM1
+PGAM2
+PGF
+PGK1
+PGK2
+PGLS
+PGLYRP1
+PGM1
+PGM2
+PGM3
+PGR
+PGRMC1
+PHB
+PHF15
+PHF21A
+PHGDH
+PHKA1
+PHKB
+PHKG1
+PHKG2
+PIAS1
+PIAS2
+PIAS3
+PIDD
+PIGB
+PIH1D1
+PIK3C2A
+PIK3C2B
+PIK3C3
+PIK3CA
+PIK3CB
+PIK3CD
+PIK3CG
+PIK3R1
+PIK3R2
+PIK3R3
+PIK3R4
+PIK3R5
+PIM1
+PIN1
+PINX1
+PIP4K2B
+PIP5K1A
+PIP5K1C
+PITX2
+PKD2
+PKIG
+PKLR
+PKM
+PKP1
+PKP2
+PKP3
+PKP4
+PLA2G10
+PLA2G12A
+PLA2G12B
+PLA2G15
+PLA2G1B
+PLA2G2A
+PLA2G2D
+PLA2G2E
+PLA2G2F
+PLA2G3
+PLA2G4A
+PLA2G4B
+PLA2G5
+PLA2G6
+PLAU
+PLAUR
+PLCB1
+PLCB2
+PLCB3
+PLCG1
+PLCG2
+PLD1
+PLD2
+PLD3
+PLEC
+PLEK2
+PLEKHF1
+PLEKHJ1
+PLEKHM1
+PLG
+PLK1
+PLOD3
+PLP2
+PLS1
+PLSCR1
+PLSCR3
+PLTP
+PMAIP1
+PMCH
+PMEPA1
+PML
+PMM2
+PMS1
+PMS2
+PMVK
+PNKP
+PNN
+PNP
+PNPLA3
+POFUT1
+POFUT2
+POLB
+POLD2
+POLD3
+POLD4
+POLE2
+POLG2
+POLL
+POLR1C
+POLR2F
+POLR2I
+POLR2K
+POMC
+POMGNT1
+POMT1
+POMT2
+PON1
+POP4
+PORCN
+POSTN
+POTEF
+POU2F1
+POU2F2
+POU5F1
+PPA1
+PPAP2B
+PPARA
+PPARD
+PPARG
+PPARGC1A
+PPARGC1B
+PPAT
+PPBP
+PPIC
+PPIE
+PPL
+PPM1D
+PPOX
+PPP1CA
+PPP1R13B
+PPP1R15A
+PPP2CA
+PPP2CB
+PPP2R1A
+PPP2R1B
+PPP2R2B
+PPP2R2D
+PPP2R3C
+PPP2R4
+PPP2R5A
+PPP2R5E
+PPP3CA
+PPP3CC
+PPP3P
+PPP3R1
+PPP3R2
+PPRC1
+PRAF2
+PRC1
+PRCP
+PRDM16
+PRDM2
+PRDX1
+PRDX6
+PRF1
+PRG2
+PRICKLE1
+PRKAA1
+PRKAA2
+PRKAB1
+PRKAB2
+PRKACA
+PRKACG
+PRKAG1
+PRKAG2
+PRKAG3
+PRKAP
+PRKAR1A
+PRKCA
+PRKCB
+PRKCD
+PRKCE
+PRKCG
+PRKCH
+PRKCI
+PRKCQ
+PRKCSH
+PRKCZ
+PRKDC
+PRKG1
+PRKG2
+PRKRA
+PRKX
+PRL
+PRLHR
+PRLR
+PRMT6
+PROK1
+PROK2
+PROS1
+PRPF4
+PRPF8
+PRPS1
+PRPS1L1
+PRPS2
+PRR15L
+PRR7
+PRSS23
+PRUNE
+PSD3
+PSIP1
+PSMB10
+PSMB8
+PSMB9
+PSMC1
+PSMC5
+PSMD10
+PSMD2
+PSMD4
+PSMD5
+PSMD9
+PSME1
+PSME2
+PSMF1
+PSMG1
+PSMG2
+PSRC1
+PSTPIP1
+PTCH1
+PTDSS2
+PTEN
+PTGDR
+PTGDR2
+PTGER2
+PTGS1
+PTGS2
+PTH1R
+PTHLH
+PTK2
+PTK2B
+PTK6
+PTK7
+PTN
+PTP4A1
+PTPLAD1
+PTPN1
+PTPN11
+PTPN12
+PTPN14
+PTPN6
+PTPRC
+PTPRF
+PTPRK
+PTTG1
+PUF60
+PVR
+PVRL1
+PVRL2
+PVRL3
+PVRL4
+PWP1
+PXMP2
+PXN
+PYCARD
+PYCR1
+PYDC1
+PYGL
+PYGM
+PYGO1
+PYY
+QARS
+QPRT
+RAB11FIP2
+RAB21
+RAB24
+RAB25
+RAB27A
+RAB31
+RAB4A
+RAB5A
+RAB7A
+RAC1
+RAC2
+RAD1
+RAD17
+RAD18
+RAD21
+RAD23A
+RAD23B
+RAD50
+RAD51
+RAD51B
+RAD51C
+RAD51D
+RAD52
+RAD54L
+RAD9A
+RAE1
+RAF1
+RAG1
+RAI14
+RALA
+RALB
+RALGDS
+RAMP3
+RAP1A
+RAP1B
+RAP1GAP
+RAPGEF1
+RAPGEF3
+RARA
+RARB
+RASA1
+RASA3
+RASGRF1
+RASSF1
+RASSF2
+RASSF4
+RB1
+RBBP8
+RBKS
+RBL1
+RBL2
+RBM15B
+RBM23
+RBM34
+RBM39
+RBM6
+RBP1
+RBP4
+RDX
+REEP5
+REL
+RELA
+RELB
+RELN
+RELT
+RERE
+RET
+RETN
+RETNLB
+REV1
+RFC1
+RFC2
+RFC5
+RFNG
+RFX5
+RGS12
+RGS19
+RGS2
+RHEB
+RHO
+RHOA
+RHOB
+RHOC
+RHOJ
+RHOU
+RICTOR
+RIPK1
+RIPK2
+RNASE2
+RNASE3
+RND3
+RNF128
+RNF167
+RNF168
+RNF19A
+RNF8
+RNH1
+RNMT
+RNPS1
+ROCK1
+ROCK2
+ROR1
+ROR2
+RORA
+RORB
+RORC
+ROS1
+RPA1
+RPA2
+RPA3
+RPE
+RPIA
+RPL13
+RPL18
+RPL19
+RPL23
+RPL27A
+RPL37
+RPL39L
+RPL5
+RPN1
+RPP38
+RPRM
+RPS5
+RPS6
+RPS6KA1
+RPS6KA2
+RPS6KA5
+RPS6KB1
+RPS6KB2
+RPSA
+RPTOR
+RRAGA
+RRAGB
+RRAGC
+RRAGD
+RRAS
+RRAS2
+RRP12
+RRP1B
+RRP8
+RRS1
+RSU1
+RTN2
+RUNX1
+RUNX1T1
+RUNX2
+RUNX3
+RUVBL1
+RXRA
+RXRB
+RXRG
+RYBP
+RYK
+S100A13
+S100A4
+S100A7A
+S100A8
+S1PR1
+S1PR2
+S1PR3
+SACM1L
+SARM1
+SARS
+SATB1
+SAV1
+SCAF11
+SCAND1
+SCAP
+SCARB1
+SCARF1
+SCCPDH
+SCD
+SCGB1A1
+SCNM1
+SCP2
+SCRIB
+SCRN1
+SCTR
+SCYL3
+SDHA
+SDHB
+SDHC
+SDHD
+SELE
+SELL
+SELP
+SENP6
+2024-09-07 00:00:00
+SEPW1
+SERPINA3
+SERPINB2
+SERPINB5
+SERPINC1
+SERPINE1
+SERPINF1
+SERTAD1
+SESN1
+SESN2
+SET
+SF3A3
+SFN
+SFRP1
+SFRP2
+SFRP4
+SFRP5
+SFTA3
+SFTPC
+SFTPD
+SGCB
+SGCE
+SGK1
+SH2B1
+SH2D2A
+SH3BP5
+SH3GL1
+SH3PXD2A
+SHB
+SHBG
+SHC1
+SHC2
+SHH
+SHMT1
+SIAH1
+SIGIRR
+SIGLEC1
+SIGLEC11
+SIGLEC8
+SIGMAR1
+SIRPB1
+SIRT1
+SIRT2
+SIRT3
+SIX1
+SKIV2L
+SKP1
+SKP2
+SLC10A6
+SLC11A1
+SLC11A2
+SLC19A1
+SLC19A2
+SLC1A4
+SLC22A5
+SLC25A13
+SLC25A14
+SLC25A4
+SLC25A46
+SLC27A1
+SLC27A2
+SLC27A3
+SLC27A4
+SLC27A5
+SLC27A6
+SLC2A1
+SLC2A2
+SLC2A4
+SLC2A6
+SLC31A1
+SLC35A1
+SLC35A3
+SLC35B1
+SLC35F2
+SLC37A4
+SLC39A6
+SLC4A1AP
+SLC5A3
+SLC5A6
+SLC5A8
+SLC7A11
+SLIT2
+SLK
+SMAD1
+SMAD2
+SMAD3
+SMAD4
+SMAD5
+SMAD6
+SMAD7
+SMARCA4
+SMARCC1
+SMARCD2
+SMARCD3
+SMC1A
+SMC3
+SMC4
+SMNDC1
+SMO
+SMUG1
+SMURF1
+SNAI1
+SNAI2
+SNAI3
+SNAP23
+SNAP25
+SNCA
+SNRPB
+SNTA1
+SNX11
+SNX13
+SNX17
+SNX25
+SNX6
+SNX7
+SOAT1
+SOCS1
+SOCS2
+SOCS3
+SOCS4
+SOCS5
+SOD1
+SOD2
+SORBS1
+SORBS3
+SORL1
+SORT1
+SOS1
+SOS2
+SOSTDC1
+SOX10
+SOX17
+SOX2
+SOX4
+SOX9
+SP1
+SPAG4
+SPAG7
+SPARC
+SPATA2
+SPDEF
+SPEN
+SPG7
+SPHK1
+SPHK2
+SPI1
+SPINK1
+SPINK5
+SPOP
+SPP1
+SPR
+SPRED2
+SPRR1A
+SPSB1
+SPTA1
+SPTAN1
+SPTB
+SPTLC2
+SQRDL
+SQSTM1
+SRC
+SREBF1
+SREBF2
+SRF
+SRM
+SRMS
+SRSF1
+SRXN1
+SSBP2
+SST
+SSTR2
+SSX2IP
+ST3GAL1
+ST3GAL2
+ST3GAL5
+ST6GAL1
+ST6GALNAC1
+ST6GALNAC2
+ST7
+ST8SIA2
+ST8SIA3
+ST8SIA4
+ST8SIA6
+STAB1
+STAB2
+STAM
+STAMBP
+STAP2
+STARD3
+STAT1
+STAT2
+STAT3
+STAT4
+STAT5A
+STAT5B
+STAT6
+STEAP1
+STK10
+STK11
+STK24
+STK25
+STK3
+STK38L
+STK4
+STMN1
+STRADB
+STUB1
+STX11
+STX18
+STX1A
+STX4
+STXBP1
+STXBP2
+SUCLA2
+SUCLG1
+SUCLG2
+SUGT1
+SUMO1
+SUMO4
+SUPT7L
+SUPV3L1
+SUV39H1
+SUZ12
+SVIL
+SYCP2
+SYK
+SYMPK
+SYNE2
+SYNGR3
+SYPL1
+T
+TAB1
+TAB2
+TADA3
+TAL1
+TALDO1
+TAOK1
+TAOK2
+TAOK3
+TAP1
+TAP2
+TAPBP
+TARBP1
+TATDN2
+TAZ
+TBC1D9B
+TBK1
+TBL1XR1
+TBP
+TBPL1
+TBX2
+TBX21
+TBXA2R
+TCEA2
+TCEAL4
+TCERG1
+TCF20
+TCF21
+TCF3
+TCF4
+TCF7
+TCF7L1
+TCF7L2
+TCFL5
+TCL1A
+TCTA
+TCTN1
+TDG
+TEAD1
+TEAD2
+TEAD3
+TEAD4
+TEC
+TEK
+TELO2
+TEP1
+TERF1
+TERF2IP
+TERT
+TES
+TESK1
+TEX10
+TFAP2A
+TFDP1
+TFDP2
+TFF3
+TFPI2
+TG
+TGFA
+TGFB1
+TGFB1I1
+TGFB2
+TGFB3
+TGFBI
+TGFBR1
+TGFBR2
+TGFBR3
+TGFBRAP1
+TGIF1
+TGM2
+TGS1
+THAP11
+THBS1
+THBS2
+THBS3
+THPO
+THRB
+TIAM1
+TICAM1
+TIE1
+TIFA
+TIMELESS
+TIMM10B
+TIMM17B
+TIMM22
+TIMM9
+TIMP1
+TIMP2
+TIMP3
+TIMP4
+TINF2
+TIPARP
+TIRAP
+TJAP1
+TJP1
+TJP2
+TJP3
+TKT
+TLE1
+TLK2
+TLN1
+TLN2
+TLR1
+TLR10
+TLR2
+TLR3
+TLR4
+TLR5
+TLR6
+TLR7
+TLR8
+TLR9
+TLX1
+TLX3
+TM6SF1
+TM7SF2
+TM9SF2
+TM9SF3
+TMCO1
+TMED10
+TMEFF1
+TMEM109
+TMEM110
+TMEM132A
+TMEM2
+TMEM5
+TMEM50A
+TMEM57
+TMEM74
+TMEM97
+TMPRSS2
+TNC
+TNF
+TNFAIP2
+TNFAIP3
+TNFRSF10A
+TNFRSF10B
+TNFRSF10C
+TNFRSF10D
+TNFRSF11A
+TNFRSF11B
+TNFRSF12A
+TNFRSF13B
+TNFRSF13C
+TNFRSF14
+TNFRSF17
+TNFRSF18
+TNFRSF19
+TNFRSF1A
+TNFRSF1B
+TNFRSF21
+TNFRSF25
+TNFRSF4
+TNFRSF8
+TNFRSF9
+TNFSF10
+TNFSF11
+TNFSF12
+TNFSF13
+TNFSF13B
+TNFSF14
+TNFSF15
+TNFSF18
+TNFSF4
+TNFSF8
+TNFSF9
+TNIP1
+TNK1
+TNK2
+TNKS
+TNKS2
+TNNI2
+TNNI3
+TNS1
+TOLLIP
+TOMM34
+TOMM70A
+TOP1
+TOP2A
+TOP3A
+TOP3B
+TOPBP1
+TOR1A
+TOX
+TOX3
+TP53
+TP53AIP1
+TP53BP1
+TP53BP2
+TP53INP1
+TP63
+TP73
+TPD52L2
+TPI1
+TPM1
+TPSAB1
+TRADD
+TRAF1
+TRAF2
+TRAF3
+TRAF6
+TRAK2
+TRAM2
+TRAP1
+TRAPPC3
+TRAPPC6A
+TRERF1
+TRH
+TRIB1
+TRIB3
+TRIM10
+TRIM13
+TRIM2
+TRPM1
+TSC1
+TSC2
+TSC22D1
+TSC22D3
+TSEN2
+TSHR
+TSHZ1
+TSHZ2
+TSHZ3
+TSKU
+TSLP
+TSNAX
+TSPAN13
+TSPAN3
+TSPAN4
+TSPAN6
+TSTA3
+TTC1
+TTC9
+TUBA1C
+TUBA3C
+TUBA4A
+TUBA4B
+TUBA8
+TUBAL3
+TUBB
+TUBB1
+TUBB2A
+TUBB2B
+TUBB3
+TUBB4A
+TUBB4B
+TUBB6
+TUBB8
+TWF2
+TWIST1
+TXK
+TXLNA
+TXN
+TXNDC9
+TXNIP
+TXNL4B
+TXNRD1
+TYK2
+TYMP
+TYMS
+TYRO3
+UBASH3B
+UBE2A
+UBE2C
+UBE2J1
+UBE2L6
+UBE2N
+UBE3B
+UBE3C
+UBQLN2
+UBR7
+UCN
+UCP1
+UFM1
+UGDH
+UGGT1
+UGGT2
+UGP2
+UGT1A1
+ULK1
+ULK2
+UNG
+UQCRFS1
+USF1
+USP1
+USP14
+USP2
+USP22
+USP5
+USP54
+USP6NL
+USP7
+UTP14A
+UTS2
+UVRAG
+VAMP3
+VAMP7
+VANGL2
+VAPA
+VAPB
+VASP
+VAT1
+VAV1
+VAV2
+VAV3
+VCAM1
+VCAN
+VCL
+VDAC1
+VDR
+VEGFA
+VEGFB
+VEGFC
+VEZT
+VGLL4
+VHL
+VIM
+VLDLR
+VMP1
+VPREB1
+VPS13A
+VPS28
+VPS72
+VTN
+WAS
+WASF1
+WASF2
+WASF3
+WASL
+WDR61
+WDR67
+WDR7
+WDTC1
+WEE1
+WFS1
+WIF1
+WIPF1
+WIPF2
+WIPI1
+WISP1
+WISP2
+WNT1
+WNT10A
+WNT10B
+WNT11
+WNT16
+WNT2
+WNT2B
+WNT3
+WNT3A
+WNT4
+WNT5A
+WNT5B
+WNT6
+WNT7A
+WNT7B
+WNT8A
+WNT9A
+WRB
+WT1
+WTIP
+WWC1
+WWOX
+WWTR1
+XAB2
+XBP1
+XCL1
+XCL2
+XCR1
+XDH
+XIAP
+XPA
+XPC
+XPNPEP1
+XPO7
+XRCC1
+XRCC2
+XRCC3
+XRCC4
+XRCC5
+XRCC6
+XRCC6BP1
+YAP1
+YBX3
+YES1
+YKT6
+YME1L1
+YTHDF1
+YWHAB
+YWHAE
+YWHAH
+YWHAQ
+YWHAZ
+YY1
+ZAK
+ZAP70
+ZBTB7B
+ZC3H7A
+ZDHHC14
+ZDHHC18
+ZDHHC6
+ZEB1
+ZEB2
+ZFP36
+ZFP36L1
+ZFP91
+ZHX2
+ZHX3
+ZMIZ1
+ZMYM2
+ZNF131
+ZNF185
+ZNF274
+ZNF281
+ZNF318
+ZNF395
+ZNF451
+ZNF586
+ZNF589
+ZW10
+ZYX

CIGS/Gene_names.ipynb

@@ -0,0 +1,253 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "5c5b0c64",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>0</th>\n",
+       "      <th>1</th>\n",
+       "      <th>2</th>\n",
+       "      <th>3</th>\n",
+       "      <th>4</th>\n",
+       "      <th>5</th>\n",
+       "      <th>6</th>\n",
+       "      <th>7</th>\n",
+       "      <th>8</th>\n",
+       "      <th>9</th>\n",
+       "      <th>...</th>\n",
+       "      <th>3398</th>\n",
+       "      <th>3399</th>\n",
+       "      <th>3400</th>\n",
+       "      <th>3401</th>\n",
+       "      <th>3402</th>\n",
+       "      <th>3403</th>\n",
+       "      <th>3404</th>\n",
+       "      <th>3405</th>\n",
+       "      <th>3406</th>\n",
+       "      <th>3407</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Note: All these datasets are for non-commercia...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Sample_id</td>\n",
+       "      <td>A2M</td>\n",
+       "      <td>A4GNT</td>\n",
+       "      <td>AARS</td>\n",
+       "      <td>ABCA1</td>\n",
+       "      <td>ABCB1</td>\n",
+       "      <td>ABCB6</td>\n",
+       "      <td>ABCC5</td>\n",
+       "      <td>ABCC8</td>\n",
+       "      <td>ABCF1</td>\n",
+       "      <td>...</td>\n",
+       "      <td>ZNF185</td>\n",
+       "      <td>ZNF274</td>\n",
+       "      <td>ZNF281</td>\n",
+       "      <td>ZNF318</td>\n",
+       "      <td>ZNF395</td>\n",
+       "      <td>ZNF451</td>\n",
+       "      <td>ZNF586</td>\n",
+       "      <td>ZNF589</td>\n",
+       "      <td>ZW10</td>\n",
+       "      <td>ZYX</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>2 rows × 3408 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                0    1      2     3      4     \\\n",
+       "0  Note: All these datasets are for non-commercia...  NaN    NaN   NaN    NaN   \n",
+       "1                                          Sample_id  A2M  A4GNT  AARS  ABCA1   \n",
+       "\n",
+       "    5      6      7      8      9     ...    3398    3399    3400    3401  \\\n",
+       "0    NaN    NaN    NaN    NaN    NaN  ...     NaN     NaN     NaN     NaN   \n",
+       "1  ABCB1  ABCB6  ABCC5  ABCC8  ABCF1  ...  ZNF185  ZNF274  ZNF281  ZNF318   \n",
+       "\n",
+       "     3402    3403    3404    3405  3406 3407  \n",
+       "0     NaN     NaN     NaN     NaN   NaN  NaN  \n",
+       "1  ZNF395  ZNF451  ZNF586  ZNF589  ZW10  ZYX  \n",
+       "\n",
+       "[2 rows x 3408 columns]"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import pandas as pd\n",
+    "\n",
+    "file_path = \"/data/boom/CIGS/MCE_Bioactive_Compounds_MDA_MB_231_10μM_Counts.xlsx\"\n",
+    "\n",
+    "# 第 1 步：只把前两行当表头读进来，不读数据\n",
+    "header = pd.read_excel(file_path, nrows=0)          # 0 行数据，只解析列名\n",
+    "# 第 2 步：再读一次，跳过中间所有行，只拿前两行数据\n",
+    "df = pd.read_excel(file_path,\n",
+    "                   skiprows=0,       # 从第 0 行开始\n",
+    "                   nrows=2,          # 只要 2 行\n",
+    "                   header=None)      # 不再用第一行当表头\n",
+    "df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "c660f0f6",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "3407"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "gene_name = df.iloc[1].to_list()[1:]\n",
+    "\n",
+    "len(gene_name)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "25708d2e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "978"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "LINCS = pd.read_csv('/home/bob/boom/VCBench/data/LINCS2020/landmark_geneinfo.csv')\n",
+    "gene_info_LINCS = LINCS['gene_symbol'].astype(str)\n",
+    "len(gene_info_LINCS)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "ee20f28a",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "942"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# 取交集：gene_name & gene_info_LINCS\n",
+    "gene_name = set(gene_name)\n",
+    "gene_info_LINCS = set(gene_info_LINCS)\n",
+    "commen = gene_name & gene_info_LINCS\n",
+    "\n",
+    "len(commen) # 就用原始的"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "663f0a39",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "92377cdc",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

Get_Molecular_Embeddings.ipynb

@@ -0,0 +1,1662 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# look data_example"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "with open('./LINCS2020/idx2smi.pickle', 'rb') as f:\n",
+    "    idx2smi = pickle.load(f)\n",
+    "        \n",
+    "smiles_list = list(idx2smi.values())  # 假设idx2smi是一个字典，其值为SMILES字符串"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 BrC1C(Br)C(Br)C(Br)C(Br)C1Br\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(idx2smi), idx2smi[0])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "huggingface-cli upload Boom5426/MVCBench ./LINCS2020 ./data/ --repo-type dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'BrC1C(Br)C(Br)C(Br)C(Br)C1Br'"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "idx2smi[0]  # to path_to_input_smiles.csv"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "SMILES已成功保存到LINCS2020_smiles.csv文件中。\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "import csv\n",
+    "\n",
+    "# 将SMILES字符串写入CSV文件\n",
+    "with open('./LINCS2020/LINCS2020_smiles.csv', 'w', newline='') as csvfile:\n",
+    "    writer = csv.writer(csvfile)\n",
+    "    # 写入标题行（可选）\n",
+    "    writer.writerow(['SMILES'])\n",
+    "    # 遍历字典，写入SMILES\n",
+    "    for idx, smi in idx2smi.items():\n",
+    "        writer.writerow([smi])\n",
+    "\n",
+    "print(\"SMILES已成功保存到LINCS2020_smiles.csv文件中。\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate ECFP4 embedding for all smiles\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from rdkit import Chem\n",
+    "from rdkit.Chem import AllChem\n",
+    "import numpy as np\n",
+    "import pickle\n",
+    "\n",
+    "def smiles_to_ecfp4(smiles_list, radius=2, n_bits=2048):\n",
+    "    \"\"\"\n",
+    "    将SMILES列表转换为ECFP4特征向量。\n",
+    "\n",
+    "    参数:\n",
+    "    smiles_list (list): SMILES字符串列表。\n",
+    "    radius (int): ECFP的半径，默认为2（即ECFP4）。\n",
+    "    n_bits (int): 特征向量的长度，默认为2048。\n",
+    "\n",
+    "    返回:\n",
+    "    dict: 一个字典，键为SMILES字符串，值为对应的ECFP4特征向量。\n",
+    "    \"\"\"\n",
+    "    ecfp4_dict = {}\n",
+    "    \n",
+    "    for smiles in smiles_list:\n",
+    "        mol = Chem.MolFromSmiles(smiles)\n",
+    "        if mol is not None:\n",
+    "            # 替换编码器\n",
+    "            ecfp4 = AllChem.GetMorganFingerprintAsBitVect(mol, radius=radius, nBits=n_bits)\n",
+    "            ecfp4_dict[smiles] = np.array(ecfp4, dtype=np.float32)\n",
+    "        else:\n",
+    "            print(\"Error in \", smiles)\n",
+    "            # 如果SMILES无法解析为分子，则填充一个全零的向量\n",
+    "            ecfp4_dict[smiles] = np.zeros(n_bits, dtype=np.float32)\n",
+    "    \n",
+    "    return ecfp4_dict\n",
+    "\n",
+    "# 从字典中提取SMILES列表\n",
+    "smiles_list = list(idx2smi.values())  # 假设idx2smi是一个字典，其值为SMILES字符串\n",
+    "\n",
+    "# 计算ECFP4特征\n",
+    "ecfp4_features_dict = smiles_to_ecfp4(smiles_list)\n",
+    "\n",
+    "# ecfp4_features_dict\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 44,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 8316\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(smiles_list), len(ecfp4_features_dict))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ecfp4_features_dict\n",
+    "\n",
+    "# np.str_('BrC1C(Br)C(Br)C(Br)C(Br)C1Br'): array([0., 0., 0., ..., 0., 0., 0.], dtype=float32),"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 保存为 pickle 文件\n",
+    "with open('./LINCS2020/ECFP4_emb2048.pickle', 'wb') as f:\n",
+    "    pickle.dump(ecfp4_features_dict, f)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 50,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(2048,)"
+      ]
+     },
+     "execution_count": 50,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "ecfp4_features_dict['CN1CCN(CCCN2c3ccccc3Sc3ccc(cc23)C(F)(F)F)CC1'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# import pickle, h5py\n",
+    "# import numpy as np\n",
+    "# import pandas as pd\n",
+    "\n",
+    "# with open('./embeddings/ChemBERTa2_emb384.pickle', 'rb') as f:\n",
+    "#     KPGT_emb2304 = pickle.load(f)\n",
+    "# KPGT_emb2304"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(300,)"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# KPGT_emb2304['BrC1C(Br)C(Br)C(Br)C(Br)C1Br'].shape"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate KGPT embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "with open('./embeddings/KPGT_emb2304.pickle', 'rb') as f:\n",
+    "    KPGT_emb2304 = pickle.load(f)\n",
+    "KPGT_emb2304"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 111,755,936\n",
+      "111.8 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/KPGT/base.pth', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate InfoAlign embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# pip install infoalign\n",
+    "\n",
+    "# infoalign_predict --input LINCS2020/LINCS2020_smiles.csv \\\n",
+    "#                   --output InfoAlign_output.npy \\\n",
+    "#                   --output-to-input-column  # Adds the representation as an additional column in the input CSV\n",
+    "\n",
+    "# 是否存在数据泄露，因为这个是基于表征学习的"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 13,808,553\n",
+      "13.8 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/1463163983.py:2: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/infoalign_model/pretrain.pt', map_location='cpu')\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/infoalign_model/pretrain.pt', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')      # → 21.2 M"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import csv\n",
+    "import numpy as np\n",
+    "import pickle\n",
+    "\n",
+    "# CSV文件路径\n",
+    "csv_path = 'infoalign_model/LINCS2020_smiles.csv'\n",
+    "# NPY文件路径\n",
+    "npy_path = 'infoalign_model/InfoAlign_output.npy'\n",
+    "\n",
+    "# 初始化一个空字典来存储SMILES和对应的表征\n",
+    "smiles_to_representation = {}\n",
+    "\n",
+    "# 读取NPY文件\n",
+    "info_align_output = np.load(npy_path)\n",
+    "\n",
+    "# 读取CSV文件并同时替换字典中的值\n",
+    "with open(csv_path, mode='r', newline='', encoding='utf-8') as csvfile:\n",
+    "    reader = csv.DictReader(csvfile)\n",
+    "    for idx, row in enumerate(reader):\n",
+    "        smiles = row['SMILES']\n",
+    "        # 直接从info_align_output中获取对应的数值\n",
+    "        representation = info_align_output[idx]\n",
+    "        smiles_to_representation[smiles] = representation\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "8316"
+      ]
+     },
+     "execution_count": 34,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(smiles_to_representation)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 embeddings/InfoAlign_emb300.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/InfoAlign_emb300.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(smiles_to_representation, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate ChemBERTa-2 embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
+     ]
+    }
+   ],
+   "source": [
+    "from transformers import AutoModel, AutoTokenizer\n",
+    "model = AutoModel.from_pretrained(\"DeepChem/ChemBERTa-77M-MLM\")   # 77M SMILES（完整PubChem数据集)\n",
+    "tokenizer = AutoTokenizer.from_pretrained(\"DeepChem/ChemBERTa-77M-MLM\")\n",
+    "inputs = tokenizer(\"CCO\", return_tensors=\"pt\")\n",
+    "outputs = model(**inputs)  # 嵌入位于 outputs.last_hidden_state"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "torch.Size([1, 5, 384])"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "outputs.last_hidden_state.shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "ChemBERTa-77M 可训练参数量: 3,427,440\n",
+      "3.4 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "# from transformers import AutoModel\n",
+    "\n",
+    "# model = AutoModel.from_pretrained(\"DeepChem/ChemBERTa-77M-MLM\")\n",
+    "\n",
+    "# trainable = sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
+    "# print(f\"ChemBERTa-77M 可训练参数量: {trainable:,}\")\n",
+    "# print(f\"{trainable/1e6:.1f} M\")   # → 85.7 M"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# from transformers import AutoModel, AutoTokenizer\n",
+    "# import torch, os, shutil\n",
+    "\n",
+    "# repo_id = \"DeepChem/ChemBERTa-77M-MLM\"\n",
+    "\n",
+    "# # 1. 磁盘大小：下载后统计整个 snapshot 目录\n",
+    "# tmp_dir = \"./tmp_snapshot\"          # 临时目录，用完即删\n",
+    "# model = AutoModel.from_pretrained(repo_id, cache_dir=tmp_dir)\n",
+    "# tokenizer = AutoTokenizer.from_pretrained(repo_id, cache_dir=tmp_dir)\n",
+    "\n",
+    "# disk_mb = sum(\n",
+    "#     os.path.getsize(os.path.join(root, f))\n",
+    "#     for root, _, files in os.walk(tmp_dir)\n",
+    "#     for f in files\n",
+    "# ) / 1024 ** 2\n",
+    "\n",
+    "# print(f\"权重文件磁盘占用：{disk_mb:.1f} MB\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n",
+      "100%|██████████| 260/260 [00:01<00:00, 168.25it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "总分子数 : 8316\n",
+      "总耗时   : 2.82 s\n",
+      "平均推理 : 0.34 ms / molecule\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "import time\n",
+    "from transformers import AutoModel, AutoTokenizer\n",
+    "import torch\n",
+    "import numpy as np\n",
+    "from rdkit import Chem\n",
+    "from tqdm import tqdm  # 可选：用于进度条\n",
+    "\n",
+    "def smiles_to_chemberta2(smiles_list, model_name=\"DeepChem/ChemBERTa-77M-MLM\", batch_size=32):\n",
+    "    \"\"\"\n",
+    "    将SMILES列表转换为ChemBERTa-2的嵌入向量（CLS token的隐藏状态）。\n",
+    "    \n",
+    "    参数:\n",
+    "        smiles_list (list): SMILES字符串列表\n",
+    "        model_name (str): Hugging Face模型ID，默认为ChemBERTa-2的MLM版本\n",
+    "        batch_size (int): 批处理大小，根据GPU内存调整\n",
+    "        \n",
+    "    返回:\n",
+    "        dict: 键为SMILES，值为768维嵌入向量（numpy数组）\n",
+    "    \"\"\"\n",
+    "    # 加载模型和tokenizer\n",
+    "    tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
+    "    model = AutoModel.from_pretrained(model_name)\n",
+    "    model.eval()  # 切换到推理模式\n",
+    "    \n",
+    "    # 检查CUDA可用性\n",
+    "    device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
+    "    model.to(device)\n",
+    "    \n",
+    "    chemberta2_dict = {}\n",
+    "    \n",
+    "    # 分批处理SMILES\n",
+    "    for i in tqdm(range(0, len(smiles_list), batch_size)):\n",
+    "        batch_smiles = smiles_list[i:i+batch_size]\n",
+    "        valid_indices = []\n",
+    "        batch_inputs = []\n",
+    "        \n",
+    "        # 过滤无效SMILES并生成输入\n",
+    "        for idx, smiles in enumerate(batch_smiles):\n",
+    "            mol = Chem.MolFromSmiles(smiles)\n",
+    "            if mol is not None:\n",
+    "                valid_indices.append(idx)\n",
+    "                batch_inputs.append(smiles)\n",
+    "            else:\n",
+    "                print(f\"Invalid SMILES: {smiles}\")\n",
+    "        \n",
+    "        if not batch_inputs:\n",
+    "            continue\n",
+    "            \n",
+    "        # Tokenize并移动到设备\n",
+    "        inputs = tokenizer(\n",
+    "            batch_inputs, \n",
+    "            return_tensors=\"pt\", \n",
+    "            padding=True, \n",
+    "            truncation=True, \n",
+    "            max_length=512\n",
+    "        ).to(device)\n",
+    "        \n",
+    "        # 生成嵌入\n",
+    "        with torch.no_grad():\n",
+    "            outputs = model(**inputs)\n",
+    "            embeddings = outputs.last_hidden_state[:, 0, :]  # 取CLS token的嵌入\n",
+    "        \n",
+    "        # 将有效结果存入字典\n",
+    "        for j, idx in enumerate(valid_indices):\n",
+    "            chemberta2_dict[batch_smiles[idx]] = embeddings[j].cpu().numpy()\n",
+    "        \n",
+    "        # 处理无效SMILES（填充零向量）\n",
+    "        for idx, smiles in enumerate(batch_smiles):\n",
+    "            if idx not in valid_indices:\n",
+    "                chemberta2_dict[smiles] = np.zeros(768, dtype=np.float32)\n",
+    "    \n",
+    "    return chemberta2_dict\n",
+    "\n",
+    "# 使用示例\n",
+    "import time\n",
+    "# ----- 计时开始 -----\n",
+    "t0 = time.perf_counter()\n",
+    "smiles_list = list(idx2smi.values())  # 假设idx2smi是SMILES字典\n",
+    "chemberta2_features_dict = smiles_to_chemberta2(smiles_list)\n",
+    "\n",
+    "t1 = time.perf_counter()\n",
+    "# ----- 计时结束 -----\n",
+    "\n",
+    "# 基本统计\n",
+    "total_mols = len(chemberta2_features_dict)\n",
+    "avg_time_ms = (t1 - t0) / total_mols * 1000\n",
+    "\n",
+    "print(f\"总分子数 : {total_mols}\")\n",
+    "print(f\"总耗时   : {t1 - t0:.2f} s\")\n",
+    "print(f\"平均推理 : {avg_time_ms:.2f} ms / molecule\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 384\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(chemberta2_features_dict), len(chemberta2_features_dict['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "chemberta2_features_dict['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 embeddings/ChemBERTa2_emb384.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/ChemBERTa2_emb384.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(chemberta2_features_dict, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate MolT5 embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "torch.Size([1, 768])\n"
+     ]
+    }
+   ],
+   "source": [
+    "from transformers import T5ForConditionalGeneration, AutoTokenizer\n",
+    "model = T5ForConditionalGeneration.from_pretrained(\"laituan245/molt5-base\")\n",
+    "tokenizer = AutoTokenizer.from_pretrained(\"laituan245/molt5-base\")\n",
+    "inputs = tokenizer(\"CCO\", return_tensors=\"pt\")\n",
+    "embeddings = model.encoder(**inputs).last_hidden_state.mean(dim=1)  # 平均池化\n",
+    "print(embeddings.shape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "MolT5-base 可训练参数量: 247.6 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "trainable = sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
+    "print(f'MolT5-base 可训练参数量: {trainable/1e6:.1f} M')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "已缓存权重磁盘占用：3784.1 MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "# molt5_disk_usage.py\n",
+    "from transformers import AutoModel, AutoTokenizer\n",
+    "import os, pathlib\n",
+    "\n",
+    "repo_id = \"laituan245/molt5-base\"\n",
+    "\n",
+    "# 让 transformers 只走本地缓存，不会触发下载\n",
+    "model = AutoModel.from_pretrained(repo_id, local_files_only=True)\n",
+    "tokenizer = AutoTokenizer.from_pretrained(repo_id, local_files_only=True)\n",
+    "\n",
+    "# 找到缓存目录\n",
+    "cache_root = pathlib.Path.home() / \".cache/huggingface/hub\"\n",
+    "# 仓库目录名格式：models--{user}--{repo}\n",
+    "repo_dir = cache_root / f\"models--{repo_id.replace('/', '--')}\"\n",
+    "\n",
+    "disk_bytes = sum(f.stat().st_size for f in repo_dir.rglob(\"*\") if f.is_file())\n",
+    "disk_mb = disk_bytes / 1024 ** 2\n",
+    "\n",
+    "print(f\"已缓存权重磁盘占用：{disk_mb:.1f} MB\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|██████████| 130/130 [00:09<00:00, 13.90it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "总分子数 : 8316\n",
+      "总耗时   : 11.55 s\n",
+      "平均推理 : 1.39 ms / molecule\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "from transformers import T5ForConditionalGeneration, AutoTokenizer\n",
+    "import torch\n",
+    "import numpy as np\n",
+    "from rdkit import Chem\n",
+    "from tqdm import tqdm\n",
+    "\n",
+    "def smiles_to_molt5(smiles_list, model_name=\"laituan245/molt5-base\", batch_size=32, pooling=\"mean\"):\n",
+    "    \"\"\"\n",
+    "    使用 MolT5 生成 SMILES 的嵌入向量（基于 encoder 的隐藏状态）\n",
+    "    \n",
+    "    参数:\n",
+    "        smiles_list (list): SMILES 字符串列表\n",
+    "        model_name (str): Hugging Face 模型 ID，默认为 MolT5-base\n",
+    "        batch_size (int): 批处理大小，根据 GPU 内存调整\n",
+    "        pooling (str): 池化方法，可选 \"mean\"（平均池化）或 \"cls\"（取 CLS token）\n",
+    "        \n",
+    "    返回:\n",
+    "        dict: 键为 SMILES，值为嵌入向量（numpy 数组）\n",
+    "    \"\"\"\n",
+    "    # 加载 MolT5 的 tokenizer 和模型\n",
+    "    model = T5ForConditionalGeneration.from_pretrained(model_name)\n",
+    "    tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
+    "    model.eval()\n",
+    "    \n",
+    "    # 设备设置\n",
+    "    device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
+    "    model.to(device)\n",
+    "    \n",
+    "    molt5_dict = {}\n",
+    "    \n",
+    "    # 分批处理\n",
+    "    for i in tqdm(range(0, len(smiles_list), batch_size)):\n",
+    "        batch_smiles = smiles_list[i:i+batch_size]\n",
+    "        valid_indices = []\n",
+    "        batch_inputs = []\n",
+    "        \n",
+    "        # 过滤无效 SMILES\n",
+    "        for idx, smiles in enumerate(batch_smiles):\n",
+    "            mol = Chem.MolFromSmiles(smiles)\n",
+    "            if mol is not None:\n",
+    "                valid_indices.append(idx)\n",
+    "                batch_inputs.append(smiles)\n",
+    "            else:\n",
+    "                print(f\"Invalid SMILES: {smiles}\")\n",
+    "                molt5_dict[smiles] = np.zeros(model.config.d_model, dtype=np.float32)  # 填充零向量\n",
+    "        \n",
+    "        if not batch_inputs:\n",
+    "            continue\n",
+    "            \n",
+    "        # Tokenize 并生成嵌入\n",
+    "        inputs = tokenizer(\n",
+    "            batch_inputs,\n",
+    "            return_tensors=\"pt\",\n",
+    "            padding=True,\n",
+    "            truncation=True,\n",
+    "            max_length=512\n",
+    "        ).to(device)\n",
+    "        \n",
+    "        with torch.no_grad():\n",
+    "            # 获取 encoder 的隐藏状态（形状: [batch_size, seq_len, hidden_dim]）\n",
+    "            encoder_outputs = model.encoder(**inputs)\n",
+    "            hidden_states = encoder_outputs.last_hidden_state\n",
+    "            \n",
+    "            # 池化策略\n",
+    "            if pooling == \"mean\":\n",
+    "                embeddings = hidden_states.mean(dim=1)  # 平均池化\n",
+    "            elif pooling == \"cls\":\n",
+    "                embeddings = hidden_states[:, 0, :]     # CLS token\n",
+    "            else:\n",
+    "                raise ValueError(\"pooling 必须是 'mean' 或 'cls'\")\n",
+    "        \n",
+    "        # 存储结果\n",
+    "        for j, idx in enumerate(valid_indices):\n",
+    "            molt5_dict[batch_smiles[idx]] = embeddings[j].cpu().numpy()\n",
+    "    \n",
+    "    return molt5_dict\n",
+    "\n",
+    "\n",
+    "import time\n",
+    "# ----- 计时开始 -----\n",
+    "smiles_list = list(idx2smi.values())  # 假设 idx2smi 是 SMILES 字典\n",
+    "\n",
+    "t0 = time.perf_counter()\n",
+    "molt5_embeddings = smiles_to_molt5(smiles_list, batch_size=64, pooling=\"mean\")\n",
+    "\n",
+    "t1 = time.perf_counter()\n",
+    "# ----- 计时结束 -----\n",
+    "\n",
+    "# 基本统计\n",
+    "total_mols = len(molt5_embeddings)\n",
+    "avg_time_ms = (t1 - t0) / total_mols * 1000\n",
+    "\n",
+    "print(f\"总分子数 : {total_mols}\")\n",
+    "print(f\"总耗时   : {t1 - t0:.2f} s\")\n",
+    "print(f\"平均推理 : {avg_time_ms:.2f} ms / molecule\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 768\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(molt5_embeddings), len(molt5_embeddings['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "molt5_embeddings['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 embeddings/molt5_emb768.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/molt5_emb768.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(molt5_embeddings, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate Chemprop embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# git clone https://github.com/chemprop/chemprop.git\n",
+    "# https://chemprop.readthedocs.io/en/main/installation.html\n",
+    "# https://chemprop.readthedocs.io/en/main/tutorial/cli/fingerprint.html#fingerprint \n",
+    "\n",
+    "# chemprop fingerprint --test-path ./embeddings/LINCS2020_smiles.csv --model-path chemprop/tests/data/example_model_v2_regression_mol.ckpt --output embeddings/fps.csv --ffn-block-index -2"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训���权重总参数量: 319,505\n",
+      "0.3 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/3140832406.py:3: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  ckpt = torch.load('/home/bob/boom/VCBench/Molecule_encoder/chemprop/tests/data/example_model_v2_regression_mol.ckpt',\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "\n",
+    "ckpt = torch.load('/home/bob/boom/VCBench/Molecule_encoder/chemprop/tests/data/example_model_v2_regression_mol.ckpt',\n",
+    "                  map_location='cpu')\n",
+    "\n",
+    "# 取出模型参数部分\n",
+    "model_state = ckpt['state_dict']\n",
+    "\n",
+    "total = sum(v.numel() for v in model_state.values() if isinstance(v, torch.Tensor))\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "8316"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import csv\n",
+    "import numpy as np\n",
+    "\n",
+    "# CSV文件路径\n",
+    "csv1_path = 'infoalign_model/LINCS2020_smiles.csv'\n",
+    "csv2_path = 'embeddings/fps_0.csv'\n",
+    "\n",
+    "# 初始化一个空字典来存储SMILES和对应的表征\n",
+    "smiles_to_representation = {}\n",
+    "\n",
+    "# 读取第一个CSV文件（SMILES信息）\n",
+    "smiles_list = []\n",
+    "with open(csv1_path, mode='r', newline='', encoding='utf-8') as csvfile:\n",
+    "    reader = csv.DictReader(csvfile)\n",
+    "    for row in reader:\n",
+    "        smiles = row['SMILES']  # 假设CSV文件中列名为'SMILES'\n",
+    "        smiles_list.append(smiles)\n",
+    "\n",
+    "# 读取第二个CSV文件（表征信息），跳过第一行\n",
+    "representation_list = []\n",
+    "with open(csv2_path, mode='r', newline='', encoding='utf-8') as csvfile:\n",
+    "    reader = csv.reader(csvfile)\n",
+    "    next(reader)  # 跳过第一行\n",
+    "    for row in reader:\n",
+    "        # 将字符串转换为浮点数\n",
+    "        representation = [np.float32(x)  for x in row]  # 假设表征信息是整行数据\n",
+    "        representation_list.append(representation)\n",
+    "\n",
+    "# 将列表转换为numpy数组\n",
+    "representation_array = np.array(representation_list)\n",
+    "\n",
+    "# 将SMILES和对应的表征合并到字典中\n",
+    "for smiles, representation in zip(smiles_list, representation_array):\n",
+    "    smiles_to_representation[smiles] = representation\n",
+    "\n",
+    "len(smiles_to_representation)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "300 <class 'numpy.ndarray'>\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "dtype('float32')"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "print(len(smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']), type(smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))\n",
+    "\n",
+    "smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br'].dtype"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "300\n",
+      "字典已成功保存到 embeddings/Chemprop_emb300.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))\n",
+    "\n",
+    "import pickle\n",
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/Chemprop_emb300.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(smiles_to_representation, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate MolCLR embedding for all smiles\n",
+    "\n",
+    "## see /home/bob/boom/DrugIM/TranSiGen/data/MolCLR/Get_emd.ipynb\n",
+    "\n",
+    "### We use GIN model, which has a better performance in their paper"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 2,407,201\n",
+      "2.4 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/3871607093.py:2: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/MolCLR/ckpt/pretrained_gin/checkpoints/model.pth', map_location='cpu')\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/MolCLR/ckpt/pretrained_gin/checkpoints/model.pth', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')      # → 21.2 M"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate Mole-BERT embedding for all smiles\n",
+    "\n",
+    "## see /home/bob/boom/DrugIM/TranSiGen/data/Mole-BERT/Get_emd.ipynb\n",
+    "\n",
+    "### We use GNN_graphpred model, which has a better performance in their paper"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 1,860,905\n",
+      "1.9 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/2103728332.py:2: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/Mole-BERT/model_gin/Mole-BERT.pth', map_location='cpu')\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/Mole-BERT/model_gin/Mole-BERT.pth', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')      # → 21.2 M"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 3D  UniMol"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "8316"
+      ]
+     },
+     "execution_count": 1,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "with open('./LINCS2020/idx2smi.pickle', 'rb') as f:\n",
+    "    idx2smi = pickle.load(f)\\\n",
+    "\n",
+    "unique_smiles = list(idx2smi.values())  # 假设idx2smi是一个字典，其值为SMILES字符串\n",
+    "len(unique_smiles)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2025-09-15 11:17:31 | unimol_tools/models/unimolv2.py | 161 | INFO | Uni-Mol Tools | Loading pretrained weights from /home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/unimol_tools/weights/modelzoo/310M/checkpoint.pt\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "开始生成32个SMILES的嵌入表示...\n",
+      "开始\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "处理批次:   0%|          | 0/1 [00:00<?, ?it/s]2025-09-15 11:17:32 | unimol_tools/data/conformer.py | 437 | INFO | Uni-Mol Tools | Start generating conformers...\n",
+      "32it [00:00, 82.76it/s]\n",
+      "2025-09-15 11:17:33 | unimol_tools/data/conformer.py | 452 | INFO | Uni-Mol Tools | Succeeded in generating conformers for 100.00% of molecules.\n",
+      "2025-09-15 11:17:33 | unimol_tools/data/conformer.py | 469 | INFO | Uni-Mol Tools | Succeeded in generating 3d conformers for 100.00% of molecules.\n",
+      "2025-09-15 11:17:33 | unimol_tools/tasks/trainer.py | 78 | INFO | Uni-Mol Tools | Number of GPUs available: 1\n",
+      "2025-09-15 11:17:33 | unimol_tools/tasks/trainer.py | 98 | INFO | Uni-Mol Tools | Using single GPU.\n",
+      "100%|██████████| 1/1 [00:00<00:00,  2.32it/s]\n",
+      "处理批次: 100%|██████████| 1/1 [00:01<00:00,  1.25s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "嵌入生成完成！共处理 32 个SMILES，32 个成功，0 个失败。\n",
+      "嵌入结果已保存至 embeddings/UniMolV2_emb1024.pkl\n",
+      "总分子数 : 32\n",
+      "总耗时   : 3.50 s\n",
+      "平均推理 : 109.37 ms / molecule\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import pickle\n",
+    "import numpy as np\n",
+    "from tqdm import tqdm\n",
+    "from unimol_tools import UniMolRepr\n",
+    "import os, sys, contextlib\n",
+    "\n",
+    "@contextlib.contextmanager\n",
+    "def suppress_stdout_stderr():\n",
+    "    \"\"\"with 块内所有 print / tqdm / warning 都不会显示\"\"\"\n",
+    "    with open(os.devnull, 'w') as devnull:\n",
+    "        old_out, old_err = sys.stdout, sys.stderr\n",
+    "        sys.stdout, sys.stderr = devnull, devnull\n",
+    "        try:\n",
+    "            yield\n",
+    "        finally:\n",
+    "            sys.stdout, sys.stderr = old_out, old_err\n",
+    "\n",
+    "def get_unimol_embeddings(smiles_list, output_file=\"UniMol_emb512.pkl\", \n",
+    "                         model_name='unimolv1', model_size='84m', \n",
+    "                         remove_hs=False, batch_size=32):\n",
+    "    \"\"\"\n",
+    "    使用Uni-Mol模型为SMILES列表生成分子嵌入，并保存为pickle文件\n",
+    "    \n",
+    "    参数:\n",
+    "    smiles_list (list): SMILES字符串列表\n",
+    "    output_file (str): 输出pickle文件路径\n",
+    "    model_name (str): 模型名称，可选'unimolv1'或'unimolv2'\n",
+    "    model_size (str): 模型大小，仅在使用unimolv2时有效\n",
+    "    remove_hs (bool): 是否移除氢原子\n",
+    "    batch_size (int): 批处理大小\n",
+    "    \n",
+    "    返回:\n",
+    "    dict: 包含SMILES及其对应嵌入的字典\n",
+    "    \"\"\"\n",
+    "    # 初始化模型\n",
+    "    clf = UniMolRepr(\n",
+    "        data_type='molecule',\n",
+    "        remove_hs=remove_hs,\n",
+    "        model_name=model_name,\n",
+    "        model_size=model_size\n",
+    "    )\n",
+    "    \n",
+    "    # 用于存储结果的字典\n",
+    "    embeddings_dict = {}\n",
+    "    error_smiles = []\n",
+    "    \n",
+    "    # 批处理SMILES\n",
+    "    total_batches = (len(smiles_list) + batch_size - 1) // batch_size\n",
+    "    \n",
+    "    print(f\"开始生成{len(smiles_list)}个SMILES的嵌入表示...\")\n",
+    "    \n",
+    "    print(\"开始\")\n",
+    "    # with suppress_stdout_stderr():\n",
+    "    for i in tqdm(range(total_batches), desc=\"处理批次\"):\n",
+    "        batch = smiles_list[i*batch_size : (i+1)*batch_size]\n",
+    "        \n",
+    "        try:\n",
+    "            # 获取嵌入表示\n",
+    "            batch_repr = clf.get_repr(batch, return_atomic_reprs=False)\n",
+    "            \n",
+    "            # 将结果存入字典 (使用CLS token作为分子表示)\n",
+    "            for idx, smiles in enumerate(batch):\n",
+    "                embeddings_dict[smiles] = batch_repr['cls_repr'][idx]\n",
+    "                \n",
+    "        except Exception as e:\n",
+    "            print(f\"处理批次 {i+1}/{total_batches} 时发生错误: {str(e)}\")\n",
+    "            # 记录处理失败的SMILES\n",
+    "            error_smiles.extend(batch)\n",
+    "        \n",
+    "    # 保存嵌入结果\n",
+    "    with open(output_file, 'wb') as f:\n",
+    "        pickle.dump(embeddings_dict, f)\n",
+    "    \n",
+    "    print(f\"嵌入生成完成！共处理 {len(smiles_list)} 个SMILES，\"\n",
+    "          f\"{len(smiles_list) - len(error_smiles)} 个成功，\"\n",
+    "          f\"{len(error_smiles)} 个失败。\")\n",
+    "    print(f\"嵌入结果已保存至 {output_file}\")\n",
+    "    \n",
+    "    if error_smiles:\n",
+    "        print(f\"处理失败的SMILES已记录。\")\n",
+    "    \n",
+    "    return embeddings_dict\n",
+    "\n",
+    "# 使用示例\n",
+    "if __name__ == \"__main__\":\n",
+    "    \n",
+    "    # 假设unique_smiles是你的SMILES列表\n",
+    "    unique_smiles = [\n",
+    "    \"CC(=O)OC1=CC=CC=C1C(=O)O\",  # Aspirin\n",
+    "    \"CN1C=NC2=C1C(=O)N(C(=O)N2C)C\" # Caffeine\n",
+    "    ]  # 示例SMILES列表\n",
+    "    unique_smiles = pd.read_csv('./LINCS2020/LINCS2020_smiles.csv')['SMILES'].tolist()\n",
+    "\n",
+    "    # UniMol V1\n",
+    "    # embeddings = get_unimol_embeddings(\n",
+    "    #     smiles_list=unique_smiles,\n",
+    "    #     output_file=\"embeddings/UniMol_emb512.pkl\",\n",
+    "    #     model_name='unimolv1',\n",
+    "    #     model_size='84m',\n",
+    "    #     remove_hs=False,\n",
+    "    #     batch_size=32\n",
+    "    # )\n",
+    "    \n",
+    "    # UniMol V2\n",
+    "    embeddings = get_unimol_embeddings(\n",
+    "    smiles_list=unique_smiles,\n",
+    "    output_file=\"embeddings/UniMolV2_emb1024.pkl\", # save path\n",
+    "    model_name='unimolv2',      # 修改为 v2 版本\n",
+    "    model_size='310m',          # 指定 v2 模型大小（根据需要选择）\n",
+    "    remove_hs=False,\n",
+    "    batch_size=32\n",
+    "    )\n",
+    "\n",
+    "    # 打印样例嵌入\n",
+    "    sample_smiles = unique_smiles[0]\n",
+    "    print(f\"SMILES: {sample_smiles}\")\n",
+    "    print(f\"嵌入向量: {embeddings[sample_smiles]}\")\n",
+    "    print(f\"嵌入维度: {len(embeddings[sample_smiles])}\")\n",
+    "    "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2025-10-16 21:29:18 | unimol_tools/models/unimol.py | 136 | INFO | Uni-Mol Tools | Loading pretrained weights from /home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/unimol_tools/weights/mol_pre_all_h_220816.pt\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "UniMol-unimolv1-84m 可训练参数量: 47.3 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2025-10-16 21:29:20 | unimol_tools/models/unimolv2.py | 161 | INFO | Uni-Mol Tools | Loading pretrained weights from /home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/unimol_tools/weights/modelzoo/310M/checkpoint.pt\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "UniMol-unimolv2-310m 可训练参数量: 308.9 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import pickle\n",
+    "import numpy as np\n",
+    "from tqdm import tqdm\n",
+    "from unimol_tools import UniMolRepr\n",
+    "import os, sys, contextlib\n",
+    "\n",
+    "def get_unimol_embeddings(smiles_list, output_file=\"UniMol_emb512.pkl\", \n",
+    "                         model_name='unimolv1', model_size='84m', \n",
+    "                         remove_hs=False, batch_size=32):\n",
+    "    \"\"\"\n",
+    "    使用Uni-Mol模型为SMILES列表生成分子嵌入，并保存为pickle文件\n",
+    "    \n",
+    "    参数:\n",
+    "    smiles_list (list): SMILES字符串列表\n",
+    "    output_file (str): 输出pickle文件路径\n",
+    "    model_name (str): 模型名称，可选'unimolv1'或'unimolv2'\n",
+    "    model_size (str): 模型大小，仅在使用unimolv2时有效\n",
+    "    remove_hs (bool): 是否移除氢原子\n",
+    "    batch_size (int): 批处理大小\n",
+    "    \n",
+    "    返回:\n",
+    "    dict: 包含SMILES及其对应嵌入的字典\n",
+    "    \"\"\"\n",
+    "    # 初始化模型\n",
+    "    # ===== 在 clf 初始化完成后立即统计 =====\n",
+    "    clf = UniMolRepr(\n",
+    "            data_type='molecule',\n",
+    "            remove_hs=remove_hs,\n",
+    "            model_name=model_name,\n",
+    "            model_size=model_size)\n",
+    "\n",
+    "    # 取出真正的网络\n",
+    "    real_model = clf.model          # <== 关键属性\n",
+    "\n",
+    "    total = sum(p.numel() for p in real_model.parameters() if p.requires_grad)\n",
+    "    print(f'UniMol-{model_name}-{model_size} 可训练参数量: {total/1e6:.1f} M')\n",
+    "  \n",
+    "        \n",
+    "# 使用示例\n",
+    "if __name__ == \"__main__\":\n",
+    "    # 假设unique_smiles是你的SMILES列表\n",
+    "    # 生成并保存嵌入\n",
+    "    import time\n",
+    "    # ----- 计时开始 -----\n",
+    "    unique_smiles = list(idx2smi.values())  # 假设 idx2smi 是 SMILES 字典\n",
+    "\n",
+    "    t0 = time.perf_counter()\n",
+    "    \n",
+    "    embeddings = get_unimol_embeddings(\n",
+    "        smiles_list=unique_smiles[:32],\n",
+    "        output_file=\"embeddings/UniMol_emb512.pkl\",\n",
+    "        model_name='unimolv1',\n",
+    "        model_size='84m',\n",
+    "        remove_hs=False,\n",
+    "        batch_size=32\n",
+    "    )\n",
+    "    # UniMol 2\n",
+    "    embeddings = get_unimol_embeddings(\n",
+    "    smiles_list=unique_smiles[:32],\n",
+    "    output_file=\"embeddings/UniMolV2_emb1024.pkl\",\n",
+    "    model_name='unimolv2',      # 修改为 v2 版本\n",
+    "    model_size='310m',          # 指定 v2 模型大小（根据需要选择）\n",
+    "    remove_hs=False,\n",
+    "    batch_size=32\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# output_file=\"embeddings/UniMolV2_emb1024.pkl\"\n",
+    "# with open(output_file, 'wb') as f:\n",
+    "#     pickle.dump(embeddings, f)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "path = './embeddings/UniMolV2_emb1024.pkl'\n",
+    "import pickle\n",
+    "with open(path, 'rb') as f:\n",
+    "    unimol_feat = pickle.load(f)\n",
+    "\n",
+    "unimol_feat"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 512\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(unimol_feat), len(unimol_feat['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "处理前维度: (1, 512)\n",
+      "处理后维度: (512,)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# import pickle\n",
+    "# import numpy as np\n",
+    "\n",
+    "# # 加载原始特征\n",
+    "# path = './embeddings/UniMol_emb512.pkl'\n",
+    "# with open(path, 'rb') as f:\n",
+    "#     unimol_feat = pickle.load(f)\n",
+    "\n",
+    "# # 处理特征维度：从(1,512)变为(512,)\n",
+    "# unimol_feat_processed = {k: v.squeeze() for k, v in unimol_feat.items()}\n",
+    "\n",
+    "# # 验证处理结果\n",
+    "# test_molecule = 'BrC1C(Br)C(Br)C(Br)C(Br)C1Br'\n",
+    "# if test_molecule in unimol_feat_processed:\n",
+    "#     print(f\"处理前维度: {unimol_feat[test_molecule].shape}\")\n",
+    "#     print(f\"处理后维度: {unimol_feat_processed[test_molecule].shape}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# # 保存处理后的特征\n",
+    "# # 选项1：覆盖原始文件\n",
+    "# with open(path, 'wb') as f:\n",
+    "#     pickle.dump(unimol_feat_processed, f)\n",
+    "\n",
+    "# 选项2：保存到新文件（推荐，避免覆盖原始数据）\n",
+    "# processed_path = './embeddings/UniMol_emb512_processed.pkl'\n",
+    "# with open(processed_path, 'wb') as f:\n",
+    "#     pickle.dump(unimol_feat_processed, f)\n",
+    "\n",
+    "# print(f\"处理后的特征已保存至: {processed_path}\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

LINCS2020/gene_978_to965.ipynb

@@ -0,0 +1,582 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "8d20992a",
+   "metadata": {},
+   "source": [
+    "# 找到978中的965，去掉其他的基因，得到纯种 LINCS965"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "a03cc045",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['BrC1C(Br)C(Br)C(Br)C(Br)C1Br', 'BrC1C(Br)C(Br)C(Br)C(Br)C1Br',\n",
+       "        'BrC1C(Br)C(Br)C(Br)C(Br)C1Br', ..., 'c1nc(cs1)-c1nc2ccccc2[nH]1',\n",
+       "        'c1nc(cs1)-c1nc2ccccc2[nH]1', 'c1nc(cs1)-c1nc2ccccc2[nH]1'],\n",
+       "       dtype='<U454'),\n",
+       " 'cid': array(['A549', 'ASC', 'MCF7', ..., 'THP1', 'VCAP', 'YAPC'], dtype='<U8'),\n",
+       " 'sig': array(['A549_0', 'ASC_0', 'MCF7_0', ..., 'THP1_8315', 'VCAP_8315',\n",
+       "        'YAPC_8315'], dtype='<U13'),\n",
+       " 'x1': array([[ 9.533199 ,  4.79235  ,  9.7237   , ..., 11.9874   ,  9.3001   ,\n",
+       "          8.39445  ],\n",
+       "        [ 8.509138 ,  4.8156996,  7.3642497, ..., 11.6545   , 10.211599 ,\n",
+       "          8.973351 ],\n",
+       "        [ 9.043461 ,  7.6456842,  7.459799 , ...,  6.712334 , 11.85135  ,\n",
+       "          6.3363895],\n",
+       "        ...,\n",
+       "        [ 8.86875  ,  5.9342003,  8.8046   , ...,  7.338975 , 11.65115  ,\n",
+       "          8.347799 ],\n",
+       "        [ 9.482506 ,  5.9174104,  8.216723 , ...,  6.722115 , 10.991553 ,\n",
+       "          7.3406167],\n",
+       "        [ 9.725147 ,  7.049716 ,  9.0830555, ...,  9.198456 , 11.310257 ,\n",
+       "          7.920928 ]], dtype=float32),\n",
+       " 'x2': array([[ 9.4455   ,  6.92005  , 10.2842   , ..., 12.2337   ,  9.618799 ,\n",
+       "          7.8896003],\n",
+       "        [ 8.291338 ,  5.3398876,  6.832475 , ..., 11.05195  , 11.712725 ,\n",
+       "          9.1215   ],\n",
+       "        [ 9.176098 ,  7.160512 ,  7.151138 , ...,  6.86923  , 11.213164 ,\n",
+       "          6.2830935],\n",
+       "        ...,\n",
+       "        [ 8.7068   ,  5.7699003,  8.881599 , ...,  7.468175 , 11.75805  ,\n",
+       "          8.6483   ],\n",
+       "        [ 9.617616 ,  5.709352 ,  7.277331 , ...,  7.262997 , 10.814158 ,\n",
+       "          7.3617706],\n",
+       "        [ 9.710126 ,  6.609707 ,  9.075139 , ...,  9.596974 , 10.573312 ,\n",
+       "          7.4035025]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n",
+    "\n",
+    "LINCS_data_example = load_from_HDF('./processed_data.h5')\n",
+    "LINCS_data_example"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "ae1d3985",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(78569, 978)"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "LINCS_data_example['x1'].shape  # Example shape of the data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "5cc633d4",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "978"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "LINCS = pd.read_csv('./landmark_geneinfo.csv')\n",
+    "gene_info_LINCS = LINCS['gene_symbol'].astype(str)\n",
+    "len(gene_info_LINCS)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "a41c6d65",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "0      GNPDA1\n",
+       "1        CDH3\n",
+       "2       HDAC6\n",
+       "3       PARP2\n",
+       "4      MAMLD1\n",
+       "        ...  \n",
+       "973    CDC25B\n",
+       "974     OXSR1\n",
+       "975       MVP\n",
+       "976     CDC42\n",
+       "977     DMTF1\n",
+       "Name: gene_symbol, Length: 978, dtype: object"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "gene_info_LINCS"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "0166c778",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "19034\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "0        SLC25A37\n",
+       "1           CRYGC\n",
+       "2         CCDC198\n",
+       "3        TBC1D10A\n",
+       "4           SARM1\n",
+       "           ...   \n",
+       "19029        INMT\n",
+       "19030    RAD51AP1\n",
+       "19031       ACSF3\n",
+       "19032      ZNF334\n",
+       "19033     GOLGA8H\n",
+       "Name: gene_symbol, Length: 19034, dtype: object"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "gene_info_Tahoe = pd.read_csv('./tahoe_gene_names.csv')['gene_symbol'].astype(str)\n",
+    "print(len(gene_info_Tahoe))\n",
+    "gene_info_Tahoe"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "981be0b2",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Number of common genes: 965\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 找到gene_info_LINCS和gene_info_Tahoe的交集\n",
+    "intersection_genes = set(gene_info_LINCS).intersection(set(gene_info_Tahoe))\n",
+    "print(f\"Number of common genes: {len(intersection_genes)}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "68001a70",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 找到gene_info_LINCS中的intersection_genes的ID，去掉其他的基因，得到纯种 LINCS965\n",
+    "\n",
+    "def get_pure_lincs965(gene_info_LINCS, intersection_genes):\n",
+    "    \"\"\"\n",
+    "    从gene_info_LINCS中筛选出属于intersection_genes的基因，得到纯种LINCS965\n",
+    "    \n",
+    "    参数:\n",
+    "        gene_info_LINCS: pandas Series，包含LINCS基因信息，索引为ID，值为基因符号\n",
+    "        intersection_genes: 可迭代对象，包含需要保留的基因符号集合\n",
+    "        \n",
+    "    返回:\n",
+    "        pandas Series: 筛选后的纯种LINCS965基因集\n",
+    "    \"\"\"\n",
+    "    # 将intersection_genes转换为集合以提高查找效率\n",
+    "    intersection_set = set(intersection_genes)\n",
+    "    \n",
+    "    # 筛选出属于intersection_genes的基因\n",
+    "    pure_lincs965 = gene_info_LINCS[gene_info_LINCS.isin(intersection_set)]\n",
+    "    \n",
+    "    return pure_lincs965\n",
+    "\n",
+    "pure_lincs965 = get_pure_lincs965(gene_info_LINCS, intersection_genes)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "b66f29a0",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "965\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "0      GNPDA1\n",
+       "1        CDH3\n",
+       "2       HDAC6\n",
+       "3       PARP2\n",
+       "4      MAMLD1\n",
+       "        ...  \n",
+       "973    CDC25B\n",
+       "974     OXSR1\n",
+       "975       MVP\n",
+       "976     CDC42\n",
+       "977     DMTF1\n",
+       "Name: gene_symbol, Length: 965, dtype: object"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "print(len(pure_lincs965))\n",
+    "pure_lincs965"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f9d67f9c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# save the pure LINCS965 data to a csv file\n",
+    "# pure_lincs965.to_csv('./lincs978_to965.csv', index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "9b8ccab9",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['BrC1C(Br)C(Br)C(Br)C(Br)C1Br', 'BrC1C(Br)C(Br)C(Br)C(Br)C1Br',\n",
+       "        'BrC1C(Br)C(Br)C(Br)C(Br)C1Br', ..., 'c1nc(cs1)-c1nc2ccccc2[nH]1',\n",
+       "        'c1nc(cs1)-c1nc2ccccc2[nH]1', 'c1nc(cs1)-c1nc2ccccc2[nH]1'],\n",
+       "       dtype='<U454'),\n",
+       " 'cid': array(['A549', 'ASC', 'MCF7', ..., 'THP1', 'VCAP', 'YAPC'], dtype='<U8'),\n",
+       " 'sig': array(['A549_0', 'ASC_0', 'MCF7_0', ..., 'THP1_8315', 'VCAP_8315',\n",
+       "        'YAPC_8315'], dtype='<U13'),\n",
+       " 'x1': array([[ 9.533199 ,  4.79235  ,  9.7237   , ..., 11.9874   ,  9.3001   ,\n",
+       "          8.39445  ],\n",
+       "        [ 8.509138 ,  4.8156996,  7.3642497, ..., 11.6545   , 10.211599 ,\n",
+       "          8.973351 ],\n",
+       "        [ 9.043461 ,  7.6456842,  7.459799 , ...,  6.712334 , 11.85135  ,\n",
+       "          6.3363895],\n",
+       "        ...,\n",
+       "        [ 8.86875  ,  5.9342003,  8.8046   , ...,  7.338975 , 11.65115  ,\n",
+       "          8.347799 ],\n",
+       "        [ 9.482506 ,  5.9174104,  8.216723 , ...,  6.722115 , 10.991553 ,\n",
+       "          7.3406167],\n",
+       "        [ 9.725147 ,  7.049716 ,  9.0830555, ...,  9.198456 , 11.310257 ,\n",
+       "          7.920928 ]], dtype=float32),\n",
+       " 'x2': array([[ 9.4455   ,  6.92005  , 10.2842   , ..., 12.2337   ,  9.618799 ,\n",
+       "          7.8896003],\n",
+       "        [ 8.291338 ,  5.3398876,  6.832475 , ..., 11.05195  , 11.712725 ,\n",
+       "          9.1215   ],\n",
+       "        [ 9.176098 ,  7.160512 ,  7.151138 , ...,  6.86923  , 11.213164 ,\n",
+       "          6.2830935],\n",
+       "        ...,\n",
+       "        [ 8.7068   ,  5.7699003,  8.881599 , ...,  7.468175 , 11.75805  ,\n",
+       "          8.6483   ],\n",
+       "        [ 9.617616 ,  5.709352 ,  7.277331 , ...,  7.262997 , 10.814158 ,\n",
+       "          7.3617706],\n",
+       "        [ 9.710126 ,  6.609707 ,  9.075139 , ...,  9.596974 , 10.573312 ,\n",
+       "          7.4035025]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 19,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "LINCS_data_example"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "4ea59ef7",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Indices to keep: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 357, 358, 359, 360, 361, 362, 363, 364, 365, 366, 367, 368, 369, 370, 371, 372, 373, 374, 375, 376, 377, 378, 379, 380, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415, 416, 417, 418, 419, 420, 421, 422, 423, 424, 425, 426, 427, 428, 429, 430, 431, 432, 433, 434, 435, 436, 437, 438, 439, 440, 441, 442, 443, 444, 445, 446, 447, 448, 449, 450, 451, 452, 453, 454, 455, 456, 457, 458, 459, 460, 461, 462, 463, 464, 465, 466, 467, 468, 469, 470, 471, 472, 473, 474, 475, 476, 477, 478, 479, 480, 481, 482, 483, 484, 485, 486, 487, 488, 489, 490, 491, 492, 493, 494, 495, 496, 497, 498, 499, 500, 501, 502, 503, 504, 505, 506, 507, 508, 509, 510, 511, 512, 513, 514, 515, 516, 517, 518, 519, 520, 521, 522, 523, 524, 525, 526, 527, 528, 529, 530, 531, 532, 533, 534, 535, 536, 537, 538, 539, 540, 541, 542, 543, 544, 545, 546, 547, 548, 549, 550, 551, 552, 553, 554, 555, 556, 557, 558, 559, 560, 561, 562, 563, 564, 565, 566, 567, 568, 569, 570, 571, 572, 573, 574, 575, 576, 577, 578, 579, 580, 581, 582, 583, 584, 585, 586, 587, 588, 589, 590, 591, 592, 593, 594, 595, 596, 597, 598, 599, 600, 601, 602, 603, 604, 605, 606, 607, 608, 609, 610, 611, 612, 613, 614, 615, 616, 617, 618, 619, 620, 621, 622, 623, 624, 625, 626, 627, 628, 629, 630, 631, 632, 633, 634, 635, 636, 637, 638, 639, 640, 641, 642, 643, 644, 645, 646, 647, 648, 649, 650, 651, 652, 653, 654, 655, 656, 657, 658, 659, 660, 661, 662, 663, 664, 665, 666, 667, 668, 669, 670, 671, 672, 673, 674, 675, 676, 677, 678, 679, 680, 681, 682, 683, 684, 685, 686, 687, 688, 689, 690, 691, 692, 693, 694, 695, 696, 698, 699, 700, 701, 702, 703, 704, 705, 706, 707, 708, 709, 710, 711, 712, 713, 714, 715, 716, 717, 718, 719, 720, 721, 722, 723, 724, 725, 726, 727, 728, 729, 730, 731, 732, 733, 734, 735, 736, 737, 738, 739, 740, 741, 742, 743, 744, 745, 746, 747, 748, 749, 750, 751, 752, 753, 754, 755, 756, 757, 758, 759, 760, 761, 763, 764, 765, 766, 767, 768, 769, 770, 772, 773, 774, 775, 776, 777, 778, 779, 780, 781, 782, 783, 784, 785, 786, 787, 788, 789, 790, 791, 792, 793, 794, 795, 796, 797, 799, 800, 801, 802, 803, 804, 805, 806, 808, 809, 810, 811, 812, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 825, 826, 828, 829, 830, 831, 832, 833, 834, 835, 836, 837, 838, 839, 840, 841, 842, 843, 844, 845, 846, 848, 849, 850, 851, 852, 853, 854, 855, 856, 857, 858, 859, 860, 861, 862, 863, 864, 865, 866, 867, 868, 869, 870, 871, 872, 873, 874, 875, 876, 877, 878, 879, 880, 881, 882, 883, 884, 885, 886, 887, 888, 889, 890, 891, 892, 893, 894, 895, 896, 897, 898, 899, 900, 901, 902, 903, 904, 905, 906, 907, 908, 909, 910, 911, 912, 913, 914, 915, 916, 917, 919, 920, 921, 922, 923, 924, 925, 926, 927, 928, 929, 930, 931, 932, 933, 934, 935, 936, 937, 938, 939, 940, 941, 942, 943, 945, 947, 948, 949, 950, 951, 952, 953, 954, 955, 956, 957, 958, 959, 960, 961, 962, 963, 964, 965, 966, 967, 968, 969, 970, 971, 972, 973, 974, 975, 976, 977]\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['BrC1C(Br)C(Br)C(Br)C(Br)C1Br', 'BrC1C(Br)C(Br)C(Br)C(Br)C1Br',\n",
+       "        'BrC1C(Br)C(Br)C(Br)C(Br)C1Br', ..., 'c1nc(cs1)-c1nc2ccccc2[nH]1',\n",
+       "        'c1nc(cs1)-c1nc2ccccc2[nH]1', 'c1nc(cs1)-c1nc2ccccc2[nH]1'],\n",
+       "       dtype='<U454'),\n",
+       " 'cid': array(['A549', 'ASC', 'MCF7', ..., 'THP1', 'VCAP', 'YAPC'], dtype='<U8'),\n",
+       " 'sig': array(['A549_0', 'ASC_0', 'MCF7_0', ..., 'THP1_8315', 'VCAP_8315',\n",
+       "        'YAPC_8315'], dtype='<U13'),\n",
+       " 'x1': array([[ 9.533199 ,  4.79235  ,  9.7237   , ..., 11.9874   ,  9.3001   ,\n",
+       "          8.39445  ],\n",
+       "        [ 8.509138 ,  4.8156996,  7.3642497, ..., 11.6545   , 10.211599 ,\n",
+       "          8.973351 ],\n",
+       "        [ 9.043461 ,  7.6456842,  7.459799 , ...,  6.712334 , 11.85135  ,\n",
+       "          6.3363895],\n",
+       "        ...,\n",
+       "        [ 8.86875  ,  5.9342003,  8.8046   , ...,  7.338975 , 11.65115  ,\n",
+       "          8.347799 ],\n",
+       "        [ 9.482506 ,  5.9174104,  8.216723 , ...,  6.722115 , 10.991553 ,\n",
+       "          7.3406167],\n",
+       "        [ 9.725147 ,  7.049716 ,  9.0830555, ...,  9.198456 , 11.310257 ,\n",
+       "          7.920928 ]], dtype=float32),\n",
+       " 'x2': array([[ 9.4455   ,  6.92005  , 10.2842   , ..., 12.2337   ,  9.618799 ,\n",
+       "          7.8896003],\n",
+       "        [ 8.291338 ,  5.3398876,  6.832475 , ..., 11.05195  , 11.712725 ,\n",
+       "          9.1215   ],\n",
+       "        [ 9.176098 ,  7.160512 ,  7.151138 , ...,  6.86923  , 11.213164 ,\n",
+       "          6.2830935],\n",
+       "        ...,\n",
+       "        [ 8.7068   ,  5.7699003,  8.881599 , ...,  7.468175 , 11.75805  ,\n",
+       "          8.6483   ],\n",
+       "        [ 9.617616 ,  5.709352 ,  7.277331 , ...,  7.262997 , 10.814158 ,\n",
+       "          7.3617706],\n",
+       "        [ 9.710126 ,  6.609707 ,  9.075139 , ...,  9.596974 , 10.573312 ,\n",
+       "          7.4035025]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 22,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# 根据纯种LINCS965的基因信息，在LINCS_data_example中的x1和x2中筛选出对应的基因数据，都是978维度的，减去对应位置不存在的基因数据\n",
+    "# 'x1': array([[ 9.533199 ,  4.79235  ,  9.7237   , ..., 11.9874   ,  9.3001   , 8.39445  ],\n",
+    "def filter_lincs_dictionary(lincs_data, pure_lincs965):\n",
+    "    \"\"\"\n",
+    "    筛选LINCS数据字典中的x1和x2，仅保留pure_lincs965对应的基因位置数据\n",
+    "    \n",
+    "    参数:\n",
+    "        lincs_data: 包含LINCS数据的字典，至少包含'x1'和'x2'键\n",
+    "        pure_lincs965: pandas Series，纯种LINCS965基因集，包含需要保留的基因索引\n",
+    "        \n",
+    "    返回:\n",
+    "        字典: 处理后的LINCS数据，x1和x2已筛选，其他键值对保持不变\n",
+    "    \"\"\"\n",
+    "    # 获取需要保留的基因索引并排序\n",
+    "    keep_indices = sorted(pure_lincs965.index.tolist())\n",
+    "    # print(f\"Indices to keep: {keep_indices}\")\n",
+    "    \n",
+    "    # 创建一个新字典，复制不需要修改的键值对\n",
+    "    filtered_data = {\n",
+    "        key: value for key, value in lincs_data.items() \n",
+    "        if key not in ['x1', 'x2']\n",
+    "    }\n",
+    "    \n",
+    "    # 筛选x1和x2数据\n",
+    "    filtered_data['x1'] = lincs_data['x1'][:, keep_indices]\n",
+    "    filtered_data['x2'] = lincs_data['x2'][:, keep_indices]\n",
+    "    \n",
+    "    return filtered_data\n",
+    "\n",
+    "filtered_lincs_data = filter_lincs_dictionary(LINCS_data_example, pure_lincs965)\n",
+    "# 现在filtered_lincs_data中只包含纯种LINCS965的基因数据\n",
+    "# 你可以根据需要进一步处理filtered_lincs_data，例如保存到文件或进行分析\n",
+    "filtered_lincs_data\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "9f4238d1",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(78569, 965)"
+      ]
+     },
+     "execution_count": 23,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "filtered_lincs_data['x1'].shape  # 查看筛选后的x1数据形状"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "db14e8d8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# save filtered_lincs_data.h5\n",
+    "\n",
+    "smiles_data = filtered_lincs_data['canonical_smiles'].astype('|S')  # 转换为字节字符串\n",
+    "\n",
+    "with h5py.File('lincs_data965.h5', 'w') as f:\n",
+    "    f.create_dataset('canonical_smiles', data=smiles_data)\n",
+    "    # 保存 control 数据，转换为 float32\n",
+    "    # f.create_dataset('cid', data=filtered_lincs_data['cid'])\n",
+    "    # # 保存 target 数据，转换为 float32\n",
+    "    # f.create_dataset('sig', data=filtered_lincs_data['sig'])\n",
+    "    # 保存 target 数据，转换为 float32\n",
+    "    f.create_dataset('control', data=filtered_lincs_data['x1'].astype(np.float32))\n",
+    "    # 保存 target 数据，转换为 float32\n",
+    "    f.create_dataset('target', data=filtered_lincs_data['x2'].astype(np.float32))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 31,
+   "id": "74a24715",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(78569, 965)"
+      ]
+     },
+     "execution_count": 31,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# load filtered_lincs_data\n",
+    "filtered_lincs_data_loaded = load_from_HDF('./lincs_data965.h5')\n",
+    "filtered_lincs_data_loaded['control'].shape  # 查看加载后的x1数据形"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 32,
+   "id": "a2b8c37e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['BrC1C(Br)C(Br)C(Br)C(Br)C1Br', 'BrC1C(Br)C(Br)C(Br)C(Br)C1Br',\n",
+       "        'BrC1C(Br)C(Br)C(Br)C(Br)C1Br', ..., 'c1nc(cs1)-c1nc2ccccc2[nH]1',\n",
+       "        'c1nc(cs1)-c1nc2ccccc2[nH]1', 'c1nc(cs1)-c1nc2ccccc2[nH]1'],\n",
+       "       dtype='<U454'),\n",
+       " 'control': array([[ 9.533199 ,  4.79235  ,  9.7237   , ..., 11.9874   ,  9.3001   ,\n",
+       "          8.39445  ],\n",
+       "        [ 8.509138 ,  4.8156996,  7.3642497, ..., 11.6545   , 10.211599 ,\n",
+       "          8.973351 ],\n",
+       "        [ 9.043461 ,  7.6456842,  7.459799 , ...,  6.712334 , 11.85135  ,\n",
+       "          6.3363895],\n",
+       "        ...,\n",
+       "        [ 8.86875  ,  5.9342003,  8.8046   , ...,  7.338975 , 11.65115  ,\n",
+       "          8.347799 ],\n",
+       "        [ 9.482506 ,  5.9174104,  8.216723 , ...,  6.722115 , 10.991553 ,\n",
+       "          7.3406167],\n",
+       "        [ 9.725147 ,  7.049716 ,  9.0830555, ...,  9.198456 , 11.310257 ,\n",
+       "          7.920928 ]], dtype=float32),\n",
+       " 'target': array([[ 9.4455   ,  6.92005  , 10.2842   , ..., 12.2337   ,  9.618799 ,\n",
+       "          7.8896003],\n",
+       "        [ 8.291338 ,  5.3398876,  6.832475 , ..., 11.05195  , 11.712725 ,\n",
+       "          9.1215   ],\n",
+       "        [ 9.176098 ,  7.160512 ,  7.151138 , ...,  6.86923  , 11.213164 ,\n",
+       "          6.2830935],\n",
+       "        ...,\n",
+       "        [ 8.7068   ,  5.7699003,  8.881599 , ...,  7.468175 , 11.75805  ,\n",
+       "          8.6483   ],\n",
+       "        [ 9.617616 ,  5.709352 ,  7.277331 , ...,  7.262997 , 10.814158 ,\n",
+       "          7.3617706],\n",
+       "        [ 9.710126 ,  6.609707 ,  9.075139 , ...,  9.596974 , 10.573312 ,\n",
+       "          7.4035025]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 32,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "filtered_lincs_data_loaded"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

LINCS2020/landmark_geneinfo.csv

@@ -0,0 +1,979 @@
+gene_id,gene_symbol,ensembl_id,gene_title,gene_type,src,feature_space
+10007,GNPDA1,ENSG00000113552,glucosamine-6-phosphate deaminase 1,protein-coding,NCBI,landmark
+1001,CDH3,ENSG00000062038,cadherin 3,protein-coding,NCBI,landmark
+10013,HDAC6,ENSG00000094631,histone deacetylase 6,protein-coding,NCBI,landmark
+10038,PARP2,ENSG00000129484,poly(ADP-ribose) polymerase 2,protein-coding,NCBI,landmark
+10046,MAMLD1,ENSG00000013619,mastermind like domain containing 1,protein-coding,NCBI,landmark
+10049,DNAJB6,ENSG00000105993,DnaJ heat shock protein family (Hsp40) member B6,protein-coding,NCBI,landmark
+10051,SMC4,ENSG00000113810,structural maintenance of chromosomes 4,protein-coding,NCBI,landmark
+10057,ABCC5,ENSG00000114770,ATP binding cassette subfamily C member 5,protein-coding,NCBI,landmark
+10058,ABCB6,ENSG00000115657,ATP binding cassette subfamily B member 6 (Langereis blood group),protein-coding,NCBI,landmark
+10059,DNM1L,ENSG00000087470,dynamin 1 like,protein-coding,NCBI,landmark
+10099,TSPAN3,ENSG00000140391,tetraspanin 3,protein-coding,NCBI,landmark
+10112,KIF20A,ENSG00000112984,kinesin family member 20A,protein-coding,NCBI,landmark
+10123,ARL4C,ENSG00000188042,ADP ribosylation factor like GTPase 4C,protein-coding,NCBI,landmark
+10131,TRAP1,ENSG00000126602,TNF receptor associated protein 1,protein-coding,NCBI,landmark
+10146,G3BP1,ENSG00000145907,G3BP stress granule assembly factor 1,protein-coding,NCBI,landmark
+10150,MBNL2,ENSG00000139793,muscleblind like splicing regulator 2,protein-coding,NCBI,landmark
+10153,CEBPZ,ENSG00000115816,CCAAT enhancer binding protein zeta,protein-coding,NCBI,landmark
+10165,SLC25A13,ENSG00000004864,solute carrier family 25 member 13,protein-coding,NCBI,landmark
+1017,CDK2,ENSG00000123374,cyclin dependent kinase 2,protein-coding,NCBI,landmark
+10174,SORBS3,ENSG00000120896,sorbin and SH3 domain containing 3,protein-coding,NCBI,landmark
+10180,RBM6,ENSG00000004534,RNA binding motif protein 6,protein-coding,NCBI,landmark
+1019,CDK4,ENSG00000135446,cyclin dependent kinase 4,protein-coding,NCBI,landmark
+10190,TXNDC9,ENSG00000115514,thioredoxin domain containing 9,protein-coding,NCBI,landmark
+102,ADAM10,ENSG00000137845,ADAM metallopeptidase domain 10,protein-coding,NCBI,landmark
+10206,TRIM13,ENSG00000204977,tripartite motif containing 13,protein-coding,NCBI,landmark
+1021,CDK6,ENSG00000105810,cyclin dependent kinase 6,protein-coding,NCBI,landmark
+1022,CDK7,ENSG00000134058,cyclin dependent kinase 7,protein-coding,NCBI,landmark
+10221,TRIB1,ENSG00000173334,tribbles pseudokinase 1,protein-coding,NCBI,landmark
+10227,MFSD10,ENSG00000109736,major facilitator superfamily domain containing 10,protein-coding,NCBI,landmark
+10237,SLC35B1,ENSG00000121073,solute carrier family 35 member B1,protein-coding,NCBI,landmark
+10245,TIMM17B,ENSG00000126768,translocase of inner mitochondrial membrane 17B,protein-coding,NCBI,landmark
+1026,CDKN1A,ENSG00000124762,cyclin dependent kinase inhibitor 1A,protein-coding,NCBI,landmark
+1027,CDKN1B,ENSG00000111276,cyclin dependent kinase inhibitor 1B,protein-coding,NCBI,landmark
+10270,AKAP8,ENSG00000105127,A-kinase anchoring protein 8,protein-coding,NCBI,landmark
+10273,STUB1,ENSG00000103266,STIP1 homology and U-box containing protein 1,protein-coding,NCBI,landmark
+10276,NET1,ENSG00000173848,neuroepithelial cell transforming 1,protein-coding,NCBI,landmark
+10285,SMNDC1,ENSG00000119953,survival motor neuron domain containing 1,protein-coding,NCBI,landmark
+1029,CDKN2A,ENSG00000147889,cyclin dependent kinase inhibitor 2A,protein-coding,NCBI,landmark
+10298,PAK4,ENSG00000130669,p21 (RAC1) activated kinase 4,protein-coding,NCBI,landmark
+10318,TNIP1,ENSG00000145901,TNFAIP3 interacting protein 1,protein-coding,NCBI,landmark
+10320,IKZF1,ENSG00000185811,IKAROS family zinc finger 1,protein-coding,NCBI,landmark
+10329,RXYLT1,ENSG00000118600,ribitol xylosyltransferase 1,protein-coding,NCBI,landmark
+10362,HMG20B,ENSG00000064961,high mobility group 20B,protein-coding,NCBI,landmark
+10398,MYL9,ENSG00000101335,myosin light chain 9,protein-coding,NCBI,landmark
+10434,LYPLA1,ENSG00000120992,lysophospholipase 1,protein-coding,NCBI,landmark
+10450,PPIE,ENSG00000084072,peptidylprolyl isomerase E,protein-coding,NCBI,landmark
+10451,VAV3,ENSG00000134215,vav guanine nucleotide exchange factor 3,protein-coding,NCBI,landmark
+10489,LRRC41,ENSG00000132128,leucine rich repeat containing 41,protein-coding,NCBI,landmark
+10491,CRTAP,ENSG00000170275,cartilage associated protein,protein-coding,NCBI,landmark
+10493,VAT1,ENSG00000108828,vesicle amine transport 1,protein-coding,NCBI,landmark
+10494,STK25,ENSG00000115694,serine/threonine kinase 25,protein-coding,NCBI,landmark
+1050,CEBPA,ENSG00000245848,CCAAT enhancer binding protein alpha,protein-coding,NCBI,landmark
+10513,APPBP2,ENSG00000062725,amyloid beta precursor protein binding protein 2,protein-coding,NCBI,landmark
+1052,CEBPD,ENSG00000221869,CCAAT enhancer binding protein delta,protein-coding,NCBI,landmark
+10523,CHERP,ENSG00000085872,calcium homeostasis endoplasmic reticulum protein,protein-coding,NCBI,landmark
+10525,HYOU1,ENSG00000149428,hypoxia up-regulated 1,protein-coding,NCBI,landmark
+10557,RPP38,ENSG00000152464,ribonuclease P/MRP subunit p38,protein-coding,NCBI,landmark
+10559,SLC35A1,ENSG00000164414,solute carrier family 35 member A1,protein-coding,NCBI,landmark
+10589,DRAP1,ENSG00000175550,DR1 associated protein 1,protein-coding,NCBI,landmark
+10606,PAICS,ENSG00000128050,phosphoribosylaminoimidazole carboxylase and phosphoribosylaminoimidazolesuccinocarboxamide synthase,protein-coding,NCBI,landmark
+10610,ST6GALNAC2,ENSG00000070731,"ST6 N-acetylgalactosaminide alpha-2,6-sialyltransferase 2",protein-coding,NCBI,landmark
+10617,STAMBP,ENSG00000124356,STAM binding protein,protein-coding,NCBI,landmark
+1062,CENPE,ENSG00000138778,centromere protein E,protein-coding,NCBI,landmark
+10641,NPRL2,ENSG00000114388,"NPR2 like, GATOR1 complex subunit",protein-coding,NCBI,landmark
+10644,IGF2BP2,ENSG00000073792,insulin like growth factor 2 mRNA binding protein 2,protein-coding,NCBI,landmark
+10652,YKT6,ENSG00000106636,YKT6 v-SNARE homolog,protein-coding,NCBI,landmark
+10668,CGRRF1,ENSG00000100532,cell growth regulator with ring finger domain 1,protein-coding,NCBI,landmark
+10670,RRAGA,ENSG00000155876,Ras related GTP binding A,protein-coding,NCBI,landmark
+10681,GNB5,ENSG00000069966,G protein subunit beta 5,protein-coding,NCBI,landmark
+10682,EBP,ENSG00000147155,EBP cholestenol delta-isomerase,protein-coding,NCBI,landmark
+10695,CNPY3,ENSG00000137161,canopy FGF signaling regulator 3,protein-coding,NCBI,landmark
+1070,CETN3,ENSG00000153140,centrin 3,protein-coding,NCBI,landmark
+10730,YME1L1,ENSG00000136758,YME1 like 1 ATPase,protein-coding,NCBI,landmark
+10732,TCFL5,ENSG00000101190,transcription factor like 5,protein-coding,NCBI,landmark
+10765,KDM5B,ENSG00000117139,lysine demethylase 5B,protein-coding,NCBI,landmark
+10775,POP4,ENSG00000105171,"POP4 homolog, ribonuclease P/MRP subunit",protein-coding,NCBI,landmark
+10776,ARPP19,ENSG00000128989,cAMP regulated phosphoprotein 19,protein-coding,NCBI,landmark
+10782,ZNF274,ENSG00000171606,zinc finger protein 274,protein-coding,NCBI,landmark
+10797,MTHFD2,ENSG00000065911,"methylenetetrahydrofolate dehydrogenase (NADP+ dependent) 2, methenyltetrahydrofolate cyclohydrolase",protein-coding,NCBI,landmark
+10810,WASF3,ENSG00000132970,WASP family member 3,protein-coding,NCBI,landmark
+10813,UTP14A,ENSG00000156697,UTP14A small subunit processome component,protein-coding,NCBI,landmark
+10818,FRS2,ENSG00000166225,fibroblast growth factor receptor substrate 2,protein-coding,NCBI,landmark
+10845,CLPX,ENSG00000166855,caseinolytic mitochondrial matrix peptidase chaperone subunit,protein-coding,NCBI,landmark
+10857,PGRMC1,ENSG00000101856,progesterone receptor membrane component 1,protein-coding,NCBI,landmark
+10892,MALT1,ENSG00000172175,MALT1 paracaspase,protein-coding,NCBI,landmark
+10898,CPSF4,ENSG00000160917,cleavage and polyadenylation specific factor 4,protein-coding,NCBI,landmark
+10904,BLCAP,ENSG00000166619,BLCAP apoptosis inducing factor,protein-coding,NCBI,landmark
+10915,TCERG1,ENSG00000113649,transcription elongation regulator 1,protein-coding,NCBI,landmark
+10921,RNPS1,ENSG00000205937,RNA binding protein with serine rich domain 1,protein-coding,NCBI,landmark
+10953,TOMM34,ENSG00000025772,translocase of outer mitochondrial membrane 34,protein-coding,NCBI,landmark
+10954,PDIA5,ENSG00000065485,protein disulfide isomerase family A member 5,protein-coding,NCBI,landmark
+10962,MLLT11,ENSG00000213190,MLLT11 transcription factor 7 cofactor,protein-coding,NCBI,landmark
+10969,EBNA1BP2,ENSG00000117395,EBNA1 binding protein 2,protein-coding,NCBI,landmark
+10972,TMED10,ENSG00000170348,transmembrane p24 trafficking protein 10,protein-coding,NCBI,landmark
+10973,ASCC3,ENSG00000112249,activating signal cointegrator 1 complex subunit 3,protein-coding,NCBI,landmark
+11000,SLC27A3,ENSG00000143554,solute carrier family 27 member 3,protein-coding,NCBI,landmark
+11004,KIF2C,ENSG00000142945,kinesin family member 2C,protein-coding,NCBI,landmark
+11007,CCDC85B,ENSG00000175602,coiled-coil domain containing 85B,protein-coding,NCBI,landmark
+11011,TLK2,ENSG00000146872,tousled like kinase 2,protein-coding,NCBI,landmark
+11014,KDELR2,ENSG00000136240,KDEL endoplasmic reticulum protein retention receptor 2,protein-coding,NCBI,landmark
+11031,RAB31,ENSG00000168461,"RAB31, member RAS oncogene family",protein-coding,NCBI,landmark
+11041,B4GAT1,ENSG00000174684,"beta-1,4-glucuronyltransferase 1",protein-coding,NCBI,landmark
+11044,TENT4A,ENSG00000112941,terminal nucleotidyltransferase 4A,protein-coding,NCBI,landmark
+11065,UBE2C,ENSG00000175063,ubiquitin conjugating enzyme E2 C,protein-coding,NCBI,landmark
+11072,DUSP14,ENSG00000276023,dual specificity phosphatase 14,protein-coding,NCBI,landmark
+11073,TOPBP1,ENSG00000163781,DNA topoisomerase II binding protein 1,protein-coding,NCBI,landmark
+11098,PRSS23,ENSG00000150687,serine protease 23,protein-coding,NCBI,landmark
+1111,CHEK1,ENSG00000149554,checkpoint kinase 1,protein-coding,NCBI,landmark
+11137,PWP1,ENSG00000136045,"PWP1 homolog, endonuclein",protein-coding,NCBI,landmark
+11142,PKIG,ENSG00000168734,cAMP-dependent protein kinase inhibitor gamma,protein-coding,NCBI,landmark
+11151,CORO1A,ENSG00000102879,coronin 1A,protein-coding,NCBI,landmark
+11157,LSM6,ENSG00000164167,"LSM6 homolog, U6 small nuclear RNA and mRNA degradation associated",protein-coding,NCBI,landmark
+11168,PSIP1,ENSG00000164985,PC4 and SFRS1 interacting protein 1,protein-coding,NCBI,landmark
+11182,SLC2A6,ENSG00000160326,solute carrier family 2 member 6,protein-coding,NCBI,landmark
+11188,NISCH,ENSG00000010322,nischarin,protein-coding,NCBI,landmark
+11200,CHEK2,ENSG00000183765,checkpoint kinase 2,protein-coding,NCBI,landmark
+1123,CHN1,ENSG00000128656,chimerin 1,protein-coding,NCBI,landmark
+11230,PRAF2,ENSG00000243279,PRA1 domain family member 2,protein-coding,NCBI,landmark
+11232,POLG2,ENSG00000256525,"DNA polymerase gamma 2, accessory subunit",protein-coding,NCBI,landmark
+11261,CHP1,ENSG00000187446,calcineurin like EF-hand protein 1,protein-coding,NCBI,landmark
+11284,PNKP,ENSG00000039650,polynucleotide kinase 3'-phosphatase,protein-coding,NCBI,landmark
+11319,ECD,ENSG00000122882,ecdysoneless cell cycle regulator,protein-coding,NCBI,landmark
+11325,DDX42,ENSG00000198231,DEAD-box helicase 42,protein-coding,NCBI,landmark
+11344,TWF2,ENSG00000247596,twinfilin actin binding protein 2,protein-coding,NCBI,landmark
+1153,CIRBP,ENSG00000099622,cold inducible RNA binding protein,protein-coding,NCBI,landmark
+116832,RPL39L,ENSG00000163923,ribosomal protein L39 like,protein-coding,NCBI,landmark
+1212,CLTB,ENSG00000175416,clathrin light chain B,protein-coding,NCBI,landmark
+1213,CLTC,ENSG00000141367,clathrin heavy chain,protein-coding,NCBI,landmark
+124583,CANT1,ENSG00000171302,calcium activated nucleotidase 1,protein-coding,NCBI,landmark
+1277,COL1A1,ENSG00000108821,collagen type I alpha 1 chain,protein-coding,NCBI,landmark
+128,ADH5,ENSG00000197894,"alcohol dehydrogenase 5 (class III), chi polypeptide",protein-coding,NCBI,landmark
+1282,COL4A1,ENSG00000187498,collagen type IV alpha 1 chain,protein-coding,NCBI,landmark
+1385,CREB1,ENSG00000118260,cAMP responsive element binding protein 1,protein-coding,NCBI,landmark
+1398,CRK,ENSG00000167193,"CRK proto-oncogene, adaptor protein",protein-coding,NCBI,landmark
+1399,CRKL,ENSG00000099942,"CRK like proto-oncogene, adaptor protein",protein-coding,NCBI,landmark
+142,PARP1,ENSG00000143799,poly(ADP-ribose) polymerase 1,protein-coding,NCBI,landmark
+1429,CRYZ,ENSG00000116791,crystallin zeta,protein-coding,NCBI,landmark
+1445,CSK,ENSG00000103653,C-terminal Src kinase,protein-coding,NCBI,landmark
+1452,CSNK1A1,ENSG00000113712,casein kinase 1 alpha 1,protein-coding,NCBI,landmark
+1454,CSNK1E,ENSG00000213923,casein kinase 1 epsilon,protein-coding,NCBI,landmark
+1459,CSNK2A2,ENSG00000070770,casein kinase 2 alpha 2,protein-coding,NCBI,landmark
+1465,CSRP1,ENSG00000159176,cysteine and glycine rich protein 1,protein-coding,NCBI,landmark
+147179,WIPF2,ENSG00000171475,WAS/WASL interacting protein family member 2,protein-coding,NCBI,landmark
+148022,TICAM1,ENSG00000127666,toll like receptor adaptor molecule 1,protein-coding,NCBI,landmark
+1500,CTNND1,ENSG00000198561,catenin delta 1,protein-coding,NCBI,landmark
+1509,CTSD,ENSG00000117984,cathepsin D,protein-coding,NCBI,landmark
+1514,CTSL,ENSG00000135047,cathepsin L,protein-coding,NCBI,landmark
+1534,CYB561,ENSG00000008283,cytochrome b561,protein-coding,NCBI,landmark
+154,ADRB2,ENSG00000169252,adrenoceptor beta 2,protein-coding,NCBI,landmark
+16,AARS,ENSG00000090861,alanyl-tRNA synthetase,protein-coding,NCBI,landmark
+1605,DAG1,ENSG00000173402,dystroglycan 1,protein-coding,NCBI,landmark
+1616,DAXX,ENSG00000204209,death domain associated protein,protein-coding,NCBI,landmark
+1633,DCK,ENSG00000156136,deoxycytidine kinase,protein-coding,NCBI,landmark
+1635,DCTD,ENSG00000129187,dCMP deaminase,protein-coding,NCBI,landmark
+1643,DDB2,ENSG00000134574,damage specific DNA binding protein 2,protein-coding,NCBI,landmark
+1647,GADD45A,ENSG00000116717,growth arrest and DNA damage inducible alpha,protein-coding,NCBI,landmark
+1662,DDX10,ENSG00000178105,DEAD-box helicase 10,protein-coding,NCBI,landmark
+1666,DECR1,ENSG00000104325,"2,4-dienoyl-CoA reductase 1",protein-coding,NCBI,landmark
+1676,DFFA,ENSG00000160049,DNA fragmentation factor subunit alpha,protein-coding,NCBI,landmark
+1677,DFFB,ENSG00000169598,DNA fragmentation factor subunit beta,protein-coding,NCBI,landmark
+1738,DLD,ENSG00000091140,dihydrolipoamide dehydrogenase,protein-coding,NCBI,landmark
+1759,DNM1,ENSG00000106976,dynamin 1,protein-coding,NCBI,landmark
+178,AGL,ENSG00000162688,"amylo-alpha-1, 6-glucosidase, 4-alpha-glucanotransferase",protein-coding,NCBI,landmark
+1786,DNMT1,ENSG00000130816,DNA methyltransferase 1,protein-coding,NCBI,landmark
+1788,DNMT3A,ENSG00000119772,DNA methyltransferase 3 alpha,protein-coding,NCBI,landmark
+1802,DPH2,ENSG00000132768,diphthamide biosynthesis 2,protein-coding,NCBI,landmark
+1829,DSG2,ENSG00000046604,desmoglein 2,protein-coding,NCBI,landmark
+1831,TSC22D3,ENSG00000157514,TSC22 domain family member 3,protein-coding,NCBI,landmark
+1845,DUSP3,ENSG00000108861,dual specificity phosphatase 3,protein-coding,NCBI,landmark
+1846,DUSP4,ENSG00000120875,dual specificity phosphatase 4,protein-coding,NCBI,landmark
+1848,DUSP6,ENSG00000139318,dual specificity phosphatase 6,protein-coding,NCBI,landmark
+1861,TOR1A,ENSG00000136827,torsin family 1 member A,protein-coding,NCBI,landmark
+1870,E2F2,ENSG00000007968,E2F transcription factor 2,protein-coding,NCBI,landmark
+1891,ECH1,ENSG00000104823,enoyl-CoA hydratase 1,protein-coding,NCBI,landmark
+1906,EDN1,ENSG00000078401,endothelin 1,protein-coding,NCBI,landmark
+1950,EGF,ENSG00000138798,epidermal growth factor,protein-coding,NCBI,landmark
+1956,EGFR,ENSG00000146648,epidermal growth factor receptor,protein-coding,NCBI,landmark
+1958,EGR1,ENSG00000120738,early growth response 1,protein-coding,NCBI,landmark
+1978,EIF4EBP1,ENSG00000187840,eukaryotic translation initiation factor 4E binding protein 1,protein-coding,NCBI,landmark
+1981,EIF4G1,ENSG00000114867,eukaryotic translation initiation factor 4 gamma 1,protein-coding,NCBI,landmark
+1983,EIF5,ENSG00000100664,eukaryotic translation initiation factor 5,protein-coding,NCBI,landmark
+1994,ELAVL1,ENSG00000066044,ELAV like RNA binding protein 1,protein-coding,NCBI,landmark
+200081,TXLNA,ENSG00000084652,taxilin alpha,protein-coding,NCBI,landmark
+200734,SPRED2,ENSG00000198369,sprouty related EVH1 domain containing 2,protein-coding,NCBI,landmark
+2017,CTTN,ENSG00000085733,cortactin,protein-coding,NCBI,landmark
+2037,EPB41L2,ENSG00000079819,erythrocyte membrane protein band 4.1 like 2,protein-coding,NCBI,landmark
+2042,EPHA3,ENSG00000044524,EPH receptor A3,protein-coding,NCBI,landmark
+2048,EPHB2,ENSG00000133216,EPH receptor B2,protein-coding,NCBI,landmark
+2058,EPRS,ENSG00000136628,glutamyl-prolyl-tRNA synthetase,protein-coding,NCBI,landmark
+2063,NR2F6,ENSG00000160113,nuclear receptor subfamily 2 group F member 6,protein-coding,NCBI,landmark
+2064,ERBB2,ENSG00000141736,erb-b2 receptor tyrosine kinase 2,protein-coding,NCBI,landmark
+2065,ERBB3,ENSG00000065361,erb-b2 receptor tyrosine kinase 3,protein-coding,NCBI,landmark
+207,AKT1,ENSG00000142208,AKT serine/threonine kinase 1,protein-coding,NCBI,landmark
+2109,ETFB,ENSG00000105379,electron transfer flavoprotein subunit beta,protein-coding,NCBI,landmark
+211,ALAS1,ENSG00000023330,5'-aminolevulinate synthase 1,protein-coding,NCBI,landmark
+2113,ETS1,ENSG00000134954,"ETS proto-oncogene 1, transcription factor",protein-coding,NCBI,landmark
+2115,ETV1,ENSG00000006468,ETS variant 1,protein-coding,NCBI,landmark
+2131,EXT1,ENSG00000182197,exostosin glycosyltransferase 1,protein-coding,NCBI,landmark
+2146,EZH2,ENSG00000106462,enhancer of zeste 2 polycomb repressive complex 2 subunit,protein-coding,NCBI,landmark
+2184,FAH,ENSG00000103876,fumarylacetoacetate hydrolase,protein-coding,NCBI,landmark
+2185,PTK2B,ENSG00000120899,protein tyrosine kinase 2 beta,protein-coding,NCBI,landmark
+2195,FAT1,ENSG00000083857,FAT atypical cadherin 1,protein-coding,NCBI,landmark
+2222,FDFT1,ENSG00000079459,farnesyl-diphosphate farnesyltransferase 1,protein-coding,NCBI,landmark
+226,ALDOA,ENSG00000149925,"aldolase, fructose-bisphosphate A",protein-coding,NCBI,landmark
+2263,FGFR2,ENSG00000066468,fibroblast growth factor receptor 2,protein-coding,NCBI,landmark
+2264,FGFR4,ENSG00000160867,fibroblast growth factor receptor 4,protein-coding,NCBI,landmark
+2274,FHL2,ENSG00000115641,four and a half LIM domains 2,protein-coding,NCBI,landmark
+22794,CASC3,ENSG00000108349,CASC3 exon junction complex subunit,protein-coding,NCBI,landmark
+22796,COG2,ENSG00000135775,component of oligomeric golgi complex 2,protein-coding,NCBI,landmark
+22809,ATF5,ENSG00000169136,activating transcription factor 5,protein-coding,NCBI,landmark
+22823,MTF2,ENSG00000143033,metal response element binding transcription factor 2,protein-coding,NCBI,landmark
+22827,PUF60,ENSG00000179950,poly(U) binding splicing factor 60,protein-coding,NCBI,landmark
+22841,RAB11FIP2,ENSG00000107560,RAB11 family interacting protein 2,protein-coding,NCBI,landmark
+2288,FKBP4,ENSG00000004478,FKBP prolyl isomerase 4,protein-coding,NCBI,landmark
+22883,CLSTN1,ENSG00000171603,calsyntenin 1,protein-coding,NCBI,landmark
+22887,FOXJ3,ENSG00000198815,forkhead box J3,protein-coding,NCBI,landmark
+22889,KHDC4,ENSG00000132680,"KH domain containing 4, pre-mRNA splicing factor",protein-coding,NCBI,landmark
+22905,EPN2,ENSG00000072134,epsin 2,protein-coding,NCBI,landmark
+22908,SACM1L,ENSG00000211456,SAC1 like phosphatidylinositide phosphatase,protein-coding,NCBI,landmark
+22926,ATF6,ENSG00000118217,activating transcription factor 6,protein-coding,NCBI,landmark
+22934,RPIA,ENSG00000153574,ribose 5-phosphate isomerase A,protein-coding,NCBI,landmark
+23,ABCF1,ENSG00000204574,ATP binding cassette subfamily F member 1,protein-coding,NCBI,landmark
+230,ALDOC,ENSG00000109107,"aldolase, fructose-bisphosphate C",protein-coding,NCBI,landmark
+23011,RAB21,ENSG00000080371,"RAB21, member RAS oncogene family",protein-coding,NCBI,landmark
+23013,SPEN,ENSG00000065526,spen family transcriptional repressor,protein-coding,NCBI,landmark
+23014,FBXO21,ENSG00000135108,F-box protein 21,protein-coding,NCBI,landmark
+23029,RBM34,ENSG00000188739,RNA binding motif protein 34,protein-coding,NCBI,landmark
+23038,WDTC1,ENSG00000142784,WD and tetratricopeptide repeats 1,protein-coding,NCBI,landmark
+23039,XPO7,ENSG00000130227,exportin 7,protein-coding,NCBI,landmark
+23061,TBC1D9B,ENSG00000197226,TBC1 domain family member 9B,protein-coding,NCBI,landmark
+23076,RRP1B,ENSG00000160208,ribosomal RNA processing 1B,protein-coding,NCBI,landmark
+23077,MYCBP2,ENSG00000005810,"MYC binding protein 2, E3 ubiquitin protein ligase",protein-coding,NCBI,landmark
+2309,FOXO3,ENSG00000118689,forkhead box O3,protein-coding,NCBI,landmark
+23097,CDK19,ENSG00000155111,cyclin dependent kinase 19,protein-coding,NCBI,landmark
+23131,GPATCH8,ENSG00000186566,G-patch domain containing 8,protein-coding,NCBI,landmark
+23139,MAST2,ENSG00000086015,microtubule associated serine/threonine kinase 2,protein-coding,NCBI,landmark
+23142,DCUN1D4,ENSG00000109184,defective in cullin neddylation 1 domain containing 4,protein-coding,NCBI,landmark
+23149,FCHO1,ENSG00000130475,FCH domain only 1,protein-coding,NCBI,landmark
+23161,SNX13,ENSG00000071189,sorting nexin 13,protein-coding,NCBI,landmark
+23200,ATP11B,ENSG00000058063,ATPase phospholipid transporting 11B (putative),protein-coding,NCBI,landmark
+23210,JMJD6,ENSG00000070495,"jumonji domain containing 6, arginine demethylase and lysine hydroxylase",protein-coding,NCBI,landmark
+23212,RRS1,ENSG00000179041,ribosome biogenesis regulator homolog,protein-coding,NCBI,landmark
+23223,RRP12,ENSG00000052749,ribosomal RNA processing 12 homolog,protein-coding,NCBI,landmark
+23224,SYNE2,ENSG00000054654,spectrin repeat containing nuclear envelope protein 2,protein-coding,NCBI,landmark
+23244,PDS5A,ENSG00000121892,PDS5 cohesin associated factor A,protein-coding,NCBI,landmark
+23271,CAMSAP2,ENSG00000118200,calmodulin regulated spectrin associated protein family member 2,protein-coding,NCBI,landmark
+23300,ATMIN,ENSG00000166454,ATM interactor,protein-coding,NCBI,landmark
+23321,TRIM2,ENSG00000109654,tripartite motif containing 2,protein-coding,NCBI,landmark
+23325,WASHC4,ENSG00000136051,WASH complex subunit 4,protein-coding,NCBI,landmark
+23326,USP22,ENSG00000124422,ubiquitin specific peptidase 22,protein-coding,NCBI,landmark
+23335,WDR7,ENSG00000091157,WD repeat domain 7,protein-coding,NCBI,landmark
+23338,JADE2,ENSG00000043143,jade family PHD finger 2,protein-coding,NCBI,landmark
+23365,ARHGEF12,ENSG00000196914,Rho guanine nucleotide exchange factor 12,protein-coding,NCBI,landmark
+23368,PPP1R13B,ENSG00000088808,protein phosphatase 1 regulatory subunit 13B,protein-coding,NCBI,landmark
+23378,RRP8,ENSG00000132275,ribosomal RNA processing 8,protein-coding,NCBI,landmark
+23386,NUDCD3,ENSG00000015676,NudC domain containing 3,protein-coding,NCBI,landmark
+23410,SIRT3,ENSG00000142082,sirtuin 3,protein-coding,NCBI,landmark
+23443,SLC35A3,ENSG00000117620,solute carrier family 35 member A3,protein-coding,NCBI,landmark
+23463,ICMT,ENSG00000116237,isoprenylcysteine carboxyl methyltransferase,protein-coding,NCBI,landmark
+23499,MACF1,ENSG00000127603,microtubule actin crosslinking factor 1,protein-coding,NCBI,landmark
+23512,SUZ12,ENSG00000178691,SUZ12 polycomb repressive complex 2 subunit,protein-coding,NCBI,landmark
+23522,KAT6B,ENSG00000156650,lysine acetyltransferase 6B,protein-coding,NCBI,landmark
+2353,FOS,ENSG00000170345,"Fos proto-oncogene, AP-1 transcription factor subunit",protein-coding,NCBI,landmark
+23530,NNT,ENSG00000112992,nicotinamide nucleotide transhydrogenase,protein-coding,NCBI,landmark
+23536,ADAT1,ENSG00000065457,adenosine deaminase tRNA specific 1,protein-coding,NCBI,landmark
+2356,FPGS,ENSG00000136877,folylpolyglutamate synthase,protein-coding,NCBI,landmark
+23585,TMEM50A,ENSG00000183726,transmembrane protein 50A,protein-coding,NCBI,landmark
+23588,KLHDC2,ENSG00000165516,kelch domain containing 2,protein-coding,NCBI,landmark
+23597,ACOT9,ENSG00000123130,acyl-CoA thioesterase 9,protein-coding,NCBI,landmark
+23635,SSBP2,ENSG00000145687,single stranded DNA binding protein 2,protein-coding,NCBI,landmark
+23636,NUP62,ENSG00000213024,nucleoporin 62,protein-coding,NCBI,landmark
+23647,ARFIP2,ENSG00000132254,ADP ribosylation factor interacting protein 2,protein-coding,NCBI,landmark
+23658,LSM5,ENSG00000106355,"LSM5 homolog, U6 small nuclear RNA and mRNA degradation associated",protein-coding,NCBI,landmark
+23659,PLA2G15,ENSG00000103066,phospholipase A2 group XV,protein-coding,NCBI,landmark
+23670,CEMIP2,ENSG00000135048,cell migration inducing hyaluronidase 2,protein-coding,NCBI,landmark
+24149,ZNF318,ENSG00000171467,zinc finger protein 318,protein-coding,NCBI,landmark
+25,ABL1,ENSG00000097007,"ABL proto-oncogene 1, non-receptor tyrosine kinase",protein-coding,NCBI,landmark
+2523,FUT1,ENSG00000174951,fucosyltransferase 1 (H blood group),protein-coding,NCBI,landmark
+2534,FYN,ENSG00000010810,"FYN proto-oncogene, Src family tyrosine kinase",protein-coding,NCBI,landmark
+2542,SLC37A4,ENSG00000137700,solute carrier family 37 member 4,protein-coding,NCBI,landmark
+2548,GAA,ENSG00000171298,"glucosidase alpha, acid",protein-coding,NCBI,landmark
+2553,GABPB1,ENSG00000104064,GA binding protein transcription factor subunit beta 1,protein-coding,NCBI,landmark
+256364,EML3,ENSG00000149499,EMAP like 3,protein-coding,NCBI,landmark
+25793,FBXO7,ENSG00000100225,F-box protein 7,protein-coding,NCBI,landmark
+25803,SPDEF,ENSG00000124664,SAM pointed domain containing ETS transcription factor,protein-coding,NCBI,landmark
+25805,BAMBI,ENSG00000095739,BMP and activin membrane bound inhibitor,protein-coding,NCBI,landmark
+2582,GALE,ENSG00000117308,UDP-galactose-4-epimerase,protein-coding,NCBI,landmark
+25825,BACE2,ENSG00000182240,beta-secretase 2,protein-coding,NCBI,landmark
+25839,COG4,ENSG00000103051,component of oligomeric golgi complex 4,protein-coding,NCBI,landmark
+25874,MPC2,ENSG00000143158,mitochondrial pyruvate carrier 2,protein-coding,NCBI,landmark
+25932,CLIC4,ENSG00000169504,chloride intracellular channel 4,protein-coding,NCBI,landmark
+25966,C2CD2,ENSG00000157617,C2 calcium dependent domain containing 2,protein-coding,NCBI,landmark
+2597,GAPDH,ENSG00000111640,glyceraldehyde-3-phosphate dehydrogenase,protein-coding,NCBI,landmark
+25976,TIPARP,ENSG00000163659,TCDD inducible poly(ADP-ribose) polymerase,protein-coding,NCBI,landmark
+25987,TSKU,ENSG00000182704,"tsukushi, small leucine rich proteoglycan",protein-coding,NCBI,landmark
+26001,RNF167,ENSG00000108523,ring finger protein 167,protein-coding,NCBI,landmark
+26020,LRP10,ENSG00000197324,LDL receptor related protein 10,protein-coding,NCBI,landmark
+26036,ZNF451,ENSG00000112200,zinc finger protein 451,protein-coding,NCBI,landmark
+26054,SENP6,ENSG00000112701,SUMO specific peptidase 6,protein-coding,NCBI,landmark
+26064,RAI14,ENSG00000039560,retinoic acid induced 14,protein-coding,NCBI,landmark
+26128,KIF1BP,ENSG00000198954,KIF1 binding protein,protein-coding,NCBI,landmark
+26136,TES,ENSG00000135269,testin LIM domain protein,protein-coding,NCBI,landmark
+26227,PHGDH,ENSG00000092621,phosphoglycerate dehydrogenase,protein-coding,NCBI,landmark
+2624,GATA2,ENSG00000179348,GATA binding protein 2,protein-coding,NCBI,landmark
+2625,GATA3,ENSG00000107485,GATA binding protein 3,protein-coding,NCBI,landmark
+26292,MYCBP,ENSG00000214114,MYC binding protein,protein-coding,NCBI,landmark
+26511,CHIC2,ENSG00000109220,cysteine rich hydrophobic domain 2,protein-coding,NCBI,landmark
+26520,TIMM9,ENSG00000100575,translocase of inner mitochondrial membrane 9,protein-coding,NCBI,landmark
+2673,GFPT1,ENSG00000198380,glutamine--fructose-6-phosphate transaminase 1,protein-coding,NCBI,landmark
+2690,GHR,ENSG00000112964,growth hormone receptor,protein-coding,NCBI,landmark
+26993,AKAP8L,ENSG00000011243,A-kinase anchoring protein 8 like,protein-coding,NCBI,landmark
+27032,ATP2C1,ENSG00000017260,ATPase secretory pathway Ca2+ transporting 1,protein-coding,NCBI,landmark
+27095,TRAPPC3,ENSG00000054116,trafficking protein particle complex 3,protein-coding,NCBI,landmark
+27109,DMAC2L,ENSG00000125375,distal membrane arm assembly complex 2 like,protein-coding,NCBI,landmark
+27242,TNFRSF21,ENSG00000146072,TNF receptor superfamily member 21,protein-coding,NCBI,landmark
+27244,SESN1,ENSG00000080546,sestrin 1,protein-coding,NCBI,landmark
+27336,HTATSF1,ENSG00000102241,HIV-1 Tat specific factor 1,protein-coding,NCBI,landmark
+27346,TMEM97,ENSG00000109084,transmembrane protein 97,protein-coding,NCBI,landmark
+2736,GLI2,ENSG00000074047,GLI family zinc finger 2,protein-coding,NCBI,landmark
+2745,GLRX,ENSG00000173221,glutaredoxin,protein-coding,NCBI,landmark
+2767,GNA11,ENSG00000088256,G protein subunit alpha 11,protein-coding,NCBI,landmark
+2769,GNA15,ENSG00000060558,G protein subunit alpha 15,protein-coding,NCBI,landmark
+2770,GNAI1,ENSG00000127955,G protein subunit alpha i1,protein-coding,NCBI,landmark
+2771,GNAI2,ENSG00000114353,G protein subunit alpha i2,protein-coding,NCBI,landmark
+2778,GNAS,ENSG00000087460,GNAS complex locus,protein-coding,NCBI,landmark
+2810,SFN,ENSG00000175793,stratifin,protein-coding,NCBI,landmark
+2817,GPC1,ENSG00000063660,glypican 1,protein-coding,NCBI,landmark
+2852,GPER1,ENSG00000164850,G protein-coupled estrogen receptor 1,protein-coding,NCBI,landmark
+2886,GRB7,ENSG00000141738,growth factor receptor bound protein 7,protein-coding,NCBI,landmark
+2887,GRB10,ENSG00000106070,growth factor receptor bound protein 10,protein-coding,NCBI,landmark
+2896,GRN,ENSG00000030582,granulin precursor,protein-coding,NCBI,landmark
+28969,BZW2,ENSG00000136261,basic leucine zipper and W2 domains 2,protein-coding,NCBI,landmark
+2908,NR3C1,ENSG00000113580,nuclear receptor subfamily 3 group C member 1,protein-coding,NCBI,landmark
+29082,CHMP4A,ENSG00000254505,charged multivesicular body protein 4A,protein-coding,NCBI,landmark
+29083,GTPBP8,ENSG00000163607,GTP binding protein 8 (putative),protein-coding,NCBI,landmark
+291,SLC25A4,ENSG00000151729,solute carrier family 25 member 4,protein-coding,NCBI,landmark
+29103,DNAJC15,ENSG00000120675,DnaJ heat shock protein family (Hsp40) member C15,protein-coding,NCBI,landmark
+2920,CXCL2,ENSG00000081041,C-X-C motif chemokine ligand 2,protein-coding,NCBI,landmark
+2946,GSTM2,ENSG00000213366,glutathione S-transferase mu 2,protein-coding,NCBI,landmark
+2954,GSTZ1,ENSG00000100577,glutathione S-transferase zeta 1,protein-coding,NCBI,landmark
+2956,MSH6,ENSG00000116062,mutS homolog 6,protein-coding,NCBI,landmark
+2958,GTF2A2,ENSG00000140307,general transcription factor IIA subunit 2,protein-coding,NCBI,landmark
+2961,GTF2E2,ENSG00000197265,general transcription factor IIE subunit 2,protein-coding,NCBI,landmark
+29763,PACSIN3,ENSG00000165912,protein kinase C and casein kinase substrate in neurons 3,protein-coding,NCBI,landmark
+29890,RBM15B,ENSG00000259956,RNA binding motif protein 15B,protein-coding,NCBI,landmark
+29911,HOOK2,ENSG00000095066,hook microtubule tethering protein 2,protein-coding,NCBI,landmark
+29916,SNX11,ENSG00000002919,sorting nexin 11,protein-coding,NCBI,landmark
+29928,TIMM22,ENSG00000177370,translocase of inner mitochondrial membrane 22,protein-coding,NCBI,landmark
+29937,NENF,ENSG00000117691,neudesin neurotrophic factor,protein-coding,NCBI,landmark
+29978,UBQLN2,ENSG00000188021,ubiquilin 2,protein-coding,NCBI,landmark
+30,ACAA1,ENSG00000060971,acetyl-CoA acyltransferase 1,protein-coding,NCBI,landmark
+30001,ERO1A,ENSG00000197930,endoplasmic reticulum oxidoreductase 1 alpha,protein-coding,NCBI,landmark
+3028,HSD17B10,ENSG00000072506,hydroxysteroid 17-beta dehydrogenase 10,protein-coding,NCBI,landmark
+3033,HADH,ENSG00000138796,hydroxyacyl-CoA dehydrogenase,protein-coding,NCBI,landmark
+3066,HDAC2,ENSG00000196591,histone deacetylase 2,protein-coding,NCBI,landmark
+30836,DNTTIP2,ENSG00000067334,deoxynucleotidyltransferase terminal interacting protein 2,protein-coding,NCBI,landmark
+30849,PIK3R4,ENSG00000196455,phosphoinositide-3-kinase regulatory subunit 4,protein-coding,NCBI,landmark
+3091,HIF1A,ENSG00000100644,hypoxia inducible factor 1 subunit alpha,protein-coding,NCBI,landmark
+3098,HK1,ENSG00000156515,hexokinase 1,protein-coding,NCBI,landmark
+310,ANXA7,ENSG00000138279,annexin A7,protein-coding,NCBI,landmark
+3108,HLA-DMA,ENSG00000204257,"major histocompatibility complex, class II, DM alpha",protein-coding,NCBI,landmark
+3122,HLA-DRA,ENSG00000204287,"major histocompatibility complex, class II, DR alpha",protein-coding,NCBI,landmark
+3156,HMGCR,ENSG00000113161,3-hydroxy-3-methylglutaryl-CoA reductase,protein-coding,NCBI,landmark
+3157,HMGCS1,ENSG00000112972,3-hydroxy-3-methylglutaryl-CoA synthase 1,protein-coding,NCBI,landmark
+3162,HMOX1,ENSG00000100292,heme oxygenase 1,protein-coding,NCBI,landmark
+3202,HOXA5,ENSG00000106004,homeobox A5,protein-coding,NCBI,landmark
+3206,HOXA10,ENSG00000253293,homeobox A10,protein-coding,NCBI,landmark
+323,APBB2,ENSG00000163697,amyloid beta precursor protein binding family B member 2,protein-coding,NCBI,landmark
+3251,HPRT1,ENSG00000165704,hypoxanthine phosphoribosyltransferase 1,protein-coding,NCBI,landmark
+3280,HES1,ENSG00000114315,hes family bHLH transcription factor 1,protein-coding,NCBI,landmark
+329,BIRC2,ENSG00000110330,baculoviral IAP repeat containing 2,protein-coding,NCBI,landmark
+3300,DNAJB2,ENSG00000135924,DnaJ heat shock protein family (Hsp40) member B2,protein-coding,NCBI,landmark
+3303,HSPA1A,ENSG00000204389,heat shock protein family A (Hsp70) member 1A,protein-coding,NCBI,landmark
+3308,HSPA4,ENSG00000170606,heat shock protein family A (Hsp70) member 4,protein-coding,NCBI,landmark
+3312,HSPA8,ENSG00000109971,heat shock protein family A (Hsp70) member 8,protein-coding,NCBI,landmark
+3315,HSPB1,ENSG00000106211,heat shock protein family B (small) member 1,protein-coding,NCBI,landmark
+332,BIRC5,ENSG00000089685,baculoviral IAP repeat containing 5,protein-coding,NCBI,landmark
+3329,HSPD1,ENSG00000144381,heat shock protein family D (Hsp60) member 1,protein-coding,NCBI,landmark
+3337,DNAJB1,ENSG00000132002,DnaJ heat shock protein family (Hsp40) member B1,protein-coding,NCBI,landmark
+3383,ICAM1,ENSG00000090339,intercellular adhesion molecule 1,protein-coding,NCBI,landmark
+3385,ICAM3,ENSG00000076662,intercellular adhesion molecule 3,protein-coding,NCBI,landmark
+3398,ID2,ENSG00000115738,inhibitor of DNA binding 2,protein-coding,NCBI,landmark
+3416,IDE,ENSG00000119912,insulin degrading enzyme,protein-coding,NCBI,landmark
+3454,IFNAR1,ENSG00000142166,interferon alpha and beta receptor subunit 1,protein-coding,NCBI,landmark
+348,APOE,ENSG00000130203,apolipoprotein E,protein-coding,NCBI,landmark
+3480,IGF1R,ENSG00000140443,insulin like growth factor 1 receptor,protein-coding,NCBI,landmark
+3482,IGF2R,ENSG00000197081,insulin like growth factor 2 receptor,protein-coding,NCBI,landmark
+3486,IGFBP3,ENSG00000146674,insulin like growth factor binding protein 3,protein-coding,NCBI,landmark
+3508,IGHMBP2,ENSG00000132740,immunoglobulin mu DNA binding protein 2,protein-coding,NCBI,landmark
+351,APP,ENSG00000142192,amyloid beta precursor protein,protein-coding,NCBI,landmark
+355,FAS,ENSG00000026103,Fas cell surface death receptor,protein-coding,NCBI,landmark
+3551,IKBKB,ENSG00000104365,inhibitor of nuclear factor kappa B kinase subunit beta,protein-coding,NCBI,landmark
+3553,IL1B,ENSG00000125538,interleukin 1 beta,protein-coding,NCBI,landmark
+3566,IL4R,ENSG00000077238,interleukin 4 receptor,protein-coding,NCBI,landmark
+3597,IL13RA1,ENSG00000131724,interleukin 13 receptor subunit alpha 1,protein-coding,NCBI,landmark
+3611,ILK,ENSG00000166333,integrin linked kinase,protein-coding,NCBI,landmark
+3628,INPP1,ENSG00000151689,inositol polyphosphate-1-phosphatase,protein-coding,NCBI,landmark
+3638,INSIG1,ENSG00000186480,insulin induced gene 1,protein-coding,NCBI,landmark
+3682,ITGAE,ENSG00000083457,integrin subunit alpha E,protein-coding,NCBI,landmark
+3693,ITGB5,ENSG00000082781,integrin subunit beta 5,protein-coding,NCBI,landmark
+3725,JUN,ENSG00000177606,"Jun proto-oncogene, AP-1 transcription factor subunit",protein-coding,NCBI,landmark
+375346,STIMATE,ENSG00000213533,STIM activating enhancer,protein-coding,NCBI,landmark
+3775,KCNK1,ENSG00000135750,potassium two pore domain channel subfamily K member 1,protein-coding,NCBI,landmark
+3800,KIF5C,ENSG00000168280,kinesin family member 5C,protein-coding,NCBI,landmark
+3815,KIT,ENSG00000157404,"KIT proto-oncogene, receptor tyrosine kinase",protein-coding,NCBI,landmark
+387,RHOA,ENSG00000067560,ras homolog family member A,protein-coding,NCBI,landmark
+388650,DIPK1A,ENSG00000154511,divergent protein kinase domain 1A,protein-coding,NCBI,landmark
+3895,KTN1,ENSG00000126777,kinectin 1,protein-coding,NCBI,landmark
+39,ACAT2,ENSG00000120437,acetyl-CoA acetyltransferase 2,protein-coding,NCBI,landmark
+3909,LAMA3,ENSG00000053747,laminin subunit alpha 3,protein-coding,NCBI,landmark
+392,ARHGAP1,ENSG00000175220,Rho GTPase activating protein 1,protein-coding,NCBI,landmark
+3925,STMN1,ENSG00000117632,stathmin 1,protein-coding,NCBI,landmark
+3930,LBR,ENSG00000143815,lamin B receptor,protein-coding,NCBI,landmark
+3964,LGALS8,ENSG00000116977,galectin 8,protein-coding,NCBI,landmark
+3978,LIG1,ENSG00000105486,DNA ligase 1,protein-coding,NCBI,landmark
+3988,LIPA,ENSG00000107798,"lipase A, lysosomal acid type",protein-coding,NCBI,landmark
+4016,LOXL1,ENSG00000129038,lysyl oxidase like 1,protein-coding,NCBI,landmark
+4043,LRPAP1,ENSG00000163956,LDL receptor related protein associated protein 1,protein-coding,NCBI,landmark
+4067,LYN,ENSG00000254087,"LYN proto-oncogene, Src family tyrosine kinase",protein-coding,NCBI,landmark
+4088,SMAD3,ENSG00000166949,SMAD family member 3,protein-coding,NCBI,landmark
+4125,MAN2B1,ENSG00000104774,mannosidase alpha class 2B member 1,protein-coding,NCBI,landmark
+4144,MAT2A,ENSG00000168906,methionine adenosyltransferase 2A,protein-coding,NCBI,landmark
+4154,MBNL1,ENSG00000152601,muscleblind like splicing regulator 1,protein-coding,NCBI,landmark
+4172,MCM3,ENSG00000112118,minichromosome maintenance complex component 3,protein-coding,NCBI,landmark
+4200,ME2,ENSG00000082212,malic enzyme 2,protein-coding,NCBI,landmark
+4208,MEF2C,ENSG00000081189,myocyte enhancer factor 2C,protein-coding,NCBI,landmark
+4216,MAP3K4,ENSG00000085511,mitogen-activated protein kinase kinase kinase 4,protein-coding,NCBI,landmark
+4232,MEST,ENSG00000106484,mesoderm specific transcript,protein-coding,NCBI,landmark
+427,ASAH1,ENSG00000104763,N-acylsphingosine amidohydrolase 1,protein-coding,NCBI,landmark
+4282,MIF,ENSG00000240972,macrophage migration inhibitory factor,protein-coding,NCBI,landmark
+4303,FOXO4,ENSG00000184481,forkhead box O4,protein-coding,NCBI,landmark
+4312,MMP1,ENSG00000196611,matrix metallopeptidase 1,protein-coding,NCBI,landmark
+4313,MMP2,ENSG00000087245,matrix metallopeptidase 2,protein-coding,NCBI,landmark
+4331,MNAT1,ENSG00000020426,MNAT1 component of CDK activating kinase,protein-coding,NCBI,landmark
+4482,MSRA,ENSG00000175806,methionine sulfoxide reductase A,protein-coding,NCBI,landmark
+4582,MUC1,ENSG00000185499,"mucin 1, cell surface associated",protein-coding,NCBI,landmark
+4605,MYBL2,ENSG00000101057,MYB proto-oncogene like 2,protein-coding,NCBI,landmark
+4609,MYC,ENSG00000136997,"MYC proto-oncogene, bHLH transcription factor",protein-coding,NCBI,landmark
+4616,GADD45B,ENSG00000099860,growth arrest and DNA damage inducible beta,protein-coding,NCBI,landmark
+4638,MYLK,ENSG00000065534,myosin light chain kinase,protein-coding,NCBI,landmark
+4651,MYO10,ENSG00000145555,myosin X,protein-coding,NCBI,landmark
+466,ATF1,ENSG00000123268,activating transcription factor 1,protein-coding,NCBI,landmark
+4690,NCK1,ENSG00000158092,NCK adaptor protein 1,protein-coding,NCBI,landmark
+47,ACLY,ENSG00000131473,ATP citrate lyase,protein-coding,NCBI,landmark
+4775,NFATC3,ENSG00000072736,nuclear factor of activated T cells 3,protein-coding,NCBI,landmark
+4776,NFATC4,ENSG00000100968,nuclear factor of activated T cells 4,protein-coding,NCBI,landmark
+4780,NFE2L2,ENSG00000116044,"nuclear factor, erythroid 2 like 2",protein-coding,NCBI,landmark
+4783,NFIL3,ENSG00000165030,"nuclear factor, interleukin 3 regulated",protein-coding,NCBI,landmark
+4791,NFKB2,ENSG00000077150,nuclear factor kappa B subunit 2,protein-coding,NCBI,landmark
+4792,NFKBIA,ENSG00000100906,NFKB inhibitor alpha,protein-coding,NCBI,landmark
+4793,NFKBIB,ENSG00000104825,NFKB inhibitor beta,protein-coding,NCBI,landmark
+4794,NFKBIE,ENSG00000146232,NFKB inhibitor epsilon,protein-coding,NCBI,landmark
+481,ATP1B1,ENSG00000143153,ATPase Na+/K+ transporting subunit beta 1,protein-coding,NCBI,landmark
+4817,NIT1,ENSG00000158793,nitrilase 1,protein-coding,NCBI,landmark
+4836,NMT1,ENSG00000136448,N-myristoyltransferase 1,protein-coding,NCBI,landmark
+4846,NOS3,ENSG00000164867,nitric oxide synthase 3,protein-coding,NCBI,landmark
+4850,CNOT4,ENSG00000080802,CCR4-NOT transcription complex subunit 4,protein-coding,NCBI,landmark
+4851,NOTCH1,ENSG00000148400,notch receptor 1,protein-coding,NCBI,landmark
+4860,PNP,ENSG00000198805,purine nucleoside phosphorylase,protein-coding,NCBI,landmark
+4864,NPC1,ENSG00000141458,NPC intracellular cholesterol transporter 1,protein-coding,NCBI,landmark
+4891,SLC11A2,ENSG00000110911,solute carrier family 11 member 2,protein-coding,NCBI,landmark
+4893,NRAS,ENSG00000213281,"NRAS proto-oncogene, GTPase",protein-coding,NCBI,landmark
+4925,NUCB2,ENSG00000070081,nucleobindin 2,protein-coding,NCBI,landmark
+4927,NUP88,ENSG00000108559,nucleoporin 88,protein-coding,NCBI,landmark
+4931,NVL,ENSG00000143748,nuclear VCP like,protein-coding,NCBI,landmark
+4998,ORC1,ENSG00000085840,origin recognition complex subunit 1,protein-coding,NCBI,landmark
+501,ALDH7A1,ENSG00000164904,aldehyde dehydrogenase 7 family member A1,protein-coding,NCBI,landmark
+5018,OXA1L,ENSG00000155463,OXA1L mitochondrial inner membrane protein,protein-coding,NCBI,landmark
+5019,OXCT1,ENSG00000083720,3-oxoacid CoA-transferase 1,protein-coding,NCBI,landmark
+5048,PAFAH1B1,ENSG00000007168,platelet activating factor acetylhydrolase 1b regulatory subunit 1,protein-coding,NCBI,landmark
+5050,PAFAH1B3,ENSG00000079462,platelet activating factor acetylhydrolase 1b catalytic subunit 3,protein-coding,NCBI,landmark
+5054,SERPINE1,ENSG00000106366,serpin family E member 1,protein-coding,NCBI,landmark
+5058,PAK1,ENSG00000149269,p21 (RAC1) activated kinase 1,protein-coding,NCBI,landmark
+50810,HDGFL3,ENSG00000166503,HDGF like 3,protein-coding,NCBI,landmark
+50813,COPS7A,ENSG00000111652,COP9 signalosome subunit 7A,protein-coding,NCBI,landmark
+50814,NSDHL,ENSG00000147383,NAD(P) dependent steroid dehydrogenase-like,protein-coding,NCBI,landmark
+50865,HEBP1,ENSG00000013583,heme binding protein 1,protein-coding,NCBI,landmark
+5092,PCBD1,ENSG00000166228,pterin-4 alpha-carbinolamine dehydratase 1,protein-coding,NCBI,landmark
+5096,PCCB,ENSG00000114054,propionyl-CoA carboxylase subunit beta,protein-coding,NCBI,landmark
+51001,MTERF3,ENSG00000156469,mitochondrial transcription termination factor 3,protein-coding,NCBI,landmark
+51005,AMDHD2,ENSG00000162066,amidohydrolase domain containing 2,protein-coding,NCBI,landmark
+51015,ISOC1,ENSG00000066583,isochorismatase domain containing 1,protein-coding,NCBI,landmark
+51021,MRPS16,ENSG00000182180,mitochondrial ribosomal protein S16,protein-coding,NCBI,landmark
+51024,FIS1,ENSG00000214253,"fission, mitochondrial 1",protein-coding,NCBI,landmark
+51026,GOLT1B,ENSG00000111711,golgi transport 1B,protein-coding,NCBI,landmark
+51031,GLOD4,ENSG00000167699,glyoxalase domain containing 4,protein-coding,NCBI,landmark
+51053,GMNN,ENSG00000112312,geminin DNA replication inhibitor,protein-coding,NCBI,landmark
+51056,LAP3,ENSG00000002549,leucine aminopeptidase 3,protein-coding,NCBI,landmark
+5106,PCK2,ENSG00000100889,"phosphoenolpyruvate carboxykinase 2, mitochondrial",protein-coding,NCBI,landmark
+51070,NOSIP,ENSG00000142546,nitric oxide synthase interacting protein,protein-coding,NCBI,landmark
+51071,DERA,ENSG00000023697,deoxyribose-phosphate aldolase,protein-coding,NCBI,landmark
+5108,PCM1,ENSG00000078674,pericentriolar material 1,protein-coding,NCBI,landmark
+51097,SCCPDH,ENSG00000143653,saccharopine dehydrogenase (putative),protein-coding,NCBI,landmark
+5110,PCMT1,ENSG00000120265,protein-L-isoaspartate (D-aspartate) O-methyltransferase,protein-coding,NCBI,landmark
+5111,PCNA,ENSG00000132646,proliferating cell nuclear antigen,protein-coding,NCBI,landmark
+51116,MRPS2,ENSG00000122140,mitochondrial ribosomal protein S2,protein-coding,NCBI,landmark
+51160,VPS28,ENSG00000160948,VPS28 subunit of ESCRT-I,protein-coding,NCBI,landmark
+51170,HSD17B11,ENSG00000198189,hydroxysteroid 17-beta dehydrogenase 11,protein-coding,NCBI,landmark
+51203,NUSAP1,ENSG00000137804,nucleolar and spindle associated protein 1,protein-coding,NCBI,landmark
+51282,SCAND1,ENSG00000171222,SCAN domain containing 1,protein-coding,NCBI,landmark
+51293,CD320,ENSG00000167775,CD320 molecule,protein-coding,NCBI,landmark
+51335,NGRN,ENSG00000182768,"neugrin, neurite outgrowth associated",protein-coding,NCBI,landmark
+51375,SNX7,ENSG00000162627,sorting nexin 7,protein-coding,NCBI,landmark
+51382,ATP6V1D,ENSG00000100554,ATPase H+ transporting V1 subunit D,protein-coding,NCBI,landmark
+51385,ZNF589,ENSG00000164048,zinc finger protein 589,protein-coding,NCBI,landmark
+51422,PRKAG2,ENSG00000106617,protein kinase AMP-activated non-catalytic subunit gamma 2,protein-coding,NCBI,landmark
+51465,UBE2J1,ENSG00000198833,ubiquitin conjugating enzyme E2 J1,protein-coding,NCBI,landmark
+51466,EVL,ENSG00000196405,Enah/Vasp-like,protein-coding,NCBI,landmark
+51495,HACD3,ENSG00000074696,3-hydroxyacyl-CoA dehydratase 3,protein-coding,NCBI,landmark
+5154,PDGFA,ENSG00000197461,platelet derived growth factor subunit A,protein-coding,NCBI,landmark
+51569,UFM1,ENSG00000120686,ubiquitin fold modifier 1,protein-coding,NCBI,landmark
+51599,LSR,ENSG00000105699,lipolysis stimulated lipoprotein receptor,protein-coding,NCBI,landmark
+51635,DHRS7,ENSG00000100612,dehydrogenase/reductase 7,protein-coding,NCBI,landmark
+51719,CAB39,ENSG00000135932,calcium binding protein 39,protein-coding,NCBI,landmark
+51742,ARID4B,ENSG00000054267,AT-rich interaction domain 4B,protein-coding,NCBI,landmark
+5211,PFKL,ENSG00000141959,"phosphofructokinase, liver type",protein-coding,NCBI,landmark
+5223,PGAM1,ENSG00000171314,phosphoglycerate mutase 1,protein-coding,NCBI,landmark
+5236,PGM1,ENSG00000079739,phosphoglucomutase 1,protein-coding,NCBI,landmark
+5255,PHKA1,ENSG00000067177,phosphorylase kinase regulatory subunit alpha 1,protein-coding,NCBI,landmark
+5257,PHKB,ENSG00000102893,phosphorylase kinase regulatory subunit beta,protein-coding,NCBI,landmark
+5261,PHKG2,ENSG00000156873,phosphorylase kinase catalytic subunit gamma 2,protein-coding,NCBI,landmark
+5287,PIK3C2B,ENSG00000133056,phosphatidylinositol-4-phosphate 3-kinase catalytic subunit type 2 beta,protein-coding,NCBI,landmark
+5289,PIK3C3,ENSG00000078142,phosphatidylinositol 3-kinase catalytic subunit type 3,protein-coding,NCBI,landmark
+5290,PIK3CA,ENSG00000121879,"phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha",protein-coding,NCBI,landmark
+5300,PIN1,ENSG00000127445,"peptidylprolyl cis/trans isomerase, NIMA-interacting 1",protein-coding,NCBI,landmark
+5321,PLA2G4A,ENSG00000116711,phospholipase A2 group IVA,protein-coding,NCBI,landmark
+533,ATP6V0B,ENSG00000117410,ATPase H+ transporting V0 subunit b,protein-coding,NCBI,landmark
+5331,PLCB3,ENSG00000149782,phospholipase C beta 3,protein-coding,NCBI,landmark
+53343,NUDT9,ENSG00000170502,nudix hydrolase 9,protein-coding,NCBI,landmark
+5347,PLK1,ENSG00000166851,polo like kinase 1,protein-coding,NCBI,landmark
+5355,PLP2,ENSG00000102007,proteolipid protein 2,protein-coding,NCBI,landmark
+5357,PLS1,ENSG00000120756,plastin 1,protein-coding,NCBI,landmark
+5359,PLSCR1,ENSG00000188313,phospholipid scramblase 1,protein-coding,NCBI,landmark
+5366,PMAIP1,ENSG00000141682,phorbol-12-myristate-13-acetate-induced protein 1,protein-coding,NCBI,landmark
+5373,PMM2,ENSG00000140650,phosphomannomutase 2,protein-coding,NCBI,landmark
+54205,CYCS,ENSG00000172115,"cytochrome c, somatic",protein-coding,NCBI,landmark
+5423,POLB,ENSG00000070501,DNA polymerase beta,protein-coding,NCBI,landmark
+5427,POLE2,ENSG00000100479,"DNA polymerase epsilon 2, accessory subunit",protein-coding,NCBI,landmark
+5438,POLR2I,ENSG00000105258,RNA polymerase II subunit I,protein-coding,NCBI,landmark
+54386,TERF2IP,ENSG00000166848,TERF2 interacting protein,protein-coding,NCBI,landmark
+5440,POLR2K,ENSG00000147669,RNA polymerase II subunit K,protein-coding,NCBI,landmark
+54438,GFOD1,ENSG00000145990,glucose-fructose oxidoreductase domain containing 1,protein-coding,NCBI,landmark
+54442,KCTD5,ENSG00000167977,potassium channel tetramerization domain containing 5,protein-coding,NCBI,landmark
+54499,TMCO1,ENSG00000143183,transmembrane and coiled-coil domains 1,protein-coding,NCBI,landmark
+54505,DHX29,ENSG00000067248,DExH-box helicase 29,protein-coding,NCBI,landmark
+54512,EXOSC4,ENSG00000178896,exosome component 4,protein-coding,NCBI,landmark
+54541,DDIT4,ENSG00000168209,DNA damage inducible transcript 4,protein-coding,NCBI,landmark
+54623,PAF1,ENSG00000006712,"PAF1 homolog, Paf1/RNA polymerase II complex component",protein-coding,NCBI,landmark
+5467,PPARD,ENSG00000112033,peroxisome proliferator activated receptor delta,protein-coding,NCBI,landmark
+5468,PPARG,ENSG00000132170,peroxisome proliferator activated receptor gamma,protein-coding,NCBI,landmark
+54681,P4HTM,ENSG00000178467,"prolyl 4-hydroxylase, transmembrane",protein-coding,NCBI,landmark
+54733,SLC35F2,ENSG00000110660,solute carrier family 35 member F2,protein-coding,NCBI,landmark
+5480,PPIC,ENSG00000168938,peptidylprolyl isomerase C,protein-coding,NCBI,landmark
+54807,ZNF586,ENSG00000083828,zinc finger protein 586,protein-coding,NCBI,landmark
+54850,FBXL12,ENSG00000127452,F-box and leucine rich repeat protein 12,protein-coding,NCBI,landmark
+54881,TEX10,ENSG00000136891,testis expressed 10,protein-coding,NCBI,landmark
+54915,YTHDF1,ENSG00000149658,YTH N6-methyladenosine RNA binding protein 1,protein-coding,NCBI,landmark
+54957,TXNL4B,ENSG00000140830,thioredoxin like 4B,protein-coding,NCBI,landmark
+5498,PPOX,ENSG00000143224,protoporphyrinogen oxidase,protein-coding,NCBI,landmark
+55008,HERC6,ENSG00000138642,HECT and RLD domain containing E3 ubiquitin protein ligase family member 6,protein-coding,NCBI,landmark
+55011,PIH1D1,ENSG00000104872,PIH1 domain containing 1,protein-coding,NCBI,landmark
+55012,PPP2R3C,ENSG00000092020,protein phosphatase 2 regulatory subunit B''gamma,protein-coding,NCBI,landmark
+55033,FKBP14,ENSG00000106080,FKBP prolyl isomerase 14,protein-coding,NCBI,landmark
+55038,CDCA4,ENSG00000170779,cell division cycle associated 4,protein-coding,NCBI,landmark
+55111,PLEKHJ1,ENSG00000104886,pleckstrin homology domain containing J1,protein-coding,NCBI,landmark
+55127,HEATR1,ENSG00000119285,HEAT repeat containing 1,protein-coding,NCBI,landmark
+55129,ANO10,ENSG00000160746,anoctamin 10,protein-coding,NCBI,landmark
+55148,UBR7,ENSG00000012963,ubiquitin protein ligase E3 component n-recognin 7 (putative),protein-coding,NCBI,landmark
+55179,FAIM,ENSG00000158234,Fas apoptotic inhibitory molecule,protein-coding,NCBI,landmark
+5525,PPP2R5A,ENSG00000066027,protein phosphatase 2 regulatory subunit B'alpha,protein-coding,NCBI,landmark
+55256,ADI1,ENSG00000182551,acireductone dioxygenase 1,protein-coding,NCBI,landmark
+5529,PPP2R5E,ENSG00000154001,protein phosphatase 2 regulatory subunit B'epsilon,protein-coding,NCBI,landmark
+55324,ABCF3,ENSG00000161204,ATP binding cassette subfamily F member 3,protein-coding,NCBI,landmark
+5547,PRCP,ENSG00000137509,prolylcarboxypeptidase,protein-coding,NCBI,landmark
+55556,ENOSF1,ENSG00000132199,enolase superfamily member 1,protein-coding,NCBI,landmark
+55604,CARMIL1,ENSG00000079691,capping protein regulator and myosin 1 linker 1,protein-coding,NCBI,landmark
+55608,ANKRD10,ENSG00000088448,ankyrin repeat domain 10,protein-coding,NCBI,landmark
+55620,STAP2,ENSG00000178078,signal transducing adaptor family member 2,protein-coding,NCBI,landmark
+5566,PRKACA,ENSG00000072062,protein kinase cAMP-activated catalytic subunit alpha,protein-coding,NCBI,landmark
+55699,IARS2,ENSG00000067704,"isoleucyl-tRNA synthetase 2, mitochondrial",protein-coding,NCBI,landmark
+55746,NUP133,ENSG00000069248,nucleoporin 133,protein-coding,NCBI,landmark
+55748,CNDP2,ENSG00000133313,carnosine dipeptidase 2,protein-coding,NCBI,landmark
+55793,MINDY1,ENSG00000143409,MINDY lysine 48 deubiquitinase 1,protein-coding,NCBI,landmark
+5580,PRKCD,ENSG00000163932,protein kinase C delta,protein-coding,NCBI,landmark
+55818,KDM3A,ENSG00000115548,lysine demethylase 3A,protein-coding,NCBI,landmark
+55825,PECR,ENSG00000115425,peroxisomal trans-2-enoyl-CoA reductase,protein-coding,NCBI,landmark
+5583,PRKCH,ENSG00000027075,protein kinase C eta,protein-coding,NCBI,landmark
+55837,EAPP,ENSG00000129518,E2F associated phosphoprotein,protein-coding,NCBI,landmark
+55847,CISD1,ENSG00000122873,CDGSH iron sulfur domain 1,protein-coding,NCBI,landmark
+5588,PRKCQ,ENSG00000065675,protein kinase C theta,protein-coding,NCBI,landmark
+55893,ZNF395,ENSG00000186918,zinc finger protein 395,protein-coding,NCBI,landmark
+55958,KLHL9,ENSG00000198642,kelch like family member 9,protein-coding,NCBI,landmark
+5601,MAPK9,ENSG00000050748,mitogen-activated protein kinase 9,protein-coding,NCBI,landmark
+5603,MAPK13,ENSG00000156711,mitogen-activated protein kinase 13,protein-coding,NCBI,landmark
+5607,MAP2K5,ENSG00000137764,mitogen-activated protein kinase kinase 5,protein-coding,NCBI,landmark
+5613,PRKX,ENSG00000183943,protein kinase X-linked,protein-coding,NCBI,landmark
+5627,PROS1,ENSG00000184500,protein S,protein-coding,NCBI,landmark
+5641,LGMN,ENSG00000100600,legumain,protein-coding,NCBI,landmark
+5654,HTRA1,ENSG00000166033,HtrA serine peptidase 1,protein-coding,NCBI,landmark
+56654,NPDC1,ENSG00000107281,"neural proliferation, differentiation and control 1",protein-coding,NCBI,landmark
+56889,TM9SF3,ENSG00000077147,transmembrane 9 superfamily member 3,protein-coding,NCBI,landmark
+56924,PAK6,ENSG00000137843,p21 (RAC1) activated kinase 6,protein-coding,NCBI,landmark
+56940,DUSP22,ENSG00000112679,dual specificity phosphatase 22,protein-coding,NCBI,landmark
+5696,PSMB8,ENSG00000204264,proteasome subunit beta 8,protein-coding,NCBI,landmark
+5699,PSMB10,ENSG00000205220,proteasome subunit beta 10,protein-coding,NCBI,landmark
+56997,COQ8A,ENSG00000163050,coenzyme Q8A,protein-coding,NCBI,landmark
+57019,CIAPIN1,ENSG00000005194,cytokine induced apoptosis inhibitor 1,protein-coding,NCBI,landmark
+57048,PLSCR3,ENSG00000187838,phospholipid scramblase 3,protein-coding,NCBI,landmark
+5708,PSMD2,ENSG00000175166,"proteasome 26S subunit, non-ATPase 2",protein-coding,NCBI,landmark
+5710,PSMD4,ENSG00000159352,"proteasome 26S subunit, non-ATPase 4",protein-coding,NCBI,landmark
+57147,SCYL3,ENSG00000000457,SCY1 like pseudokinase 3,protein-coding,NCBI,landmark
+57149,LYRM1,ENSG00000102897,LYR motif containing 1,protein-coding,NCBI,landmark
+5715,PSMD9,ENSG00000110801,"proteasome 26S subunit, non-ATPase 9",protein-coding,NCBI,landmark
+5716,PSMD10,ENSG00000101843,"proteasome 26S subunit, non-ATPase 10",protein-coding,NCBI,landmark
+57178,ZMIZ1,ENSG00000108175,zinc finger MIZ-type containing 1,protein-coding,NCBI,landmark
+57192,MCOLN1,ENSG00000090674,mucolipin 1,protein-coding,NCBI,landmark
+572,BAD,ENSG00000002330,BCL2 associated agonist of cell death,protein-coding,NCBI,landmark
+5720,PSME1,ENSG00000092010,proteasome activator subunit 1,protein-coding,NCBI,landmark
+5721,PSME2,ENSG00000100911,proteasome activator subunit 2,protein-coding,NCBI,landmark
+57215,THAP11,ENSG00000168286,THAP domain containing 11,protein-coding,NCBI,landmark
+57406,ABHD6,ENSG00000163686,abhydrolase domain containing 6,protein-coding,NCBI,landmark
+5743,PTGS2,ENSG00000073756,prostaglandin-endoperoxide synthase 2,protein-coding,NCBI,landmark
+5747,PTK2,ENSG00000169398,protein tyrosine kinase 2,protein-coding,NCBI,landmark
+5770,PTPN1,ENSG00000196396,protein tyrosine phosphatase non-receptor type 1,protein-coding,NCBI,landmark
+57761,TRIB3,ENSG00000101255,tribbles pseudokinase 3,protein-coding,NCBI,landmark
+5777,PTPN6,ENSG00000111679,protein tyrosine phosphatase non-receptor type 6,protein-coding,NCBI,landmark
+57804,POLD4,ENSG00000175482,"DNA polymerase delta 4, accessory subunit",protein-coding,NCBI,landmark
+5782,PTPN12,ENSG00000127947,protein tyrosine phosphatase non-receptor type 12,protein-coding,NCBI,landmark
+5788,PTPRC,ENSG00000081237,protein tyrosine phosphatase receptor type C,protein-coding,NCBI,landmark
+5792,PTPRF,ENSG00000142949,protein tyrosine phosphatase receptor type F,protein-coding,NCBI,landmark
+5796,PTPRK,ENSG00000152894,protein tyrosine phosphatase receptor type K,protein-coding,NCBI,landmark
+581,BAX,ENSG00000087088,"BCL2 associated X, apoptosis regulator",protein-coding,NCBI,landmark
+5827,PXMP2,ENSG00000176894,peroxisomal membrane protein 2,protein-coding,NCBI,landmark
+5829,PXN,ENSG00000089159,paxillin,protein-coding,NCBI,landmark
+5831,PYCR1,ENSG00000183010,pyrroline-5-carboxylate reductase 1,protein-coding,NCBI,landmark
+5836,PYGL,ENSG00000100504,glycogen phosphorylase L,protein-coding,NCBI,landmark
+58472,SQOR,ENSG00000137767,sulfide quinone oxidoreductase,protein-coding,NCBI,landmark
+58478,ENOPH1,ENSG00000145293,enolase-phosphatase 1,protein-coding,NCBI,landmark
+58497,PRUNE1,ENSG00000143363,prune exopolyphosphatase 1,protein-coding,NCBI,landmark
+58533,SNX6,ENSG00000129515,sorting nexin 6,protein-coding,NCBI,landmark
+5867,RAB4A,ENSG00000168118,"RAB4A, member RAS oncogene family",protein-coding,NCBI,landmark
+5873,RAB27A,ENSG00000069974,"RAB27A, member RAS oncogene family",protein-coding,NCBI,landmark
+5880,RAC2,ENSG00000128340,Rac family small GTPase 2,protein-coding,NCBI,landmark
+5883,RAD9A,ENSG00000172613,RAD9 checkpoint clamp component A,protein-coding,NCBI,landmark
+5889,RAD51C,ENSG00000108384,RAD51 paralog C,protein-coding,NCBI,landmark
+5891,MOK,ENSG00000080823,MOK protein kinase,protein-coding,NCBI,landmark
+5898,RALA,ENSG00000006451,RAS like proto-oncogene A,protein-coding,NCBI,landmark
+5899,RALB,ENSG00000144118,RAS like proto-oncogene B,protein-coding,NCBI,landmark
+5900,RALGDS,ENSG00000160271,ral guanine nucleotide dissociation stimulator,protein-coding,NCBI,landmark
+5909,RAP1GAP,ENSG00000076864,RAP1 GTPase activating protein,protein-coding,NCBI,landmark
+5921,RASA1,ENSG00000145715,RAS p21 protein activator 1,protein-coding,NCBI,landmark
+5925,RB1,ENSG00000139687,RB transcriptional corepressor 1,protein-coding,NCBI,landmark
+5927,KDM5A,ENSG00000073614,lysine demethylase 5A,protein-coding,NCBI,landmark
+595,CCND1,ENSG00000110092,cyclin D1,protein-coding,NCBI,landmark
+596,BCL2,ENSG00000171791,BCL2 apoptosis regulator,protein-coding,NCBI,landmark
+5971,RELB,ENSG00000104856,"RELB proto-oncogene, NF-kB subunit",protein-coding,NCBI,landmark
+5982,RFC2,ENSG00000049541,replication factor C subunit 2,protein-coding,NCBI,landmark
+5985,RFC5,ENSG00000111445,replication factor C subunit 5,protein-coding,NCBI,landmark
+5986,RFNG,ENSG00000169733,RFNG O-fucosylpeptide 3-beta-N-acetylglucosaminyltransferase,protein-coding,NCBI,landmark
+5993,RFX5,ENSG00000143390,regulatory factor X5,protein-coding,NCBI,landmark
+5997,RGS2,ENSG00000116741,regulator of G protein signaling 2,protein-coding,NCBI,landmark
+6009,RHEB,ENSG00000106615,"Ras homolog, mTORC1 binding",protein-coding,NCBI,landmark
+60493,FASTKD5,ENSG00000215251,FAST kinase domains 5,protein-coding,NCBI,landmark
+6050,RNH1,ENSG00000023191,ribonuclease/angiogenin inhibitor 1,protein-coding,NCBI,landmark
+60528,ELAC2,ENSG00000006744,elaC ribonuclease Z 2,protein-coding,NCBI,landmark
+6117,RPA1,ENSG00000132383,replication protein A1,protein-coding,NCBI,landmark
+6118,RPA2,ENSG00000117748,replication protein A2,protein-coding,NCBI,landmark
+6119,RPA3,ENSG00000106399,replication protein A3,protein-coding,NCBI,landmark
+6182,MRPL12,ENSG00000262814,mitochondrial ribosomal protein L12,protein-coding,NCBI,landmark
+6184,RPN1,ENSG00000163902,ribophorin I,protein-coding,NCBI,landmark
+6193,RPS5,ENSG00000083845,ribosomal protein S5,protein-coding,NCBI,landmark
+6194,RPS6,ENSG00000137154,ribosomal protein S6,protein-coding,NCBI,landmark
+6195,RPS6KA1,ENSG00000117676,ribosomal protein S6 kinase A1,protein-coding,NCBI,landmark
+622,BDH1,ENSG00000161267,3-hydroxybutyrate dehydrogenase 1,protein-coding,NCBI,landmark
+6251,RSU1,ENSG00000148484,Ras suppressor protein 1,protein-coding,NCBI,landmark
+6253,RTN2,ENSG00000125744,reticulon 2,protein-coding,NCBI,landmark
+6275,S100A4,ENSG00000196154,S100 calcium binding protein A4,protein-coding,NCBI,landmark
+6284,S100A13,ENSG00000189171,S100 calcium binding protein A13,protein-coding,NCBI,landmark
+6304,SATB1,ENSG00000182568,SATB homeobox 1,protein-coding,NCBI,landmark
+6342,SCP2,ENSG00000116171,sterol carrier protein 2,protein-coding,NCBI,landmark
+6347,CCL2,ENSG00000108691,C-C motif chemokine ligand 2,protein-coding,NCBI,landmark
+637,BID,ENSG00000015475,BH3 interacting domain death agonist,protein-coding,NCBI,landmark
+63874,ABHD4,ENSG00000100439,abhydrolase domain containing 4,protein-coding,NCBI,landmark
+6390,SDHB,ENSG00000117118,succinate dehydrogenase complex iron sulfur subunit B,protein-coding,NCBI,landmark
+63933,MCUR1,ENSG00000050393,mitochondrial calcium uniporter regulator 1,protein-coding,NCBI,landmark
+64080,RBKS,ENSG00000171174,ribokinase,protein-coding,NCBI,landmark
+642,BLMH,ENSG00000108578,bleomycin hydrolase,protein-coding,NCBI,landmark
+644,BLVRA,ENSG00000106605,biliverdin reductase A,protein-coding,NCBI,landmark
+64422,ATG3,ENSG00000144848,autophagy related 3,protein-coding,NCBI,landmark
+64428,CIAO3,ENSG00000103245,cytosolic iron-sulfur assembly component 3,protein-coding,NCBI,landmark
+64429,ZDHHC6,ENSG00000023041,zinc finger DHHC-type containing 6,protein-coding,NCBI,landmark
+6443,SGCB,ENSG00000163069,sarcoglycan beta,protein-coding,NCBI,landmark
+6461,SHB,ENSG00000107338,SH2 domain containing adaptor protein B,protein-coding,NCBI,landmark
+6464,SHC1,ENSG00000160691,SHC adaptor protein 1,protein-coding,NCBI,landmark
+64746,ACBD3,ENSG00000182827,acyl-CoA binding domain containing 3,protein-coding,NCBI,landmark
+64781,CERK,ENSG00000100422,ceramide kinase,protein-coding,NCBI,landmark
+64943,NT5DC2,ENSG00000168268,5'-nucleotidase domain containing 2,protein-coding,NCBI,landmark
+6499,SKIV2L,ENSG00000204351,Ski2 like RNA helicase,protein-coding,NCBI,landmark
+6500,SKP1,ENSG00000113558,S-phase kinase associated protein 1,protein-coding,NCBI,landmark
+65057,ACD,ENSG00000102977,ACD shelterin complex subunit and telomerase recruitment factor,protein-coding,NCBI,landmark
+6509,SLC1A4,ENSG00000115902,solute carrier family 1 member 4,protein-coding,NCBI,landmark
+65123,INTS3,ENSG00000143624,integrator complex subunit 3,protein-coding,NCBI,landmark
+652,BMP4,ENSG00000125378,bone morphogenetic protein 4,protein-coding,NCBI,landmark
+6597,SMARCA4,ENSG00000127616,"SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily a, member 4",protein-coding,NCBI,landmark
+6599,SMARCC1,ENSG00000173473,"SWI/SNF related, matrix associated, actin dependent regulator of chromatin subfamily c member 1",protein-coding,NCBI,landmark
+66008,TRAK2,ENSG00000115993,trafficking kinesin protein 2,protein-coding,NCBI,landmark
+6603,SMARCD2,ENSG00000108604,"SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily d, member 2",protein-coding,NCBI,landmark
+6616,SNAP25,ENSG00000132639,synaptosome associated protein 25,protein-coding,NCBI,landmark
+6622,SNCA,ENSG00000145335,synuclein alpha,protein-coding,NCBI,landmark
+664,BNIP3,ENSG00000176171,BCL2 interacting protein 3,protein-coding,NCBI,landmark
+665,BNIP3L,ENSG00000104765,BCL2 interacting protein 3 like,protein-coding,NCBI,landmark
+6657,SOX2,ENSG00000181449,SRY-box 2,protein-coding,NCBI,landmark
+6659,SOX4,ENSG00000124766,SRY-box 4,protein-coding,NCBI,landmark
+6676,SPAG4,ENSG00000061656,sperm associated antigen 4,protein-coding,NCBI,landmark
+6696,SPP1,ENSG00000118785,secreted phosphoprotein 1,protein-coding,NCBI,landmark
+6697,SPR,ENSG00000116096,sepiapterin reductase,protein-coding,NCBI,landmark
+670,BPHL,ENSG00000137274,biphenyl hydrolase like,protein-coding,NCBI,landmark
+6709,SPTAN1,ENSG00000197694,"spectrin alpha, non-erythrocytic 1",protein-coding,NCBI,landmark
+6714,SRC,ENSG00000197122,"SRC proto-oncogene, non-receptor tyrosine kinase",protein-coding,NCBI,landmark
+672,BRCA1,ENSG00000012048,BRCA1 DNA repair associated,protein-coding,NCBI,landmark
+6772,STAT1,ENSG00000115415,signal transducer and activator of transcription 1,protein-coding,NCBI,landmark
+6774,STAT3,ENSG00000168610,signal transducer and activator of transcription 3,protein-coding,NCBI,landmark
+6777,STAT5B,ENSG00000173757,signal transducer and activator of transcription 5B,protein-coding,NCBI,landmark
+6790,AURKA,ENSG00000087586,aurora kinase A,protein-coding,NCBI,landmark
+6793,STK10,ENSG00000072786,serine/threonine kinase 10,protein-coding,NCBI,landmark
+6804,STX1A,ENSG00000106089,syntaxin 1A,protein-coding,NCBI,landmark
+6810,STX4,ENSG00000103496,syntaxin 4,protein-coding,NCBI,landmark
+6812,STXBP1,ENSG00000136854,syntaxin binding protein 1,protein-coding,NCBI,landmark
+6813,STXBP2,ENSG00000076944,syntaxin binding protein 2,protein-coding,NCBI,landmark
+6832,SUPV3L1,ENSG00000156502,Suv3 like RNA helicase,protein-coding,NCBI,landmark
+6839,SUV39H1,ENSG00000101945,suppressor of variegation 3-9 homolog 1,protein-coding,NCBI,landmark
+6850,SYK,ENSG00000165025,spleen associated tyrosine kinase,protein-coding,NCBI,landmark
+6856,SYPL1,ENSG00000008282,synaptophysin like 1,protein-coding,NCBI,landmark
+6894,TARBP1,ENSG00000059588,TAR (HIV-1) RNA binding protein 1,protein-coding,NCBI,landmark
+6908,TBP,ENSG00000112592,TATA-box binding protein,protein-coding,NCBI,landmark
+6909,TBX2,ENSG00000121068,T-box 2,protein-coding,NCBI,landmark
+6915,TBXA2R,ENSG00000006638,thromboxane A2 receptor,protein-coding,NCBI,landmark
+6919,TCEA2,ENSG00000171703,transcription elongation factor A2,protein-coding,NCBI,landmark
+6944,VPS72,ENSG00000163159,vacuolar protein sorting 72 homolog,protein-coding,NCBI,landmark
+695,BTK,ENSG00000010671,Bruton tyrosine kinase,protein-coding,NCBI,landmark
+6988,TCTA,ENSG00000145022,T cell leukemia translocation altered,protein-coding,NCBI,landmark
+6990,DYNLT3,ENSG00000165169,dynein light chain Tctex-type 3,protein-coding,NCBI,landmark
+701,BUB1B,ENSG00000156970,BUB1 mitotic checkpoint serine/threonine kinase B,protein-coding,NCBI,landmark
+7015,TERT,ENSG00000164362,telomerase reverse transcriptase,protein-coding,NCBI,landmark
+7016,TESK1,ENSG00000107140,testis associated actin remodelling kinase 1,protein-coding,NCBI,landmark
+7020,TFAP2A,ENSG00000137203,transcription factor AP-2 alpha,protein-coding,NCBI,landmark
+7027,TFDP1,ENSG00000198176,transcription factor Dp-1,protein-coding,NCBI,landmark
+7043,TGFB3,ENSG00000119699,transforming growth factor beta 3,protein-coding,NCBI,landmark
+7048,TGFBR2,ENSG00000163513,transforming growth factor beta receptor 2,protein-coding,NCBI,landmark
+7074,TIAM1,ENSG00000156299,T cell lymphoma invasion and metastasis 1,protein-coding,NCBI,landmark
+7077,TIMP2,ENSG00000035862,TIMP metallopeptidase inhibitor 2,protein-coding,NCBI,landmark
+7082,TJP1,ENSG00000104067,tight junction protein 1,protein-coding,NCBI,landmark
+7088,TLE1,ENSG00000196781,"TLE family member 1, transcriptional corepressor",protein-coding,NCBI,landmark
+7099,TLR4,ENSG00000136869,toll like receptor 4,protein-coding,NCBI,landmark
+7105,TSPAN6,ENSG00000000003,tetraspanin 6,protein-coding,NCBI,landmark
+7106,TSPAN4,ENSG00000214063,tetraspanin 4,protein-coding,NCBI,landmark
+7153,TOP2A,ENSG00000131747,DNA topoisomerase II alpha,protein-coding,NCBI,landmark
+7157,TP53,ENSG00000141510,tumor protein p53,protein-coding,NCBI,landmark
+7158,TP53BP1,ENSG00000067369,tumor protein p53 binding protein 1,protein-coding,NCBI,landmark
+7159,TP53BP2,ENSG00000143514,tumor protein p53 binding protein 2,protein-coding,NCBI,landmark
+7165,TPD52L2,ENSG00000101150,TPD52 like 2,protein-coding,NCBI,landmark
+7168,TPM1,ENSG00000140416,tropomyosin 1,protein-coding,NCBI,landmark
+7264,TSTA3,ENSG00000104522,tissue specific transplantation antigen P35B,protein-coding,NCBI,landmark
+727,C5,ENSG00000106804,complement C5,protein-coding,NCBI,landmark
+7296,TXNRD1,ENSG00000198431,thioredoxin reductase 1,protein-coding,NCBI,landmark
+7319,UBE2A,ENSG00000077721,ubiquitin conjugating enzyme E2 A,protein-coding,NCBI,landmark
+7358,UGDH,ENSG00000109814,UDP-glucose 6-dehydrogenase,protein-coding,NCBI,landmark
+7376,NR1H2,ENSG00000131408,nuclear receptor subfamily 1 group H member 2,protein-coding,NCBI,landmark
+7398,USP1,ENSG00000162607,ubiquitin specific peptidase 1,protein-coding,NCBI,landmark
+7416,VDAC1,ENSG00000213585,voltage dependent anion channel 1,protein-coding,NCBI,landmark
+7466,WFS1,ENSG00000109501,wolframin ER transmembrane glycoprotein,protein-coding,NCBI,landmark
+7485,WRB,ENSG00000182093,tryptophan rich basic protein,protein-coding,NCBI,landmark
+7494,XBP1,ENSG00000100219,X-box binding protein 1,protein-coding,NCBI,landmark
+7511,XPNPEP1,ENSG00000108039,X-prolyl aminopeptidase 1,protein-coding,NCBI,landmark
+7538,ZFP36,ENSG00000128016,ZFP36 ring finger protein,protein-coding,NCBI,landmark
+7690,ZNF131,ENSG00000172262,zinc finger protein 131,protein-coding,NCBI,landmark
+7750,ZMYM2,ENSG00000121741,zinc finger MYM-type containing 2,protein-coding,NCBI,landmark
+780,DDR1,ENSG00000204580,discoidin domain receptor tyrosine kinase 1,protein-coding,NCBI,landmark
+7849,PAX8,ENSG00000125618,paired box 8,protein-coding,NCBI,landmark
+7852,CXCR4,ENSG00000121966,C-X-C motif chemokine receptor 4,protein-coding,NCBI,landmark
+7866,IFRD2,ENSG00000214706,interferon related developmental regulator 2,protein-coding,NCBI,landmark
+7867,MAPKAPK3,ENSG00000114738,mitogen-activated protein kinase-activated protein kinase 3,protein-coding,NCBI,landmark
+7874,USP7,ENSG00000187555,ubiquitin specific peptidase 7,protein-coding,NCBI,landmark
+79006,METRN,ENSG00000103260,"meteorin, glial cell differentiation regulator",protein-coding,NCBI,landmark
+7905,REEP5,ENSG00000129625,receptor accessory protein 5,protein-coding,NCBI,landmark
+79071,ELOVL6,ENSG00000170522,ELOVL fatty acid elongase 6,protein-coding,NCBI,landmark
+79073,TMEM109,ENSG00000110108,transmembrane protein 109,protein-coding,NCBI,landmark
+79080,CCDC86,ENSG00000110104,coiled-coil domain containing 86,protein-coding,NCBI,landmark
+79090,TRAPPC6A,ENSG00000007255,trafficking protein particle complex 6A,protein-coding,NCBI,landmark
+79094,CHAC1,ENSG00000128965,ChaC glutathione specific gamma-glutamylcyclotransferase 1,protein-coding,NCBI,landmark
+79143,MBOAT7,ENSG00000125505,membrane bound O-acyltransferase domain containing 7,protein-coding,NCBI,landmark
+79170,PRR15L,ENSG00000167183,proline rich 15 like,protein-coding,NCBI,landmark
+79174,CRELD2,ENSG00000184164,cysteine rich with EGF like domains 2,protein-coding,NCBI,landmark
+79187,FSD1,ENSG00000105255,fibronectin type III and SPRY domain containing 1,protein-coding,NCBI,landmark
+79600,TCTN1,ENSG00000204852,tectonic family member 1,protein-coding,NCBI,landmark
+79643,CHMP6,ENSG00000176108,charged multivesicular body protein 6,protein-coding,NCBI,landmark
+79716,NPEPL1,ENSG00000215440,aminopeptidase like 1,protein-coding,NCBI,landmark
+7982,ST7,ENSG00000004866,suppression of tumorigenicity 7,protein-coding,NCBI,landmark
+79850,FAM57A,ENSG00000167695,family with sequence similarity 57 member A,protein-coding,NCBI,landmark
+79902,NUP85,ENSG00000125450,nucleoporin 85,protein-coding,NCBI,landmark
+79921,TCEAL4,ENSG00000133142,transcription elongation factor A like 4,protein-coding,NCBI,landmark
+7994,KAT6A,ENSG00000083168,lysine acetyltransferase 6A,protein-coding,NCBI,landmark
+79947,DHDDS,ENSG00000117682,dehydrodolichyl diphosphate synthase subunit,protein-coding,NCBI,landmark
+79961,DENND2D,ENSG00000162777,DENN domain containing 2D,protein-coding,NCBI,landmark
+80204,FBXO11,ENSG00000138081,F-box protein 11,protein-coding,NCBI,landmark
+80212,CCDC92,ENSG00000119242,coiled-coil domain containing 92,protein-coding,NCBI,landmark
+80347,COASY,ENSG00000068120,Coenzyme A synthase,protein-coding,NCBI,landmark
+80349,WDR61,ENSG00000140395,WD repeat domain 61,protein-coding,NCBI,landmark
+8050,PDHX,ENSG00000110435,pyruvate dehydrogenase complex component X,protein-coding,NCBI,landmark
+8061,FOSL1,ENSG00000175592,"FOS like 1, AP-1 transcription factor subunit",protein-coding,NCBI,landmark
+80746,TSEN2,ENSG00000154743,tRNA splicing endonuclease subunit 2,protein-coding,NCBI,landmark
+80758,PRR7,ENSG00000131188,"proline rich 7, synaptic",protein-coding,NCBI,landmark
+808,CALM3,ENSG00000160014,calmodulin 3,protein-coding,NCBI,landmark
+8091,HMGA2,ENSG00000149948,high mobility group AT-hook 2,protein-coding,NCBI,landmark
+813,CALU,ENSG00000128595,calumenin,protein-coding,NCBI,landmark
+81533,ITFG1,ENSG00000129636,integrin alpha FG-GAP repeat containing 1,protein-coding,NCBI,landmark
+81544,GDPD5,ENSG00000158555,glycerophosphodiester phosphodiesterase domain containing 5,protein-coding,NCBI,landmark
+8202,NCOA3,ENSG00000124151,nuclear receptor coactivator 3,protein-coding,NCBI,landmark
+8204,NRIP1,ENSG00000180530,nuclear receptor interacting protein 1,protein-coding,NCBI,landmark
+823,CAPN1,ENSG00000014216,calpain 1,protein-coding,NCBI,landmark
+8243,SMC1A,ENSG00000072501,structural maintenance of chromosomes 1A,protein-coding,NCBI,landmark
+8270,LAGE3,ENSG00000196976,L antigen family member 3,protein-coding,NCBI,landmark
+831,CAST,ENSG00000153113,calpastatin,protein-coding,NCBI,landmark
+8312,AXIN1,ENSG00000103126,axin 1,protein-coding,NCBI,landmark
+8318,CDC45,ENSG00000093009,cell division cycle 45,protein-coding,NCBI,landmark
+8321,FZD1,ENSG00000157240,frizzled class receptor 1,protein-coding,NCBI,landmark
+8324,FZD7,ENSG00000155760,frizzled class receptor 7,protein-coding,NCBI,landmark
+8349,HIST2H2BE,ENSG00000184678,histone cluster 2 H2B family member e,protein-coding,NCBI,landmark
+835,CASP2,ENSG00000106144,caspase 2,protein-coding,NCBI,landmark
+836,CASP3,ENSG00000164305,caspase 3,protein-coding,NCBI,landmark
+83743,GRWD1,ENSG00000105447,glutamate rich WD repeat containing 1,protein-coding,NCBI,landmark
+8396,PIP4K2B,ENSG00000276293,phosphatidylinositol-5-phosphate 4-kinase type 2 beta,protein-coding,NCBI,landmark
+840,CASP7,ENSG00000165806,caspase 7,protein-coding,NCBI,landmark
+84159,ARID5B,ENSG00000150347,AT-rich interaction domain 5B,protein-coding,NCBI,landmark
+843,CASP10,ENSG00000003400,caspase 10,protein-coding,NCBI,landmark
+8440,NCK2,ENSG00000071051,NCK adaptor protein 2,protein-coding,NCBI,landmark
+8444,DYRK3,ENSG00000143479,dual specificity tyrosine phosphorylation regulated kinase 3,protein-coding,NCBI,landmark
+8446,DUSP11,ENSG00000144048,dual specificity phosphatase 11,protein-coding,NCBI,landmark
+84617,TUBB6,ENSG00000176014,tubulin beta 6 class V,protein-coding,NCBI,landmark
+847,CAT,ENSG00000121691,catalase,protein-coding,NCBI,landmark
+84722,PSRC1,ENSG00000134222,proline and serine rich coiled-coil 1,protein-coding,NCBI,landmark
+8480,RAE1,ENSG00000101146,ribonucleic acid export 1,protein-coding,NCBI,landmark
+84890,ADO,ENSG00000181915,2-aminoethanethiol dioxygenase,protein-coding,NCBI,landmark
+8503,PIK3R3,ENSG00000117461,phosphoinositide-3-kinase regulatory subunit 3,protein-coding,NCBI,landmark
+8508,NIPSNAP1,ENSG00000184117,nipsnap homolog 1,protein-coding,NCBI,landmark
+8518,ELP1,ENSG00000070061,elongator complex protein 1,protein-coding,NCBI,landmark
+8520,HAT1,ENSG00000128708,histone acetyltransferase 1,protein-coding,NCBI,landmark
+85236,HIST1H2BK,ENSG00000197903,histone cluster 1 H2B family member k,protein-coding,NCBI,landmark
+85377,MICALL1,ENSG00000100139,MICAL like 1,protein-coding,NCBI,landmark
+8550,MAPKAPK5,ENSG00000089022,mitogen-activated protein kinase-activated protein kinase 5,protein-coding,NCBI,landmark
+8553,BHLHE40,ENSG00000134107,basic helix-loop-helix family member e40,protein-coding,NCBI,landmark
+8569,MKNK1,ENSG00000079277,MAP kinase interacting serine/threonine kinase 1,protein-coding,NCBI,landmark
+8573,CASK,ENSG00000147044,calcium/calmodulin dependent serine protein kinase,protein-coding,NCBI,landmark
+8574,AKR7A2,ENSG00000053371,aldo-keto reductase family 7 member A2,protein-coding,NCBI,landmark
+8607,RUVBL1,ENSG00000175792,RuvB like AAA ATPase 1,protein-coding,NCBI,landmark
+8624,PSMG1,ENSG00000183527,proteasome assembly chaperone 1,protein-coding,NCBI,landmark
+8678,BECN1,ENSG00000126581,beclin 1,protein-coding,NCBI,landmark
+868,CBLB,ENSG00000114423,Cbl proto-oncogene B,protein-coding,NCBI,landmark
+8720,MBTPS1,ENSG00000140943,"membrane bound transcription factor peptidase, site 1",protein-coding,NCBI,landmark
+8726,EED,ENSG00000074266,embryonic ectoderm development,protein-coding,NCBI,landmark
+8727,CTNNAL1,ENSG00000119326,catenin alpha like 1,protein-coding,NCBI,landmark
+873,CBR1,ENSG00000159228,carbonyl reductase 1,protein-coding,NCBI,landmark
+8731,RNMT,ENSG00000101654,RNA guanine-7 methyltransferase,protein-coding,NCBI,landmark
+874,CBR3,ENSG00000159231,carbonyl reductase 3,protein-coding,NCBI,landmark
+8800,PEX11A,ENSG00000166821,peroxisomal biogenesis factor 11 alpha,protein-coding,NCBI,landmark
+8804,CREG1,ENSG00000143162,cellular repressor of E1A stimulated genes 1,protein-coding,NCBI,landmark
+8821,INPP4B,ENSG00000109452,inositol polyphosphate-4-phosphatase type II B,protein-coding,NCBI,landmark
+8826,IQGAP1,ENSG00000140575,IQ motif containing GTPase activating protein 1,protein-coding,NCBI,landmark
+8835,SOCS2,ENSG00000120833,suppressor of cytokine signaling 2,protein-coding,NCBI,landmark
+8837,CFLAR,ENSG00000003402,CASP8 and FADD like apoptosis regulator,protein-coding,NCBI,landmark
+8851,CDK5R1,ENSG00000176749,cyclin dependent kinase 5 regulatory subunit 1,protein-coding,NCBI,landmark
+8869,ST3GAL5,ENSG00000115525,"ST3 beta-galactoside alpha-2,3-sialyltransferase 5",protein-coding,NCBI,landmark
+8870,IER3,ENSG00000137331,immediate early response 3,protein-coding,NCBI,landmark
+8878,SQSTM1,ENSG00000161011,sequestosome 1,protein-coding,NCBI,landmark
+8884,SLC5A6,ENSG00000138074,solute carrier family 5 member 6,protein-coding,NCBI,landmark
+8895,CPNE3,ENSG00000085719,copine 3,protein-coding,NCBI,landmark
+890,CCNA2,ENSG00000145386,cyclin A2,protein-coding,NCBI,landmark
+8900,CCNA1,ENSG00000133101,cyclin A1,protein-coding,NCBI,landmark
+891,CCNB1,ENSG00000134057,cyclin B1,protein-coding,NCBI,landmark
+8914,TIMELESS,ENSG00000111602,timeless circadian regulator,protein-coding,NCBI,landmark
+896,CCND3,ENSG00000112576,cyclin D3,protein-coding,NCBI,landmark
+8974,P4HA2,ENSG00000072682,prolyl 4-hydroxylase subunit alpha 2,protein-coding,NCBI,landmark
+8985,PLOD3,ENSG00000106397,"procollagen-lysine,2-oxoglutarate 5-dioxygenase 3",protein-coding,NCBI,landmark
+899,CCNF,ENSG00000162063,cyclin F,protein-coding,NCBI,landmark
+89910,UBE3B,ENSG00000151148,ubiquitin protein ligase E3B,protein-coding,NCBI,landmark
+8996,NOL3,ENSG00000140939,nucleolar protein 3,protein-coding,NCBI,landmark
+9016,SLC25A14,ENSG00000102078,solute carrier family 25 member 14,protein-coding,NCBI,landmark
+9019,MPZL1,ENSG00000197965,myelin protein zero like 1,protein-coding,NCBI,landmark
+902,CCNH,ENSG00000134480,cyclin H,protein-coding,NCBI,landmark
+9053,MAP7,ENSG00000135525,microtubule associated protein 7,protein-coding,NCBI,landmark
+90861,JPT2,ENSG00000206053,Jupiter microtubule associated homolog 2,protein-coding,NCBI,landmark
+9093,DNAJA3,ENSG00000103423,DnaJ heat shock protein family (Hsp40) member A3,protein-coding,NCBI,landmark
+9097,USP14,ENSG00000101557,ubiquitin specific peptidase 14,protein-coding,NCBI,landmark
+9112,MTA1,ENSG00000182979,metastasis associated 1,protein-coding,NCBI,landmark
+91137,SLC25A46,ENSG00000164209,solute carrier family 25 member 46,protein-coding,NCBI,landmark
+9124,PDLIM1,ENSG00000107438,PDZ and LIM domain 1,protein-coding,NCBI,landmark
+9126,SMC3,ENSG00000108055,structural maintenance of chromosomes 3,protein-coding,NCBI,landmark
+9128,PRPF4,ENSG00000136875,pre-mRNA processing factor 4,protein-coding,NCBI,landmark
+9133,CCNB2,ENSG00000157456,cyclin B2,protein-coding,NCBI,landmark
+9134,CCNE2,ENSG00000175305,cyclin E2,protein-coding,NCBI,landmark
+9143,SYNGR3,ENSG00000127561,synaptogyrin 3,protein-coding,NCBI,landmark
+9170,LPAR2,ENSG00000064547,lysophosphatidic acid receptor 2,protein-coding,NCBI,landmark
+9181,ARHGEF2,ENSG00000116584,Rho/Rac guanine nucleotide exchange factor 2,protein-coding,NCBI,landmark
+9183,ZW10,ENSG00000086827,zw10 kinetochore protein,protein-coding,NCBI,landmark
+91949,COG7,ENSG00000168434,component of oligomeric golgi complex 7,protein-coding,NCBI,landmark
+9212,AURKB,ENSG00000178999,aurora kinase B,protein-coding,NCBI,landmark
+9217,VAPB,ENSG00000124164,VAMP associated protein B and C,protein-coding,NCBI,landmark
+9221,NOLC1,ENSG00000166197,nucleolar and coiled-body phosphoprotein 1,protein-coding,NCBI,landmark
+9246,UBE2L6,ENSG00000156587,ubiquitin conjugating enzyme E2 L6,protein-coding,NCBI,landmark
+9261,MAPKAPK2,ENSG00000162889,mitogen-activated protein kinase-activated protein kinase 2,protein-coding,NCBI,landmark
+9267,CYTH1,ENSG00000108669,cytohesin 1,protein-coding,NCBI,landmark
+9270,ITGB1BP1,ENSG00000119185,integrin subunit beta 1 binding protein 1,protein-coding,NCBI,landmark
+9275,BCL7B,ENSG00000106635,BAF chromatin remodeling complex subunit BCL7B,protein-coding,NCBI,landmark
+9276,COPB2,ENSG00000184432,coatomer protein complex subunit beta 2,protein-coding,NCBI,landmark
+9289,ADGRG1,ENSG00000205336,adhesion G protein-coupled receptor G1,protein-coding,NCBI,landmark
+93487,MAPK1IP1L,ENSG00000168175,mitogen-activated protein kinase 1 interacting protein 1 like,protein-coding,NCBI,landmark
+93594,TBC1D31,ENSG00000156787,TBC1 domain family member 31,protein-coding,NCBI,landmark
+9375,TM9SF2,ENSG00000125304,transmembrane 9 superfamily member 2,protein-coding,NCBI,landmark
+94239,H2AFV,ENSG00000105968,H2A histone family member V,protein-coding,NCBI,landmark
+9448,MAP4K4,ENSG00000071054,mitogen-activated protein kinase kinase kinase kinase 4,protein-coding,NCBI,landmark
+9455,HOMER2,ENSG00000103942,homer scaffold protein 2,protein-coding,NCBI,landmark
+9467,SH3BP5,ENSG00000131370,SH3 domain binding protein 5,protein-coding,NCBI,landmark
+9488,PIGB,ENSG00000069943,phosphatidylinositol glycan anchor biosynthesis class B,protein-coding,NCBI,landmark
+949,SCARB1,ENSG00000073060,scavenger receptor class B member 1,protein-coding,NCBI,landmark
+9491,PSMF1,ENSG00000125818,proteasome inhibitor subunit 1,protein-coding,NCBI,landmark
+9517,SPTLC2,ENSG00000100596,serine palmitoyltransferase long chain base subunit 2,protein-coding,NCBI,landmark
+9519,TBPL1,ENSG00000028839,TATA-box binding protein like 1,protein-coding,NCBI,landmark
+9531,BAG3,ENSG00000151929,BCL2 associated athanogene 3,protein-coding,NCBI,landmark
+9533,POLR1C,ENSG00000171453,RNA polymerase I and III subunit C,protein-coding,NCBI,landmark
+9552,SPAG7,ENSG00000091640,sperm associated antigen 7,protein-coding,NCBI,landmark
+958,CD40,ENSG00000101017,CD40 molecule,protein-coding,NCBI,landmark
+960,CD44,ENSG00000026508,CD44 molecule (Indian blood group),protein-coding,NCBI,landmark
+9637,FEZ2,ENSG00000171055,fasciculation and elongation protein zeta 2,protein-coding,NCBI,landmark
+9641,IKBKE,ENSG00000263528,inhibitor of nuclear factor kappa B kinase subunit epsilon,protein-coding,NCBI,landmark
+965,CD58,ENSG00000116815,CD58 molecule,protein-coding,NCBI,landmark
+9650,MTFR1,ENSG00000066855,mitochondrial fission regulator 1,protein-coding,NCBI,landmark
+9653,HS2ST1,ENSG00000153936,heparan sulfate 2-O-sulfotransferase 1,protein-coding,NCBI,landmark
+9670,IPO13,ENSG00000117408,importin 13,protein-coding,NCBI,landmark
+9686,VGLL4,ENSG00000144560,vestigial like family member 4,protein-coding,NCBI,landmark
+9688,NUP93,ENSG00000102900,nucleoporin 93,protein-coding,NCBI,landmark
+9690,UBE3C,ENSG00000009335,ubiquitin protein ligase E3C,protein-coding,NCBI,landmark
+9695,EDEM1,ENSG00000134109,ER degradation enhancing alpha-mannosidase like protein 1,protein-coding,NCBI,landmark
+9697,TRAM2,ENSG00000065308,translocation associated membrane protein 2,protein-coding,NCBI,landmark
+9702,CEP57,ENSG00000166037,centrosomal protein 57,protein-coding,NCBI,landmark
+9703,KIAA0100,ENSG00000007202,KIAA0100,protein-coding,NCBI,landmark
+9709,HERPUD1,ENSG00000051108,homocysteine inducible ER protein with ubiquitin like domain 1,protein-coding,NCBI,landmark
+9710,KIAA0355,ENSG00000166398,KIAA0355,protein-coding,NCBI,landmark
+9712,USP6NL,ENSG00000148429,USP6 N-terminal like,protein-coding,NCBI,landmark
+9738,CCP110,ENSG00000103540,centriolar coiled-coil protein 110,protein-coding,NCBI,landmark
+976,ADGRE5,ENSG00000123146,adhesion G protein-coupled receptor E5,protein-coding,NCBI,landmark
+9761,MLEC,ENSG00000110917,malectin,protein-coding,NCBI,landmark
+9797,TATDN2,ENSG00000157014,TatD DNase domain containing 2,protein-coding,NCBI,landmark
+9801,MRPL19,ENSG00000115364,mitochondrial ribosomal protein L19,protein-coding,NCBI,landmark
+9805,SCRN1,ENSG00000136193,secernin 1,protein-coding,NCBI,landmark
+9813,EFCAB14,ENSG00000159658,EF-hand calcium binding domain 14,protein-coding,NCBI,landmark
+9817,KEAP1,ENSG00000079999,kelch like ECH associated protein 1,protein-coding,NCBI,landmark
+983,CDK1,ENSG00000170312,cyclin dependent kinase 1,protein-coding,NCBI,landmark
+9833,MELK,ENSG00000165304,maternal embryonic leucine zipper kinase,protein-coding,NCBI,landmark
+9842,PLEKHM1,ENSG00000225190,pleckstrin homology and RUN domain containing M1,protein-coding,NCBI,landmark
+9847,C2CD5,ENSG00000111731,C2 calcium dependent domain containing 5,protein-coding,NCBI,landmark
+9851,KIAA0753,ENSG00000198920,KIAA0753,protein-coding,NCBI,landmark
+9854,C2CD2L,ENSG00000172375,C2CD2 like,protein-coding,NCBI,landmark
+9868,TOMM70,ENSG00000154174,translocase of outer mitochondrial membrane 70,protein-coding,NCBI,landmark
+9897,WASHC5,ENSG00000164961,WASH complex subunit 5,protein-coding,NCBI,landmark
+9903,KLHL21,ENSG00000162413,kelch like family member 21,protein-coding,NCBI,landmark
+991,CDC20,ENSG00000117399,cell division cycle 20,protein-coding,NCBI,landmark
+9915,ARNT2,ENSG00000172379,aryl hydrocarbon receptor nuclear translocator 2,protein-coding,NCBI,landmark
+9917,FAM20B,ENSG00000116199,FAM20B glycosaminoglycan xylosylkinase,protein-coding,NCBI,landmark
+9918,NCAPD2,ENSG00000010292,non-SMC condensin I complex subunit D2,protein-coding,NCBI,landmark
+9924,PAN2,ENSG00000135473,poly(A) specific ribonuclease subunit PAN2,protein-coding,NCBI,landmark
+9926,LPGAT1,ENSG00000123684,lysophosphatidylglycerol acyltransferase 1,protein-coding,NCBI,landmark
+9928,KIF14,ENSG00000118193,kinesin family member 14,protein-coding,NCBI,landmark
+993,CDC25A,ENSG00000164045,cell division cycle 25A,protein-coding,NCBI,landmark
+994,CDC25B,ENSG00000101224,cell division cycle 25B,protein-coding,NCBI,landmark
+9943,OXSR1,ENSG00000172939,oxidative stress responsive kinase 1,protein-coding,NCBI,landmark
+9961,MVP,ENSG00000013364,major vault protein,protein-coding,NCBI,landmark
+998,CDC42,ENSG00000070831,cell division cycle 42,protein-coding,NCBI,landmark
+9988,DMTF1,ENSG00000135164,cyclin D binding myb like transcription factor 1,protein-coding,NCBI,landmark

LINCS2020/lincs978_to965.csv

@@ -0,0 +1,966 @@
+gene_symbol
+GNPDA1
+CDH3
+HDAC6
+PARP2
+MAMLD1
+DNAJB6
+SMC4
+ABCC5
+ABCB6
+DNM1L
+TSPAN3
+KIF20A
+ARL4C
+TRAP1
+G3BP1
+MBNL2
+CEBPZ
+SLC25A13
+CDK2
+SORBS3
+RBM6
+CDK4
+TXNDC9
+ADAM10
+TRIM13
+CDK6
+CDK7
+TRIB1
+MFSD10
+SLC35B1
+TIMM17B
+CDKN1A
+CDKN1B
+AKAP8
+STUB1
+NET1
+SMNDC1
+CDKN2A
+PAK4
+TNIP1
+IKZF1
+RXYLT1
+HMG20B
+MYL9
+LYPLA1
+PPIE
+VAV3
+LRRC41
+CRTAP
+VAT1
+STK25
+CEBPA
+APPBP2
+CEBPD
+CHERP
+HYOU1
+RPP38
+SLC35A1
+DRAP1
+PAICS
+ST6GALNAC2
+STAMBP
+CENPE
+NPRL2
+IGF2BP2
+YKT6
+CGRRF1
+RRAGA
+GNB5
+EBP
+CNPY3
+CETN3
+YME1L1
+TCFL5
+KDM5B
+POP4
+ARPP19
+ZNF274
+MTHFD2
+WASF3
+UTP14A
+FRS2
+CLPX
+PGRMC1
+MALT1
+CPSF4
+BLCAP
+TCERG1
+RNPS1
+TOMM34
+PDIA5
+MLLT11
+EBNA1BP2
+TMED10
+ASCC3
+SLC27A3
+KIF2C
+CCDC85B
+TLK2
+KDELR2
+RAB31
+B4GAT1
+TENT4A
+UBE2C
+DUSP14
+TOPBP1
+PRSS23
+CHEK1
+PWP1
+PKIG
+CORO1A
+LSM6
+PSIP1
+SLC2A6
+NISCH
+CHEK2
+CHN1
+PRAF2
+POLG2
+CHP1
+PNKP
+ECD
+DDX42
+TWF2
+CIRBP
+RPL39L
+CLTB
+CLTC
+CANT1
+COL1A1
+ADH5
+COL4A1
+CREB1
+CRK
+CRKL
+PARP1
+CRYZ
+CSK
+CSNK1A1
+CSNK1E
+CSNK2A2
+CSRP1
+WIPF2
+TICAM1
+CTNND1
+CTSD
+CTSL
+CYB561
+ADRB2
+DAG1
+DAXX
+DCK
+DCTD
+DDB2
+GADD45A
+DDX10
+DECR1
+DFFA
+DFFB
+DLD
+DNM1
+AGL
+DNMT1
+DNMT3A
+DPH2
+DSG2
+TSC22D3
+DUSP3
+DUSP4
+DUSP6
+TOR1A
+E2F2
+ECH1
+EDN1
+EGF
+EGFR
+EGR1
+EIF4EBP1
+EIF4G1
+EIF5
+ELAVL1
+TXLNA
+SPRED2
+CTTN
+EPB41L2
+EPHA3
+EPHB2
+NR2F6
+ERBB2
+ERBB3
+AKT1
+ETFB
+ALAS1
+ETS1
+ETV1
+EXT1
+EZH2
+FAH
+PTK2B
+FAT1
+FDFT1
+ALDOA
+FGFR2
+FGFR4
+FHL2
+CASC3
+COG2
+ATF5
+MTF2
+PUF60
+RAB11FIP2
+FKBP4
+CLSTN1
+FOXJ3
+KHDC4
+EPN2
+SACM1L
+ATF6
+RPIA
+ABCF1
+ALDOC
+RAB21
+SPEN
+FBXO21
+RBM34
+WDTC1
+XPO7
+TBC1D9B
+RRP1B
+MYCBP2
+FOXO3
+CDK19
+GPATCH8
+MAST2
+DCUN1D4
+FCHO1
+SNX13
+ATP11B
+JMJD6
+RRS1
+RRP12
+SYNE2
+PDS5A
+CAMSAP2
+ATMIN
+TRIM2
+WASHC4
+USP22
+WDR7
+JADE2
+ARHGEF12
+PPP1R13B
+RRP8
+NUDCD3
+SIRT3
+SLC35A3
+ICMT
+MACF1
+SUZ12
+KAT6B
+FOS
+NNT
+ADAT1
+FPGS
+TMEM50A
+KLHDC2
+ACOT9
+SSBP2
+NUP62
+ARFIP2
+LSM5
+PLA2G15
+CEMIP2
+ZNF318
+ABL1
+FUT1
+FYN
+SLC37A4
+GAA
+GABPB1
+EML3
+FBXO7
+SPDEF
+BAMBI
+GALE
+BACE2
+COG4
+MPC2
+CLIC4
+C2CD2
+GAPDH
+TIPARP
+TSKU
+RNF167
+LRP10
+ZNF451
+SENP6
+RAI14
+TES
+PHGDH
+GATA2
+GATA3
+MYCBP
+CHIC2
+TIMM9
+GFPT1
+GHR
+AKAP8L
+ATP2C1
+TRAPPC3
+DMAC2L
+TNFRSF21
+SESN1
+HTATSF1
+TMEM97
+GLI2
+GLRX
+GNA11
+GNA15
+GNAI1
+GNAI2
+GNAS
+SFN
+GPC1
+GPER1
+GRB7
+GRB10
+GRN
+BZW2
+NR3C1
+CHMP4A
+GTPBP8
+SLC25A4
+DNAJC15
+CXCL2
+GSTM2
+GSTZ1
+MSH6
+GTF2A2
+GTF2E2
+PACSIN3
+RBM15B
+HOOK2
+SNX11
+TIMM22
+NENF
+UBQLN2
+ACAA1
+ERO1A
+HSD17B10
+HADH
+HDAC2
+DNTTIP2
+PIK3R4
+HIF1A
+HK1
+ANXA7
+HLA-DMA
+HLA-DRA
+HMGCR
+HMGCS1
+HMOX1
+HOXA5
+HOXA10
+APBB2
+HPRT1
+HES1
+BIRC2
+DNAJB2
+HSPA1A
+HSPA4
+HSPA8
+HSPB1
+BIRC5
+HSPD1
+DNAJB1
+ICAM1
+ICAM3
+ID2
+IDE
+IFNAR1
+APOE
+IGF1R
+IGF2R
+IGFBP3
+IGHMBP2
+APP
+FAS
+IKBKB
+IL1B
+IL4R
+IL13RA1
+ILK
+INPP1
+INSIG1
+ITGAE
+ITGB5
+JUN
+STIMATE
+KCNK1
+KIF5C
+KIT
+RHOA
+DIPK1A
+KTN1
+ACAT2
+LAMA3
+ARHGAP1
+STMN1
+LBR
+LGALS8
+LIG1
+LIPA
+LOXL1
+LRPAP1
+LYN
+SMAD3
+MAN2B1
+MAT2A
+MBNL1
+MCM3
+ME2
+MEF2C
+MAP3K4
+MEST
+ASAH1
+MIF
+FOXO4
+MMP1
+MMP2
+MNAT1
+MSRA
+MUC1
+MYBL2
+MYC
+GADD45B
+MYLK
+MYO10
+ATF1
+NCK1
+ACLY
+NFATC3
+NFATC4
+NFE2L2
+NFIL3
+NFKB2
+NFKBIA
+NFKBIB
+NFKBIE
+ATP1B1
+NIT1
+NMT1
+NOS3
+CNOT4
+NOTCH1
+PNP
+NPC1
+SLC11A2
+NRAS
+NUCB2
+NUP88
+NVL
+ORC1
+ALDH7A1
+OXA1L
+OXCT1
+PAFAH1B1
+PAFAH1B3
+SERPINE1
+PAK1
+HDGFL3
+COPS7A
+NSDHL
+HEBP1
+PCBD1
+PCCB
+MTERF3
+AMDHD2
+ISOC1
+MRPS16
+FIS1
+GOLT1B
+GLOD4
+GMNN
+LAP3
+PCK2
+NOSIP
+DERA
+PCM1
+SCCPDH
+PCMT1
+PCNA
+MRPS2
+VPS28
+HSD17B11
+NUSAP1
+SCAND1
+CD320
+NGRN
+SNX7
+ATP6V1D
+ZNF589
+PRKAG2
+UBE2J1
+EVL
+HACD3
+PDGFA
+UFM1
+LSR
+DHRS7
+CAB39
+ARID4B
+PFKL
+PGAM1
+PGM1
+PHKA1
+PHKB
+PHKG2
+PIK3C2B
+PIK3C3
+PIK3CA
+PIN1
+PLA2G4A
+ATP6V0B
+PLCB3
+NUDT9
+PLK1
+PLP2
+PLS1
+PLSCR1
+PMAIP1
+PMM2
+CYCS
+POLB
+POLE2
+POLR2I
+TERF2IP
+POLR2K
+GFOD1
+KCTD5
+TMCO1
+DHX29
+EXOSC4
+DDIT4
+PAF1
+PPARD
+PPARG
+P4HTM
+SLC35F2
+PPIC
+ZNF586
+FBXL12
+TEX10
+YTHDF1
+TXNL4B
+PPOX
+HERC6
+PIH1D1
+PPP2R3C
+FKBP14
+CDCA4
+PLEKHJ1
+HEATR1
+ANO10
+UBR7
+FAIM
+PPP2R5A
+ADI1
+PPP2R5E
+ABCF3
+PRCP
+ENOSF1
+CARMIL1
+ANKRD10
+STAP2
+PRKACA
+IARS2
+NUP133
+CNDP2
+MINDY1
+PRKCD
+KDM3A
+PECR
+PRKCH
+EAPP
+CISD1
+PRKCQ
+ZNF395
+KLHL9
+MAPK9
+MAPK13
+MAP2K5
+PRKX
+PROS1
+LGMN
+HTRA1
+NPDC1
+TM9SF3
+PAK6
+DUSP22
+PSMB8
+PSMB10
+COQ8A
+CIAPIN1
+PLSCR3
+PSMD2
+PSMD4
+SCYL3
+LYRM1
+PSMD9
+PSMD10
+ZMIZ1
+MCOLN1
+BAD
+PSME1
+PSME2
+THAP11
+ABHD6
+PTGS2
+PTK2
+PTPN1
+TRIB3
+PTPN6
+POLD4
+PTPN12
+PTPRC
+PTPRF
+PTPRK
+BAX
+PXMP2
+PXN
+PYCR1
+PYGL
+SQOR
+ENOPH1
+PRUNE1
+SNX6
+RAB4A
+RAB27A
+RAC2
+RAD9A
+RAD51C
+MOK
+RALA
+RALB
+RALGDS
+RAP1GAP
+RASA1
+RB1
+KDM5A
+CCND1
+BCL2
+RELB
+RFC2
+RFC5
+RFNG
+RFX5
+RGS2
+RHEB
+FASTKD5
+RNH1
+ELAC2
+RPA1
+RPA2
+RPA3
+MRPL12
+RPN1
+RPS5
+RPS6
+RPS6KA1
+BDH1
+RSU1
+RTN2
+S100A4
+S100A13
+SATB1
+SCP2
+CCL2
+BID
+ABHD4
+SDHB
+MCUR1
+RBKS
+BLMH
+BLVRA
+ATG3
+CIAO3
+ZDHHC6
+SGCB
+SHB
+SHC1
+ACBD3
+CERK
+NT5DC2
+SKP1
+ACD
+SLC1A4
+INTS3
+BMP4
+SMARCA4
+SMARCC1
+TRAK2
+SMARCD2
+SNAP25
+SNCA
+BNIP3
+BNIP3L
+SOX2
+SOX4
+SPAG4
+SPP1
+SPR
+BPHL
+SPTAN1
+SRC
+BRCA1
+STAT1
+STAT3
+STAT5B
+AURKA
+STK10
+STX1A
+STX4
+STXBP1
+STXBP2
+SUPV3L1
+SUV39H1
+SYK
+SYPL1
+TARBP1
+TBP
+TBX2
+TBXA2R
+TCEA2
+VPS72
+BTK
+TCTA
+DYNLT3
+BUB1B
+TERT
+TESK1
+TFAP2A
+TFDP1
+TGFB3
+TGFBR2
+TIAM1
+TIMP2
+TJP1
+TLE1
+TLR4
+TSPAN6
+TSPAN4
+TOP2A
+TP53
+TP53BP1
+TP53BP2
+TPD52L2
+TPM1
+C5
+TXNRD1
+UBE2A
+UGDH
+NR1H2
+USP1
+VDAC1
+WFS1
+XBP1
+XPNPEP1
+ZFP36
+ZNF131
+ZMYM2
+DDR1
+PAX8
+CXCR4
+IFRD2
+MAPKAPK3
+USP7
+METRN
+REEP5
+ELOVL6
+TMEM109
+CCDC86
+TRAPPC6A
+CHAC1
+MBOAT7
+PRR15L
+CRELD2
+FSD1
+TCTN1
+CHMP6
+NPEPL1
+ST7
+NUP85
+TCEAL4
+KAT6A
+DHDDS
+DENND2D
+FBXO11
+CCDC92
+COASY
+PDHX
+FOSL1
+TSEN2
+PRR7
+CALM3
+HMGA2
+CALU
+ITFG1
+GDPD5
+NCOA3
+NRIP1
+CAPN1
+SMC1A
+LAGE3
+CAST
+AXIN1
+CDC45
+FZD1
+FZD7
+CASP2
+CASP3
+GRWD1
+PIP4K2B
+CASP7
+ARID5B
+CASP10
+NCK2
+DYRK3
+DUSP11
+TUBB6
+CAT
+PSRC1
+RAE1
+ADO
+PIK3R3
+NIPSNAP1
+ELP1
+HAT1
+MICALL1
+MAPKAPK5
+BHLHE40
+MKNK1
+CASK
+AKR7A2
+RUVBL1
+PSMG1
+BECN1
+CBLB
+MBTPS1
+EED
+CTNNAL1
+CBR1
+RNMT
+CBR3
+PEX11A
+CREG1
+INPP4B
+IQGAP1
+SOCS2
+CFLAR
+CDK5R1
+ST3GAL5
+IER3
+SQSTM1
+SLC5A6
+CPNE3
+CCNA2
+CCNA1
+CCNB1
+TIMELESS
+CCND3
+P4HA2
+PLOD3
+CCNF
+UBE3B
+NOL3
+SLC25A14
+MPZL1
+CCNH
+MAP7
+JPT2
+DNAJA3
+USP14
+MTA1
+SLC25A46
+PDLIM1
+SMC3
+PRPF4
+CCNB2
+CCNE2
+SYNGR3
+LPAR2
+ARHGEF2
+ZW10
+COG7
+AURKB
+VAPB
+NOLC1
+UBE2L6
+MAPKAPK2
+CYTH1
+ITGB1BP1
+BCL7B
+COPB2
+ADGRG1
+MAPK1IP1L
+TBC1D31
+TM9SF2
+MAP4K4
+HOMER2
+SH3BP5
+PIGB
+SCARB1
+PSMF1
+SPTLC2
+TBPL1
+BAG3
+POLR1C
+SPAG7
+CD40
+CD44
+FEZ2
+IKBKE
+CD58
+MTFR1
+HS2ST1
+IPO13
+VGLL4
+NUP93
+UBE3C
+EDEM1
+TRAM2
+CEP57
+HERPUD1
+USP6NL
+CCP110
+ADGRE5
+MLEC
+TATDN2
+MRPL19
+SCRN1
+EFCAB14
+KEAP1
+CDK1
+MELK
+PLEKHM1
+C2CD5
+KIAA0753
+C2CD2L
+TOMM70
+WASHC5
+KLHL21
+CDC20
+ARNT2
+FAM20B
+NCAPD2
+PAN2
+LPGAT1
+KIF14
+CDC25A
+CDC25B
+OXSR1
+MVP
+CDC42
+DMTF1

MVC/CDRP-BBBC047-Bray/Step3_aggreate_2modality.ipynb

@@ -0,0 +1,476 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 数据聚合：「输入药物 SMILES，输出基因表达 + 细胞形态」的多模态虚拟细胞模型，配对不齐，训练集不均衡。\n",
+    "\n",
+    "## 策略有3种：\n",
+    "\n",
+    "### 1 是按 SMILES 聚合，对每个 SMILES 在每个模态上 取平均或中位数（或更复杂的 embedding 聚合）。\n",
+    "\n",
+    "### 2 是多实例学习（Multiple Instance Learning, MIL）：思路：把一个 SMILES 的所有 replicate 当作一个 bag，模型学习 bag-level 对齐。\n",
+    "\n",
+    "### 3 是不对齐 replicate，随机采样：思路：训练时，每个 batch 从两个模态中随机采样该 SMILES 的一个 replicate，逐步逼近两个分布。\n",
+    "\n",
+    "## 此处用1，最为简单"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "root = '/home/bob/boom/VCBench/data/MVC/CDRP-BBBC047-Bray/'\n",
+    "\n",
+    "CP_path = root + './Processed_Paired_CP.h5'\n",
+    "GE_path = root + './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(86796, 745)"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV['target'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/numpy/core/_methods.py:118: RuntimeWarning: invalid value encountered in reduce\n",
+      "  ret = umr_sum(arr, axis, dtype, out, keepdims, where=where)\n",
+      "/home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/numpy/core/_methods.py:152: RuntimeWarning: invalid value encountered in reduce\n",
+      "  arrmean = umr_sum(arr, axis, dtype, keepdims=True, where=where)\n",
+      "/home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/numpy/core/_methods.py:173: RuntimeWarning: invalid value encountered in subtract\n",
+      "  x = asanyarray(arr - arrmean)\n",
+      "/home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/numpy/core/_methods.py:118: RuntimeWarning: invalid value encountered in reduce\n",
+      "  ret = umr_sum(arr, axis, dtype, out, keepdims, where=where)\n",
+      "/home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/numpy/core/_methods.py:152: RuntimeWarning: invalid value encountered in reduce\n",
+      "  arrmean = umr_sum(arr, axis, dtype, keepdims=True, where=where)\n",
+      "/home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/numpy/core/_methods.py:173: RuntimeWarning: invalid value encountered in subtract\n",
+      "  x = asanyarray(arr - arrmean)\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      ">>> 前 10 列明细 <<<\n",
+      "    set  col       min        max      mean      std\n",
+      "control    0 -0.003959   0.078018  0.024671 0.013316\n",
+      "control    1 -0.039080   0.070839 -0.008352 0.008747\n",
+      "control    2 -0.078927   0.220322 -0.003043 0.016821\n",
+      "control    3 -0.015957   0.186684  0.053808 0.031368\n",
+      "control    4 -0.011257   0.261045  0.047642 0.031550\n",
+      "control    5 -0.063300   0.086711 -0.000350 0.011541\n",
+      "control    6 -0.046063   0.093030 -0.015935 0.011592\n",
+      "control    7 -0.013254   0.019495 -0.001661 0.003557\n",
+      "control    8 -0.061549   0.004743 -0.017370 0.011673\n",
+      "control    9 -0.032541   0.013303  0.000285 0.003969\n",
+      " target    0 -1.313544  20.910727  0.110194 0.240215\n",
+      " target    1 -3.000090   5.296432 -0.042150 0.181244\n",
+      " target    2 -3.415679   5.304324  0.027545 0.308696\n",
+      " target    3 -3.018066  40.497417  0.149676 0.529733\n",
+      " target    4 -3.267396 130.057816  0.086563 0.833441\n",
+      " target    5 -8.066486   3.052383  0.010240 0.222079\n",
+      " target    6 -2.858146   3.832927 -0.080404 0.217838\n",
+      " target    7 -1.747062   2.119149 -0.017756 0.084344\n",
+      " target    8 -5.563820   1.317170 -0.092266 0.185681\n",
+      " target    9 -2.521358   2.361632  0.007691 0.096224\n",
+      "\n",
+      ">>> 全局信息 <<<\n",
+      "全局最小值 : inf\n",
+      "全局最大值 : -inf\n",
+      "动态范围   : -inf\n",
+      "\n",
+      ">>> 可能异常列（max > μ+3σ 或 min < μ-3σ）<<<\n",
+      "[('control', 0), ('control', 1), ('control', 2), ('control', 3), ('control', 4), ('control', 5), ('control', 6), ('control', 7), ('control', 8), ('control', 9), ('control', 10), ('control', 11), ('control', 12), ('control', 13), ('control', 14), ('control', 15), ('control', 16), ('control', 17), ('control', 18), ('control', 19), ('control', 20), ('control', 21), ('control', 22), ('control', 23), ('control', 24), ('control', 25), ('control', 26), ('control', 27), ('control', 28), ('control', 29), ('control', 30), ('control', 31), ('control', 32), ('control', 33), ('control', 34), ('control', 35), ('control', 42), ('control', 43), ('control', 44), ('control', 45), ('control', 46), ('control', 47), ('control', 48), ('control', 49), ('control', 50), ('control', 51), ('control', 54), ('control', 55), ('control', 56), ('control', 57), ('control', 58), ('control', 59), ('control', 60), ('control', 61), ('control', 62), ('control', 63), ('control', 64), ('control', 65), ('control', 66), ('control', 67), ('control', 68), ('control', 69), ('control', 70), ('control', 71), ('control', 72), ('control', 73), ('control', 74), ('control', 75), ('control', 76), ('control', 77), ('control', 78), ('control', 79), ('control', 80), ('control', 81), ('control', 82), ('control', 83), ('control', 84), ('control', 85), ('control', 86), ('control', 87), ('control', 88), ('control', 89), ('control', 90), ('control', 93), ('control', 94), ('control', 95), ('control', 96), ('control', 97), ('control', 98), ('control', 99), ('control', 100), ('control', 101), ('control', 102), ('control', 103), ('control', 104), ('control', 105), ('control', 106), ('control', 107), ('control', 108), ('control', 109), ('control', 110), ('control', 111), ('control', 112), ('control', 113), ('control', 114), ('control', 115), ('control', 116), ('control', 117), ('control', 118), ('control', 119), ('control', 120), ('control', 121), ('control', 122), ('control', 123), ('control', 124), ('control', 125), ('control', 126), ('control', 127), ('control', 128), ('control', 129), ('control', 130), ('control', 131), ('control', 132), ('control', 133), ('control', 134), ('control', 135), ('control', 137), ('control', 138), ('control', 139), ('control', 140), ('control', 141), ('control', 142), ('control', 143), ('control', 144), ('control', 145), ('control', 146), ('control', 147), ('control', 148), ('control', 149), ('control', 150), ('control', 151), ('control', 152), ('control', 153), ('control', 154), ('control', 155), ('control', 156), ('control', 157), ('control', 158), ('control', 159), ('control', 160), ('control', 161), ('control', 162), ('control', 163), ('control', 164), ('control', 165), ('control', 166), ('control', 167), ('control', 168), ('control', 169), ('control', 170), ('control', 171), ('control', 172), ('control', 173), ('control', 174), ('control', 175), ('control', 176), ('control', 177), ('control', 178), ('control', 179), ('control', 180), ('control', 181), ('control', 182), ('control', 183), ('control', 184), ('control', 185), ('control', 186), ('control', 187), ('control', 188), ('control', 189), ('control', 190), ('control', 191), ('control', 192), ('control', 193), ('control', 194), ('control', 195), ('control', 196), ('control', 197), ('control', 198), ('control', 199), ('control', 200), ('control', 201), ('control', 202), ('control', 203), ('control', 204), ('control', 205), ('control', 206), ('control', 207), ('control', 208), ('control', 210), ('control', 211), ('control', 212), ('control', 213), ('control', 214), ('control', 215), ('control', 216), ('control', 217), ('control', 218), ('control', 219), ('control', 220), ('control', 221), ('control', 222), ('control', 223), ('control', 224), ('control', 225), ('control', 226), ('control', 227), ('control', 228), ('control', 229), ('control', 230), ('control', 231), ('control', 232), ('control', 233), ('control', 234), ('control', 235), ('control', 236), ('control', 237), ('control', 238), ('control', 239), ('control', 240), ('control', 241), ('control', 242), ('control', 243), ('control', 244), ('control', 245), ('control', 246), ('control', 247), ('control', 248), ('control', 249), ('control', 257), ('control', 258), ('control', 259), ('control', 260), ('control', 261), ('control', 262), ('control', 263), ('control', 264), ('control', 265), ('control', 266), ('control', 267), ('control', 268), ('control', 269), ('control', 270), ('control', 271), ('control', 272), ('control', 273), ('control', 274), ('control', 275), ('control', 276), ('control', 277), ('control', 278), ('control', 279), ('control', 280), ('control', 281), ('control', 282), ('control', 283), ('control', 284), ('control', 286), ('control', 287), ('control', 288), ('control', 289), ('control', 290), ('control', 292), ('control', 293), ('control', 294), ('control', 295), ('control', 296), ('control', 297), ('control', 310), ('control', 311), ('control', 312), ('control', 313), ('control', 314), ('control', 315), ('control', 316), ('control', 317), ('control', 318), ('control', 319), ('control', 320), ('control', 321), ('control', 322), ('control', 323), ('control', 324), ('control', 325), ('control', 326), ('control', 327), ('control', 328), ('control', 329), ('control', 330), ('control', 332), ('control', 333), ('control', 334), ('control', 336), ('control', 339), ('control', 340), ('control', 341), ('control', 342), ('control', 343), ('control', 344), ('control', 345), ('control', 346), ('control', 347), ('control', 348), ('control', 349), ('control', 350), ('control', 351), ('control', 352), ('control', 353), ('control', 354), ('control', 355), ('control', 356), ('control', 357), ('control', 358), ('control', 359), ('control', 360), ('control', 361), ('control', 362), ('control', 363), ('control', 364), ('control', 365), ('control', 366), ('control', 367), ('control', 368), ('control', 369), ('control', 370), ('control', 371), ('control', 372), ('control', 373), ('control', 374), ('control', 375), ('control', 376), ('control', 377), ('control', 378), ('control', 379), ('control', 380), ('control', 381), ('control', 382), ('control', 383), ('control', 384), ('control', 385), ('control', 386), ('control', 387), ('control', 388), ('control', 389), ('control', 390), ('control', 391), ('control', 392), ('control', 393), ('control', 394), ('control', 395), ('control', 396), ('control', 397), ('control', 398), ('control', 399), ('control', 400), ('control', 401), ('control', 402), ('control', 403), ('control', 404), ('control', 405), ('control', 406), ('control', 407), ('control', 408), ('control', 409), ('control', 410), ('control', 411), ('control', 412), ('control', 413), ('control', 414), ('control', 415), ('control', 416), ('control', 417), ('control', 418), ('control', 419), ('control', 420), ('control', 421), ('control', 422), ('control', 423), ('control', 424), ('control', 425), ('control', 426), ('control', 427), ('control', 428), ('control', 429), ('control', 430), ('control', 431), ('control', 432), ('control', 433), ('control', 434), ('control', 435), ('control', 436), ('control', 437), ('control', 438), ('control', 439), ('control', 440), ('control', 441), ('control', 442), ('control', 443), ('control', 444), ('control', 445), ('control', 446), ('control', 447), ('control', 448), ('control', 449), ('control', 450), ('control', 451), ('control', 452), ('control', 453), ('control', 454), ('control', 455), ('control', 456), ('control', 457), ('control', 458), ('control', 459), ('control', 460), ('control', 461), ('control', 462), ('control', 463), ('control', 464), ('control', 465), ('control', 466), ('control', 467), ('control', 468), ('control', 469), ('control', 470), ('control', 471), ('control', 472), ('control', 473), ('control', 474), ('control', 475), ('control', 476), ('control', 477), ('control', 478), ('control', 479), ('control', 480), ('control', 481), ('control', 482), ('control', 483), ('control', 484), ('control', 485), ('control', 486), ('control', 487), ('control', 488), ('control', 489), ('control', 490), ('control', 491), ('control', 492), ('control', 493), ('control', 494), ('control', 495), ('control', 496), ('control', 497), ('control', 498), ('control', 499), ('control', 500), ('control', 501), ('control', 502), ('control', 503), ('control', 504), ('control', 505), ('control', 506), ('control', 507), ('control', 508), ('control', 509), ('control', 510), ('control', 511), ('control', 512), ('control', 513), ('control', 514), ('control', 515), ('control', 516), ('control', 517), ('control', 518), ('control', 519), ('control', 520), ('control', 521), ('control', 522), ('control', 523), ('control', 524), ('control', 525), ('control', 526), ('control', 527), ('control', 528), ('control', 529), ('control', 530), ('control', 531), ('control', 533), ('control', 534), ('control', 535), ('control', 536), ('control', 537), ('control', 538), ('control', 539), ('control', 540), ('control', 541), ('control', 542), ('control', 543), ('control', 544), ('control', 545), ('control', 546), ('control', 547), ('control', 548), ('control', 549), ('control', 550), ('control', 551), ('control', 552), ('control', 553), ('control', 554), ('control', 555), ('control', 556), ('control', 557), ('control', 558), ('control', 559), ('control', 560), ('control', 561), ('control', 563), ('control', 566), ('control', 567), ('control', 568), ('control', 569), ('control', 570), ('control', 571), ('control', 572), ('control', 573), ('control', 574), ('control', 575), ('control', 576), ('control', 577), ('control', 578), ('control', 579), ('control', 580), ('control', 581), ('control', 582), ('control', 583), ('control', 584), ('control', 585), ('control', 586), ('control', 587), ('control', 588), ('control', 589), ('control', 590), ('control', 591), ('control', 592), ('control', 593), ('control', 594), ('control', 595), ('control', 596), ('control', 597), ('control', 598), ('control', 599), ('control', 600), ('control', 601), ('control', 602), ('control', 603), ('control', 604), ('control', 605), ('control', 606), ('control', 607), ('control', 608), ('control', 609), ('control', 610), ('control', 611), ('control', 612), ('control', 613), ('control', 614), ('control', 615), ('control', 616), ('control', 617), ('control', 618), ('control', 619), ('control', 620), ('control', 621), ('control', 622), ('control', 623), ('control', 624), ('control', 625), ('control', 626), ('control', 627), ('control', 628), ('control', 629), ('control', 630), ('control', 631), ('control', 632), ('control', 633), ('control', 634), ('control', 635), ('control', 636), ('control', 637), ('control', 638), ('control', 639), ('control', 640), ('control', 641), ('control', 642), ('control', 643), ('control', 644), ('control', 645), ('control', 646), ('control', 647), ('control', 648), ('control', 649), ('control', 650), ('control', 651), ('control', 652), ('control', 653), ('control', 654), ('control', 655), ('control', 656), ('control', 657), ('control', 658), ('control', 659), ('control', 660), ('control', 661), ('control', 662), ('control', 663), ('control', 664), ('control', 665), ('control', 666), ('control', 667), ('control', 668), ('control', 669), ('control', 670), ('control', 671), ('control', 672), ('control', 673), ('control', 674), ('control', 675), ('control', 676), ('control', 677), ('control', 678), ('control', 679), ('control', 680), ('control', 681), ('control', 682), ('control', 683), ('control', 684), ('control', 685), ('control', 686), ('control', 687), ('control', 688), ('control', 689), ('control', 690), ('control', 691), ('control', 692), ('control', 693), ('control', 694), ('control', 695), ('control', 696), ('control', 697), ('control', 698), ('control', 699), ('control', 700), ('control', 701), ('control', 702), ('control', 703), ('control', 704), ('control', 705), ('control', 706), ('control', 707), ('control', 708), ('control', 709), ('control', 710), ('control', 711), ('control', 712), ('control', 713), ('control', 714), ('control', 716), ('control', 717), ('control', 718), ('control', 719), ('control', 720), ('control', 721), ('control', 722), ('control', 723), ('control', 724), ('control', 725), ('control', 726), ('control', 727), ('control', 728), ('control', 729), ('control', 730), ('control', 731), ('control', 732), ('control', 733), ('control', 734), ('control', 735), ('control', 736), ('control', 737), ('control', 738), ('control', 739), ('control', 740), ('control', 741), ('control', 742), ('control', 743), ('control', 744), ('target', 0), ('target', 1), ('target', 2), ('target', 3), ('target', 4), ('target', 5), ('target', 6), ('target', 7), ('target', 8), ('target', 9), ('target', 10), ('target', 11), ('target', 12), ('target', 13), ('target', 14), ('target', 15), ('target', 16), ('target', 17), ('target', 18), ('target', 19), ('target', 20), ('target', 21), ('target', 22), ('target', 23), ('target', 24), ('target', 25), ('target', 26), ('target', 27), ('target', 28), ('target', 29), ('target', 30), ('target', 31), ('target', 32), ('target', 33), ('target', 34), ('target', 35), ('target', 42), ('target', 43), ('target', 44), ('target', 45), ('target', 46), ('target', 47), ('target', 48), ('target', 49), ('target', 50), ('target', 51), ('target', 54), ('target', 55), ('target', 56), ('target', 57), ('target', 58), ('target', 59), ('target', 60), ('target', 61), ('target', 62), ('target', 63), ('target', 64), ('target', 65), ('target', 66), ('target', 67), ('target', 68), ('target', 69), ('target', 70), ('target', 71), ('target', 72), ('target', 73), ('target', 74), ('target', 75), ('target', 76), ('target', 77), ('target', 78), ('target', 79), ('target', 80), ('target', 81), ('target', 82), ('target', 83), ('target', 84), ('target', 85), ('target', 86), ('target', 87), ('target', 88), ('target', 89), ('target', 90), ('target', 94), ('target', 95), ('target', 96), ('target', 97), ('target', 98), ('target', 99), ('target', 100), ('target', 101), ('target', 102), ('target', 103), ('target', 104), ('target', 105), ('target', 106), ('target', 107), ('target', 108), ('target', 109), ('target', 110), ('target', 111), ('target', 112), ('target', 113), ('target', 114), ('target', 115), ('target', 116), ('target', 117), ('target', 118), ('target', 119), ('target', 120), ('target', 121), ('target', 122), ('target', 123), ('target', 124), ('target', 125), ('target', 126), ('target', 127), ('target', 128), ('target', 129), ('target', 130), ('target', 131), ('target', 132), ('target', 133), ('target', 134), ('target', 135), ('target', 136), ('target', 137), ('target', 138), ('target', 139), ('target', 140), ('target', 141), ('target', 142), ('target', 143), ('target', 144), ('target', 145), ('target', 146), ('target', 147), ('target', 148), ('target', 149), ('target', 150), ('target', 151), ('target', 152), ('target', 153), ('target', 154), ('target', 155), ('target', 156), ('target', 157), ('target', 158), ('target', 159), ('target', 160), ('target', 161), ('target', 162), ('target', 163), ('target', 164), ('target', 165), ('target', 166), ('target', 167), ('target', 168), ('target', 169), ('target', 170), ('target', 171), ('target', 172), ('target', 173), ('target', 174), ('target', 175), ('target', 176), ('target', 177), ('target', 178), ('target', 179), ('target', 180), ('target', 181), ('target', 182), ('target', 183), ('target', 184), ('target', 185), ('target', 186), ('target', 187), ('target', 188), ('target', 189), ('target', 190), ('target', 191), ('target', 192), ('target', 193), ('target', 194), ('target', 195), ('target', 196), ('target', 197), ('target', 198), ('target', 199), ('target', 200), ('target', 201), ('target', 202), ('target', 203), ('target', 204), ('target', 205), ('target', 206), ('target', 207), ('target', 208), ('target', 209), ('target', 210), ('target', 211), ('target', 212), ('target', 213), ('target', 214), ('target', 215), ('target', 216), ('target', 217), ('target', 218), ('target', 219), ('target', 220), ('target', 221), ('target', 222), ('target', 223), ('target', 224), ('target', 225), ('target', 226), ('target', 227), ('target', 228), ('target', 229), ('target', 230), ('target', 231), ('target', 232), ('target', 233), ('target', 234), ('target', 235), ('target', 236), ('target', 237), ('target', 238), ('target', 239), ('target', 240), ('target', 241), ('target', 242), ('target', 243), ('target', 244), ('target', 245), ('target', 246), ('target', 247), ('target', 248), ('target', 249), ('target', 257), ('target', 258), ('target', 259), ('target', 260), ('target', 261), ('target', 262), ('target', 263), ('target', 264), ('target', 265), ('target', 266), ('target', 267), ('target', 268), ('target', 269), ('target', 270), ('target', 271), ('target', 272), ('target', 273), ('target', 274), ('target', 275), ('target', 276), ('target', 277), ('target', 278), ('target', 279), ('target', 280), ('target', 281), ('target', 282), ('target', 283), ('target', 284), ('target', 286), ('target', 287), ('target', 288), ('target', 289), ('target', 290), ('target', 292), ('target', 293), ('target', 294), ('target', 295), ('target', 296), ('target', 297), ('target', 310), ('target', 311), ('target', 312), ('target', 313), ('target', 314), ('target', 315), ('target', 316), ('target', 317), ('target', 318), ('target', 319), ('target', 320), ('target', 321), ('target', 322), ('target', 323), ('target', 324), ('target', 325), ('target', 326), ('target', 327), ('target', 328), ('target', 329), ('target', 330), ('target', 332), ('target', 333), ('target', 334), ('target', 336), ('target', 337), ('target', 339), ('target', 340), ('target', 341), ('target', 342), ('target', 343), ('target', 344), ('target', 345), ('target', 346), ('target', 347), ('target', 348), ('target', 349), ('target', 350), ('target', 351), ('target', 352), ('target', 353), ('target', 354), ('target', 355), ('target', 356), ('target', 357), ('target', 358), ('target', 359), ('target', 360), ('target', 361), ('target', 362), ('target', 363), ('target', 364), ('target', 365), ('target', 366), ('target', 367), ('target', 368), ('target', 369), ('target', 370), ('target', 371), ('target', 372), ('target', 373), ('target', 374), ('target', 375), ('target', 376), ('target', 377), ('target', 378), ('target', 379), ('target', 380), ('target', 381), ('target', 382), ('target', 384), ('target', 385), ('target', 386), ('target', 388), ('target', 389), ('target', 391), ('target', 392), ('target', 393), ('target', 394), ('target', 395), ('target', 396), ('target', 397), ('target', 398), ('target', 399), ('target', 400), ('target', 401), ('target', 402), ('target', 403), ('target', 404), ('target', 405), ('target', 406), ('target', 407), ('target', 408), ('target', 409), ('target', 410), ('target', 411), ('target', 412), ('target', 413), ('target', 414), ('target', 415), ('target', 416), ('target', 417), ('target', 418), ('target', 419), ('target', 420), ('target', 421), ('target', 422), ('target', 423), ('target', 424), ('target', 425), ('target', 426), ('target', 427), ('target', 428), ('target', 429), ('target', 430), ('target', 431), ('target', 432), ('target', 433), ('target', 434), ('target', 435), ('target', 436), ('target', 437), ('target', 438), ('target', 439), ('target', 440), ('target', 441), ('target', 442), ('target', 443), ('target', 444), ('target', 445), ('target', 446), ('target', 447), ('target', 448), ('target', 449), ('target', 450), ('target', 451), ('target', 452), ('target', 453), ('target', 454), ('target', 455), ('target', 456), ('target', 457), ('target', 458), ('target', 459), ('target', 460), ('target', 461), ('target', 462), ('target', 463), ('target', 464), ('target', 465), ('target', 466), ('target', 467), ('target', 468), ('target', 469), ('target', 470), ('target', 471), ('target', 472), ('target', 473), ('target', 474), ('target', 475), ('target', 476), ('target', 477), ('target', 478), ('target', 479), ('target', 480), ('target', 481), ('target', 482), ('target', 483), ('target', 484), ('target', 485), ('target', 486), ('target', 487), ('target', 488), ('target', 489), ('target', 490), ('target', 491), ('target', 492), ('target', 493), ('target', 494), ('target', 495), ('target', 496), ('target', 497), ('target', 498), ('target', 499), ('target', 500), ('target', 501), ('target', 502), ('target', 503), ('target', 504), ('target', 505), ('target', 506), ('target', 507), ('target', 508), ('target', 509), ('target', 510), ('target', 511), ('target', 512), ('target', 513), ('target', 514), ('target', 515), ('target', 516), ('target', 517), ('target', 518), ('target', 519), ('target', 520), ('target', 521), ('target', 522), ('target', 523), ('target', 524), ('target', 525), ('target', 526), ('target', 527), ('target', 528), ('target', 529), ('target', 530), ('target', 531), ('target', 533), ('target', 534), ('target', 535), ('target', 536), ('target', 537), ('target', 538), ('target', 539), ('target', 540), ('target', 541), ('target', 542), ('target', 543), ('target', 544), ('target', 545), ('target', 546), ('target', 547), ('target', 548), ('target', 549), ('target', 550), ('target', 551), ('target', 552), ('target', 553), ('target', 554), ('target', 555), ('target', 556), ('target', 557), ('target', 558), ('target', 559), ('target', 560), ('target', 561), ('target', 563), ('target', 566), ('target', 567), ('target', 568), ('target', 569), ('target', 570), ('target', 571), ('target', 572), ('target', 573), ('target', 574), ('target', 575), ('target', 576), ('target', 577), ('target', 578), ('target', 579), ('target', 580), ('target', 581), ('target', 582), ('target', 583), ('target', 584), ('target', 585), ('target', 586), ('target', 587), ('target', 588), ('target', 589), ('target', 590), ('target', 591), ('target', 592), ('target', 593), ('target', 594), ('target', 595), ('target', 596), ('target', 597), ('target', 598), ('target', 599), ('target', 600), ('target', 601), ('target', 602), ('target', 603), ('target', 604), ('target', 605), ('target', 606), ('target', 607), ('target', 608), ('target', 609), ('target', 610), ('target', 611), ('target', 612), ('target', 613), ('target', 614), ('target', 615), ('target', 616), ('target', 617), ('target', 618), ('target', 619), ('target', 620), ('target', 621), ('target', 622), ('target', 623), ('target', 624), ('target', 625), ('target', 626), ('target', 627), ('target', 628), ('target', 629), ('target', 630), ('target', 631), ('target', 632), ('target', 633), ('target', 634), ('target', 635), ('target', 636), ('target', 637), ('target', 638), ('target', 639), ('target', 640), ('target', 641), ('target', 642), ('target', 643), ('target', 644), ('target', 645), ('target', 646), ('target', 647), ('target', 648), ('target', 649), ('target', 650), ('target', 651), ('target', 652), ('target', 653), ('target', 654), ('target', 655), ('target', 656), ('target', 657), ('target', 658), ('target', 659), ('target', 660), ('target', 661), ('target', 662), ('target', 663), ('target', 664), ('target', 665), ('target', 666), ('target', 667), ('target', 668), ('target', 669), ('target', 670), ('target', 671), ('target', 672), ('target', 673), ('target', 674), ('target', 675), ('target', 676), ('target', 677), ('target', 678), ('target', 679), ('target', 680), ('target', 681), ('target', 682), ('target', 683), ('target', 684), ('target', 685), ('target', 686), ('target', 687), ('target', 688), ('target', 689), ('target', 690), ('target', 691), ('target', 692), ('target', 693), ('target', 694), ('target', 695), ('target', 696), ('target', 697), ('target', 698), ('target', 699), ('target', 700), ('target', 701), ('target', 702), ('target', 703), ('target', 704), ('target', 705), ('target', 706), ('target', 707), ('target', 708), ('target', 709), ('target', 710), ('target', 711), ('target', 712), ('target', 713), ('target', 714), ('target', 715), ('target', 716), ('target', 717), ('target', 718), ('target', 719), ('target', 720), ('target', 721), ('target', 722), ('target', 723), ('target', 724), ('target', 725), ('target', 726), ('target', 727), ('target', 728), ('target', 729), ('target', 730), ('target', 731), ('target', 732), ('target', 733), ('target', 734), ('target', 735), ('target', 736), ('target', 737), ('target', 738), ('target', 739), ('target', 740), ('target', 741), ('target', 742), ('target', 743), ('target', 744)]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "def scan_range(dic, keys=('control', 'target'), head=10):\n",
+    "    \"\"\"\n",
+    "    快速扫描字典中指定 key 的数值范围\n",
+    "    dic : dict{'control':(N,745), 'target':(N,745), ...}\n",
+    "    head: 只打印前 head 列，省屏；想看全部设 745\n",
+    "    \"\"\"\n",
+    "    stats = []                       # 收集每列统计\n",
+    "    global_min, global_max = np.inf, -np.inf\n",
+    "\n",
+    "    for k in keys:\n",
+    "        X = dic[k]\n",
+    "        amin  = X.min(axis=0)        # (745,)\n",
+    "        amax  = X.max(axis=0)\n",
+    "        mean  = X.mean(axis=0)\n",
+    "        std   = X.std(axis=0)\n",
+    "        global_min = min(global_min, amin.min())\n",
+    "        global_max = max(global_max, amax.max())\n",
+    "\n",
+    "        # 前 head 列明细\n",
+    "        for c in range(min(head, X.shape[1])):\n",
+    "            stats.append({'set':k, 'col':c,\n",
+    "                          'min':amin[c], 'max':amax[c],\n",
+    "                          'mean':mean[c], 'std':std[c]})\n",
+    "\n",
+    "    df = pd.DataFrame(stats)\n",
+    "    print('>>> 前 {} 列明细 <<<'.format(head))\n",
+    "    print(df.to_string(index=False))\n",
+    "\n",
+    "    print('\\n>>> 全局信息 <<<')\n",
+    "    print('全局最小值 :', global_min)\n",
+    "    print('全局最大值 :', global_max)\n",
+    "    print('动态范围   :', global_max - global_min)\n",
+    "\n",
+    "    # 3σ 异常列检测（任意 set 中只要有一列异常就报）\n",
+    "    print('\\n>>> 可能异常列（max > μ+3σ 或 min < μ-3σ）<<<')\n",
+    "    abnormal = []\n",
+    "    for k in keys:\n",
+    "        X = dic[k]\n",
+    "        mean = X.mean(axis=0)\n",
+    "        std  = X.std(axis=0)\n",
+    "        amin, amax = X.min(axis=0), X.max(axis=0)\n",
+    "        mask = (amax > mean + 3*std) | (amin < mean - 3*std)\n",
+    "        if mask.any():\n",
+    "            cols = np.where(mask)[0]\n",
+    "            abnormal.extend([(k, int(c)) for c in cols])\n",
+    "    if abnormal:\n",
+    "        print(abnormal)\n",
+    "    else:\n",
+    "        print('无异常列')\n",
+    "\n",
+    "# -------------- 运行 --------------\n",
+    "scan_range(CP_CSV, keys=('control', 'target'), head=10)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "control → 均值绝对最大值: nan\n",
+      "control → 标准差最大值: nan\n",
+      "target → 均值绝对最大值: nan\n",
+      "target → 标准差最大值: nan\n"
+     ]
+    }
+   ],
+   "source": [
+    "for k in ('control', 'target'):\n",
+    "    X = CP_CSV[k]\n",
+    "    print(k, '→ 均值绝对最大值:', np.abs(X.mean(axis=0)).max())\n",
+    "    print(k, '→ 标准差最大值:', X.std(axis=0).max())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_207036/3124577792.py:19: RuntimeWarning: Mean of empty slice\n",
+      "  CP_grouped = CP_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+      "/tmp/ipykernel_207036/3124577792.py:19: RuntimeWarning: Mean of empty slice\n",
+      "  CP_grouped = CP_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ 聚合后的数据已保存到 /home/bob/boom/VCBench/data/MVC/CDRP-BBBC047-Bray/./Aggregated_CP_GE.h5，所有 NaN 已处理为 0.0\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "def aggregate_modalities(CP, GE, out_path):\n",
+    "    # 转成 DataFrame 方便聚合\n",
+    "    CP_df = pd.DataFrame({\n",
+    "        \"SMILES\": CP[\"canonical_smiles\"],\n",
+    "        \"control\": list(CP[\"control\"]),\n",
+    "        \"target\": list(CP[\"target\"])\n",
+    "    })\n",
+    "    GE_df = pd.DataFrame({\n",
+    "        \"SMILES\": GE[\"canonical_smiles\"],\n",
+    "        \"control\": list(GE[\"control\"]),\n",
+    "        \"target\": list(GE[\"target\"])\n",
+    "    })\n",
+    "\n",
+    "    # groupby + mean pooling\n",
+    "    CP_grouped = CP_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "    GE_grouped = GE_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "\n",
+    "    # merge 保证对齐\n",
+    "    merged = pd.merge(CP_grouped, GE_grouped, on=\"SMILES\", how=\"inner\", suffixes=(\"_CP\", \"_GE\"))\n",
+    "\n",
+    "    # 转 numpy array 并替换剩余 NaN\n",
+    "    def clean_array(arr_list):\n",
+    "        arr = np.stack(arr_list.to_numpy())\n",
+    "        arr = np.nan_to_num(arr, nan=0.0, posinf=0.0, neginf=0.0)  # 替换 NaN/Inf\n",
+    "        return arr.astype(np.float32)\n",
+    "\n",
+    "    smiles_array     = merged[\"SMILES\"].to_numpy(dtype=\"S\")\n",
+    "    control_CP_array = clean_array(merged[\"control_CP\"])\n",
+    "    target_CP_array  = clean_array(merged[\"target_CP\"])\n",
+    "    control_GE_array = clean_array(merged[\"control_GE\"])\n",
+    "    target_GE_array  = clean_array(merged[\"target_GE\"])\n",
+    "\n",
+    "    # 存 HDF5\n",
+    "    with h5py.File(out_path, \"w\") as f:\n",
+    "        f.create_dataset(\"canonical_smiles\", data=smiles_array)\n",
+    "        f.create_dataset(\"control_CP\", data=control_CP_array)\n",
+    "        f.create_dataset(\"target_CP\", data=target_CP_array)\n",
+    "        f.create_dataset(\"control_GE\", data=control_GE_array)\n",
+    "        f.create_dataset(\"target_GE\", data=target_GE_array)\n",
+    "\n",
+    "    print(f\"✅ 聚合后的数据已保存到 {out_path}，所有 NaN 已处理为 0.0\")\n",
+    "\n",
+    "# 用法\n",
+    "root = '/home/bob/boom/VCBench/data/MVC/CDRP-BBBC047-Bray/'\n",
+    "CP_path = root + './Processed_Paired_CP.h5'\n",
+    "GE_path = root + './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)\n",
+    "\n",
+    "out_path = root + './Aggregated_CP_GE.h5'\n",
+    "aggregate_modalities(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 38,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['Brc1c(CSc2nc3ccccc3s2)nc2ncccn12', 'Brc1c(NC2=NCCN2)ccc2nccnc12',\n",
+       "        'Brc1ccc(CSc2nnc(-c3ccccn3)n2Cc2ccco2)cc1', ...,\n",
+       "        'c1nc(CC2CCNCC2)c[nH]1', 'c1nc(CCCC2CCNCC2)c[nH]1',\n",
+       "        'c1ncn(CCOCCOc2ccc3c(c2)CCC3)n1'], dtype='<U227'),\n",
+       " 'control_CP': array([[ 3.2161873e-02, -1.7372804e-02, -1.0016474e-02, ...,\n",
+       "          4.8336964e-03,  9.2573617e-05,  1.3814055e-02],\n",
+       "        [ 3.0686287e-02, -1.1331222e-02, -3.4925649e-03, ...,\n",
+       "          2.5207591e-03,  1.7474355e-03,  8.4845992e-03],\n",
+       "        [ 3.6669757e-02, -1.6801117e-02, -5.4300507e-03, ...,\n",
+       "          5.7292758e-03, -8.5048133e-04,  3.2608695e-02],\n",
+       "        ...,\n",
+       "        [ 2.9445987e-02, -1.6279269e-02, -9.0824757e-03, ...,\n",
+       "          3.4394469e-03,  2.6764136e-03,  1.5279663e-02],\n",
+       "        [ 3.0686287e-02, -1.1331222e-02, -3.4925649e-03, ...,\n",
+       "          2.5207591e-03,  1.7474355e-03,  8.4845992e-03],\n",
+       "        [ 2.2611173e-02, -1.0286282e-02,  1.6986302e-03, ...,\n",
+       "          6.0957628e-03,  6.0581490e-03,  1.8283216e-02]], dtype=float32),\n",
+       " 'control_GE': array([[-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        ...,\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297]], dtype=float32),\n",
+       " 'target_CP': array([[ 1.5516217e-01, -1.2652491e-01, -1.5108880e-01, ...,\n",
+       "          5.1517323e-02,  4.3028165e-02, -3.2104313e-01],\n",
+       "        [ 1.3919285e-01, -7.0041016e-02,  5.6847539e-02, ...,\n",
+       "         -6.6846021e-02,  2.0045139e-02,  3.1137601e-01],\n",
+       "        [-4.1009387e-01,  4.8901412e-01,  9.5608789e-01, ...,\n",
+       "          5.4136354e-01,  4.3033558e-01, -1.5829408e-01],\n",
+       "        ...,\n",
+       "        [ 1.6184693e-02, -3.3477028e-03,  6.8760978e-04, ...,\n",
+       "         -1.0370575e-01, -9.1575049e-02, -7.9115279e-02],\n",
+       "        [ 1.4948754e-01, -8.7920934e-02,  5.6048077e-02, ...,\n",
+       "         -1.9580002e-01, -8.4258631e-02, -5.9720445e-02],\n",
+       "        [ 4.1791059e-02,  2.5397958e-02,  2.0725157e-02, ...,\n",
+       "          2.7885816e-01,  9.2793293e-02,  1.4922591e-02]], dtype=float32),\n",
+       " 'target_GE': array([[ 0.09065   ,  0.0786    , -0.06155   , ..., -0.1333    ,\n",
+       "          0.097875  ,  0.15165   ],\n",
+       "        [-0.07302775,  0.0460825 , -0.02222175, ...,  0.17655624,\n",
+       "         -0.154201  ,  0.033125  ],\n",
+       "        [-0.03446667, -0.01266667,  0.04358333, ...,  0.27253333,\n",
+       "         -0.06081   ,  0.28026667],\n",
+       "        ...,\n",
+       "        [ 0.224725  ,  0.2090425 , -0.2872    , ..., -0.2203515 ,\n",
+       "          0.5861085 ,  0.278331  ],\n",
+       "        [ 0.01292225,  0.0713975 , -0.13613375, ..., -0.16344374,\n",
+       "          0.053449  , -0.200645  ],\n",
+       "        [ 0.14866666, -0.054     ,  0.04613333, ...,  0.22286667,\n",
+       "         -0.09675   , -0.01868333]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 38,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out_path = './Aggregated_CP_GE.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 42,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(20341, 745)"
+      ]
+     },
+     "execution_count": 42,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(20341, 977)"
+      ]
+     },
+     "execution_count": 28,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_GE'].shape  # 查看前5个 SMILES"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 21,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "'O=C(C[C@@H]1CC[C@@H]2[C@H](COC[C@H](O)CN2Cc2cc(F)cc(F)c2)O1)Nc1nccs1' in agg_data['canonical_smiles']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

MVC/CDRP-BBBC047-Bray/Step4_outlier_process.ipynb

@@ -0,0 +1,462 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "389cd136",
+   "metadata": {},
+   "source": [
+    "# 处理异常数据\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "917fa110",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "cddb976e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "path47 = 'Aggregated_CP_GE.h5'\n",
+    "\n",
+    "# data36 = load_from_HDF(path36)\n",
+    "data47 = load_from_HDF(path47)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "54e51b05",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-20.614704132080078, 2870874.0]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-1249.9088134765625, 19152610.0]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['control_CP'].min()\n",
+    "data_max = data47['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data47['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['target_CP'].min()\n",
+    "data_max = data47['target_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['target_CP'], axis=0)\n",
+    "max_per_column = np.max(data47['target_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "bde80b4c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-2.230109453201294, 1.6778349876403809]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-6.4108500480651855, 9.534767150878906]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['control_GE'].min()\n",
+    "data_max = data47['control_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['control_GE'], axis=0)\n",
+    "max_per_column = np.max(data47['control_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['target_GE'].min()\n",
+    "data_max = data47['target_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['target_GE'], axis=0)\n",
+    "max_per_column = np.max(data47['target_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "053cfaa7",
+   "metadata": {},
+   "source": [
+    "## 看看 cpg0016 的数值范围"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "c81440ab",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-1.2491892576217651, 1.3490736484527588]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-6.057624816894531, 6.188350677490234]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "cpg0016_data = '/home/bob/boom/VCBench/data/Image/cpg0016_data.h5'\n",
+    "\n",
+    "# data36 = load_from_HDF(path36)\n",
+    "cpg0016_data = load_from_HDF(cpg0016_data)\n",
+    "\n",
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = cpg0016_data['control'].min()\n",
+    "data_max = cpg0016_data['control'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(cpg0016_data['control'], axis=0)\n",
+    "max_per_column = np.max(cpg0016_data['control'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = cpg0016_data['target'].min()\n",
+    "data_max = cpg0016_data['target'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(cpg0016_data['target'], axis=0)\n",
+    "max_per_column = np.max(cpg0016_data['target'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "2b2f331e",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "标准化后的整个数组的数据范围：[-11.38830852508545, 10.123730659484863]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 假设 data47['target_CP'] 是你的原始数据, target_CP, \n",
+    "data_with_outlier = data47['control_CP']\n",
+    "\n",
+    "# 1. 异常值处理 - 使用截断法 (Clipping)\n",
+    "# 设定一个更保守的上限，例如基于分位数\n",
+    "q99 = np.percentile(data_with_outlier, 99)\n",
+    "clip_max = q99 * 2 # 或者直接用一个固定值，比如 5\n",
+    "data_clipped = np.clip(data_with_outlier, a_min=None, a_max=clip_max)\n",
+    "# 2. 数据标准化 - Z-score Normalization\n",
+    "# 计算每列（特征）的均值和标准差\n",
+    "mu = data_clipped.mean(axis=0)\n",
+    "sigma = data_clipped.std(axis=0)\n",
+    "\n",
+    "# 初始化一个与数据形状相同的零数组\n",
+    "data_normalized = np.zeros_like(data_clipped, dtype=float)\n",
+    "\n",
+    "# 找到标准差不为0的列\n",
+    "# 为了避免浮点数误差，使用一个非常小的阈值来检查\n",
+    "epsilon = 1e-8\n",
+    "nonzero_sigma_cols = sigma > epsilon\n",
+    "\n",
+    "# 对这些列进行零均值化\n",
+    "data_normalized[:, nonzero_sigma_cols] = (data_clipped[:, nonzero_sigma_cols] - mu[nonzero_sigma_cols]) / sigma[nonzero_sigma_cols]\n",
+    "\n",
+    "# 检查结果\n",
+    "data_min = data_normalized.min()\n",
+    "data_max = data_normalized.max()\n",
+    "\n",
+    "print(f\"\\n标准化后的整个数组的数据范围：[{data_min}, {data_max}]\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "9ccda958",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "control_CP = data_normalized.copy()\n",
+    "# print(control_CP.shape)\n",
+    "# target_CP = data_normalized.copy()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "6e515034",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-11.38830852508545, 10.123730659484863]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-129.0272979736328, 64.47639465332031]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data_min = control_CP.min()\n",
+    "data_max = control_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = target_CP.min()\n",
+    "data_max = target_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "id": "481b57e7",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "float64 float64\n",
+      "float32 float32\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(control_CP.dtype, target_CP.dtype)\n",
+    "\n",
+    "control_CP_float32 = control_CP.astype(np.float32)\n",
+    "target_CP_float32 = target_CP.astype(np.float32)\n",
+    "print(control_CP_float32.dtype, target_CP_float32.dtype)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "6e2d6030",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "已删除旧的 'control_CP' 数据集。\n",
+      "已删除旧的 'target_CP' 数据集。\n",
+      "已成��写入新的 'control_CP' 和 'target_CP' 数据集。\n",
+      "操作完成。\n"
+     ]
+    }
+   ],
+   "source": [
+    "file_path = 'Aggregated_CP_GE.h5'\n",
+    "# 使用 'a' (append) 模式打开文件\n",
+    "with h5py.File(file_path, 'a') as f:\n",
+    "    \n",
+    "    # 1. 删除旧的数据集（如果存在）\n",
+    "    if \"control_CP\" in f:\n",
+    "        del f[\"control_CP\"]\n",
+    "        print(\"已删除旧的 'control_CP' 数据集。\")\n",
+    "        \n",
+    "    if \"target_CP\" in f:\n",
+    "        del f[\"target_CP\"]\n",
+    "        print(\"已删除旧的 'target_CP' 数据集。\")\n",
+    "    \n",
+    "    # 2. 创建并写入新的数据集\n",
+    "    f.create_dataset(\"control_CP\", data=control_CP_float32)\n",
+    "    f.create_dataset(\"target_CP\", data=target_CP_float32)\n",
+    "    \n",
+    "    print(\"已成功写入新的 'control_CP' 和 'target_CP' 数据集。\")\n",
+    "\n",
+    "print(\"操作完成。\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b131c13e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5a82372f",
+   "metadata": {},
+   "source": [
+    "# 看基因数据是否异常 -- 没有"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "a81ddc86",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-11.38830852508545, 10.123730659484863]\n",
+      "------------------------------\n",
+      "每列数据的最小值（前10个）：\n",
+      "[-1.7736303 -2.8932922 -1.9792434 -1.3063283 -1.5805442 -2.5623124\n",
+      " -2.9260125 -3.1737845 -3.493395  -5.5112543]\n",
+      "------------------------------\n",
+      "每列数据的最大值（前10个）：\n",
+      "[3.8115907 4.765787  9.468964  3.2014632 2.3736608 4.2486434 4.5805535\n",
+      " 4.513591  1.7128    2.9639645]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data47 = load_from_HDF( 'Aggregated_CP_GE.h5')\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['control_CP'].min()\n",
+    "data_max = data47['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data47['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最小值（前10个）：\")\n",
+    "print(min_per_column[:10])\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最大值（前10个）：\")\n",
+    "print(max_per_column[:10])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "770f3727",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 看看方差\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "10693804",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c072eea5",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

MVC/CDRPBIO-BBBC036-Bray/Step3_aggreate_2modality.ipynb

@@ -0,0 +1,230 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 数据聚合：「输入药物 SMILES，输出基因表达 + 细胞形态」的多模态虚拟细胞模型，配对不齐，训练集不均衡。\n",
+    "\n",
+    "## 策略有3种：\n",
+    "\n",
+    "### 1 是按 SMILES 聚合，对每个 SMILES 在每个模态上 取平均或中位数（或更复杂的 embedding 聚合）。\n",
+    "\n",
+    "### 2 是多实例学习（Multiple Instance Learning, MIL）：思路：把一个 SMILES 的所有 replicate 当作一个 bag，模型学习 bag-level 对齐。\n",
+    "\n",
+    "### 3 是不对齐 replicate，随机采样：思路：训练时，每个 batch 从两个模态中随机采样该 SMILES 的一个 replicate，逐步逼近两个分布。\n",
+    "\n",
+    "## 此处用1，最为简单"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "root = '/home/bob/boom/VCBench/data/MVC/CDRPBIO-BBBC036-Bray/'\n",
+    "\n",
+    "CP_path = root + './Processed_Paired_CP.h5'\n",
+    "GE_path = root + './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ 聚合后的数据已保存到 /home/bob/boom/VCBench/data/MVC/CDRPBIO-BBBC036-Bray/./Aggregated_CP_GE.h5，所有 NaN 已处理为 0.0\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "def aggregate_modalities(CP, GE, out_path):\n",
+    "    # 转成 DataFrame 方便聚合\n",
+    "    CP_df = pd.DataFrame({\n",
+    "        \"SMILES\": CP[\"canonical_smiles\"],\n",
+    "        \"control\": list(CP[\"control\"]),\n",
+    "        \"target\": list(CP[\"target\"])\n",
+    "    })\n",
+    "    GE_df = pd.DataFrame({\n",
+    "        \"SMILES\": GE[\"canonical_smiles\"],\n",
+    "        \"control\": list(GE[\"control\"]),\n",
+    "        \"target\": list(GE[\"target\"])\n",
+    "    })\n",
+    "\n",
+    "    # groupby + mean pooling\n",
+    "    CP_grouped = CP_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "    GE_grouped = GE_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "\n",
+    "    # merge 保证对齐\n",
+    "    merged = pd.merge(CP_grouped, GE_grouped, on=\"SMILES\", how=\"inner\", suffixes=(\"_CP\", \"_GE\"))\n",
+    "\n",
+    "    # 转 numpy array 并替换剩余 NaN\n",
+    "    def clean_array(arr_list):\n",
+    "        arr = np.stack(arr_list.to_numpy())\n",
+    "        arr = np.nan_to_num(arr, nan=0.0, posinf=0.0, neginf=0.0)  # 替换 NaN/Inf\n",
+    "        return arr.astype(np.float32)\n",
+    "\n",
+    "    smiles_array     = merged[\"SMILES\"].to_numpy(dtype=\"S\")\n",
+    "    control_CP_array = clean_array(merged[\"control_CP\"])\n",
+    "    target_CP_array  = clean_array(merged[\"target_CP\"])\n",
+    "    control_GE_array = clean_array(merged[\"control_GE\"])\n",
+    "    target_GE_array  = clean_array(merged[\"target_GE\"])\n",
+    "\n",
+    "    # 存 HDF5\n",
+    "    with h5py.File(out_path, \"w\") as f:\n",
+    "        f.create_dataset(\"canonical_smiles\", data=smiles_array)\n",
+    "        f.create_dataset(\"control_CP\", data=control_CP_array)\n",
+    "        f.create_dataset(\"target_CP\", data=target_CP_array)\n",
+    "        f.create_dataset(\"control_GE\", data=control_GE_array)\n",
+    "        f.create_dataset(\"target_GE\", data=target_GE_array)\n",
+    "\n",
+    "    print(f\"✅ 聚合后的数据已保存到 {out_path}，所有 NaN 已处理为 0.0\")\n",
+    "\n",
+    "# 用法\n",
+    "# root = '/home/bob/boom/VCBench/data/MVC/CDRP-BBBC047-Bray/'\n",
+    "CP_path = root + './Processed_Paired_CP.h5'\n",
+    "GE_path = root + './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)\n",
+    "\n",
+    "out_path = root + './Aggregated_CP_GE.h5'\n",
+    "aggregate_modalities(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['Brc1c(NC2=NCCN2)ccc2nccnc12', 'C#CCN(C)CCCOc1ccc(Cl)cc1Cl',\n",
+       "        'C#CCN1C(C)=C(C(=O)CCO)C(c2cccc(-n3oo3)c2)C(C(=O)OC)=C1C', ...,\n",
+       "        'c1coc(CNc2ncnc3[nH]cnc23)c1', 'c1nc(CC2CCNCC2)c[nH]1',\n",
+       "        'c1nc(CCCC2CCNCC2)c[nH]1'], dtype='<U227'),\n",
+       " 'control_CP': array([[ 0.00068819,  0.03068629, -0.01133122, ..., -0.00035833,\n",
+       "         -0.0068023 ,  0.00832675],\n",
+       "        [ 0.00068819,  0.03068629, -0.01133122, ..., -0.00035833,\n",
+       "         -0.0068023 ,  0.00832675],\n",
+       "        [-0.00018166,  0.01862261, -0.00605156, ...,  0.01070234,\n",
+       "          0.01840025,  0.03512387],\n",
+       "        ...,\n",
+       "        [ 0.00068819,  0.03068629, -0.01133122, ..., -0.00035833,\n",
+       "         -0.0068023 ,  0.00832675],\n",
+       "        [-0.00193635,  0.02944599, -0.01627927, ...,  0.0022546 ,\n",
+       "         -0.00520131,  0.01978289],\n",
+       "        [ 0.00068819,  0.03068629, -0.01133122, ..., -0.00035833,\n",
+       "         -0.0068023 ,  0.00832675]]),\n",
+       " 'control_GE': array([[-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        ...,\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297],\n",
+       "        [-0.00217228, -0.00637228, -0.00293138, ..., -0.00425048,\n",
+       "         -0.01134733, -0.01150297]]),\n",
+       " 'target_CP': array([[ 0.06232562,  0.13919284, -0.07004101, ...,  0.00523225,\n",
+       "         -0.0840081 ,  0.3163032 ],\n",
+       "        [-0.02893378, -0.0963472 ,  0.02897743, ...,  0.03984685,\n",
+       "         -0.06590394, -0.04551435],\n",
+       "        [ 0.05013672,  0.00352105,  0.0064943 , ..., -0.05595761,\n",
+       "         -0.08462357, -0.1268212 ],\n",
+       "        ...,\n",
+       "        [-0.01697117, -0.06073245,  0.04498612, ...,  0.06319024,\n",
+       "          0.03119937,  0.02412124],\n",
+       "        [ 0.02098823,  0.01618469, -0.0033477 , ..., -0.07113561,\n",
+       "         -0.06343231, -0.06900362],\n",
+       "        [-0.05296022,  0.14948754, -0.08792093, ..., -0.06916199,\n",
+       "         -0.13310869, -0.03078984]]),\n",
+       " 'target_GE': array([[-0.07302775,  0.0460825 , -0.02222175, ...,  0.17655625,\n",
+       "         -0.154201  ,  0.033125  ],\n",
+       "        [ 0.1019155 , -0.2480317 ,  0.119545  , ..., -0.16921   ,\n",
+       "          0.02239785, -0.0529015 ],\n",
+       "        [-0.034165  ,  0.048892  ,  0.1231085 , ...,  0.42318   ,\n",
+       "          0.034605  ,  0.21026   ],\n",
+       "        ...,\n",
+       "        [ 0.0316805 , -0.1364167 ,  0.02315   , ...,  0.10307   ,\n",
+       "          0.17449785,  0.1486385 ],\n",
+       "        [ 0.224725  ,  0.2090425 , -0.2872    , ..., -0.2203515 ,\n",
+       "          0.5861085 ,  0.278331  ],\n",
+       "        [ 0.01292225,  0.0713975 , -0.13613375, ..., -0.16344375,\n",
+       "          0.053449  , -0.200645  ]])}"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# load aggregated data\n",
+    "# agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

MVC/CDRPBIO-BBBC036-Bray/Step4_outlier_process.ipynb

@@ -0,0 +1,491 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "389cd136",
+   "metadata": {},
+   "source": [
+    "# 处理异常数据\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "917fa110",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "cddb976e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "path47 = 'Aggregated_CP_GE.h5'\n",
+    "\n",
+    "# data36 = load_from_HDF(path36)\n",
+    "data36 = load_from_HDF(path47)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "92260f6e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(1916, 602)"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "data36['control_CP'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "54e51b05",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-0.15427954494953156, 32876.36328125]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-8.007139205932617, 370889.125]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['control_CP'].min()\n",
+    "data_max = data36['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data36['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['target_CP'].min()\n",
+    "data_max = data36['target_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['target_CP'], axis=0)\n",
+    "max_per_column = np.max(data36['target_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "bde80b4c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-0.059715718030929565, 0.10131344199180603]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-4.208789825439453, 6.420098304748535]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['control_GE'].min()\n",
+    "data_max = data36['control_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['control_GE'], axis=0)\n",
+    "max_per_column = np.max(data36['control_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['target_GE'].min()\n",
+    "data_max = data36['target_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['target_GE'], axis=0)\n",
+    "max_per_column = np.max(data36['target_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "053cfaa7",
+   "metadata": {},
+   "source": [
+    "## 看看 cpg0016 的数值范围"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c81440ab",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-1.2491892576217651, 1.3490736484527588]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-6.057624816894531, 6.188350677490234]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "# cpg0016_data = '/home/bob/boom/VCBench/data/Image/cpg0016_data.h5'\n",
+    "\n",
+    "# # data36 = load_from_HDF(path36)\n",
+    "# cpg0016_data = load_from_HDF(cpg0016_data)\n",
+    "\n",
+    "# import numpy as np\n",
+    "\n",
+    "# # 计算整个数组的最小值和最大值\n",
+    "# data_min = cpg0016_data['control'].min()\n",
+    "# data_max = cpg0016_data['control'].max()\n",
+    "\n",
+    "# # 计算每列（即每个特征）的最小值和最大值\n",
+    "# # axis=0 表示对每一列进行操作\n",
+    "# min_per_column = np.min(cpg0016_data['control'], axis=0)\n",
+    "# max_per_column = np.max(cpg0016_data['control'], axis=0)\n",
+    "\n",
+    "# print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "# print(\"-\" * 30)\n",
+    "\n",
+    "# # 计算整个数组的最小值和最大值\n",
+    "# data_min = cpg0016_data['target'].min()\n",
+    "# data_max = cpg0016_data['target'].max()\n",
+    "\n",
+    "# # 计算每列（即每个特征）的最小值和最大值\n",
+    "# # axis=0 表示对每一列进行操作\n",
+    "# min_per_column = np.min(cpg0016_data['target'], axis=0)\n",
+    "# max_per_column = np.max(cpg0016_data['target'], axis=0)\n",
+    "\n",
+    "# print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "# print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c7005461",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "2b2f331e",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "标准化后的整个数组的数据范围：[-19.69989776611328, 31.564712524414062]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 假设 data36['target_CP'] 是你的原始数据, target_CP, control_CP\n",
+    "data_with_outlier = data36['target_CP']\n",
+    "\n",
+    "# 1. 异常值处理 - 使用截断法 (Clipping)\n",
+    "# 设定一个更保守的上限，例如基于分位数\n",
+    "q99 = np.percentile(data_with_outlier, 99)\n",
+    "clip_max = q99 * 2 # 或者直接用一个固定值，比如 5\n",
+    "data_clipped = np.clip(data_with_outlier, a_min=None, a_max=clip_max)\n",
+    "# 2. 数据标准化 - Z-score Normalization\n",
+    "# 计算每列（特征）的均值和标准差\n",
+    "mu = data_clipped.mean(axis=0)\n",
+    "sigma = data_clipped.std(axis=0)\n",
+    "\n",
+    "# 初始化一个与数据形状相同的零数组\n",
+    "data_normalized = np.zeros_like(data_clipped, dtype=float)\n",
+    "\n",
+    "# 找到标准差不为0的列\n",
+    "# 为了避免浮点数误差，使用一个非常小的阈值来检查\n",
+    "epsilon = 1e-8\n",
+    "nonzero_sigma_cols = sigma > epsilon\n",
+    "\n",
+    "# 对这些列进行零均值化\n",
+    "data_normalized[:, nonzero_sigma_cols] = (data_clipped[:, nonzero_sigma_cols] - mu[nonzero_sigma_cols]) / sigma[nonzero_sigma_cols]\n",
+    "\n",
+    "# 检查结果\n",
+    "data_min = data_normalized.min()\n",
+    "data_max = data_normalized.max()\n",
+    "\n",
+    "print(f\"\\n标准化后的整个数组的数据范围：[{data_min}, {data_max}]\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "9ccda958",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# control_CP = data_normalized.copy()\n",
+    "# print(control_CP.shape)\n",
+    "target_CP = data_normalized.copy()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "6e515034",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-43.761436462402344, 6.348904132843018]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-19.69989776611328, 31.564712524414062]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data_min = control_CP.min()\n",
+    "data_max = control_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = target_CP.min()\n",
+    "data_max = target_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "481b57e7",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "float64 float64\n",
+      "float32 float32\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(control_CP.dtype, target_CP.dtype)\n",
+    "\n",
+    "control_CP_float32 = control_CP.astype(np.float32)\n",
+    "target_CP_float32 = target_CP.astype(np.float32)\n",
+    "print(control_CP_float32.dtype, target_CP_float32.dtype)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "6e2d6030",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "已删除旧的 'control_CP' 数据集。\n",
+      "已删除旧的 'target_CP' 数据集。\n",
+      "已成功写入新的 'control_CP' 和 'target_CP' 数据集。\n",
+      "操作完成。\n"
+     ]
+    }
+   ],
+   "source": [
+    "file_path = 'Aggregated_CP_GE.h5'\n",
+    "# 使用 'a' (append) 模式打开文件\n",
+    "with h5py.File(file_path, 'a') as f:\n",
+    "    \n",
+    "    # 1. 删除旧的数据集（如果存在）\n",
+    "    if \"control_CP\" in f:\n",
+    "        del f[\"control_CP\"]\n",
+    "        print(\"已删除旧的 'control_CP' 数据集。\")\n",
+    "        \n",
+    "    if \"target_CP\" in f:\n",
+    "        del f[\"target_CP\"]\n",
+    "        print(\"已删除旧的 'target_CP' 数据集。\")\n",
+    "    \n",
+    "    # 2. 创建并写入新的数据集\n",
+    "    f.create_dataset(\"control_CP\", data=control_CP_float32)\n",
+    "    f.create_dataset(\"target_CP\", data=target_CP_float32)\n",
+    "    \n",
+    "    print(\"已成功写入新的 'control_CP' 和 'target_CP' 数据集。\")\n",
+    "\n",
+    "print(\"操作完成。\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b131c13e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5a82372f",
+   "metadata": {},
+   "source": [
+    "# 看基因数据是否异常 -- 没有"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "a81ddc86",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-43.761436462402344, 6.348904132843018]\n",
+      "------------------------------\n",
+      "每列数据的最小值（前10个）：\n",
+      "[-1.6301312 -1.4819461 -1.9675876 -2.272238  -3.393242  -1.1765975\n",
+      " -2.6176555 -1.2396002 -2.462905  -2.8029377]\n",
+      "------------------------------\n",
+      "每列数据的最大值（前10个）：\n",
+      "[3.0697236 1.985853  1.09526   1.1315886 2.1168044 2.9211075 2.1078165\n",
+      " 2.3139157 2.1509798 1.5131917]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data36 = load_from_HDF( 'Aggregated_CP_GE.h5')\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['control_CP'].min()\n",
+    "data_max = data36['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data36['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最小值（前10个）：\")\n",
+    "print(min_per_column[:10])\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最大值（前10个）：\")\n",
+    "print(max_per_column[:10])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "770f3727",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 看看方差\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "10693804",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c072eea5",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

MVC/MVC_BBBC036/Step3_aggreate_2modality copy.ipynb

@@ -0,0 +1,786 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 数据聚合：「输入药物 SMILES，输出基因表达 + 细胞形态」的多模态虚拟细胞模型，配对不齐，训练集不均衡。\n",
+    "\n",
+    "## 策略有3种：\n",
+    "\n",
+    "### 1 是按 SMILES 聚合，对每个 SMILES 在每个模态上 取平均或中位数（或更复杂的 embedding 聚合）。\n",
+    "\n",
+    "### 2 是多实例学习（Multiple Instance Learning, MIL）：思路：把一个 SMILES 的所有 replicate 当作一个 bag，模型学习 bag-level 对齐。\n",
+    "\n",
+    "### 3 是不对齐 replicate，随机采样：思路：训练时，每个 batch 从两个模态中随机采样该 SMILES 的一个 replicate，逐步逼近两个分布。"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "CP_path =  './Processed_Paired_CP.h5'\n",
+    "GE_path = './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['C#CCOC(=O)C1=CCCN(C)C1',\n",
+       "        'C#C[C@]1(O)CC[C@H]2[C@@H]3CCC4=CC(=O)CC[C@@H]4[C@H]3CC[C@@]21CC',\n",
+       "        'C#C[C@]1(O)CC[C@H]2[C@@H]3CCc4cc(OC)ccc4[C@H]3CC[C@@]21C', ...,\n",
+       "        'c1cncc(C2=NCCC2)c1', 'c1coc(CNc2ncnc3[nH]cnc23)c1',\n",
+       "        'c1nc(CCCC2CCNCC2)c[nH]1'], dtype='<U227'),\n",
+       " 'control': array([[-0.06009524, -0.2170542 ,  0.21140526, ...,  0.1283901 ,\n",
+       "          0.2584292 ,  0.20108473],\n",
+       "        [-0.06009524, -0.2170542 ,  0.21140526, ...,  0.1283901 ,\n",
+       "          0.2584292 ,  0.20108473],\n",
+       "        [-0.06009524, -0.2170542 ,  0.21140526, ...,  0.1283901 ,\n",
+       "          0.2584292 ,  0.20108473],\n",
+       "        ...,\n",
+       "        [-0.04660993, -0.15566814,  0.16524817, ...,  0.13523042,\n",
+       "          0.2184239 ,  0.12808448],\n",
+       "        [-0.04660993, -0.15566814,  0.16524817, ...,  0.13523042,\n",
+       "          0.2184239 ,  0.12808448],\n",
+       "        [-0.04660993, -0.15566814,  0.16524817, ...,  0.13523042,\n",
+       "          0.2184239 ,  0.12808448]], dtype=float32),\n",
+       " 'target': array([[ 0.20242606,  0.37789968, -0.12808509, ..., -0.40345412,\n",
+       "         -0.02612734,  0.8973323 ],\n",
+       "        [-0.40302846, -0.21528135,  0.54286873, ..., -0.7699544 ,\n",
+       "         -0.06759822,  0.1704398 ],\n",
+       "        [-0.10030121, -0.23437162,  0.40763763, ...,  0.08733866,\n",
+       "          0.31274182,  0.99544084],\n",
+       "        ...,\n",
+       "        [-0.9691039 , -0.2935919 ,  0.6202567 , ..., -0.3105628 ,\n",
+       "         -0.07092339, -0.30973706],\n",
+       "        [ 0.23442163, -0.12433557, -0.10770472, ..., -0.13565809,\n",
+       "          0.6637088 , -0.52422494],\n",
+       "        [-0.7016537 ,  0.645092  , -0.4753387 , ..., -0.77215654,\n",
+       "         -0.44080254,  0.06545502]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((14996, 602), (3451, 977))"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV['target'].shape, GE_CSV['target'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 按照第一种方式直接 mean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ 聚合后的数据已保存到 ./Aggregated_CP_GE.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "def aggregate_modalities(CP, GE, out_path):\n",
+    "    # 转成 DataFrame 方便聚合\n",
+    "    CP_df = pd.DataFrame({\n",
+    "        \"SMILES\": CP[\"canonical_smiles\"],\n",
+    "        \"control\": list(CP[\"control\"]),\n",
+    "        \"target\": list(CP[\"target\"])\n",
+    "    })\n",
+    "    GE_df = pd.DataFrame({\n",
+    "        \"SMILES\": GE[\"canonical_smiles\"],\n",
+    "        \"control\": list(GE[\"control\"]),\n",
+    "        \"target\": list(GE[\"target\"])\n",
+    "    })\n",
+    "\n",
+    "    # groupby + mean pooling\n",
+    "    CP_grouped = CP_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "    GE_grouped = GE_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "\n",
+    "    # merge 保证对齐\n",
+    "    merged = pd.merge(CP_grouped, GE_grouped, on=\"SMILES\", how=\"inner\", suffixes=(\"_CP\", \"_GE\"))\n",
+    "\n",
+    "    # 转 numpy array 并替换剩余 NaN\n",
+    "    def clean_array(arr_list):\n",
+    "        arr = np.stack(arr_list.to_numpy())\n",
+    "        arr = np.nan_to_num(arr, nan=0.0, posinf=0.0, neginf=0.0)  # 替换 NaN/Inf\n",
+    "        return arr.astype(np.float32)\n",
+    "\n",
+    "    smiles_array     = merged[\"SMILES\"].to_numpy(dtype=\"S\")\n",
+    "    control_CP_array = clean_array(merged[\"control_CP\"])\n",
+    "    target_CP_array  = clean_array(merged[\"target_CP\"])\n",
+    "    control_GE_array = clean_array(merged[\"control_GE\"])\n",
+    "    target_GE_array  = clean_array(merged[\"target_GE\"])\n",
+    "\n",
+    "    # 存 HDF5\n",
+    "    with h5py.File(out_path, \"w\") as f:\n",
+    "        f.create_dataset(\"canonical_smiles\", data=smiles_array)\n",
+    "        f.create_dataset(\"control_CP\", data=control_CP_array)\n",
+    "        f.create_dataset(\"target_CP\", data=target_CP_array)\n",
+    "        f.create_dataset(\"control_GE\", data=control_GE_array)\n",
+    "        f.create_dataset(\"target_GE\", data=target_GE_array)\n",
+    "\n",
+    "    print(f\"✅ 聚合后的数据已保存到 {out_path}\")\n",
+    "\n",
+    "# 用法\n",
+    "# root = '/home/bob/boom/VCBench/data/MVC/CDRP-BBBC047-Bray/'\n",
+    "CP_path = './Processed_Paired_CP.h5'\n",
+    "GE_path = './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)\n",
+    "\n",
+    "out_path = './Aggregated_CP_GE.h5'\n",
+    "aggregate_modalities(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['Brc1c(NC2=NCCN2)ccc2nccnc12', 'C#CCN(C)CCCOc1ccc(Cl)cc1Cl',\n",
+       "        'C#CCN1C(C)=C(C(=O)CCO)C(c2cccc(-n3oo3)c2)C(C(=O)OC)=C1C', ...,\n",
+       "        'c1coc(CNc2ncnc3[nH]cnc23)c1', 'c1nc(CC2CCNCC2)c[nH]1',\n",
+       "        'c1nc(CCCC2CCNCC2)c[nH]1'], dtype='<U227'),\n",
+       " 'control_CP': array([[-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298],\n",
+       "        [-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298],\n",
+       "        [-0.03419905, -0.1922472 ,  0.18566999, ...,  0.17574853,\n",
+       "          0.33123744,  0.21626002],\n",
+       "        ...,\n",
+       "        [-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298],\n",
+       "        [-0.04489063, -0.14671806,  0.12026446, ...,  0.12773524,\n",
+       "          0.21260148,  0.16793586],\n",
+       "        [-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298]], dtype=float32),\n",
+       " 'control_GE': array([[ 0.01338525, -0.02699645,  0.0147262 , ..., -0.01713243,\n",
+       "         -0.02412532,  0.08308113],\n",
+       "        [-0.05236533, -0.00944031,  0.01470257, ..., -0.1302189 ,\n",
+       "          0.1095909 , -0.00176817],\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433],\n",
+       "        ...,\n",
+       "        [-0.05236533, -0.00944031,  0.01470257, ..., -0.1302189 ,\n",
+       "          0.1095909 , -0.00176817],\n",
+       "        [ 0.02062   , -0.00183125, -0.13683268, ..., -0.17978694,\n",
+       "          0.21799538,  0.17114288],\n",
+       "        [ 0.01338525, -0.02699645,  0.0147262 , ..., -0.01713243,\n",
+       "         -0.02412532,  0.08308113]], dtype=float32),\n",
+       " 'target_CP': array([[ 0.47359726,  0.3149385 , -0.2134298 , ...,  0.14258505,\n",
+       "         -0.23095778,  1.3368101 ],\n",
+       "        [-0.2682222 , -0.6758744 ,  0.3996619 , ...,  0.35662955,\n",
+       "         -0.12905961, -0.09158537],\n",
+       "        [ 0.37451744, -0.25577283,  0.26045328, ..., -0.23579153,\n",
+       "         -0.23442191, -0.41257128],\n",
+       "        ...,\n",
+       "        [-0.1709819 , -0.5260589 ,  0.4987828 , ...,  0.5009768 ,\n",
+       "          0.41748047,  0.18332437],\n",
+       "        [ 0.13757834, -0.20250246,  0.19951464, ..., -0.3296469 ,\n",
+       "         -0.11514819, -0.1843171 ],\n",
+       "        [-0.46352574,  0.35824373, -0.32413673, ..., -0.31744272,\n",
+       "         -0.5073174 , -0.03345545]], dtype=float32),\n",
+       " 'target_GE': array([[-0.07302775,  0.0460825 , -0.02222175, ...,  0.17655626,\n",
+       "         -0.154201  ,  0.033125  ],\n",
+       "        [ 0.10191549, -0.24803169,  0.11954501, ..., -0.16921   ,\n",
+       "          0.02239785, -0.0529015 ],\n",
+       "        [-0.034165  ,  0.048892  ,  0.1231085 , ...,  0.42318   ,\n",
+       "          0.034605  ,  0.21026   ],\n",
+       "        ...,\n",
+       "        [ 0.0316805 , -0.1364167 ,  0.02315   , ...,  0.10307   ,\n",
+       "          0.17449784,  0.1486385 ],\n",
+       "        [ 0.22472501,  0.20904249, -0.2872    , ..., -0.22035149,\n",
+       "          0.5861085 ,  0.278331  ],\n",
+       "        [ 0.01292225,  0.07139751, -0.13613375, ..., -0.16344374,\n",
+       "          0.053449  , -0.200645  ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out_path = './Aggregated_CP_GE.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((1916, 602), (1916, 602), (1916, 977), (1916, 977))"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape, agg_data['target_CP'].shape, agg_data['control_GE'].shape, agg_data['target_GE'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 21,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "'O=C(C[C@@H]1CC[C@@H]2[C@H](COC[C@H](O)CN2Cc2cc(F)cc(F)c2)O1)Nc1nccs1' in agg_data['canonical_smiles']"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 复杂点，样本级别的配对"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "共有化合物: 1916\n",
+      "总配对样本数: 26948\n",
+      "数据集形状:\n",
+      "  control_CP: (26948, 602)\n",
+      "  target_CP: (26948, 602)\n",
+      "  control_GE: (26948, 977)\n",
+      "  target_GE: (26948, 977)\n",
+      "✅ 配对数据集已保存到: ./Paired_CP_GE_Dataset.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "def create_paired_dataset(CP_data, GE_data, out_path):\n",
+    "    \"\"\"\n",
+    "    创建样本级配对的多模态数据集\n",
+    "    保持原始样本的完整性，只配对真正对应的样本\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    # 构建SMILES到索引列表的映射\n",
+    "    cp_smiles_to_indices = defaultdict(list)\n",
+    "    ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "        cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "        ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    # 找到共有的SMILES\n",
+    "    common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "    print(f\"共有化合物: {len(common_smiles)}\")\n",
+    "    \n",
+    "    # 收集所有有效配对\n",
+    "    paired_samples = []\n",
+    "    \n",
+    "    for smiles in common_smiles:\n",
+    "        cp_indices = cp_smiles_to_indices[smiles]\n",
+    "        ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "        # 为每个SMILES创建所有可能的CP-GE配对\n",
+    "        for cp_idx in cp_indices:\n",
+    "            for ge_idx in ge_indices:\n",
+    "                paired_samples.append({\n",
+    "                    'canonical_smiles': smiles,\n",
+    "                    'cp_index': cp_idx,\n",
+    "                    'ge_index': ge_idx\n",
+    "                })\n",
+    "    \n",
+    "    print(f\"总配对样本数: {len(paired_samples)}\")\n",
+    "    \n",
+    "    # 创建最终的字典形数据集\n",
+    "    dataset_dict = {\n",
+    "        'canonical_smiles': [],\n",
+    "        'control_CP': [],\n",
+    "        'target_CP': [], \n",
+    "        'control_GE': [],\n",
+    "        'target_GE': []\n",
+    "    }\n",
+    "    \n",
+    "    for sample in paired_samples:\n",
+    "        dataset_dict['canonical_smiles'].append(sample['canonical_smiles'])\n",
+    "        dataset_dict['control_CP'].append(CP_data['control'][sample['cp_index']])\n",
+    "        dataset_dict['target_CP'].append(CP_data['target'][sample['cp_index']])\n",
+    "        dataset_dict['control_GE'].append(GE_data['control'][sample['ge_index']]) \n",
+    "        dataset_dict['target_GE'].append(GE_data['target'][sample['ge_index']])\n",
+    "    \n",
+    "    # 转换为numpy数组\n",
+    "    final_dataset = {}\n",
+    "    final_dataset['canonical_smiles'] = np.array(dataset_dict['canonical_smiles'], dtype='S')\n",
+    "    final_dataset['control_CP'] = np.array(dataset_dict['control_CP'], dtype=np.float32)\n",
+    "    final_dataset['target_CP'] = np.array(dataset_dict['target_CP'], dtype=np.float32)\n",
+    "    final_dataset['control_GE'] = np.array(dataset_dict['control_GE'], dtype=np.float32)\n",
+    "    final_dataset['target_GE'] = np.array(dataset_dict['target_GE'], dtype=np.float32)\n",
+    "    \n",
+    "    print(f\"数据集形状:\")\n",
+    "    print(f\"  control_CP: {final_dataset['control_CP'].shape}\")\n",
+    "    print(f\"  target_CP: {final_dataset['target_CP'].shape}\")\n",
+    "    print(f\"  control_GE: {final_dataset['control_GE'].shape}\")\n",
+    "    print(f\"  target_GE: {final_dataset['target_GE'].shape}\")\n",
+    "    \n",
+    "    # 保存为HDF5\n",
+    "    with h5py.File(out_path, 'w') as f:\n",
+    "        for key, value in final_dataset.items():\n",
+    "            f.create_dataset(key, data=value)\n",
+    "    \n",
+    "    print(f\"✅ 配对数据集已保存到: {out_path}\")\n",
+    "    return final_dataset\n",
+    "\n",
+    "# 使用方案\n",
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "paired_dataset = create_paired_dataset(CP_CSV, GE_CSV, out_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['CC(C)(C)c1ccc(S(=O)(=O)/C=C/C#N)cc1',\n",
+       "        'CC(C)(C)c1ccc(S(=O)(=O)/C=C/C#N)cc1',\n",
+       "        'CC(C)(C)c1ccc(S(=O)(=O)/C=C/C#N)cc1', ...,\n",
+       "        'CCC(Cc1c(I)cc(I)c(N)c1I)C(=O)O', 'CCC(Cc1c(I)cc(I)c(N)c1I)C(=O)O',\n",
+       "        'CCC(Cc1c(I)cc(I)c(N)c1I)C(=O)O'], dtype='<U227'),\n",
+       " 'control_CP': array([[-0.05367128, -0.22861896,  0.19134662, ...,  0.09901066,\n",
+       "          0.23015772,  0.0870431 ],\n",
+       "        [-0.05367128, -0.22861896,  0.19134662, ...,  0.09901066,\n",
+       "          0.23015772,  0.0870431 ],\n",
+       "        [-0.04425929, -0.20467344,  0.19937934, ...,  0.10667558,\n",
+       "          0.25934276,  0.11115717],\n",
+       "        ...,\n",
+       "        [-0.00746798, -0.14845599,  0.17686495, ...,  0.18434836,\n",
+       "          0.3559847 ,  0.20929421],\n",
+       "        [-0.0670732 , -0.11723144,  0.11542749, ...,  0.15660104,\n",
+       "          0.31949052,  0.1973684 ],\n",
+       "        [-0.10378247, -0.11118906,  0.11418191, ...,  0.20794444,\n",
+       "          0.3261717 ,  0.20840377]], dtype=float32),\n",
+       " 'control_GE': array([[ 0.008735  , -0.0015245 , -0.00072016, ...,  0.044401  ,\n",
+       "         -0.2373012 , -0.019055  ],\n",
+       "        [-0.16669   ,  0.03942075, -0.1140243 , ..., -0.3791175 ,\n",
+       "          0.35399476, -0.2483125 ],\n",
+       "        [ 0.008735  , -0.0015245 , -0.00072016, ...,  0.044401  ,\n",
+       "         -0.2373012 , -0.019055  ],\n",
+       "        ...,\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433],\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433],\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433]], dtype=float32),\n",
+       " 'target_CP': array([[-2.475722  , -1.0180573 , -2.4777994 , ..., -1.6663419 ,\n",
+       "          6.1886635 ,  6.246197  ],\n",
+       "        [-2.475722  , -1.0180573 , -2.4777994 , ..., -1.6663419 ,\n",
+       "          6.1886635 ,  6.246197  ],\n",
+       "        [-2.475722  ,  3.6353707 , -1.1177131 , ...,  3.3711612 ,\n",
+       "          6.1886635 ,  6.246197  ],\n",
+       "        ...,\n",
+       "        [-0.7142158 ,  2.8495595 , -2.0671365 , ..., -0.34750465,\n",
+       "         -0.56493837, -0.8109943 ],\n",
+       "        [ 1.1229682 ,  0.68502927, -0.7133282 , ...,  0.237633  ,\n",
+       "          0.43848613, -0.48886108],\n",
+       "        [-0.6347912 ,  1.2815117 , -0.40972736, ..., -0.59789634,\n",
+       "         -0.81882066, -0.80796003]], dtype=float32),\n",
+       " 'target_GE': array([[-0.217655  , -0.0130605 , -0.32715666, ..., -0.279299  ,\n",
+       "          0.4241275 ,  1.109114  ],\n",
+       "        [-0.918455  ,  0.3162145 , -0.4372258 , ..., -2.098005  ,\n",
+       "          0.3910935 ,  0.438925  ],\n",
+       "        [-0.217655  , -0.0130605 , -0.32715666, ..., -0.279299  ,\n",
+       "          0.4241275 ,  1.109114  ],\n",
+       "        ...,\n",
+       "        [ 0.213685  ,  0.115822  , -0.3632915 , ...,  0.15678   ,\n",
+       "          0.091715  ,  0.10284   ],\n",
+       "        [ 0.213685  ,  0.115822  , -0.3632915 , ...,  0.15678   ,\n",
+       "          0.091715  ,  0.10284   ],\n",
+       "        [ 0.213685  ,  0.115822  , -0.3632915 , ...,  0.15678   ,\n",
+       "          0.091715  ,  0.10284   ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((26948, 602), (26948, 602), (26948, 977), (26948, 977))"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape, agg_data['target_CP'].shape, agg_data['control_GE'].shape, agg_data['target_GE'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 限制配对的数量，更好收敛。"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "开始创建稳健配对数据集...\n",
+      "共有化合物: 20341\n",
+      "处理进度: 0/20341\n",
+      "处理进度: 1000/20341\n",
+      "处理进度: 2000/20341\n",
+      "处理进度: 3000/20341\n",
+      "处理进度: 4000/20341\n",
+      "处理进度: 5000/20341\n",
+      "处理进度: 6000/20341\n",
+      "处理进度: 7000/20341\n",
+      "处理进度: 8000/20341\n",
+      "处理进度: 9000/20341\n",
+      "处理进度: 10000/20341\n",
+      "处理进度: 11000/20341\n",
+      "处理进度: 12000/20341\n",
+      "处理进度: 13000/20341\n",
+      "处理进度: 14000/20341\n",
+      "处理进度: 15000/20341\n",
+      "处理进度: 16000/20341\n",
+      "处理进度: 17000/20341\n",
+      "处理进度: 18000/20341\n",
+      "处理进度: 19000/20341\n",
+      "处理进度: 20000/20341\n",
+      "总配对样本数: 223689\n",
+      "数据收集进度: 0/223689\n",
+      "数据收集进度: 10000/223689\n",
+      "数据收集进度: 20000/223689\n",
+      "数据收集进度: 30000/223689\n",
+      "数据收集进度: 40000/223689\n",
+      "数据收集进度: 50000/223689\n",
+      "数据收集进度: 60000/223689\n",
+      "数据收集进度: 70000/223689\n",
+      "数据收集进度: 80000/223689\n",
+      "数据收集进度: 90000/223689\n",
+      "数据收集进度: 100000/223689\n",
+      "数据收集进度: 110000/223689\n",
+      "数据收集进度: 120000/223689\n",
+      "数据收集进度: 130000/223689\n",
+      "数据收集进度: 140000/223689\n",
+      "数据收集进度: 150000/223689\n",
+      "数据收集进度: 160000/223689\n",
+      "数据收集进度: 170000/223689\n",
+      "数据收集进度: 180000/223689\n",
+      "数据收集进度: 190000/223689\n",
+      "数据收集进度: 200000/223689\n",
+      "数据收集进度: 210000/223689\n",
+      "数据收集进度: 220000/223689\n",
+      "control_CP: 均值=0.0174, 标准差=0.3573, 形状=(223689, 775)\n",
+      "target_CP: 均值=-0.0105, 标准差=1.1399, 形状=(223689, 775)\n",
+      "control_GE: 均值=-0.0002, 标准差=0.1332, 形状=(223689, 977)\n",
+      "target_GE: 均值=-0.0007, 标准差=0.2795, 形状=(223689, 977)\n",
+      "✅ 稳健配对数据集已保存到: ./Paired_CP_GE_Dataset.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "def create_robust_paired_dataset(CP_data, GE_data, out_path, max_pairs_per_smiles=20):\n",
+    "    \"\"\"\n",
+    "    创建稳健的样本级配对数据集\n",
+    "    添加样本限制和数据处理\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    print(\"开始创建稳健配对数据集...\")\n",
+    "    \n",
+    "    # 构建SMILES到索引列表的映射\n",
+    "    cp_smiles_to_indices = defaultdict(list)\n",
+    "    ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "        cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "        ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    # 找到共有的SMILES\n",
+    "    common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "    print(f\"共有化合物: {len(common_smiles)}\")\n",
+    "    \n",
+    "    # 收集所有有效配对（带采样限制）\n",
+    "    paired_samples = []\n",
+    "    \n",
+    "    for i, smiles in enumerate(common_smiles):\n",
+    "        if i % 1000 == 0:\n",
+    "            print(f\"处理进度: {i}/{len(common_smiles)}\")\n",
+    "            \n",
+    "        cp_indices = cp_smiles_to_indices[smiles]\n",
+    "        ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "        # 限制每个SMILES的最大配对数，避免样本爆炸\n",
+    "        max_cp_samples = min(5, len(cp_indices))\n",
+    "        max_ge_samples = min(5, len(ge_indices))\n",
+    "        \n",
+    "        # 随机采样\n",
+    "        sampled_cp = np.random.choice(cp_indices, max_cp_samples, replace=False)\n",
+    "        sampled_ge = np.random.choice(ge_indices, max_ge_samples, replace=False)\n",
+    "        \n",
+    "        # 创建配对\n",
+    "        for cp_idx in sampled_cp:\n",
+    "            for ge_idx in sampled_ge:\n",
+    "                paired_samples.append({\n",
+    "                    'canonical_smiles': smiles,\n",
+    "                    'cp_index': cp_idx,\n",
+    "                    'ge_index': ge_idx\n",
+    "                })\n",
+    "    \n",
+    "    print(f\"总配对样本数: {len(paired_samples)}\")\n",
+    "    \n",
+    "    # 创建最终的字典形数据集\n",
+    "    dataset_dict = {\n",
+    "        'canonical_smiles': [],\n",
+    "        'control_CP': [],\n",
+    "        'target_CP': [], \n",
+    "        'control_GE': [],\n",
+    "        'target_GE': []\n",
+    "    }\n",
+    "    \n",
+    "    for i, sample in enumerate(paired_samples):\n",
+    "        if i % 10000 == 0:\n",
+    "            print(f\"数据收集进度: {i}/{len(paired_samples)}\")\n",
+    "            \n",
+    "        dataset_dict['canonical_smiles'].append(sample['canonical_smiles'])\n",
+    "        dataset_dict['control_CP'].append(CP_data['control'][sample['cp_index']])\n",
+    "        dataset_dict['target_CP'].append(CP_data['target'][sample['cp_index']])\n",
+    "        dataset_dict['control_GE'].append(GE_data['control'][sample['ge_index']]) \n",
+    "        dataset_dict['target_GE'].append(GE_data['target'][sample['ge_index']])\n",
+    "    \n",
+    "    # 转换为numpy数组并进行数据清洗\n",
+    "    final_dataset = {}\n",
+    "    final_dataset['canonical_smiles'] = np.array(dataset_dict['canonical_smiles'], dtype='S')\n",
+    "    \n",
+    "    # 数据清洗函数\n",
+    "    def clean_and_validate_data(data_list, name):\n",
+    "        data_array = np.array(data_list, dtype=np.float32)\n",
+    "        \n",
+    "        # 检查NaN和Inf\n",
+    "        nan_count = np.isnan(data_array).sum()\n",
+    "        inf_count = np.isinf(data_array).sum()\n",
+    "        \n",
+    "        if nan_count > 0 or inf_count > 0:\n",
+    "            print(f\"警告: {name} 包含 {nan_count} NaN 和 {inf_count} Inf\")\n",
+    "            data_array = np.nan_to_num(data_array, nan=0.0, posinf=1.0, neginf=-1.0)\n",
+    "        \n",
+    "        # 数据标准化\n",
+    "        mean = data_array.mean()\n",
+    "        std = data_array.std()\n",
+    "        \n",
+    "        print(f\"{name}: 均值={mean:.4f}, 标准差={std:.4f}, 形状={data_array.shape}\")\n",
+    "        \n",
+    "        return data_array\n",
+    "    \n",
+    "    final_dataset['control_CP'] = clean_and_validate_data(dataset_dict['control_CP'], 'control_CP')\n",
+    "    final_dataset['target_CP'] = clean_and_validate_data(dataset_dict['target_CP'], 'target_CP')\n",
+    "    final_dataset['control_GE'] = clean_and_validate_data(dataset_dict['control_GE'], 'control_GE')\n",
+    "    final_dataset['target_GE'] = clean_and_validate_data(dataset_dict['target_GE'], 'target_GE')\n",
+    "    \n",
+    "    # 保存为HDF5\n",
+    "    with h5py.File(out_path, 'w') as f:\n",
+    "        for key, value in final_dataset.items():\n",
+    "            f.create_dataset(key, data=value)\n",
+    "    \n",
+    "    print(f\"✅ 稳健配对数据集已保存到: {out_path}\")\n",
+    "    return final_dataset\n",
+    "\n",
+    "\n",
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "paired_dataset = create_robust_paired_dataset(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

MVC/MVC_BBBC036/Step3_aggreate_2modality.ipynb

@@ -0,0 +1,1153 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 数据聚合：「输入药物 SMILES，输出基因表达 + 细胞形态」的多模态虚拟细胞模型，配对不齐，训练集不均衡。\n",
+    "\n",
+    "## 策略有3种：\n",
+    "\n",
+    "### 1 是按 SMILES 聚合，对每个 SMILES 在每个模态上 取平均或中位数（或更复杂的 embedding 聚合）。\n",
+    "\n",
+    "### 2 是多实例学习（Multiple Instance Learning, MIL）：思路：把一个 SMILES 的所有 replicate 当作一个 bag，模型学习 bag-level 对齐。\n",
+    "\n",
+    "### 3 是不对齐 replicate，随机采样：思路：训练时，每个 batch 从两个模态中随机采样该 SMILES 的一个 replicate，逐步逼近两个分布。"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "CP_path =  './Processed_Paired_CP_KeepRep.h5'\n",
+    "GE_path = './Processed_Paired_GE_KeepRep.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((15012, 602), (3451, 977))"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV['target'].shape, GE_CSV['target'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['CCCOc1cc(N)ccc1C(=O)OCCN(CC)CC',\n",
+       "        'C=C(C)[C@@H]1CCC(C)=C[C@H]1c1c(OC)cc(C)cc1OC',\n",
+       "        'COc1cc(O)cc(/C=C/c2ccccc2)c1', ...,\n",
+       "        'CC(C)[C@H]1c2ccc(F)cc2CC[C@@]1(CCN(C)CCCc1nc2ccccc2[nH]1)OC(=O)C1CC1',\n",
+       "        'CCCCC/C=C\\\\C/C=C\\\\C/C=C\\\\CCCCCCC(=O)NCCO',\n",
+       "        'COc1cc(OC)c(C(=O)c2ccc(O)cc2)c(OC)c1'], dtype='<U227'),\n",
+       " 'control': array([[ 0.05088183, -0.16616029,  0.19908461, ...,  0.2275372 ,\n",
+       "          0.0942523 ,  0.31641835],\n",
+       "        [ 0.05088183, -0.16616029,  0.19908461, ...,  0.2275372 ,\n",
+       "          0.0942523 ,  0.31641835],\n",
+       "        [ 0.05088183, -0.16616029,  0.19908461, ...,  0.2275372 ,\n",
+       "          0.0942523 ,  0.31641835],\n",
+       "        ...,\n",
+       "        [ 0.08850026,  0.15368527, -0.09534353, ...,  0.06442614,\n",
+       "          0.06431134,  0.1678016 ],\n",
+       "        [ 0.08850026,  0.15368527, -0.09534353, ...,  0.06442614,\n",
+       "          0.06431134,  0.1678016 ],\n",
+       "        [ 0.08850026,  0.15368527, -0.09534353, ...,  0.06442614,\n",
+       "          0.06431134,  0.1678016 ]], dtype=float32),\n",
+       " 'target': array([[-0.8649911 ,  1.8669271 , -1.0574424 , ..., -1.7174782 ,\n",
+       "         -2.8599036 ,  0.525918  ],\n",
+       "        [ 0.50881827,  0.30993956,  0.5643427 , ..., -1.3119696 ,\n",
+       "         -0.96582466,  1.2100515 ],\n",
+       "        [-1.6027776 , -0.8187241 ,  0.49752134, ..., -0.10791475,\n",
+       "         -0.19130155,  0.8326575 ],\n",
+       "        ...,\n",
+       "        [-7.535165  , -8.924246  , 10.        , ..., 10.        ,\n",
+       "          0.82093847,  0.943286  ],\n",
+       "        [-1.7700051 ,  1.3930917 , -2.890018  , ..., -0.98175496,\n",
+       "         -3.2825868 , -2.1236746 ],\n",
+       "        [-4.9813    ,  3.261372  , -2.212747  , ..., -0.1282133 ,\n",
+       "         -2.6824005 , -1.7763886 ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['CCCOc1cc(N)ccc1C(=O)OCCN(CC)CC',\n",
+       "        'C=C(C)[C@@H]1CCC(C)=C[C@H]1c1c(OC)cc(C)cc1OC',\n",
+       "        'COc1cc(O)cc(/C=C/c2ccccc2)c1', ...,\n",
+       "        'CC(C)[C@H]1c2ccc(F)cc2CC[C@@]1(CCN(C)CCCc1nc2ccccc2[nH]1)OC(=O)C1CC1',\n",
+       "        'CCCCC/C=C\\\\C/C=C\\\\C/C=C\\\\CCCCCCC(=O)NCCO',\n",
+       "        'COc1cc(OC)c(C(=O)c2ccc(O)cc2)c(OC)c1'], dtype='<U227'),\n",
+       " 'control': array([[ 0.05088183, -0.16616029,  0.19908461, ...,  0.2275372 ,\n",
+       "          0.0942523 ,  0.31641835],\n",
+       "        [ 0.05088183, -0.16616029,  0.19908461, ...,  0.2275372 ,\n",
+       "          0.0942523 ,  0.31641835],\n",
+       "        [ 0.05088183, -0.16616029,  0.19908461, ...,  0.2275372 ,\n",
+       "          0.0942523 ,  0.31641835],\n",
+       "        ...,\n",
+       "        [ 0.08850026,  0.15368527, -0.09534353, ...,  0.06442614,\n",
+       "          0.06431134,  0.1678016 ],\n",
+       "        [ 0.08850026,  0.15368527, -0.09534353, ...,  0.06442614,\n",
+       "          0.06431134,  0.1678016 ],\n",
+       "        [ 0.08850026,  0.15368527, -0.09534353, ...,  0.06442614,\n",
+       "          0.06431134,  0.1678016 ]], dtype=float32),\n",
+       " 'target': array([[-0.8649911 ,  1.8669271 , -1.0574424 , ..., -1.7174782 ,\n",
+       "         -2.8599036 ,  0.525918  ],\n",
+       "        [ 0.50881827,  0.30993956,  0.5643427 , ..., -1.3119696 ,\n",
+       "         -0.96582466,  1.2100515 ],\n",
+       "        [-1.6027776 , -0.8187241 ,  0.49752134, ..., -0.10791475,\n",
+       "         -0.19130155,  0.8326575 ],\n",
+       "        ...,\n",
+       "        [-7.535165  , -8.924246  , 10.        , ..., 10.        ,\n",
+       "          0.82093847,  0.943286  ],\n",
+       "        [-1.7700051 ,  1.3930917 , -2.890018  , ..., -0.98175496,\n",
+       "         -3.2825868 , -2.1236746 ],\n",
+       "        [-4.9813    ,  3.261372  , -2.212747  , ..., -0.1282133 ,\n",
+       "         -2.6824005 , -1.7763886 ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((14996, 602), (3451, 977))"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV['target'].shape, GE_CSV['target'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 按照第一种方式直接 mean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ 聚合后的数据已保存到 ./Aggregated_CP_GE.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "def aggregate_modalities(CP, GE, out_path):\n",
+    "    # 转成 DataFrame 方便聚合\n",
+    "    CP_df = pd.DataFrame({\n",
+    "        \"SMILES\": CP[\"canonical_smiles\"],\n",
+    "        \"control\": list(CP[\"control\"]),\n",
+    "        \"target\": list(CP[\"target\"])\n",
+    "    })\n",
+    "    GE_df = pd.DataFrame({\n",
+    "        \"SMILES\": GE[\"canonical_smiles\"],\n",
+    "        \"control\": list(GE[\"control\"]),\n",
+    "        \"target\": list(GE[\"target\"])\n",
+    "    })\n",
+    "\n",
+    "    # groupby + mean pooling\n",
+    "    CP_grouped = CP_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "    GE_grouped = GE_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "\n",
+    "    # merge 保证对齐\n",
+    "    merged = pd.merge(CP_grouped, GE_grouped, on=\"SMILES\", how=\"inner\", suffixes=(\"_CP\", \"_GE\"))\n",
+    "\n",
+    "    # 转 numpy array 并替换剩余 NaN\n",
+    "    def clean_array(arr_list):\n",
+    "        arr = np.stack(arr_list.to_numpy())\n",
+    "        arr = np.nan_to_num(arr, nan=0.0, posinf=0.0, neginf=0.0)  # 替换 NaN/Inf\n",
+    "        return arr.astype(np.float32)\n",
+    "\n",
+    "    smiles_array     = merged[\"SMILES\"].to_numpy(dtype=\"S\")\n",
+    "    control_CP_array = clean_array(merged[\"control_CP\"])\n",
+    "    target_CP_array  = clean_array(merged[\"target_CP\"])\n",
+    "    control_GE_array = clean_array(merged[\"control_GE\"])\n",
+    "    target_GE_array  = clean_array(merged[\"target_GE\"])\n",
+    "\n",
+    "    # 存 HDF5\n",
+    "    with h5py.File(out_path, \"w\") as f:\n",
+    "        f.create_dataset(\"canonical_smiles\", data=smiles_array)\n",
+    "        f.create_dataset(\"control_CP\", data=control_CP_array)\n",
+    "        f.create_dataset(\"target_CP\", data=target_CP_array)\n",
+    "        f.create_dataset(\"control_GE\", data=control_GE_array)\n",
+    "        f.create_dataset(\"target_GE\", data=target_GE_array)\n",
+    "\n",
+    "    print(f\"✅ 聚合后的数据已保存到 {out_path}\")\n",
+    "\n",
+    "# 用法\n",
+    "# root = '/home/bob/boom/VCBench/data/MVC/CDRP-BBBC047-Bray/'\n",
+    "CP_path = './Processed_Paired_CP.h5'\n",
+    "GE_path = './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)\n",
+    "\n",
+    "out_path = './Aggregated_CP_GE.h5'\n",
+    "aggregate_modalities(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['Brc1c(NC2=NCCN2)ccc2nccnc12', 'C#CCN(C)CCCOc1ccc(Cl)cc1Cl',\n",
+       "        'C#CCN1C(C)=C(C(=O)CCO)C(c2cccc(-n3oo3)c2)C(C(=O)OC)=C1C', ...,\n",
+       "        'c1coc(CNc2ncnc3[nH]cnc23)c1', 'c1nc(CC2CCNCC2)c[nH]1',\n",
+       "        'c1nc(CCCC2CCNCC2)c[nH]1'], dtype='<U227'),\n",
+       " 'control_CP': array([[-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298],\n",
+       "        [-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298],\n",
+       "        [-0.03419905, -0.1922472 ,  0.18566999, ...,  0.17574853,\n",
+       "          0.33123744,  0.21626002],\n",
+       "        ...,\n",
+       "        [-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298],\n",
+       "        [-0.04489063, -0.14671806,  0.12026446, ...,  0.12773524,\n",
+       "          0.21260148,  0.16793586],\n",
+       "        [-0.03389427, -0.14150065,  0.15008315, ...,  0.11754317,\n",
+       "          0.21078211,  0.13338298]], dtype=float32),\n",
+       " 'control_GE': array([[ 0.01338525, -0.02699645,  0.0147262 , ..., -0.01713243,\n",
+       "         -0.02412532,  0.08308113],\n",
+       "        [-0.05236533, -0.00944031,  0.01470257, ..., -0.1302189 ,\n",
+       "          0.1095909 , -0.00176817],\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433],\n",
+       "        ...,\n",
+       "        [-0.05236533, -0.00944031,  0.01470257, ..., -0.1302189 ,\n",
+       "          0.1095909 , -0.00176817],\n",
+       "        [ 0.02062   , -0.00183125, -0.13683268, ..., -0.17978694,\n",
+       "          0.21799538,  0.17114288],\n",
+       "        [ 0.01338525, -0.02699645,  0.0147262 , ..., -0.01713243,\n",
+       "         -0.02412532,  0.08308113]], dtype=float32),\n",
+       " 'target_CP': array([[ 0.47359726,  0.3149385 , -0.2134298 , ...,  0.14258505,\n",
+       "         -0.23095778,  1.3368101 ],\n",
+       "        [-0.2682222 , -0.6758744 ,  0.3996619 , ...,  0.35662955,\n",
+       "         -0.12905961, -0.09158537],\n",
+       "        [ 0.37451744, -0.25577283,  0.26045328, ..., -0.23579153,\n",
+       "         -0.23442191, -0.41257128],\n",
+       "        ...,\n",
+       "        [-0.1709819 , -0.5260589 ,  0.4987828 , ...,  0.5009768 ,\n",
+       "          0.41748047,  0.18332437],\n",
+       "        [ 0.13757834, -0.20250246,  0.19951464, ..., -0.3296469 ,\n",
+       "         -0.11514819, -0.1843171 ],\n",
+       "        [-0.46352574,  0.35824373, -0.32413673, ..., -0.31744272,\n",
+       "         -0.5073174 , -0.03345545]], dtype=float32),\n",
+       " 'target_GE': array([[-0.07302775,  0.0460825 , -0.02222175, ...,  0.17655626,\n",
+       "         -0.154201  ,  0.033125  ],\n",
+       "        [ 0.10191549, -0.24803169,  0.11954501, ..., -0.16921   ,\n",
+       "          0.02239785, -0.0529015 ],\n",
+       "        [-0.034165  ,  0.048892  ,  0.1231085 , ...,  0.42318   ,\n",
+       "          0.034605  ,  0.21026   ],\n",
+       "        ...,\n",
+       "        [ 0.0316805 , -0.1364167 ,  0.02315   , ...,  0.10307   ,\n",
+       "          0.17449784,  0.1486385 ],\n",
+       "        [ 0.22472501,  0.20904249, -0.2872    , ..., -0.22035149,\n",
+       "          0.5861085 ,  0.278331  ],\n",
+       "        [ 0.01292225,  0.07139751, -0.13613375, ..., -0.16344374,\n",
+       "          0.053449  , -0.200645  ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out_path = './Aggregated_CP_GE.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((1916, 602), (1916, 602), (1916, 977), (1916, 977))"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape, agg_data['target_CP'].shape, agg_data['control_GE'].shape, agg_data['target_GE'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 21,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "'O=C(C[C@@H]1CC[C@@H]2[C@H](COC[C@H](O)CN2Cc2cc(F)cc(F)c2)O1)Nc1nccs1' in agg_data['canonical_smiles']"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 复杂点，样本级别的配对"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "加载源数据...\n",
+      "开始创建多模态配对数据集...\n",
+      "Cell Painting 唯一化合物数: 1916\n",
+      "Gene Expression 唯一化合物数: 1916\n",
+      "多模态共有化合物数 (Intersection): 1916\n",
+      "生成配对样本总数: 17253\n",
+      "正在组装最终矩阵...\n",
+      "✅ 最终数据集已保存至: ./Paired_CP_GE_Dataset.h5\n",
+      "------------------------------\n",
+      "Dataset Shapes:\n",
+      "canonical_smiles: (17253,)\n",
+      "control_CP: (17253, 602)\n",
+      "target_CP: (17253, 602)\n",
+      "control_GE: (17253, 977)\n",
+      "target_GE: (17253, 977)\n",
+      "CP_dose: (17253,)\n",
+      "GE_dose: (17253,)\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "# ==========================================\n",
+    "# 0. 辅助函数：读取 HDF5\n",
+    "# ==========================================\n",
+    "def load_from_HDF(fname):\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            # 处理字符串解码\n",
+    "            if data[key].dtype.kind == 'S':\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n",
+    "\n",
+    "# ==========================================\n",
+    "# 1. 核心逻辑：创建稳健配对数据集\n",
+    "# ==========================================\n",
+    "def create_robust_paired_dataset(CP_data, GE_data, out_path, max_samples_per_smiles=5, seed=42):\n",
+    "    \"\"\"\n",
+    "    CP_data, GE_data: 包含 'canonical_smiles', 'control', 'target', 'dose' 的字典\n",
+    "    max_samples_per_smiles: 每个模态最多采样的重复数\n",
+    "    seed: 随机种子，保证配对结果可复现\n",
+    "    \"\"\"\n",
+    "    print(\"开始创建多模态配对数据集...\")\n",
+    "    \n",
+    "    # [新增] 检查输入数据是否包含 dose\n",
+    "    if 'dose' not in CP_data or 'dose' not in GE_data:\n",
+    "        raise ValueError(\"输入数据缺少 'dose' 字段，请检查上一步是否正确保存了剂量信息。\")\n",
+    "\n",
+    "    # [新增] 设置随机种子\n",
+    "    np.random.seed(seed)\n",
+    "    \n",
+    "    # --- A. 构建索引映射 ---\n",
+    "    cp_smiles_to_indices = defaultdict(list)\n",
+    "    ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "        cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "        ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    # --- B. 找交集 ---\n",
+    "    common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "    print(f\"Cell Painting 唯一化合物数: {len(cp_smiles_to_indices)}\")\n",
+    "    print(f\"Gene Expression 唯一化合物数: {len(ge_smiles_to_indices)}\")\n",
+    "    print(f\"多模态共有化合物数 (Intersection): {len(common_smiles)}\")\n",
+    "    \n",
+    "    # --- C. 生成配对索引 ---\n",
+    "    paired_indices = []\n",
+    "    \n",
+    "    for i, smiles in enumerate(common_smiles):\n",
+    "        # 获取该 SMILES 的所有索引\n",
+    "        cp_indices = cp_smiles_to_indices[smiles]\n",
+    "        ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "        # 限制采样数量\n",
+    "        n_cp = min(max_samples_per_smiles, len(cp_indices))\n",
+    "        n_ge = min(max_samples_per_smiles, len(ge_indices))\n",
+    "        \n",
+    "        # 无放回随机采样\n",
+    "        sampled_cp_indices = np.random.choice(cp_indices, n_cp, replace=False)\n",
+    "        sampled_ge_indices = np.random.choice(ge_indices, n_ge, replace=False)\n",
+    "        \n",
+    "        # 笛卡尔积配对\n",
+    "        for cp_idx in sampled_cp_indices:\n",
+    "            for ge_idx in sampled_ge_indices:\n",
+    "                paired_indices.append({\n",
+    "                    'smiles': smiles,\n",
+    "                    'cp_idx': cp_idx,\n",
+    "                    'ge_idx': ge_idx\n",
+    "                })\n",
+    "                \n",
+    "    print(f\"生成配对样本总数: {len(paired_indices)}\")\n",
+    "\n",
+    "    # --- D. 组装数据 ---\n",
+    "    n_samples = len(paired_indices)\n",
+    "    \n",
+    "    # 获取特征维度\n",
+    "    cp_feat_dim = CP_data['target'].shape[1]\n",
+    "    ge_feat_dim = GE_data['target'].shape[1]\n",
+    "    \n",
+    "    # 预分配内存\n",
+    "    final_data = {\n",
+    "        'canonical_smiles': np.empty(n_samples, dtype='S200'),\n",
+    "        'control_CP': np.empty((n_samples, cp_feat_dim), dtype=np.float32),\n",
+    "        'target_CP':  np.empty((n_samples, cp_feat_dim), dtype=np.float32),\n",
+    "        'control_GE': np.empty((n_samples, ge_feat_dim), dtype=np.float32),\n",
+    "        'target_GE':  np.empty((n_samples, ge_feat_dim), dtype=np.float32),\n",
+    "        # [新增] 两个模态的剂量\n",
+    "        'CP_dose':    np.empty(n_samples, dtype=np.float32), \n",
+    "        'GE_dose':    np.empty(n_samples, dtype=np.float32)\n",
+    "    }\n",
+    "    \n",
+    "    print(\"正在组装最终矩阵...\")\n",
+    "    for i, item in enumerate(paired_indices):\n",
+    "        cp_idx = item['cp_idx']\n",
+    "        ge_idx = item['ge_idx']\n",
+    "        \n",
+    "        final_data['canonical_smiles'][i] = item['smiles'].encode('utf-8')\n",
+    "        \n",
+    "        # 填入 CP 数据\n",
+    "        final_data['control_CP'][i]       = CP_data['control'][cp_idx]\n",
+    "        final_data['target_CP'][i]        = CP_data['target'][cp_idx]\n",
+    "        final_data['CP_dose'][i]          = CP_data['dose'][cp_idx] # [新增]\n",
+    "        \n",
+    "        # 填入 GE 数据\n",
+    "        final_data['control_GE'][i]       = GE_data['control'][ge_idx]\n",
+    "        final_data['target_GE'][i]        = GE_data['target'][ge_idx]\n",
+    "        final_data['GE_dose'][i]          = GE_data['dose'][ge_idx] # [新增]\n",
+    "        \n",
+    "    # --- E. 数据清洗与保存 ---\n",
+    "    # 检查并处理 NaN/Inf (Dose 通常不需要处理，但如果有 NaN 也顺便处理一下安全)\n",
+    "    check_cols = ['control_CP', 'target_CP', 'control_GE', 'target_GE', 'CP_dose', 'GE_dose']\n",
+    "    \n",
+    "    for key in check_cols:\n",
+    "        if np.isnan(final_data[key]).any() or np.isinf(final_data[key]).any():\n",
+    "            print(f\"警告: {key} 包含 NaN 或 Inf，正在填充 0...\")\n",
+    "            final_data[key] = np.nan_to_num(final_data[key], nan=0.0, posinf=0.0, neginf=0.0)\n",
+    "\n",
+    "    # 保存\n",
+    "    with h5py.File(out_path, 'w') as f:\n",
+    "        for key, val in final_data.items():\n",
+    "            f.create_dataset(key, data=val)\n",
+    "            \n",
+    "    print(f\"✅ 最终数据集已保存至: {out_path}\")\n",
+    "    print(\"-\" * 30)\n",
+    "    print(f\"Dataset Shapes:\")\n",
+    "    for k, v in final_data.items():\n",
+    "        print(f\"{k}: {v.shape}\")\n",
+    "\n",
+    "# ==========================================\n",
+    "# 2. 执行流程\n",
+    "# ==========================================\n",
+    "# 定义输入输出路径\n",
+    "cp_input = './Processed_Paired_CP_KeepRep.h5'\n",
+    "ge_input = './Processed_Paired_GE_KeepRep.h5'\n",
+    "final_output = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "# 加载数据\n",
+    "print(\"加载源数据...\")\n",
+    "cp_data = load_from_HDF(cp_input)\n",
+    "ge_data = load_from_HDF(ge_input)\n",
+    "\n",
+    "# 执行配对\n",
+    "create_robust_paired_dataset(cp_data, ge_data, final_output, max_samples_per_smiles=5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['O=C(O)COC(=O)Cc1ccccc1Nc1c(Cl)cccc1Cl',\n",
+       "        'O=C(O)COC(=O)Cc1ccccc1Nc1c(Cl)cccc1Cl',\n",
+       "        'O=C(O)COC(=O)Cc1ccccc1Nc1c(Cl)cccc1Cl', ...,\n",
+       "        'CN(C(=O)c1ccc(Cl)c(Cl)c1)[C@@H]1CCCC[C@@H]1N1CCCC1',\n",
+       "        'CN(C(=O)c1ccc(Cl)c(Cl)c1)[C@@H]1CCCC[C@@H]1N1CCCC1',\n",
+       "        'CN(C(=O)c1ccc(Cl)c(Cl)c1)[C@@H]1CCCC[C@@H]1N1CCCC1'],\n",
+       "       dtype='<U200'),\n",
+       " 'control_CP': array([[-0.05779983, -0.27169958,  0.23867248, ...,  0.06572847,\n",
+       "          0.22877644,  0.16598275],\n",
+       "        [-0.05779983, -0.27169958,  0.23867248, ...,  0.06572847,\n",
+       "          0.22877644,  0.16598275],\n",
+       "        [-0.01651424, -0.13550141,  0.04095873, ..., -0.03688414,\n",
+       "          0.08280301, -0.15823033],\n",
+       "        ...,\n",
+       "        [ 0.0168145 , -0.10370001,  0.10444786, ..., -0.03251424,\n",
+       "          0.18788806, -0.09417315],\n",
+       "        [-0.09153318, -0.27400595,  0.30783677, ..., -0.01151957,\n",
+       "          0.23429662, -0.05841342],\n",
+       "        [-0.09153318, -0.27400595,  0.30783677, ..., -0.01151957,\n",
+       "          0.23429662, -0.05841342]], dtype=float32),\n",
+       " 'control_GE': array([[ 0.00299   ,  0.027625  ,  0.10459   , ..., -0.019535  ,\n",
+       "         -0.045362  ,  0.11702   ],\n",
+       "        [ 0.0519945 , -0.128145  , -0.0311875 , ...,  0.0761325 ,\n",
+       "         -0.03931   ,  0.03072   ],\n",
+       "        [ 0.00299   ,  0.027625  ,  0.10459   , ..., -0.019535  ,\n",
+       "         -0.045362  ,  0.11702   ],\n",
+       "        ...,\n",
+       "        [ 0.025065  ,  0.038105  , -0.0712    , ..., -0.08888   ,\n",
+       "          0.038267  , -0.057905  ],\n",
+       "        [ 0.00638   , -0.0219    , -0.15533999, ..., -0.203183  ,\n",
+       "          0.40017   ,  0.176652  ],\n",
+       "        [ 0.025065  ,  0.038105  , -0.0712    , ..., -0.08888   ,\n",
+       "          0.038267  , -0.057905  ]], dtype=float32),\n",
+       " 'target_CP': array([[ 0.16514239, -0.27579018, -0.5783885 , ...,  0.1167318 ,\n",
+       "          0.60331386, -0.40385294],\n",
+       "        [ 0.16514239, -0.27579018, -0.5783885 , ...,  0.1167318 ,\n",
+       "          0.60331386, -0.40385294],\n",
+       "        [-0.19817086,  0.19117425, -0.85260695, ...,  0.03670187,\n",
+       "          0.22386315, -0.41378257],\n",
+       "        ...,\n",
+       "        [ 1.1964315 , -0.270127  , -0.756721  , ..., -0.6266768 ,\n",
+       "         -1.5699604 , -1.0493801 ],\n",
+       "        [ 0.8361844 , -0.17571016, -0.70972115, ...,  0.06683134,\n",
+       "          1.4995568 ,  1.1771296 ],\n",
+       "        [ 0.8361844 , -0.17571016, -0.70972115, ...,  0.06683134,\n",
+       "          1.4995568 ,  1.1771296 ]], dtype=float32),\n",
+       " 'target_GE': array([[-0.18275  , -0.0799   ,  0.13112  , ..., -0.19384  ,  0.117488 ,\n",
+       "          0.13002  ],\n",
+       "        [ 0.1289945,  0.021015 , -0.0417475, ..., -0.1044675,  0.10721  ,\n",
+       "         -0.33066  ],\n",
+       "        [-0.18275  , -0.0799   ,  0.13112  , ..., -0.19384  ,  0.117488 ,\n",
+       "          0.13002  ],\n",
+       "        ...,\n",
+       "        [ 0.46645  , -0.22858  ,  0.32009  , ..., -0.07213  , -0.101373 ,\n",
+       "          0.24146  ],\n",
+       "        [-0.20657  ,  0.306565 ,  0.19061  , ...,  0.423967 , -0.92223  ,\n",
+       "         -0.175798 ],\n",
+       "        [ 0.46645  , -0.22858  ,  0.32009  , ..., -0.07213  , -0.101373 ,\n",
+       "          0.24146  ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_path =  './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "df_CSV = load_from_HDF(CP_path)\n",
+    "\n",
+    "df_CSV"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(17253, 602) (17253, 602) (17253, 977) (17253, 977)\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(df_CSV['target_CP'].shape, df_CSV['control_CP'].shape, df_CSV['control_GE'].shape, df_CSV['target_GE'].shape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "共有化合物: 1916\n",
+      "总配对样本数: 26948\n",
+      "数据集形状:\n",
+      "  control_CP: (26948, 602)\n",
+      "  target_CP: (26948, 602)\n",
+      "  control_GE: (26948, 977)\n",
+      "  target_GE: (26948, 977)\n",
+      "✅ 配对数据集已保存到: ./Paired_CP_GE_Dataset.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "def create_paired_dataset(CP_data, GE_data, out_path):\n",
+    "    \"\"\"\n",
+    "    创建样本级配对的多模态数据集\n",
+    "    保持原始样本的完整性，只配对真正对应的样本\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    # 构建SMILES到索引列表的映射\n",
+    "    cp_smiles_to_indices = defaultdict(list)\n",
+    "    ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "        cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "        ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    # 找到共有的SMILES\n",
+    "    common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "    print(f\"共有化合物: {len(common_smiles)}\")\n",
+    "    \n",
+    "    # 收集所有有效配对\n",
+    "    paired_samples = []\n",
+    "    \n",
+    "    for smiles in common_smiles:\n",
+    "        cp_indices = cp_smiles_to_indices[smiles]\n",
+    "        ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "        # 为每个SMILES创建���有可能的CP-GE配对\n",
+    "        for cp_idx in cp_indices:\n",
+    "            for ge_idx in ge_indices:\n",
+    "                paired_samples.append({\n",
+    "                    'canonical_smiles': smiles,\n",
+    "                    'cp_index': cp_idx,\n",
+    "                    'ge_index': ge_idx\n",
+    "                })\n",
+    "    \n",
+    "    print(f\"总配对样本数: {len(paired_samples)}\")\n",
+    "    \n",
+    "    # 创建最终的字典形数据集\n",
+    "    dataset_dict = {\n",
+    "        'canonical_smiles': [],\n",
+    "        'control_CP': [],\n",
+    "        'target_CP': [], \n",
+    "        'control_GE': [],\n",
+    "        'target_GE': []\n",
+    "    }\n",
+    "    \n",
+    "    for sample in paired_samples:\n",
+    "        dataset_dict['canonical_smiles'].append(sample['canonical_smiles'])\n",
+    "        dataset_dict['control_CP'].append(CP_data['control'][sample['cp_index']])\n",
+    "        dataset_dict['target_CP'].append(CP_data['target'][sample['cp_index']])\n",
+    "        dataset_dict['control_GE'].append(GE_data['control'][sample['ge_index']]) \n",
+    "        dataset_dict['target_GE'].append(GE_data['target'][sample['ge_index']])\n",
+    "    \n",
+    "    # 转换为numpy数组\n",
+    "    final_dataset = {}\n",
+    "    final_dataset['canonical_smiles'] = np.array(dataset_dict['canonical_smiles'], dtype='S')\n",
+    "    final_dataset['control_CP'] = np.array(dataset_dict['control_CP'], dtype=np.float32)\n",
+    "    final_dataset['target_CP'] = np.array(dataset_dict['target_CP'], dtype=np.float32)\n",
+    "    final_dataset['control_GE'] = np.array(dataset_dict['control_GE'], dtype=np.float32)\n",
+    "    final_dataset['target_GE'] = np.array(dataset_dict['target_GE'], dtype=np.float32)\n",
+    "    \n",
+    "    print(f\"数据集形状:\")\n",
+    "    print(f\"  control_CP: {final_dataset['control_CP'].shape}\")\n",
+    "    print(f\"  target_CP: {final_dataset['target_CP'].shape}\")\n",
+    "    print(f\"  control_GE: {final_dataset['control_GE'].shape}\")\n",
+    "    print(f\"  target_GE: {final_dataset['target_GE'].shape}\")\n",
+    "    \n",
+    "    # 保存为HDF5\n",
+    "    with h5py.File(out_path, 'w') as f:\n",
+    "        for key, value in final_dataset.items():\n",
+    "            f.create_dataset(key, data=value)\n",
+    "    \n",
+    "    print(f\"✅ 配对数据集已保存到: {out_path}\")\n",
+    "    return final_dataset\n",
+    "\n",
+    "# 使用方案\n",
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "paired_dataset = create_paired_dataset(CP_CSV, GE_CSV, out_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['CC(C)(C)c1ccc(S(=O)(=O)/C=C/C#N)cc1',\n",
+       "        'CC(C)(C)c1ccc(S(=O)(=O)/C=C/C#N)cc1',\n",
+       "        'CC(C)(C)c1ccc(S(=O)(=O)/C=C/C#N)cc1', ...,\n",
+       "        'CCC(Cc1c(I)cc(I)c(N)c1I)C(=O)O', 'CCC(Cc1c(I)cc(I)c(N)c1I)C(=O)O',\n",
+       "        'CCC(Cc1c(I)cc(I)c(N)c1I)C(=O)O'], dtype='<U227'),\n",
+       " 'control_CP': array([[-0.05367128, -0.22861896,  0.19134662, ...,  0.09901066,\n",
+       "          0.23015772,  0.0870431 ],\n",
+       "        [-0.05367128, -0.22861896,  0.19134662, ...,  0.09901066,\n",
+       "          0.23015772,  0.0870431 ],\n",
+       "        [-0.04425929, -0.20467344,  0.19937934, ...,  0.10667558,\n",
+       "          0.25934276,  0.11115717],\n",
+       "        ...,\n",
+       "        [-0.00746798, -0.14845599,  0.17686495, ...,  0.18434836,\n",
+       "          0.3559847 ,  0.20929421],\n",
+       "        [-0.0670732 , -0.11723144,  0.11542749, ...,  0.15660104,\n",
+       "          0.31949052,  0.1973684 ],\n",
+       "        [-0.10378247, -0.11118906,  0.11418191, ...,  0.20794444,\n",
+       "          0.3261717 ,  0.20840377]], dtype=float32),\n",
+       " 'control_GE': array([[ 0.008735  , -0.0015245 , -0.00072016, ...,  0.044401  ,\n",
+       "         -0.2373012 , -0.019055  ],\n",
+       "        [-0.16669   ,  0.03942075, -0.1140243 , ..., -0.3791175 ,\n",
+       "          0.35399476, -0.2483125 ],\n",
+       "        [ 0.008735  , -0.0015245 , -0.00072016, ...,  0.044401  ,\n",
+       "         -0.2373012 , -0.019055  ],\n",
+       "        ...,\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433],\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433],\n",
+       "        [-0.04210533,  0.04783978,  0.09477517, ...,  0.01774443,\n",
+       "         -0.04355722, -0.13316433]], dtype=float32),\n",
+       " 'target_CP': array([[-2.475722  , -1.0180573 , -2.4777994 , ..., -1.6663419 ,\n",
+       "          6.1886635 ,  6.246197  ],\n",
+       "        [-2.475722  , -1.0180573 , -2.4777994 , ..., -1.6663419 ,\n",
+       "          6.1886635 ,  6.246197  ],\n",
+       "        [-2.475722  ,  3.6353707 , -1.1177131 , ...,  3.3711612 ,\n",
+       "          6.1886635 ,  6.246197  ],\n",
+       "        ...,\n",
+       "        [-0.7142158 ,  2.8495595 , -2.0671365 , ..., -0.34750465,\n",
+       "         -0.56493837, -0.8109943 ],\n",
+       "        [ 1.1229682 ,  0.68502927, -0.7133282 , ...,  0.237633  ,\n",
+       "          0.43848613, -0.48886108],\n",
+       "        [-0.6347912 ,  1.2815117 , -0.40972736, ..., -0.59789634,\n",
+       "         -0.81882066, -0.80796003]], dtype=float32),\n",
+       " 'target_GE': array([[-0.217655  , -0.0130605 , -0.32715666, ..., -0.279299  ,\n",
+       "          0.4241275 ,  1.109114  ],\n",
+       "        [-0.918455  ,  0.3162145 , -0.4372258 , ..., -2.098005  ,\n",
+       "          0.3910935 ,  0.438925  ],\n",
+       "        [-0.217655  , -0.0130605 , -0.32715666, ..., -0.279299  ,\n",
+       "          0.4241275 ,  1.109114  ],\n",
+       "        ...,\n",
+       "        [ 0.213685  ,  0.115822  , -0.3632915 , ...,  0.15678   ,\n",
+       "          0.091715  ,  0.10284   ],\n",
+       "        [ 0.213685  ,  0.115822  , -0.3632915 , ...,  0.15678   ,\n",
+       "          0.091715  ,  0.10284   ],\n",
+       "        [ 0.213685  ,  0.115822  , -0.3632915 , ...,  0.15678   ,\n",
+       "          0.091715  ,  0.10284   ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((26948, 602), (26948, 602), (26948, 977), (26948, 977))"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape, agg_data['target_CP'].shape, agg_data['control_GE'].shape, agg_data['target_GE'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 限制配对的数量，更好收敛。"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "开始创建稳健配对数据集...\n",
+      "共有化合物: 20341\n",
+      "处理进度: 0/20341\n",
+      "处理进度: 1000/20341\n",
+      "处理进度: 2000/20341\n",
+      "处理进度: 3000/20341\n",
+      "处理进度: 4000/20341\n",
+      "处理进度: 5000/20341\n",
+      "处理进度: 6000/20341\n",
+      "处理进度: 7000/20341\n",
+      "处理进度: 8000/20341\n",
+      "处理进度: 9000/20341\n",
+      "处理进度: 10000/20341\n",
+      "处理进度: 11000/20341\n",
+      "处理进度: 12000/20341\n",
+      "处理进度: 13000/20341\n",
+      "处理进度: 14000/20341\n",
+      "处理进度: 15000/20341\n",
+      "处理进度: 16000/20341\n",
+      "处理进度: 17000/20341\n",
+      "处理进度: 18000/20341\n",
+      "处理进度: 19000/20341\n",
+      "处理进度: 20000/20341\n",
+      "总配对样本数: 223689\n",
+      "数据收集进度: 0/223689\n",
+      "数据收集进度: 10000/223689\n",
+      "数据收集进度: 20000/223689\n",
+      "数据收集进度: 30000/223689\n",
+      "数据收集进度: 40000/223689\n",
+      "数据收集进度: 50000/223689\n",
+      "数据收集进度: 60000/223689\n",
+      "数据收集进度: 70000/223689\n",
+      "数据收集进度: 80000/223689\n",
+      "数据收集进度: 90000/223689\n",
+      "数据收集进度: 100000/223689\n",
+      "数据收集进度: 110000/223689\n",
+      "数据收集进度: 120000/223689\n",
+      "数据收集进度: 130000/223689\n",
+      "数据收集进度: 140000/223689\n",
+      "数据收集进度: 150000/223689\n",
+      "数据收集进度: 160000/223689\n",
+      "数据收集进度: 170000/223689\n",
+      "数据收集进度: 180000/223689\n",
+      "数据收集进度: 190000/223689\n",
+      "数据收集进度: 200000/223689\n",
+      "数据收集进度: 210000/223689\n",
+      "数据收集进度: 220000/223689\n",
+      "control_CP: 均值=0.0174, 标准差=0.3573, 形状=(223689, 775)\n",
+      "target_CP: 均值=-0.0105, 标准差=1.1399, 形状=(223689, 775)\n",
+      "control_GE: 均值=-0.0002, 标准差=0.1332, 形状=(223689, 977)\n",
+      "target_GE: 均值=-0.0007, 标准差=0.2795, 形状=(223689, 977)\n",
+      "✅ 稳健配对数据集已保存到: ./Paired_CP_GE_Dataset.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "def create_robust_paired_dataset(CP_data, GE_data, out_path, max_pairs_per_smiles=20):\n",
+    "    \"\"\"\n",
+    "    创建稳健的样本级配对数据集\n",
+    "    添加样本限制和数据处理\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    print(\"开始创建稳健配对数据集...\")\n",
+    "    \n",
+    "    # 构建SMILES到索引列表的映射\n",
+    "    cp_smiles_to_indices = defaultdict(list)\n",
+    "    ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "        cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "        ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    # 找到共有的SMILES\n",
+    "    common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "    print(f\"共有化合物: {len(common_smiles)}\")\n",
+    "    \n",
+    "    # 收集所有有效配对（带采样限制）\n",
+    "    paired_samples = []\n",
+    "    \n",
+    "    for i, smiles in enumerate(common_smiles):\n",
+    "        if i % 1000 == 0:\n",
+    "            print(f\"处理进度: {i}/{len(common_smiles)}\")\n",
+    "            \n",
+    "        cp_indices = cp_smiles_to_indices[smiles]\n",
+    "        ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "        # 限制每个SMILES的最大配对数，避免样本爆炸\n",
+    "        max_cp_samples = min(5, len(cp_indices))\n",
+    "        max_ge_samples = min(5, len(ge_indices))\n",
+    "        \n",
+    "        # 随机采样\n",
+    "        sampled_cp = np.random.choice(cp_indices, max_cp_samples, replace=False)\n",
+    "        sampled_ge = np.random.choice(ge_indices, max_ge_samples, replace=False)\n",
+    "        \n",
+    "        # 创建配对\n",
+    "        for cp_idx in sampled_cp:\n",
+    "            for ge_idx in sampled_ge:\n",
+    "                paired_samples.append({\n",
+    "                    'canonical_smiles': smiles,\n",
+    "                    'cp_index': cp_idx,\n",
+    "                    'ge_index': ge_idx\n",
+    "                })\n",
+    "    \n",
+    "    print(f\"总配对样本数: {len(paired_samples)}\")\n",
+    "    \n",
+    "    # 创建最终的字典形数据集\n",
+    "    dataset_dict = {\n",
+    "        'canonical_smiles': [],\n",
+    "        'control_CP': [],\n",
+    "        'target_CP': [], \n",
+    "        'control_GE': [],\n",
+    "        'target_GE': []\n",
+    "    }\n",
+    "    \n",
+    "    for i, sample in enumerate(paired_samples):\n",
+    "        if i % 10000 == 0:\n",
+    "            print(f\"数据收集进度: {i}/{len(paired_samples)}\")\n",
+    "            \n",
+    "        dataset_dict['canonical_smiles'].append(sample['canonical_smiles'])\n",
+    "        dataset_dict['control_CP'].append(CP_data['control'][sample['cp_index']])\n",
+    "        dataset_dict['target_CP'].append(CP_data['target'][sample['cp_index']])\n",
+    "        dataset_dict['control_GE'].append(GE_data['control'][sample['ge_index']]) \n",
+    "        dataset_dict['target_GE'].append(GE_data['target'][sample['ge_index']])\n",
+    "    \n",
+    "    # 转换为numpy数组并进行数据清洗\n",
+    "    final_dataset = {}\n",
+    "    final_dataset['canonical_smiles'] = np.array(dataset_dict['canonical_smiles'], dtype='S')\n",
+    "    \n",
+    "    # 数据清洗函数\n",
+    "    def clean_and_validate_data(data_list, name):\n",
+    "        data_array = np.array(data_list, dtype=np.float32)\n",
+    "        \n",
+    "        # 检查NaN和Inf\n",
+    "        nan_count = np.isnan(data_array).sum()\n",
+    "        inf_count = np.isinf(data_array).sum()\n",
+    "        \n",
+    "        if nan_count > 0 or inf_count > 0:\n",
+    "            print(f\"警告: {name} 包含 {nan_count} NaN 和 {inf_count} Inf\")\n",
+    "            data_array = np.nan_to_num(data_array, nan=0.0, posinf=1.0, neginf=-1.0)\n",
+    "        \n",
+    "        # 数据标准化\n",
+    "        mean = data_array.mean()\n",
+    "        std = data_array.std()\n",
+    "        \n",
+    "        print(f\"{name}: 均值={mean:.4f}, 标准差={std:.4f}, 形状={data_array.shape}\")\n",
+    "        \n",
+    "        return data_array\n",
+    "    \n",
+    "    final_dataset['control_CP'] = clean_and_validate_data(dataset_dict['control_CP'], 'control_CP')\n",
+    "    final_dataset['target_CP'] = clean_and_validate_data(dataset_dict['target_CP'], 'target_CP')\n",
+    "    final_dataset['control_GE'] = clean_and_validate_data(dataset_dict['control_GE'], 'control_GE')\n",
+    "    final_dataset['target_GE'] = clean_and_validate_data(dataset_dict['target_GE'], 'target_GE')\n",
+    "    \n",
+    "    # 保存为HDF5\n",
+    "    with h5py.File(out_path, 'w') as f:\n",
+    "        for key, value in final_dataset.items():\n",
+    "            f.create_dataset(key, data=value)\n",
+    "    \n",
+    "    print(f\"✅ 稳健配对数据集已保存到: {out_path}\")\n",
+    "    return final_dataset\n",
+    "\n",
+    "\n",
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "paired_dataset = create_robust_paired_dataset(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

MVC/MVC_BBBC036/Step4_outlier_process.ipynb

@@ -0,0 +1,491 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "389cd136",
+   "metadata": {},
+   "source": [
+    "# 处理异常数据\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "917fa110",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "cddb976e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "path47 = 'Aggregated_CP_GE.h5'\n",
+    "\n",
+    "# data36 = load_from_HDF(path36)\n",
+    "data36 = load_from_HDF(path47)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "92260f6e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(1916, 602)"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "data36['control_CP'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "54e51b05",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-0.15427954494953156, 32876.36328125]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-8.007139205932617, 370889.125]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['control_CP'].min()\n",
+    "data_max = data36['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data36['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['target_CP'].min()\n",
+    "data_max = data36['target_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['target_CP'], axis=0)\n",
+    "max_per_column = np.max(data36['target_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "bde80b4c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-0.059715718030929565, 0.10131344199180603]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-4.208789825439453, 6.420098304748535]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['control_GE'].min()\n",
+    "data_max = data36['control_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['control_GE'], axis=0)\n",
+    "max_per_column = np.max(data36['control_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['target_GE'].min()\n",
+    "data_max = data36['target_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['target_GE'], axis=0)\n",
+    "max_per_column = np.max(data36['target_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "053cfaa7",
+   "metadata": {},
+   "source": [
+    "## 看看 cpg0016 的数值范围"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c81440ab",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-1.2491892576217651, 1.3490736484527588]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-6.057624816894531, 6.188350677490234]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "# cpg0016_data = '/home/bob/boom/VCBench/data/Image/cpg0016_data.h5'\n",
+    "\n",
+    "# # data36 = load_from_HDF(path36)\n",
+    "# cpg0016_data = load_from_HDF(cpg0016_data)\n",
+    "\n",
+    "# import numpy as np\n",
+    "\n",
+    "# # 计算整个数组的最小值和最大值\n",
+    "# data_min = cpg0016_data['control'].min()\n",
+    "# data_max = cpg0016_data['control'].max()\n",
+    "\n",
+    "# # 计算每列（即每个特征）的最小值和最大值\n",
+    "# # axis=0 表示对每一列进行操作\n",
+    "# min_per_column = np.min(cpg0016_data['control'], axis=0)\n",
+    "# max_per_column = np.max(cpg0016_data['control'], axis=0)\n",
+    "\n",
+    "# print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "# print(\"-\" * 30)\n",
+    "\n",
+    "# # 计算整个数组的最小值和最大值\n",
+    "# data_min = cpg0016_data['target'].min()\n",
+    "# data_max = cpg0016_data['target'].max()\n",
+    "\n",
+    "# # 计算每列（即每个特征）的最小值和最大值\n",
+    "# # axis=0 表示对每一列进行操作\n",
+    "# min_per_column = np.min(cpg0016_data['target'], axis=0)\n",
+    "# max_per_column = np.max(cpg0016_data['target'], axis=0)\n",
+    "\n",
+    "# print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "# print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c7005461",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "2b2f331e",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "标准化后的整个数组的数据范围：[-19.69989776611328, 31.564712524414062]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 假设 data36['target_CP'] 是你的原始数据, target_CP, control_CP\n",
+    "data_with_outlier = data36['target_CP']\n",
+    "\n",
+    "# 1. 异常值处理 - 使用截断法 (Clipping)\n",
+    "# 设定一个更保守的上限，例如基于分位数\n",
+    "q99 = np.percentile(data_with_outlier, 99)\n",
+    "clip_max = q99 * 2 # 或者直接用一个固定值，比如 5\n",
+    "data_clipped = np.clip(data_with_outlier, a_min=None, a_max=clip_max)\n",
+    "# 2. 数据标准化 - Z-score Normalization\n",
+    "# 计算每列（特征）的均值和标准差\n",
+    "mu = data_clipped.mean(axis=0)\n",
+    "sigma = data_clipped.std(axis=0)\n",
+    "\n",
+    "# 初始化一个与数据形状相同的零数组\n",
+    "data_normalized = np.zeros_like(data_clipped, dtype=float)\n",
+    "\n",
+    "# 找到标准差不为0的列\n",
+    "# 为了避免浮点数误差，使用一个非常小的阈值来检查\n",
+    "epsilon = 1e-8\n",
+    "nonzero_sigma_cols = sigma > epsilon\n",
+    "\n",
+    "# 对这些列进行零均值化\n",
+    "data_normalized[:, nonzero_sigma_cols] = (data_clipped[:, nonzero_sigma_cols] - mu[nonzero_sigma_cols]) / sigma[nonzero_sigma_cols]\n",
+    "\n",
+    "# 检查结果\n",
+    "data_min = data_normalized.min()\n",
+    "data_max = data_normalized.max()\n",
+    "\n",
+    "print(f\"\\n标准化后的整个数组的数据范围：[{data_min}, {data_max}]\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "9ccda958",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# control_CP = data_normalized.copy()\n",
+    "# print(control_CP.shape)\n",
+    "target_CP = data_normalized.copy()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "6e515034",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-43.761436462402344, 6.348904132843018]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-19.69989776611328, 31.564712524414062]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data_min = control_CP.min()\n",
+    "data_max = control_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = target_CP.min()\n",
+    "data_max = target_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "481b57e7",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "float64 float64\n",
+      "float32 float32\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(control_CP.dtype, target_CP.dtype)\n",
+    "\n",
+    "control_CP_float32 = control_CP.astype(np.float32)\n",
+    "target_CP_float32 = target_CP.astype(np.float32)\n",
+    "print(control_CP_float32.dtype, target_CP_float32.dtype)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "6e2d6030",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "已删除旧的 'control_CP' 数据集。\n",
+      "已删除旧的 'target_CP' 数据集。\n",
+      "已成功写入新的 'control_CP' 和 'target_CP' 数据集。\n",
+      "操作完成。\n"
+     ]
+    }
+   ],
+   "source": [
+    "file_path = 'Aggregated_CP_GE.h5'\n",
+    "# 使用 'a' (append) 模式打开文件\n",
+    "with h5py.File(file_path, 'a') as f:\n",
+    "    \n",
+    "    # 1. 删除旧的数据集（如果存在）\n",
+    "    if \"control_CP\" in f:\n",
+    "        del f[\"control_CP\"]\n",
+    "        print(\"已删除旧的 'control_CP' 数据集。\")\n",
+    "        \n",
+    "    if \"target_CP\" in f:\n",
+    "        del f[\"target_CP\"]\n",
+    "        print(\"已删除旧的 'target_CP' 数据集。\")\n",
+    "    \n",
+    "    # 2. 创建并写入新的数据集\n",
+    "    f.create_dataset(\"control_CP\", data=control_CP_float32)\n",
+    "    f.create_dataset(\"target_CP\", data=target_CP_float32)\n",
+    "    \n",
+    "    print(\"已成功写入新的 'control_CP' 和 'target_CP' 数据集。\")\n",
+    "\n",
+    "print(\"操作完成。\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b131c13e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5a82372f",
+   "metadata": {},
+   "source": [
+    "# 看基因数据是否异常 -- 没有"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "a81ddc86",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-43.761436462402344, 6.348904132843018]\n",
+      "------------------------------\n",
+      "每列数据的最小值（前10个）：\n",
+      "[-1.6301312 -1.4819461 -1.9675876 -2.272238  -3.393242  -1.1765975\n",
+      " -2.6176555 -1.2396002 -2.462905  -2.8029377]\n",
+      "------------------------------\n",
+      "每列数据的最大值（前10个）：\n",
+      "[3.0697236 1.985853  1.09526   1.1315886 2.1168044 2.9211075 2.1078165\n",
+      " 2.3139157 2.1509798 1.5131917]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data36 = load_from_HDF( 'Aggregated_CP_GE.h5')\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data36['control_CP'].min()\n",
+    "data_max = data36['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data36['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data36['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最小值（前10个）：\")\n",
+    "print(min_per_column[:10])\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最大值（前10个）：\")\n",
+    "print(max_per_column[:10])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "770f3727",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 看看方差\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "10693804",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c072eea5",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

MVC/MVC_BBBC047/Step3_aggreate_2modality.ipynb

@@ -0,0 +1,1155 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 数据聚合：「输入药物 SMILES，输出基因表达 + 细胞形态」的多模态虚拟细胞模型，配对不齐，训练集不均衡。\n",
+    "\n",
+    "## 策略有3种：\n",
+    "\n",
+    "### 1 是按 SMILES 聚合，对每个 SMILES 在每个模态上 取平均或中位数（或更复杂的 embedding 聚合）。\n",
+    "\n",
+    "### 2 是多实例学习（Multiple Instance Learning, MIL）：思路：把一个 SMILES 的所有 replicate 当作一个 bag，模型学习 bag-level 对齐。\n",
+    "\n",
+    "### 3 是不对齐 replicate，随机采样：思路：训练时，每个 batch 从两个模态中随机采样该 SMILES 的一个 replicate，逐步逼近两个分布。"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "CP_path =  './Processed_Paired_CP_KeepRep.h5'\n",
+    "GE_path = './Processed_Paired_GE_KeepRep.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['CCCOc1cc(N)ccc1C(=O)OCCN(CC)CC',\n",
+       "        'C=C(C)[C@@H]1CCC(C)=C[C@H]1c1c(OC)cc(C)cc1OC',\n",
+       "        'COc1cc(O)cc(/C=C/c2ccccc2)c1', ...,\n",
+       "        'C[C@@H]1CN([C@@H](C)CO)C(=O)CCCn2cc(nn2)CO[C@H]1CN(C)C(=O)c1ccncc1',\n",
+       "        'C[C@H](CO)N1C[C@H](C)[C@@H](CN(C)C(=O)c2ccncc2)OCc2cn(nn2)CCCC1=O',\n",
+       "        'C[C@@H]1CN([C@@H](C)CO)C(=O)CCCn2nncc2CO[C@H]1CN(C)Cc1ccc2c(c1)OCCN2C'],\n",
+       "       dtype='<U227'),\n",
+       " 'control': array([[-0.15425342,  0.20473444,  0.04847282, ...,  0.22247523,\n",
+       "          0.41372228,  0.09528223],\n",
+       "        [-0.15425342,  0.20473444,  0.04847282, ...,  0.22247523,\n",
+       "          0.41372228,  0.09528223],\n",
+       "        [-0.15425342,  0.20473444,  0.04847282, ...,  0.22247523,\n",
+       "          0.41372228,  0.09528223],\n",
+       "        ...,\n",
+       "        [ 0.08794986,  0.04630309,  0.0316207 , ..., -0.18811361,\n",
+       "         -0.08507317, -0.19418037],\n",
+       "        [ 0.08794986,  0.04630309,  0.0316207 , ..., -0.18811361,\n",
+       "         -0.08507317, -0.19418037],\n",
+       "        [ 0.08794986,  0.04630309,  0.0316207 , ..., -0.18811361,\n",
+       "         -0.08507317, -0.19418037]], dtype=float32),\n",
+       " 'dose': array([ 6.05, 10.  , 10.  , ..., 10.41, 10.14, 10.22], dtype=float32),\n",
+       " 'target': array([[ 1.7987275 , -1.0678272 , -1.3908694 , ..., -2.2844946 ,\n",
+       "         -1.4380481 , -0.4035441 ],\n",
+       "        [ 0.3030874 ,  0.57465357, -0.18462732, ..., -1.4371198 ,\n",
+       "         -0.8579998 ,  0.20614216],\n",
+       "        [-0.7811053 ,  0.5069796 ,  0.9922175 , ...,  0.3974406 ,\n",
+       "         -0.13609461,  0.42841423],\n",
+       "        ...,\n",
+       "        [ 1.466404  , -1.2910391 , -0.19340979, ..., -2.0167367 ,\n",
+       "         -1.2818346 , -1.4399147 ],\n",
+       "        [ 1.0264087 , -0.42999583, -1.681699  , ..., -1.4709337 ,\n",
+       "         -0.9747103 , -1.4030336 ],\n",
+       "        [ 1.4138446 , -0.93786097, -1.3868744 , ..., -1.2337868 ,\n",
+       "         -0.58811694, -0.94757   ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['CC(C)[C@H](CO)Nc1nc(Nc2cc(N)cc(Cl)c2)c2ncn(C(C)C)c2n1',\n",
+       "        'Ic1cccc(CSc2nnc(-c3ccncc3)o2)c1',\n",
+       "        'O=C(Nc1ccncc1)Nc1c(Cl)cc(Cl)cc1Cl', ...,\n",
+       "        'O=C(O)c1ccc(-c2ccc(/C=c3\\\\sc4nc5ccccc5n4c3=O)o2)cc1',\n",
+       "        'O=S(=O)(c1cccc2cncc(C(F)F)c12)N1CCCNCC1',\n",
+       "        'COc1ccc(S(=O)(=O)N(C)C[C@@H]2Oc3c(NC(=O)Nc4cccc5ccccc45)cccc3C(=O)N([C@@H](C)CO)C[C@H]2C)cc1'],\n",
+       "       dtype='<U227'),\n",
+       " 'control': array([[ 0.305876  , -0.59514517, -0.21795662, ...,  0.20644613,\n",
+       "         -0.07786681,  0.14617148],\n",
+       "        [ 0.305876  , -0.59514517, -0.21795662, ...,  0.20644613,\n",
+       "         -0.07786681,  0.14617148],\n",
+       "        [ 0.305876  , -0.59514517, -0.21795662, ...,  0.20644613,\n",
+       "         -0.07786681,  0.14617148],\n",
+       "        ...,\n",
+       "        [ 0.        ,  0.        ,  0.        , ...,  0.        ,\n",
+       "          0.        ,  0.        ],\n",
+       "        [ 0.        ,  0.        ,  0.        , ...,  0.        ,\n",
+       "          0.        ,  0.        ],\n",
+       "        [ 0.        ,  0.        ,  0.        , ...,  0.        ,\n",
+       "          0.        ,  0.        ]], dtype=float32),\n",
+       " 'dose': array([50. , 12.7, 50. , ..., 20. , 40. , 20. ], dtype=float32),\n",
+       " 'target': array([[-0.7184085 , -1.4126618 , -1.7777987 , ...,  1.1767539 ,\n",
+       "          2.0003324 ,  2.4972866 ],\n",
+       "        [ 0.25343022, -1.2717106 , -0.0366732 , ...,  1.8126812 ,\n",
+       "         -0.6341776 , -1.2102923 ],\n",
+       "        [-0.43184918,  1.0067521 , -0.00558909, ..., -0.26218557,\n",
+       "         -0.8643179 ,  1.4752281 ],\n",
+       "        ...,\n",
+       "        [-0.67632276,  0.7587029 ,  0.3990235 , ...,  2.8850307 ,\n",
+       "         -0.92035365, -1.8870968 ],\n",
+       "        [-2.1307466 ,  2.1573339 , -1.6117429 , ...,  0.566304  ,\n",
+       "         -1.1584423 , -2.332463  ],\n",
+       "        [ 2.0216608 ,  1.4105401 , -1.0605472 , ..., -0.1974064 ,\n",
+       "         -0.30219036, -1.2475965 ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 16,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "GE_CSV"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((86844, 741), (57835, 977))"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "CP_CSV['target'].shape, GE_CSV['target'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 按照第一种方式直接 mean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ 聚合后的数据已保存到 ./Aggregated_CP_GE.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "def aggregate_modalities(CP, GE, out_path):\n",
+    "    # 转成 DataFrame 方便聚合\n",
+    "    CP_df = pd.DataFrame({\n",
+    "        \"SMILES\": CP[\"canonical_smiles\"],\n",
+    "        \"control\": list(CP[\"control\"]),\n",
+    "        \"target\": list(CP[\"target\"])\n",
+    "    })\n",
+    "    GE_df = pd.DataFrame({\n",
+    "        \"SMILES\": GE[\"canonical_smiles\"],\n",
+    "        \"control\": list(GE[\"control\"]),\n",
+    "        \"target\": list(GE[\"target\"])\n",
+    "    })\n",
+    "\n",
+    "    # groupby + mean pooling\n",
+    "    CP_grouped = CP_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "    GE_grouped = GE_df.groupby(\"SMILES\").agg(lambda x: np.nanmean(np.stack(x), axis=0)).reset_index()\n",
+    "\n",
+    "    # merge 保证对齐\n",
+    "    merged = pd.merge(CP_grouped, GE_grouped, on=\"SMILES\", how=\"inner\", suffixes=(\"_CP\", \"_GE\"))\n",
+    "\n",
+    "    # 转 numpy array 并替换剩余 NaN\n",
+    "    def clean_array(arr_list):\n",
+    "        arr = np.stack(arr_list.to_numpy())\n",
+    "        arr = np.nan_to_num(arr, nan=0.0, posinf=0.0, neginf=0.0)  # 替换 NaN/Inf\n",
+    "        return arr.astype(np.float32)\n",
+    "\n",
+    "    smiles_array     = merged[\"SMILES\"].to_numpy(dtype=\"S\")\n",
+    "    control_CP_array = clean_array(merged[\"control_CP\"])\n",
+    "    target_CP_array  = clean_array(merged[\"target_CP\"])\n",
+    "    control_GE_array = clean_array(merged[\"control_GE\"])\n",
+    "    target_GE_array  = clean_array(merged[\"target_GE\"])\n",
+    "\n",
+    "    # 存 HDF5\n",
+    "    with h5py.File(out_path, \"w\") as f:\n",
+    "        f.create_dataset(\"canonical_smiles\", data=smiles_array)\n",
+    "        f.create_dataset(\"control_CP\", data=control_CP_array)\n",
+    "        f.create_dataset(\"target_CP\", data=target_CP_array)\n",
+    "        f.create_dataset(\"control_GE\", data=control_GE_array)\n",
+    "        f.create_dataset(\"target_GE\", data=target_GE_array)\n",
+    "\n",
+    "    print(f\"✅ 聚合后的数据已保存到 {out_path}\")\n",
+    "\n",
+    "# 用法\n",
+    "# root = '/home/bob/boom/VCBench/data/MVC/CDRP-BBBC047-Bray/'\n",
+    "CP_path = './Processed_Paired_CP.h5'\n",
+    "GE_path = './Processed_Paired_GE.h5'\n",
+    "\n",
+    "CP_CSV = load_from_HDF(CP_path)\n",
+    "GE_CSV = load_from_HDF(GE_path)\n",
+    "\n",
+    "out_path = './Aggregated_CP_GE.h5'\n",
+    "aggregate_modalities(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['Brc1c(CSc2nc3ccccc3s2)nc2ncccn12', 'Brc1c(NC2=NCCN2)ccc2nccnc12',\n",
+       "        'Brc1ccc(CSc2nnc(-c3ccccn3)n2Cc2ccco2)cc1', ...,\n",
+       "        'c1nc(CC2CCNCC2)c[nH]1', 'c1nc(CCCC2CCNCC2)c[nH]1',\n",
+       "        'c1ncn(CCOCCOc2ccc3c(c2)CCC3)n1'], dtype='<U227'),\n",
+       " 'control_CP': array([[-0.15487705,  0.11688805, -0.06308393, ...,  0.24665973,\n",
+       "          0.18521263,  0.24594025],\n",
+       "        [-0.1634329 ,  0.15811297, -0.03295023, ...,  0.2232287 ,\n",
+       "          0.18645084,  0.20041683],\n",
+       "        [-0.13265091,  0.11912386, -0.04087081, ...,  0.23031679,\n",
+       "          0.15957166,  0.2795726 ],\n",
+       "        ...,\n",
+       "        [-0.16390067,  0.1292826 , -0.02731108, ...,  0.2356141 ,\n",
+       "          0.20089747,  0.22413647],\n",
+       "        [-0.1634329 ,  0.15811297, -0.03295023, ...,  0.2232287 ,\n",
+       "          0.18645084,  0.20041683],\n",
+       "        [-0.1903376 ,  0.16364928, -0.02507655, ...,  0.2462861 ,\n",
+       "          0.21280092,  0.2474828 ]], dtype=float32),\n",
+       " 'control_GE': array([[-0.11589041,  0.09791137, -0.10001642, ..., -0.04151818,\n",
+       "          0.00727702, -0.21572095],\n",
+       "        [ 0.01338525, -0.02699645,  0.0147262 , ..., -0.01713243,\n",
+       "         -0.02412532,  0.08308113],\n",
+       "        [ 0.02007259,  0.03755556, -0.02430556, ..., -0.18552697,\n",
+       "         -0.0794616 , -0.020855  ],\n",
+       "        ...,\n",
+       "        [ 0.02062   , -0.00183125, -0.13683268, ..., -0.17978694,\n",
+       "          0.21799538,  0.17114288],\n",
+       "        [ 0.01338525, -0.02699645,  0.0147262 , ..., -0.01713243,\n",
+       "         -0.02412532,  0.08308113],\n",
+       "        [-0.06893609, -0.02563992,  0.06738283, ...,  0.07980002,\n",
+       "         -0.11235818, -0.03411045]], dtype=float32),\n",
+       " 'target_CP': array([[ 0.38674092, -0.62070274, -0.62418985, ...,  0.44180262,\n",
+       "          0.42434335, -1.2392346 ],\n",
+       "        [ 0.31655684, -0.23943731,  0.18836854, ..., -0.05182697,\n",
+       "          0.2953927 ,  1.5328658 ],\n",
+       "        [-1.4637322 ,  1.873155  ,  2.345407  , ...,  2.3001804 ,\n",
+       "          2.5974057 , -0.5258522 ],\n",
+       "        ...,\n",
+       "        [-0.22405519,  0.21074149, -0.0310891 , ..., -0.20554903,\n",
+       "         -0.28407168, -0.17878559],\n",
+       "        [ 0.36180133, -0.36012647,  0.18524444, ..., -0.5896245 ,\n",
+       "         -0.28982347, -0.09377167],\n",
+       "        [-0.11151703,  0.40477425,  0.04721215, ...,  1.3899186 ,\n",
+       "          0.70356005,  0.23341325]], dtype=float32),\n",
+       " 'target_GE': array([[ 0.09065   ,  0.0786    , -0.06155001, ..., -0.1333    ,\n",
+       "          0.097875  ,  0.15165001],\n",
+       "        [-0.07302775,  0.0460825 , -0.02222175, ...,  0.17655626,\n",
+       "         -0.154201  ,  0.033125  ],\n",
+       "        [-0.03446667, -0.01266666,  0.04358333, ...,  0.27253333,\n",
+       "         -0.06081   ,  0.28026667],\n",
+       "        ...,\n",
+       "        [ 0.22472501,  0.20904249, -0.2872    , ..., -0.22035149,\n",
+       "          0.5861085 ,  0.278331  ],\n",
+       "        [ 0.01292225,  0.07139751, -0.13613375, ..., -0.16344374,\n",
+       "          0.053449  , -0.200645  ],\n",
+       "        [ 0.14866668, -0.054     ,  0.04613333, ...,  0.22286667,\n",
+       "         -0.09675   , -0.01868333]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out_path = './Aggregated_CP_GE.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((20341, 775), (20341, 775), (20341, 977), (20341, 977))"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape, agg_data['target_CP'].shape, agg_data['control_GE'].shape, agg_data['target_GE'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 21,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "'O=C(C[C@@H]1CC[C@@H]2[C@H](COC[C@H](O)CN2Cc2cc(F)cc(F)c2)O1)Nc1nccs1' in agg_data['canonical_smiles']"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 复杂点，样本级别的配对"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "加载源数据...\n",
+      "开始创建多模态配对数据集...\n",
+      "Cell Painting 唯一化合物数: 20341\n",
+      "Gene Expression 唯一化合物数: 20341\n",
+      "多模态共有化合物数 (Intersection): 20341\n",
+      "生成配对样本总数: 223805\n",
+      "正在组装最终矩阵...\n",
+      "✅ 最终数据集已保存至: ./Paired_CP_GE_Dataset.h5\n",
+      "------------------------------\n",
+      "Dataset Shapes:\n",
+      "canonical_smiles: (223805,)\n",
+      "control_CP: (223805, 741)\n",
+      "target_CP: (223805, 741)\n",
+      "control_GE: (223805, 977)\n",
+      "target_GE: (223805, 977)\n",
+      "CP_dose: (223805,)\n",
+      "GE_dose: (223805,)\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "# ==========================================\n",
+    "# 0. 辅助函数：读取 HDF5\n",
+    "# ==========================================\n",
+    "def load_from_HDF(fname):\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            # 处理字符串解码\n",
+    "            if data[key].dtype.kind == 'S':\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data\n",
+    "\n",
+    "# ==========================================\n",
+    "# 1. 核心逻辑：创建稳健配对数据集\n",
+    "# ==========================================\n",
+    "def create_robust_paired_dataset(CP_data, GE_data, out_path, max_samples_per_smiles=5, seed=42):\n",
+    "    \"\"\"\n",
+    "    CP_data, GE_data: 包含 'canonical_smiles', 'control', 'target', 'dose' 的字典\n",
+    "    max_samples_per_smiles: 每个模态最多采样的重复数\n",
+    "    seed: 随机种子，保证配对结果可复现\n",
+    "    \"\"\"\n",
+    "    print(\"开始创建多模态配对数据集...\")\n",
+    "    \n",
+    "    # [新增] 检查输入数据是否包含 dose\n",
+    "    if 'dose' not in CP_data or 'dose' not in GE_data:\n",
+    "        raise ValueError(\"输入数据缺少 'dose' 字段，请检查上一步是否正确保存了剂量信息。\")\n",
+    "\n",
+    "    # [新增] 设置随机种子\n",
+    "    np.random.seed(seed)\n",
+    "    \n",
+    "    # --- A. 构建索引映射 ---\n",
+    "    cp_smiles_to_indices = defaultdict(list)\n",
+    "    ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "        cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "        ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    # --- B. 找交集 ---\n",
+    "    common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "    print(f\"Cell Painting 唯一化合物数: {len(cp_smiles_to_indices)}\")\n",
+    "    print(f\"Gene Expression 唯一化合物数: {len(ge_smiles_to_indices)}\")\n",
+    "    print(f\"多模态共有化合物数 (Intersection): {len(common_smiles)}\")\n",
+    "    \n",
+    "    # --- C. 生成配对索引 ---\n",
+    "    paired_indices = []\n",
+    "    \n",
+    "    for i, smiles in enumerate(common_smiles):\n",
+    "        # 获取该 SMILES 的所有索引\n",
+    "        cp_indices = cp_smiles_to_indices[smiles]\n",
+    "        ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "        # 限制采样数量\n",
+    "        n_cp = min(max_samples_per_smiles, len(cp_indices))\n",
+    "        n_ge = min(max_samples_per_smiles, len(ge_indices))\n",
+    "        \n",
+    "        # 无放回随机采样\n",
+    "        sampled_cp_indices = np.random.choice(cp_indices, n_cp, replace=False)\n",
+    "        sampled_ge_indices = np.random.choice(ge_indices, n_ge, replace=False)\n",
+    "        \n",
+    "        # 笛卡尔积配对\n",
+    "        for cp_idx in sampled_cp_indices:\n",
+    "            for ge_idx in sampled_ge_indices:\n",
+    "                paired_indices.append({\n",
+    "                    'smiles': smiles,\n",
+    "                    'cp_idx': cp_idx,\n",
+    "                    'ge_idx': ge_idx\n",
+    "                })\n",
+    "                \n",
+    "    print(f\"生成配对样本总数: {len(paired_indices)}\")\n",
+    "\n",
+    "    # --- D. 组装数据 ---\n",
+    "    n_samples = len(paired_indices)\n",
+    "    \n",
+    "    # 获取特征维度\n",
+    "    cp_feat_dim = CP_data['target'].shape[1]\n",
+    "    ge_feat_dim = GE_data['target'].shape[1]\n",
+    "    \n",
+    "    # 预分配内存\n",
+    "    final_data = {\n",
+    "        'canonical_smiles': np.empty(n_samples, dtype='S200'),\n",
+    "        'control_CP': np.empty((n_samples, cp_feat_dim), dtype=np.float32),\n",
+    "        'target_CP':  np.empty((n_samples, cp_feat_dim), dtype=np.float32),\n",
+    "        'control_GE': np.empty((n_samples, ge_feat_dim), dtype=np.float32),\n",
+    "        'target_GE':  np.empty((n_samples, ge_feat_dim), dtype=np.float32),\n",
+    "        # [新增] 两个模态的剂量\n",
+    "        'CP_dose':    np.empty(n_samples, dtype=np.float32), \n",
+    "        'GE_dose':    np.empty(n_samples, dtype=np.float32)\n",
+    "    }\n",
+    "    \n",
+    "    print(\"正在组装最终矩阵...\")\n",
+    "    for i, item in enumerate(paired_indices):\n",
+    "        cp_idx = item['cp_idx']\n",
+    "        ge_idx = item['ge_idx']\n",
+    "        \n",
+    "        final_data['canonical_smiles'][i] = item['smiles'].encode('utf-8')\n",
+    "        \n",
+    "        # 填入 CP 数据\n",
+    "        final_data['control_CP'][i]       = CP_data['control'][cp_idx]\n",
+    "        final_data['target_CP'][i]        = CP_data['target'][cp_idx]\n",
+    "        final_data['CP_dose'][i]          = CP_data['dose'][cp_idx] # [新增]\n",
+    "        \n",
+    "        # 填入 GE 数据\n",
+    "        final_data['control_GE'][i]       = GE_data['control'][ge_idx]\n",
+    "        final_data['target_GE'][i]        = GE_data['target'][ge_idx]\n",
+    "        final_data['GE_dose'][i]          = GE_data['dose'][ge_idx] # [新增]\n",
+    "        \n",
+    "    # --- E. 数据清洗与保存 ---\n",
+    "    # 检查并处理 NaN/Inf (Dose 通常不需要处理，但如果有 NaN 也顺便处理一下安全)\n",
+    "    check_cols = ['control_CP', 'target_CP', 'control_GE', 'target_GE', 'CP_dose', 'GE_dose']\n",
+    "    \n",
+    "    for key in check_cols:\n",
+    "        if np.isnan(final_data[key]).any() or np.isinf(final_data[key]).any():\n",
+    "            print(f\"警告: {key} 包含 NaN 或 Inf，正在填充 0...\")\n",
+    "            final_data[key] = np.nan_to_num(final_data[key], nan=0.0, posinf=0.0, neginf=0.0)\n",
+    "\n",
+    "    # 保存\n",
+    "    with h5py.File(out_path, 'w') as f:\n",
+    "        for key, val in final_data.items():\n",
+    "            f.create_dataset(key, data=val)\n",
+    "            \n",
+    "    print(f\"✅ 最终数据集已保存至: {out_path}\")\n",
+    "    print(\"-\" * 30)\n",
+    "    print(f\"Dataset Shapes:\")\n",
+    "    for k, v in final_data.items():\n",
+    "        print(f\"{k}: {v.shape}\")\n",
+    "\n",
+    "# ==========================================\n",
+    "# 2. 执行流程\n",
+    "# ==========================================\n",
+    "# 定义输入输出路径\n",
+    "cp_input = './Processed_Paired_CP_KeepRep.h5'\n",
+    "ge_input = './Processed_Paired_GE_KeepRep.h5'\n",
+    "final_output = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "# 加载数据\n",
+    "print(\"加载源数据...\")\n",
+    "cp_data = load_from_HDF(cp_input)\n",
+    "ge_data = load_from_HDF(ge_input)\n",
+    "\n",
+    "# 执行配对\n",
+    "create_robust_paired_dataset(cp_data, ge_data, final_output, max_samples_per_smiles=5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'CP_dose': array([10.65, 10.65, 10.65, ..., 10.  , 10.  , 10.  ], dtype=float32),\n",
+       " 'GE_dose': array([10.6501, 10.6501, 10.6501, ..., 10.    , 10.    , 10.    ],\n",
+       "       dtype=float32),\n",
+       " 'canonical_smiles': array(['C[C@H](CO)N1C[C@H](C)[C@H](CN(C)Cc2ccncc2)Oc2ccc(NS(C)(=O)=O)cc2CC1=O',\n",
+       "        'C[C@H](CO)N1C[C@H](C)[C@H](CN(C)Cc2ccncc2)Oc2ccc(NS(C)(=O)=O)cc2CC1=O',\n",
+       "        'C[C@H](CO)N1C[C@H](C)[C@H](CN(C)Cc2ccncc2)Oc2ccc(NS(C)(=O)=O)cc2CC1=O',\n",
+       "        ..., 'CN(C)C(=O)n1nnnc1Cc1ccc(-c2ccccc2)cc1',\n",
+       "        'CN(C)C(=O)n1nnnc1Cc1ccc(-c2ccccc2)cc1',\n",
+       "        'CN(C)C(=O)n1nnnc1Cc1ccc(-c2ccccc2)cc1'], dtype='<U200'),\n",
+       " 'control_CP': array([[-0.13953187,  0.06501261, -0.00263713, ...,  0.08332227,\n",
+       "          0.12313318, -0.02047802],\n",
+       "        [-0.13953187,  0.06501261, -0.00263713, ...,  0.08332227,\n",
+       "          0.12313318, -0.02047802],\n",
+       "        [-0.13953187,  0.06501261, -0.00263713, ...,  0.08332227,\n",
+       "          0.12313318, -0.02047802],\n",
+       "        ...,\n",
+       "        [ 0.07532343,  0.01590352,  0.11149302, ...,  0.09762042,\n",
+       "         -0.0291416 ,  0.18638068],\n",
+       "        [ 0.06642057, -0.06762539,  0.07449123, ..., -0.2643804 ,\n",
+       "         -0.08335885, -0.47207668],\n",
+       "        [ 0.06642057, -0.06762539,  0.07449123, ..., -0.2643804 ,\n",
+       "         -0.08335885, -0.47207668]], dtype=float32),\n",
+       " 'control_GE': array([[ 0.24454208,  0.60980904,  0.60357696, ...,  0.18017063,\n",
+       "         -0.88657326,  0.01677261],\n",
+       "        [-0.56268126,  0.15758337, -0.13244417, ..., -0.6394515 ,\n",
+       "          1.1245222 ,  1.1233364 ],\n",
+       "        [ 0.44537392,  0.6973298 , -1.5441636 , ..., -0.425022  ,\n",
+       "         -0.23347344,  0.57313305],\n",
+       "        ...,\n",
+       "        [ 0.65565985,  0.11522385,  0.9597033 , ...,  0.43092778,\n",
+       "         -0.15957022,  0.03049997],\n",
+       "        [ 0.7688196 ,  0.9452282 , -0.80648696, ..., -0.3607209 ,\n",
+       "          0.51043606, -1.139512  ],\n",
+       "        [ 0.65565985,  0.11522385,  0.9597033 , ...,  0.43092778,\n",
+       "         -0.15957022,  0.03049997]], dtype=float32),\n",
+       " 'target_CP': array([[-0.7322785 , -0.1324034 , -0.92490584, ...,  0.0871982 ,\n",
+       "          0.19238813, -0.2730067 ],\n",
+       "        [-0.7322785 , -0.1324034 , -0.92490584, ...,  0.0871982 ,\n",
+       "          0.19238813, -0.2730067 ],\n",
+       "        [-0.7322785 , -0.1324034 , -0.92490584, ...,  0.0871982 ,\n",
+       "          0.19238813, -0.2730067 ],\n",
+       "        ...,\n",
+       "        [-6.900763  , 10.        , 10.        , ...,  8.523304  ,\n",
+       "          8.528054  ,  9.481978  ],\n",
+       "        [-6.2097163 ,  9.116677  , 10.        , ...,  5.122239  ,\n",
+       "          5.808592  ,  1.9759494 ],\n",
+       "        [-6.2097163 ,  9.116677  , 10.        , ...,  5.122239  ,\n",
+       "          5.808592  ,  1.9759494 ]], dtype=float32),\n",
+       " 'target_GE': array([[-0.13318162, -2.3996215 , -0.60478145, ..., -1.7717572 ,\n",
+       "         -0.7603032 ,  1.6145478 ],\n",
+       "        [ 0.71842533, -2.600954  ,  0.4446499 , ..., -0.6659924 ,\n",
+       "         -1.0708426 , -0.40832376],\n",
+       "        [-1.034582  ,  1.0713423 , -4.108202  , ...,  0.01961743,\n",
+       "          1.2580668 ,  0.00507223],\n",
+       "        ...,\n",
+       "        [ 0.11101682,  1.4468411 ,  0.9668232 , ...,  0.43233392,\n",
+       "         -0.3482708 , -1.8351849 ],\n",
+       "        [-0.41256198, -0.05275529, -3.0842478 , ..., -0.2453646 ,\n",
+       "          1.7232248 ,  3.0262802 ],\n",
+       "        [ 0.11101682,  1.4468411 ,  0.9668232 , ...,  0.43233392,\n",
+       "         -0.3482708 , -1.8351849 ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 18,
+     "metadata": {},
+     "output_type": "execute_result"
+    },
+    {
+     "ename": "",
+     "evalue": "",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m在当前单元格或上一个单元格中执行代码时 Kernel 崩溃。\n",
+      "\u001b[1;31m请查看单元格中的代码，以确定故障的可能原因。\n",
+      "\u001b[1;31m单击<a href='https://aka.ms/vscodeJupyterKernelCrash'>此处</a>了解详细信息。\n",
+      "\u001b[1;31m有关更多详细信息，请查看 Jupyter <a href='command:jupyter.viewOutput'>log</a>。"
+     ]
+    }
+   ],
+   "source": [
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "共有化合物: 20341\n",
+      "总配对样本数: 242192\n",
+      "数据集形状:\n",
+      "  control_CP: (242192, 775)\n",
+      "  target_CP: (242192, 775)\n",
+      "  control_GE: (242192, 977)\n",
+      "  target_GE: (242192, 977)\n",
+      "✅ 配对数据集已保存到: ./Paired_CP_GE_Dataset.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "# import h5py\n",
+    "# import numpy as np\n",
+    "# import pandas as pd\n",
+    "# from collections import defaultdict\n",
+    "\n",
+    "# def create_paired_dataset(CP_data, GE_data, out_path):\n",
+    "#     \"\"\"\n",
+    "#     创建样本级配对的多模态数据集\n",
+    "#     保持原始样本的完整性，只配对真正对应的样本\n",
+    "#     \"\"\"\n",
+    "    \n",
+    "#     # 构建SMILES到索引列表的映射\n",
+    "#     cp_smiles_to_indices = defaultdict(list)\n",
+    "#     ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "#     for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "#         cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "#     for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "#         ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "#     # 找到共有的SMILES\n",
+    "#     common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "#     print(f\"共有化合物: {len(common_smiles)}\")\n",
+    "    \n",
+    "#     # 收集所有有效配对\n",
+    "#     paired_samples = []\n",
+    "    \n",
+    "#     for smiles in common_smiles:\n",
+    "#         cp_indices = cp_smiles_to_indices[smiles]\n",
+    "#         ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "#         # 为每个SMILES创建所有可能的CP-GE配对\n",
+    "#         for cp_idx in cp_indices:\n",
+    "#             for ge_idx in ge_indices:\n",
+    "#                 paired_samples.append({\n",
+    "#                     'canonical_smiles': smiles,\n",
+    "#                     'cp_index': cp_idx,\n",
+    "#                     'ge_index': ge_idx\n",
+    "#                 })\n",
+    "    \n",
+    "#     print(f\"总配对样本数: {len(paired_samples)}\")\n",
+    "    \n",
+    "#     # 创建最终的字典形数据集\n",
+    "#     dataset_dict = {\n",
+    "#         'canonical_smiles': [],\n",
+    "#         'control_CP': [],\n",
+    "#         'target_CP': [], \n",
+    "#         'control_GE': [],\n",
+    "#         'target_GE': []\n",
+    "#     }\n",
+    "    \n",
+    "#     for sample in paired_samples:\n",
+    "#         dataset_dict['canonical_smiles'].append(sample['canonical_smiles'])\n",
+    "#         dataset_dict['control_CP'].append(CP_data['control'][sample['cp_index']])\n",
+    "#         dataset_dict['target_CP'].append(CP_data['target'][sample['cp_index']])\n",
+    "#         dataset_dict['control_GE'].append(GE_data['control'][sample['ge_index']]) \n",
+    "#         dataset_dict['target_GE'].append(GE_data['target'][sample['ge_index']])\n",
+    "    \n",
+    "#     # 转换为numpy数组\n",
+    "#     final_dataset = {}\n",
+    "#     final_dataset['canonical_smiles'] = np.array(dataset_dict['canonical_smiles'], dtype='S')\n",
+    "#     final_dataset['control_CP'] = np.array(dataset_dict['control_CP'], dtype=np.float32)\n",
+    "#     final_dataset['target_CP'] = np.array(dataset_dict['target_CP'], dtype=np.float32)\n",
+    "#     final_dataset['control_GE'] = np.array(dataset_dict['control_GE'], dtype=np.float32)\n",
+    "#     final_dataset['target_GE'] = np.array(dataset_dict['target_GE'], dtype=np.float32)\n",
+    "    \n",
+    "#     print(f\"数据集形状:\")\n",
+    "#     print(f\"  control_CP: {final_dataset['control_CP'].shape}\")\n",
+    "#     print(f\"  target_CP: {final_dataset['target_CP'].shape}\")\n",
+    "#     print(f\"  control_GE: {final_dataset['control_GE'].shape}\")\n",
+    "#     print(f\"  target_GE: {final_dataset['target_GE'].shape}\")\n",
+    "    \n",
+    "#     # 保存为HDF5\n",
+    "#     with h5py.File(out_path, 'w') as f:\n",
+    "#         for key, value in final_dataset.items():\n",
+    "#             f.create_dataset(key, data=value)\n",
+    "    \n",
+    "#     print(f\"✅ 配对数据集已保存到: {out_path}\")\n",
+    "#     return final_dataset\n",
+    "\n",
+    "# # 使用方案\n",
+    "# out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "# paired_dataset = create_paired_dataset(CP_CSV, GE_CSV, out_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'canonical_smiles': array(['C[C@@H]1CCCCO[C@@H](CN(C)C(=O)c2ccccc2)[C@@H](C)CN([C@H](C)CO)C(=O)c2cc(N(C)C)ccc2O1',\n",
+       "        'C[C@@H]1CCCCO[C@@H](CN(C)C(=O)c2ccccc2)[C@@H](C)CN([C@H](C)CO)C(=O)c2cc(N(C)C)ccc2O1',\n",
+       "        'C[C@@H]1CCCCO[C@@H](CN(C)C(=O)c2ccccc2)[C@@H](C)CN([C@H](C)CO)C(=O)c2cc(N(C)C)ccc2O1',\n",
+       "        ...,\n",
+       "        'C/C=C/c1ccc2n(c1=O)C[C@H]1[C@H](CO)[C@@H](C(=O)N3CCCCC3)[C@@H]2N1C(=O)C1CCC1.C/C=C\\\\c1ccc2n(c1=O)C[C@H]1[C@H](CO)[C@@H](C(=O)N3CCCCC3)[C@@H]2N1C(=O)C1CCC1',\n",
+       "        'C/C=C/c1ccc2n(c1=O)C[C@H]1[C@H](CO)[C@@H](C(=O)N3CCCCC3)[C@@H]2N1C(=O)C1CCC1.C/C=C\\\\c1ccc2n(c1=O)C[C@H]1[C@H](CO)[C@@H](C(=O)N3CCCCC3)[C@@H]2N1C(=O)C1CCC1',\n",
+       "        'C/C=C/c1ccc2n(c1=O)C[C@H]1[C@H](CO)[C@@H](C(=O)N3CCCCC3)[C@@H]2N1C(=O)C1CCC1.C/C=C\\\\c1ccc2n(c1=O)C[C@H]1[C@H](CO)[C@@H](C(=O)N3CCCCC3)[C@@H]2N1C(=O)C1CCC1'],\n",
+       "       dtype='<U227'),\n",
+       " 'control_CP': array([[-0.08486467,  0.1714288 , -0.05182531, ...,  0.12218629,\n",
+       "          0.21159805,  0.24014652],\n",
+       "        [-0.08486467,  0.1714288 , -0.05182531, ...,  0.12218629,\n",
+       "          0.21159805,  0.24014652],\n",
+       "        [-0.08486467,  0.1714288 , -0.05182531, ...,  0.12218629,\n",
+       "          0.21159805,  0.24014652],\n",
+       "        ...,\n",
+       "        [-0.2378668 ,  0.23262355, -0.0475704 , ...,  0.21625724,\n",
+       "          0.23215774,  0.23364826],\n",
+       "        [-0.2378668 ,  0.23262355, -0.0475704 , ...,  0.21625724,\n",
+       "          0.23215774,  0.23364826],\n",
+       "        [-0.2378668 ,  0.23262355, -0.0475704 , ...,  0.21625724,\n",
+       "          0.23215774,  0.23364826]], dtype=float32),\n",
+       " 'control_GE': array([[-0.0330701 ,  0.01753275, -0.05136025, ...,  0.06250588,\n",
+       "         -0.0405625 , -0.19137537],\n",
+       "        [ 0.01287525,  0.03997513,  0.035275  , ..., -0.04473713,\n",
+       "         -0.1068325 ,  0.03364687],\n",
+       "        [ 0.04092   , -0.0524448 , -0.0302184 , ..., -0.045687  ,\n",
+       "          0.018924  , -0.026421  ],\n",
+       "        ...,\n",
+       "        [ 0.1941795 , -0.00547   , -0.2204975 , ..., -0.05459375,\n",
+       "          0.10817   ,  0.011832  ],\n",
+       "        [ 0.03131568,  0.1014732 , -0.214663  , ...,  0.111306  ,\n",
+       "         -0.0012838 ,  0.143236  ],\n",
+       "        [-0.0118007 , -0.124132  ,  0.111037  , ...,  0.1051161 ,\n",
+       "         -0.275516  , -0.05782   ]], dtype=float32),\n",
+       " 'target_CP': array([[ 1.1871912 , -0.0034563 ,  0.52810115, ..., -1.6746186 ,\n",
+       "         -0.9638308 ,  0.73771006],\n",
+       "        [ 1.1871912 , -0.0034563 ,  0.52810115, ..., -1.6746186 ,\n",
+       "         -0.9638308 ,  0.73771006],\n",
+       "        [ 1.1871912 , -0.0034563 ,  0.52810115, ..., -1.6746186 ,\n",
+       "         -0.9638308 ,  0.73771006],\n",
+       "        ...,\n",
+       "        [-0.45244193, -0.05774116, -0.9494869 , ..., -0.27629495,\n",
+       "         -0.4013643 , -0.337052  ],\n",
+       "        [-0.45244193, -0.05774116, -0.9494869 , ..., -0.27629495,\n",
+       "         -0.4013643 , -0.337052  ],\n",
+       "        [-0.45244193, -0.05774116, -0.9494869 , ..., -0.27629495,\n",
+       "         -0.4013643 , -0.337052  ]], dtype=float32),\n",
+       " 'target_GE': array([[ 0.10045115,  0.0189385 , -0.136369  , ..., -0.115892  ,\n",
+       "          0.021835  , -0.240325  ],\n",
+       "        [ 0.105046  , -0.199388  ,  0.0151    , ..., -0.36268   ,\n",
+       "          0.01984   ,  0.060748  ],\n",
+       "        [ 0.17609   ,  0.012886  , -0.05032   , ...,  0.045372  ,\n",
+       "          0.37968   ,  0.180597  ],\n",
+       "        ...,\n",
+       "        [-0.120883  ,  0.42693624, -0.11864   , ...,  0.14749   ,\n",
+       "         -0.361165  , -0.1950255 ],\n",
+       "        [ 0.013835  , -0.42966   ,  0.396749  , ...,  0.25792   ,\n",
+       "         -0.1543    ,  0.03015   ],\n",
+       "        [ 0.1609315 , -0.40984   ,  0.028058  , ...,  0.31154   ,\n",
+       "         -0.104255  , -0.16908   ]], dtype=float32)}"
+      ]
+     },
+     "execution_count": 26,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)\n",
+    "agg_data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((223689, 775), (223689, 775), (223689, 977), (223689, 977))"
+      ]
+     },
+     "execution_count": 27,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape, agg_data['target_CP'].shape, agg_data['control_GE'].shape, agg_data['target_GE'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 限制配对的数量，更好收敛。"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "开始创建稳健配对数据集...\n",
+      "共有化合物: 20341\n",
+      "处理进度: 0/20341\n",
+      "处理进度: 1000/20341\n",
+      "处理进度: 2000/20341\n",
+      "处理进度: 3000/20341\n",
+      "处理进度: 4000/20341\n",
+      "处理进度: 5000/20341\n",
+      "处理进度: 6000/20341\n",
+      "处理进度: 7000/20341\n",
+      "处理进度: 8000/20341\n",
+      "处理进度: 9000/20341\n",
+      "处理进度: 10000/20341\n",
+      "处理进度: 11000/20341\n",
+      "处理进度: 12000/20341\n",
+      "处理进度: 13000/20341\n",
+      "处理进度: 14000/20341\n",
+      "处理进度: 15000/20341\n",
+      "处理进度: 16000/20341\n",
+      "处理进度: 17000/20341\n",
+      "处理进度: 18000/20341\n",
+      "处理进度: 19000/20341\n",
+      "处理进度: 20000/20341\n",
+      "总配对样本数: 169287\n",
+      "数据收集进度: 0/169287\n",
+      "数据收集进度: 10000/169287\n",
+      "数据收集进度: 20000/169287\n",
+      "数据收集进度: 30000/169287\n",
+      "数据收集进度: 40000/169287\n",
+      "数据收集进度: 50000/169287\n",
+      "数据收集进度: 60000/169287\n",
+      "数据收集进度: 70000/169287\n",
+      "数据收集进度: 80000/169287\n",
+      "数据收集进度: 90000/169287\n",
+      "数据收集进度: 100000/169287\n",
+      "数据收集进度: 110000/169287\n",
+      "数据收集进度: 120000/169287\n",
+      "数据收集进度: 130000/169287\n",
+      "数据收集进度: 140000/169287\n",
+      "数据收集进度: 150000/169287\n",
+      "数据收集进度: 160000/169287\n",
+      "control_CP: 均值=0.0175, 标准差=0.3551, 形状=(169287, 775)\n",
+      "target_CP: 均值=-0.0107, 标准差=1.1325, 形状=(169287, 775)\n",
+      "control_GE: 均值=-0.0003, 标准差=0.1331, 形状=(169287, 977)\n",
+      "target_GE: 均值=-0.0007, 标准差=0.2798, 形状=(169287, 977)\n",
+      "✅ 稳健配对数据集已保存到: ./Paired_CP_GE_Dataset.h5\n"
+     ]
+    }
+   ],
+   "source": [
+    "import h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "def create_robust_paired_dataset(CP_data, GE_data, out_path, max_pairs_per_smiles=20):\n",
+    "    \"\"\"\n",
+    "    创建稳健的样本级配对数据集\n",
+    "    添加样本限制和数据处理\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    print(\"开始创建稳健配对数据集...\")\n",
+    "    \n",
+    "    # 构建SMILES到索引列表的映射\n",
+    "    cp_smiles_to_indices = defaultdict(list)\n",
+    "    ge_smiles_to_indices = defaultdict(list)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(CP_data['canonical_smiles']):\n",
+    "        cp_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    for idx, smiles in enumerate(GE_data['canonical_smiles']):\n",
+    "        ge_smiles_to_indices[smiles].append(idx)\n",
+    "    \n",
+    "    # 找到共有的SMILES\n",
+    "    common_smiles = set(cp_smiles_to_indices.keys()) & set(ge_smiles_to_indices.keys())\n",
+    "    print(f\"共有化合物: {len(common_smiles)}\")\n",
+    "    \n",
+    "    # 收集所有有效配对（带采样限制）\n",
+    "    paired_samples = []\n",
+    "    \n",
+    "    for i, smiles in enumerate(common_smiles):\n",
+    "        # if i % 1000 == 0:\n",
+    "        #     print(f\"处理进度: {i}/{len(common_smiles)}\")\n",
+    "            \n",
+    "        cp_indices = cp_smiles_to_indices[smiles]\n",
+    "        ge_indices = ge_smiles_to_indices[smiles]\n",
+    "        \n",
+    "        # 限制每个SMILES的最大配对数，避免样本爆炸\n",
+    "        max_cp_samples = min(3, len(cp_indices))\n",
+    "        max_ge_samples = min(3, len(ge_indices))\n",
+    "        \n",
+    "        # 随机采样\n",
+    "        sampled_cp = np.random.choice(cp_indices, max_cp_samples, replace=False)\n",
+    "        sampled_ge = np.random.choice(ge_indices, max_ge_samples, replace=False)\n",
+    "        \n",
+    "        # 创建配对\n",
+    "        for cp_idx in sampled_cp:\n",
+    "            for ge_idx in sampled_ge:\n",
+    "                paired_samples.append({\n",
+    "                    'canonical_smiles': smiles,\n",
+    "                    'cp_index': cp_idx,\n",
+    "                    'ge_index': ge_idx\n",
+    "                })\n",
+    "    \n",
+    "    print(f\"总配对样本数: {len(paired_samples)}\")\n",
+    "    \n",
+    "    # 创建最终的字典形数据集\n",
+    "    dataset_dict = {\n",
+    "        'canonical_smiles': [],\n",
+    "        'control_CP': [],\n",
+    "        'target_CP': [], \n",
+    "        'control_GE': [],\n",
+    "        'target_GE': []\n",
+    "    }\n",
+    "    \n",
+    "    for i, sample in enumerate(paired_samples):\n",
+    "        if i % 10000 == 0:\n",
+    "            print(f\"数据收集进度: {i}/{len(paired_samples)}\")\n",
+    "            \n",
+    "        dataset_dict['canonical_smiles'].append(sample['canonical_smiles'])\n",
+    "        dataset_dict['control_CP'].append(CP_data['control'][sample['cp_index']])\n",
+    "        dataset_dict['target_CP'].append(CP_data['target'][sample['cp_index']])\n",
+    "        dataset_dict['control_GE'].append(GE_data['control'][sample['ge_index']]) \n",
+    "        dataset_dict['target_GE'].append(GE_data['target'][sample['ge_index']])\n",
+    "    \n",
+    "    # 转换为numpy数组并进行数据清洗\n",
+    "    final_dataset = {}\n",
+    "    final_dataset['canonical_smiles'] = np.array(dataset_dict['canonical_smiles'], dtype='S')\n",
+    "    \n",
+    "    # 数据清洗函数\n",
+    "    def clean_and_validate_data(data_list, name):\n",
+    "        data_array = np.array(data_list, dtype=np.float32)\n",
+    "        \n",
+    "        # 检查NaN和Inf\n",
+    "        nan_count = np.isnan(data_array).sum()\n",
+    "        inf_count = np.isinf(data_array).sum()\n",
+    "        \n",
+    "        if nan_count > 0 or inf_count > 0:\n",
+    "            print(f\"警告: {name} 包含 {nan_count} NaN 和 {inf_count} Inf\")\n",
+    "            data_array = np.nan_to_num(data_array, nan=0.0, posinf=1.0, neginf=-1.0)\n",
+    "        \n",
+    "        # 数据标准化\n",
+    "        mean = data_array.mean()\n",
+    "        std = data_array.std()\n",
+    "        \n",
+    "        print(f\"{name}: 均值={mean:.4f}, 标准差={std:.4f}, 形状={data_array.shape}\")\n",
+    "        \n",
+    "        return data_array\n",
+    "    \n",
+    "    final_dataset['control_CP'] = clean_and_validate_data(dataset_dict['control_CP'], 'control_CP')\n",
+    "    final_dataset['target_CP'] = clean_and_validate_data(dataset_dict['target_CP'], 'target_CP')\n",
+    "    final_dataset['control_GE'] = clean_and_validate_data(dataset_dict['control_GE'], 'control_GE')\n",
+    "    final_dataset['target_GE'] = clean_and_validate_data(dataset_dict['target_GE'], 'target_GE')\n",
+    "    \n",
+    "    # 保存为HDF5\n",
+    "    with h5py.File(out_path, 'w') as f:\n",
+    "        for key, value in final_dataset.items():\n",
+    "            f.create_dataset(key, data=value)\n",
+    "    \n",
+    "    print(f\"✅ 稳健配对数据集已保存到: {out_path}\")\n",
+    "    return final_dataset\n",
+    "\n",
+    "\n",
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "paired_dataset = create_robust_paired_dataset(CP_CSV, GE_CSV, out_path)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "out_path = './Paired_CP_GE_Dataset.h5'\n",
+    "\n",
+    "agg_data = load_from_HDF(out_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "((169287, 775), (169287, 775), (169287, 977), (169287, 977))"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "agg_data['control_CP'].shape, agg_data['target_CP'].shape, agg_data['control_GE'].shape, agg_data['target_GE'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

MVC/MVC_BBBC047/Step4_outlier_process.ipynb

@@ -0,0 +1,462 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "389cd136",
+   "metadata": {},
+   "source": [
+    "# 处理异常数据\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "917fa110",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import os\n",
+    "import h5py\n",
+    "\n",
+    "def load_from_HDF(fname):\n",
+    "    \"\"\"Load data from a HDF5 file to a dictionary.\"\"\"\n",
+    "    data = dict()\n",
+    "    with h5py.File(fname, 'r') as f:\n",
+    "        for key in f:\n",
+    "            data[key] = np.asarray(f[key])\n",
+    "            if isinstance(data[key][0], np.bytes_):\n",
+    "                data[key] = data[key].astype(str)\n",
+    "    return data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "cddb976e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "path47 = 'Aggregated_CP_GE.h5'\n",
+    "\n",
+    "# data36 = load_from_HDF(path36)\n",
+    "data47 = load_from_HDF(path47)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "54e51b05",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-20.614704132080078, 2870874.0]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-1249.9088134765625, 19152610.0]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['control_CP'].min()\n",
+    "data_max = data47['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data47['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['target_CP'].min()\n",
+    "data_max = data47['target_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['target_CP'], axis=0)\n",
+    "max_per_column = np.max(data47['target_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "bde80b4c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-2.230109453201294, 1.6778349876403809]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-6.4108500480651855, 9.534767150878906]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['control_GE'].min()\n",
+    "data_max = data47['control_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['control_GE'], axis=0)\n",
+    "max_per_column = np.max(data47['control_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['target_GE'].min()\n",
+    "data_max = data47['target_GE'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['target_GE'], axis=0)\n",
+    "max_per_column = np.max(data47['target_GE'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "053cfaa7",
+   "metadata": {},
+   "source": [
+    "## 看看 cpg0016 的数值范围"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "c81440ab",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-1.2491892576217651, 1.3490736484527588]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-6.057624816894531, 6.188350677490234]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "cpg0016_data = '/home/bob/boom/VCBench/data/Image/cpg0016_data.h5'\n",
+    "\n",
+    "# data36 = load_from_HDF(path36)\n",
+    "cpg0016_data = load_from_HDF(cpg0016_data)\n",
+    "\n",
+    "import numpy as np\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = cpg0016_data['control'].min()\n",
+    "data_max = cpg0016_data['control'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(cpg0016_data['control'], axis=0)\n",
+    "max_per_column = np.max(cpg0016_data['control'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = cpg0016_data['target'].min()\n",
+    "data_max = cpg0016_data['target'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(cpg0016_data['target'], axis=0)\n",
+    "max_per_column = np.max(cpg0016_data['target'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "2b2f331e",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "标准化后的整个数组的数据范围：[-11.38830852508545, 10.123730659484863]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "# 假设 data47['target_CP'] 是你的原始数据, target_CP, \n",
+    "data_with_outlier = data47['control_CP']\n",
+    "\n",
+    "# 1. 异常值处理 - 使用截断法 (Clipping)\n",
+    "# 设定一个更保守的上限，例如基于分位数\n",
+    "q99 = np.percentile(data_with_outlier, 99)\n",
+    "clip_max = q99 * 2 # 或者直接用一个固定值，比如 5\n",
+    "data_clipped = np.clip(data_with_outlier, a_min=None, a_max=clip_max)\n",
+    "# 2. 数据标准化 - Z-score Normalization\n",
+    "# 计算每列（特征）的均值和标准差\n",
+    "mu = data_clipped.mean(axis=0)\n",
+    "sigma = data_clipped.std(axis=0)\n",
+    "\n",
+    "# 初始化一个与数据形状相同的零数组\n",
+    "data_normalized = np.zeros_like(data_clipped, dtype=float)\n",
+    "\n",
+    "# 找到标准差不为0的列\n",
+    "# 为了避免浮点数误差，使用一个非常小的阈值来检查\n",
+    "epsilon = 1e-8\n",
+    "nonzero_sigma_cols = sigma > epsilon\n",
+    "\n",
+    "# 对这些列进行零均值化\n",
+    "data_normalized[:, nonzero_sigma_cols] = (data_clipped[:, nonzero_sigma_cols] - mu[nonzero_sigma_cols]) / sigma[nonzero_sigma_cols]\n",
+    "\n",
+    "# 检查结果\n",
+    "data_min = data_normalized.min()\n",
+    "data_max = data_normalized.max()\n",
+    "\n",
+    "print(f\"\\n标准化后的整个数组的数据范围：[{data_min}, {data_max}]\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "9ccda958",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "control_CP = data_normalized.copy()\n",
+    "# print(control_CP.shape)\n",
+    "# target_CP = data_normalized.copy()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "6e515034",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-11.38830852508545, 10.123730659484863]\n",
+      "------------------------------\n",
+      "整个数组的数据范围：[-129.0272979736328, 64.47639465332031]\n",
+      "------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data_min = control_CP.min()\n",
+    "data_max = control_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = target_CP.min()\n",
+    "data_max = target_CP.max()\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "id": "481b57e7",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "float64 float64\n",
+      "float32 float32\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(control_CP.dtype, target_CP.dtype)\n",
+    "\n",
+    "control_CP_float32 = control_CP.astype(np.float32)\n",
+    "target_CP_float32 = target_CP.astype(np.float32)\n",
+    "print(control_CP_float32.dtype, target_CP_float32.dtype)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "6e2d6030",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "已删除旧的 'control_CP' 数据集。\n",
+      "已删除旧的 'target_CP' 数据集。\n",
+      "已成功��入新的 'control_CP' 和 'target_CP' 数据集。\n",
+      "操作完成。\n"
+     ]
+    }
+   ],
+   "source": [
+    "file_path = 'Aggregated_CP_GE.h5'\n",
+    "# 使用 'a' (append) 模式打开文件\n",
+    "with h5py.File(file_path, 'a') as f:\n",
+    "    \n",
+    "    # 1. 删除旧的数据集（如果存在）\n",
+    "    if \"control_CP\" in f:\n",
+    "        del f[\"control_CP\"]\n",
+    "        print(\"已删除旧的 'control_CP' 数据集。\")\n",
+    "        \n",
+    "    if \"target_CP\" in f:\n",
+    "        del f[\"target_CP\"]\n",
+    "        print(\"已删除旧的 'target_CP' 数据集。\")\n",
+    "    \n",
+    "    # 2. 创建并写入新的数据集\n",
+    "    f.create_dataset(\"control_CP\", data=control_CP_float32)\n",
+    "    f.create_dataset(\"target_CP\", data=target_CP_float32)\n",
+    "    \n",
+    "    print(\"已成功写入新的 'control_CP' 和 'target_CP' 数据集。\")\n",
+    "\n",
+    "print(\"操作完成。\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b131c13e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5a82372f",
+   "metadata": {},
+   "source": [
+    "# 看基因数据是否异常 -- 没有"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "a81ddc86",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "整个数组的数据范围：[-11.38830852508545, 10.123730659484863]\n",
+      "------------------------------\n",
+      "每列数据的最小值（前10个）：\n",
+      "[-1.7736303 -2.8932922 -1.9792434 -1.3063283 -1.5805442 -2.5623124\n",
+      " -2.9260125 -3.1737845 -3.493395  -5.5112543]\n",
+      "------------------------------\n",
+      "每列数据的最大值（前10个）：\n",
+      "[3.8115907 4.765787  9.468964  3.2014632 2.3736608 4.2486434 4.5805535\n",
+      " 4.513591  1.7128    2.9639645]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "\n",
+    "data47 = load_from_HDF( 'Aggregated_CP_GE.h5')\n",
+    "# 计算整个数组的最小值和最大值\n",
+    "data_min = data47['control_CP'].min()\n",
+    "data_max = data47['control_CP'].max()\n",
+    "\n",
+    "# 计算每列（即每个特征）的最小值和最大值\n",
+    "# axis=0 表示对每一列进行操作\n",
+    "min_per_column = np.min(data47['control_CP'], axis=0)\n",
+    "max_per_column = np.max(data47['control_CP'], axis=0)\n",
+    "\n",
+    "print(f\"整个数组的数据范围：[{data_min}, {data_max}]\")\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最小值（前10个）：\")\n",
+    "print(min_per_column[:10])\n",
+    "print(\"-\" * 30)\n",
+    "print(\"每列数据的最大值（前10个）：\")\n",
+    "print(max_per_column[:10])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "770f3727",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 看看方差\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "10693804",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c072eea5",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

Molecular_representations/Get_embeddings/MolCLR/Get_emd.ipynb

@@ -0,0 +1,291 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_29432/3546513620.py:42: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  state_dict = torch.load(model_path, map_location=self.device)\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Loaded trained model successfully.\n",
+      "Processed 1000 molecules...\n",
+      "Processed 2000 molecules...\n",
+      "Processed 3000 molecules...\n",
+      "Processed 4000 molecules...\n",
+      "Processed 5000 molecules...\n",
+      "Processed 6000 molecules...\n",
+      "Processed 7000 molecules...\n",
+      "Processed 8000 molecules...\n",
+      "Processed 9000 molecules...\n",
+      "Processed 10000 molecules...\n",
+      "Processed 11000 molecules...\n",
+      "Processed 12000 molecules...\n",
+      "Processed 13000 molecules...\n",
+      "Processed 14000 molecules...\n",
+      "Processed 15000 molecules...\n",
+      "Processed 16000 molecules...\n",
+      "Processed 17000 molecules...\n",
+      "Processed 18000 molecules...\n",
+      "Processed 19000 molecules...\n",
+      "Processed 20000 molecules...\n",
+      "Processed 21000 molecules...\n",
+      "Processed 22000 molecules...\n",
+      "Processed 23000 molecules...\n",
+      "Processed 24000 molecules...\n",
+      "Processed 25000 molecules...\n",
+      "Processed 26000 molecules...\n",
+      "Processed 27000 molecules...\n",
+      "Processed 28000 molecules...\n",
+      "Processed 29000 molecules...\n",
+      "Processed 30000 molecules...\n",
+      "Processed 31000 molecules...\n",
+      "Processed 32000 molecules...\n",
+      "Processed 33000 molecules...\n",
+      "Processed 34000 molecules...\n",
+      "Processed 35000 molecules...\n",
+      "Processed 36000 molecules...\n",
+      "Processed 37000 molecules...\n",
+      "Processed 38000 molecules...\n",
+      "Processed 39000 molecules...\n",
+      "Processed 40000 molecules...\n",
+      "Processed 41000 molecules...\n",
+      "Processed 42000 molecules...\n",
+      "Processed 43000 molecules...\n",
+      "Processed 44000 molecules...\n",
+      "Processed 45000 molecules...\n",
+      "Processed 46000 molecules...\n",
+      "Processed 47000 molecules...\n",
+      "Processed 48000 molecules...\n",
+      "Processed 49000 molecules...\n",
+      "Processed 50000 molecules...\n",
+      "Processed 51000 molecules...\n",
+      "Processed 52000 molecules...\n",
+      "Processed 53000 molecules...\n",
+      "Processed 54000 molecules...\n",
+      "\n",
+      "Completed. Success: 54201, Failed: 0\n",
+      "54201\n",
+      "总分子数 : 54201\n",
+      "成功嵌入 : 54201\n",
+      "总耗时   : 87.68 s\n",
+      "平均推理 : 1.62 ms / molecule\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import csv\n",
+    "import torch\n",
+    "import numpy as np\n",
+    "from rdkit import Chem\n",
+    "from rdkit.Chem.rdchem import BondType as BT\n",
+    "from torch_geometric.data import Data\n",
+    "import yaml\n",
+    "import time\n",
+    "from pathlib import Path\n",
+    "\n",
+    "# 禁用RDKit的日志输出\n",
+    "from rdkit import RDLogger\n",
+    "RDLogger.DisableLog('rdApp.*')\n",
+    "\n",
+    "# 原子和化学键的类型定义\n",
+    "ATOM_LIST = list(range(1, 119))\n",
+    "CHIRALITY_LIST = [\n",
+    "    Chem.rdchem.ChiralType.CHI_UNSPECIFIED,\n",
+    "    Chem.rdchem.ChiralType.CHI_TETRAHEDRAL_CW,\n",
+    "    Chem.rdchem.ChiralType.CHI_TETRAHEDRAL_CCW,\n",
+    "    Chem.rdchem.ChiralType.CHI_OTHER\n",
+    "]\n",
+    "BOND_LIST = [BT.SINGLE, BT.DOUBLE, BT.TRIPLE, BT.AROMATIC]\n",
+    "BONDDIR_LIST = [\n",
+    "    Chem.rdchem.BondDir.NONE,\n",
+    "    Chem.rdchem.BondDir.ENDUPRIGHT,\n",
+    "    Chem.rdchem.BondDir.ENDDOWNRIGHT\n",
+    "]\n",
+    "\n",
+    "class SMILESToEmbedding:\n",
+    "    def __init__(self, model_path, device='cuda' if torch.cuda.is_available() else 'cpu'):\n",
+    "        self.device = torch.device(device)\n",
+    "        self.model = self.load_pretrained_model(model_path)\n",
+    "        self.model.eval()\n",
+    "        \n",
+    "    def load_pretrained_model(self, model_path):\n",
+    "        self.config = yaml.load(open(\"config_finetune.yaml\", \"r\"), Loader=yaml.FullLoader)\n",
+    "        from models.ginet_finetune import GINet\n",
+    "        model = GINet(**self.config[\"model\"]).to(self.device)\n",
+    "        \n",
+    "        state_dict = torch.load(model_path, map_location=self.device)\n",
+    "        model.load_state_dict(state_dict, strict=False)\n",
+    "        model.to(self.device)\n",
+    "        print(\"Loaded trained model successfully.\")\n",
+    "        return model\n",
+    "    \n",
+    "    def smiles_to_graph(self, smiles):\n",
+    "        \"\"\"将SMILES字符串转换为图数据结构\"\"\"\n",
+    "        mol = Chem.MolFromSmiles(smiles)\n",
+    "        if mol is None:\n",
+    "            return None\n",
+    "            \n",
+    "        mol = Chem.AddHs(mol)\n",
+    "        \n",
+    "        # 原子特征\n",
+    "        type_idx = []\n",
+    "        chirality_idx = []\n",
+    "        atomic_number = []\n",
+    "        for atom in mol.GetAtoms():\n",
+    "            type_idx.append(ATOM_LIST.index(atom.GetAtomicNum()))\n",
+    "            chirality_idx.append(CHIRALITY_LIST.index(atom.GetChiralTag()))\n",
+    "            atomic_number.append(atom.GetAtomicNum())\n",
+    "\n",
+    "        x1 = torch.tensor(type_idx, dtype=torch.long).view(-1, 1)\n",
+    "        x2 = torch.tensor(chirality_idx, dtype=torch.long).view(-1, 1)\n",
+    "        x = torch.cat([x1, x2], dim=-1)\n",
+    "\n",
+    "        # 化学键特征\n",
+    "        row, col, edge_feat = [], [], []\n",
+    "        for bond in mol.GetBonds():\n",
+    "            start, end = bond.GetBeginAtomIdx(), bond.GetEndAtomIdx()\n",
+    "            row += [start, end]\n",
+    "            col += [end, start]\n",
+    "            edge_feat.append([\n",
+    "                BOND_LIST.index(bond.GetBondType()),\n",
+    "                BONDDIR_LIST.index(bond.GetBondDir())\n",
+    "            ])\n",
+    "            edge_feat.append([\n",
+    "                BOND_LIST.index(bond.GetBondType()),\n",
+    "                BONDDIR_LIST.index(bond.GetBondDir())\n",
+    "            ])\n",
+    "\n",
+    "        edge_index = torch.tensor([row, col], dtype=torch.long)\n",
+    "        edge_attr = torch.tensor(np.array(edge_feat), dtype=torch.long)\n",
+    "        \n",
+    "        return Data(x=x, edge_index=edge_index, edge_attr=edge_attr)\n",
+    "    \n",
+    "    def process_csv(self, csv_path, smiles_column='SMILES'):\n",
+    "        \"\"\"处理CSV文件中的SMILES字符串\"\"\"\n",
+    "        embeddings_dict = {}  # 使用字典存储结果，键为SMILES，值为embedding\n",
+    "        failed_smiles = []\n",
+    "        \n",
+    "        with open(csv_path, 'r') as f:\n",
+    "            reader = csv.DictReader(f)\n",
+    "            for i, row in enumerate(reader):\n",
+    "                smiles = row[smiles_column]\n",
+    "                try:\n",
+    "                    # 转换为图结构\n",
+    "                    graph = self.smiles_to_graph(smiles)\n",
+    "                    if graph is None:\n",
+    "                        failed_smiles.append(smiles)\n",
+    "                        continue\n",
+    "                        \n",
+    "                    # 通过模型获取嵌入\n",
+    "                    graph = graph.to(self.device)\n",
+    "                    with torch.no_grad():\n",
+    "                        embedding = self.model(graph)\n",
+    "                    embedding = embedding.squeeze(0).cpu().numpy().astype(np.float32)\n",
+    "                    \n",
+    "                    # 将结果存入字典\n",
+    "                    embeddings_dict[smiles] = embedding\n",
+    "                    \n",
+    "                except Exception as e:\n",
+    "                    print(f\"Error processing SMILES {smiles}: {str(e)}\")\n",
+    "                    failed_smiles.append(smiles)\n",
+    "                \n",
+    "                if (i+1) % 1000 == 0:\n",
+    "                    print(f\"Processed {i+1} molecules...\")\n",
+    "        \n",
+    "        print(f\"\\nCompleted. Success: {len(embeddings_dict)}, Failed: {len(failed_smiles)}\")\n",
+    "        return embeddings_dict, failed_smiles\n",
+    "\n",
+    "# 使用示例\n",
+    "if __name__ == \"__main__\":\n",
+    "    # 替换为你的模型路径和CSV文件路径\n",
+    "    processor = SMILESToEmbedding(\n",
+    "        model_path='/home/bob/boom/VCBench/Molecule_encoder/MolCLR/ckpt/pretrained_gin/checkpoints/model.pth'\n",
+    "    )  \n",
+    "    \n",
+    "    # ----- 计时开始 -----\n",
+    "    t0 = time.perf_counter()\n",
+    "    \n",
+    "    # 处理CSV文件\n",
+    "    embeddings_dict, failed = processor.process_csv(\n",
+    "        csv_path='/home/bob/boom/AIVC/VOBench/data/canonical_smiles_list.csv',\n",
+    "        smiles_column='SMILES'  # 替换为���的CSV中SMILES列的名称\n",
+    "    )\n",
+    "    \n",
+    "    t1 = time.perf_counter()\n",
+    "    # ----- 计时结束 -----    \n",
+    "    print(len(embeddings_dict))\n",
+    "    \n",
+    "    # 基本统计\n",
+    "    total_mols = len(embeddings_dict)\n",
+    "    avg_time_ms = (t1 - t0) / total_mols * 1000\n",
+    "\n",
+    "    print(f\"总分子数 : {total_mols}\")\n",
+    "    print(f\"成功嵌入 : {len(embeddings_dict)}\")\n",
+    "    print(f\"总耗时   : {t1 - t0:.2f} s\")\n",
+    "    print(f\"平均推理 : {avg_time_ms:.2f} ms / molecule\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 /data/boom/Virtual_Organoid/Molecule_encoder/MolCLR_emb512.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "# 保存字典为pickle文件\n",
+    "pickle_path = '/data/boom/Virtual_Organoid/Molecule_encoder/MolCLR_emb512.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(embeddings_dict, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

Molecular_representations/Get_embeddings/Mole-BERT/Get_emd.ipynb

@@ -0,0 +1,334 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/home/bob/boom/VCBench/Molecule_encoder/Mole-BERT/model.py:499: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  self.gnn.load_state_dict(torch.load(model_file))\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Loaded trained model successfully.\n",
+      "Processed 1000 molecules...\n",
+      "Processed 2000 molecules...\n",
+      "Processed 3000 molecules...\n",
+      "Processed 4000 molecules...\n",
+      "Processed 5000 molecules...\n",
+      "Processed 6000 molecules...\n",
+      "Processed 7000 molecules...\n",
+      "Processed 8000 molecules...\n",
+      "Processed 9000 molecules...\n",
+      "Processed 10000 molecules...\n",
+      "Processed 11000 molecules...\n",
+      "Processed 12000 molecules...\n",
+      "Processed 13000 molecules...\n",
+      "Processed 14000 molecules...\n",
+      "Processed 15000 molecules...\n",
+      "Processed 16000 molecules...\n",
+      "Processed 17000 molecules...\n",
+      "Processed 18000 molecules...\n",
+      "Processed 19000 molecules...\n",
+      "Processed 20000 molecules...\n",
+      "Processed 21000 molecules...\n",
+      "Processed 22000 molecules...\n",
+      "Processed 23000 molecules...\n",
+      "Processed 24000 molecules...\n",
+      "Processed 25000 molecules...\n",
+      "Processed 26000 molecules...\n",
+      "Processed 27000 molecules...\n",
+      "Processed 28000 molecules...\n",
+      "Processed 29000 molecules...\n",
+      "Processed 30000 molecules...\n",
+      "Processed 31000 molecules...\n",
+      "Processed 32000 molecules...\n",
+      "Processed 33000 molecules...\n",
+      "Processed 34000 molecules...\n",
+      "Processed 35000 molecules...\n",
+      "Processed 36000 molecules...\n",
+      "Processed 37000 molecules...\n",
+      "Processed 38000 molecules...\n",
+      "Processed 39000 molecules...\n",
+      "Processed 40000 molecules...\n",
+      "Processed 41000 molecules...\n",
+      "Processed 42000 molecules...\n",
+      "Processed 43000 molecules...\n",
+      "Processed 44000 molecules...\n",
+      "Processed 45000 molecules...\n",
+      "Processed 46000 molecules...\n",
+      "Processed 47000 molecules...\n",
+      "Processed 48000 molecules...\n",
+      "Processed 49000 molecules...\n",
+      "Processed 50000 molecules...\n",
+      "Processed 51000 molecules...\n",
+      "Processed 52000 molecules...\n",
+      "Processed 53000 molecules...\n",
+      "Processed 54000 molecules...\n",
+      "\n",
+      "Completed. Success: 54201, Failed: 0\n",
+      "54201\n",
+      "54201\n",
+      "总分子数 : 54201\n",
+      "成功嵌入 : 54201\n",
+      "总耗时   : 87.55 s\n",
+      "平均推理 : 1.62 ms / molecule\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import csv\n",
+    "import torch\n",
+    "import numpy as np\n",
+    "from rdkit import Chem\n",
+    "from rdkit.Chem.rdchem import BondType as BT\n",
+    "from torch_geometric.data import Data\n",
+    "import yaml\n",
+    "\n",
+    "# 禁用RDKit的日志输出\n",
+    "from rdkit import RDLogger\n",
+    "RDLogger.DisableLog('rdApp.*')\n",
+    "\n",
+    "# 原子和化学键的类型定义\n",
+    "ATOM_LIST = list(range(1, 119))\n",
+    "CHIRALITY_LIST = [\n",
+    "    Chem.rdchem.ChiralType.CHI_UNSPECIFIED,\n",
+    "    Chem.rdchem.ChiralType.CHI_TETRAHEDRAL_CW,\n",
+    "    Chem.rdchem.ChiralType.CHI_TETRAHEDRAL_CCW,\n",
+    "    Chem.rdchem.ChiralType.CHI_OTHER\n",
+    "]\n",
+    "BOND_LIST = [BT.SINGLE, BT.DOUBLE, BT.TRIPLE, BT.AROMATIC]\n",
+    "BONDDIR_LIST = [\n",
+    "    Chem.rdchem.BondDir.NONE,\n",
+    "    Chem.rdchem.BondDir.ENDUPRIGHT,\n",
+    "    Chem.rdchem.BondDir.ENDDOWNRIGHT\n",
+    "]\n",
+    "\n",
+    "class SMILESToEmbedding:\n",
+    "    def __init__(self, model_path, device='cuda' if torch.cuda.is_available() else 'cpu'):\n",
+    "        self.device = torch.device(device)\n",
+    "        self.model = self.load_pretrained_model(model_path)\n",
+    "        self.model.eval()\n",
+    "        \n",
+    "    def load_pretrained_model(self, model_path):\n",
+    "        import argparse\n",
+    "        parser = argparse.ArgumentParser(description='PyTorch implementation of pre-training of graph neural networks')\n",
+    "        parser.add_argument('--device', type=int, default=0,\n",
+    "                            help='which gpu to use if any (default: 0)')\n",
+    "        parser.add_argument('--batch_size', type=int, default=16,\n",
+    "                            help='input batch size for training (default: 32)')\n",
+    "        parser.add_argument('--epochs', type=int, default=100,\n",
+    "                            help='number of epochs to train (default: 100)')\n",
+    "        parser.add_argument('--lr', type=float, default=0.001,\n",
+    "                            help='learning rate (default: 0.001)')\n",
+    "        parser.add_argument('--lr_scale', type=float, default=1,\n",
+    "                            help='relative learning rate for the feature extraction layer (default: 1)')\n",
+    "        parser.add_argument('--decay', type=float, default=0,\n",
+    "                            help='weight decay (default: 0)')\n",
+    "        parser.add_argument('--num_layer', type=int, default=5,\n",
+    "                            help='number of GNN message passing layers (default: 5).')\n",
+    "        parser.add_argument('--emb_dim', type=int, default=300,\n",
+    "                            help='embedding dimensions (default: 300)')\n",
+    "        parser.add_argument('--dropout_ratio', type=float, default=0.5,\n",
+    "                            help='dropout ratio (default: 0.5)')\n",
+    "        parser.add_argument('--graph_pooling', type=str, default=\"mean\",\n",
+    "                            help='graph level pooling (sum, mean, max, set2set, attention)')\n",
+    "        parser.add_argument('--JK', type=str, default=\"last\",\n",
+    "                            help='how the node features across layers are combined. last, sum, max or concat')\n",
+    "        parser.add_argument('--gnn_type', type=str, default=\"gin\")\n",
+    "        parser.add_argument('--dataset', type=str, default = 'sider', help='root directory of dataset. For now, only classification.')\n",
+    "        parser.add_argument('--input_model_file', type=str, default = 'None', help='filename to read the model (if there is any)')\n",
+    "        parser.add_argument('--filename', type=str, default = '', help='output filename')\n",
+    "        parser.add_argument('--seed', type=int, default=42, help = \"Seed for splitting the dataset.\")\n",
+    "        parser.add_argument('--runseed', type=int, default=0, help = \"Seed for minibatch selection, random initialization.\")\n",
+    "        parser.add_argument('--split', type = str, default=\"scaffold\", help = \"random or scaffold or random_scaffold\")\n",
+    "        parser.add_argument('--eval_train', type=int, default = 1, help='evaluating training or not')\n",
+    "        parser.add_argument('--num_workers', type=int, default = 4, help='number of workers for dataset loading')\n",
+    "        args = parser.parse_args()\n",
+    "        \n",
+    "        from model import GNN_graphpred\n",
+    "        device = torch.device(\"cuda:\" + str(args.device)) if torch.cuda.is_available() else torch.device(\"cpu\")\n",
+    "        \n",
+    "        model = GNN_graphpred(args.num_layer, args.emb_dim, 1, JK = args.JK, drop_ratio = args.dropout_ratio, graph_pooling = args.graph_pooling, gnn_type = args.gnn_type)\n",
+    "        model.from_pretrained(model_path)\n",
+    "        model.to(device)\n",
+    "\n",
+    "        print(\"Loaded trained model successfully.\")\n",
+    "        \n",
+    "        return model\n",
+    "    \n",
+    "    def smiles_to_graph(self, smiles):\n",
+    "        \"\"\"将SMILES字符串转换为图数据结构\"\"\"\n",
+    "        mol = Chem.MolFromSmiles(smiles)\n",
+    "        if mol is None:\n",
+    "            return None\n",
+    "            \n",
+    "        mol = Chem.AddHs(mol)\n",
+    "        \n",
+    "        # 原子特征\n",
+    "        type_idx = []\n",
+    "        chirality_idx = []\n",
+    "        atomic_number = []\n",
+    "        for atom in mol.GetAtoms():\n",
+    "            type_idx.append(ATOM_LIST.index(atom.GetAtomicNum()))\n",
+    "            chirality_idx.append(CHIRALITY_LIST.index(atom.GetChiralTag()))\n",
+    "            atomic_number.append(atom.GetAtomicNum())\n",
+    "\n",
+    "        x1 = torch.tensor(type_idx, dtype=torch.long).view(-1, 1)\n",
+    "        x2 = torch.tensor(chirality_idx, dtype=torch.long).view(-1, 1)\n",
+    "        x = torch.cat([x1, x2], dim=-1)\n",
+    "\n",
+    "        # 化学键特征\n",
+    "        row, col, edge_feat = [], [], []\n",
+    "        for bond in mol.GetBonds():\n",
+    "            start, end = bond.GetBeginAtomIdx(), bond.GetEndAtomIdx()\n",
+    "            row += [start, end]\n",
+    "            col += [end, start]\n",
+    "            edge_feat.append([\n",
+    "                BOND_LIST.index(bond.GetBondType()),\n",
+    "                BONDDIR_LIST.index(bond.GetBondDir())\n",
+    "            ])\n",
+    "            edge_feat.append([\n",
+    "                BOND_LIST.index(bond.GetBondType()),\n",
+    "                BONDDIR_LIST.index(bond.GetBondDir())\n",
+    "            ])\n",
+    "\n",
+    "        edge_index = torch.tensor([row, col], dtype=torch.long)\n",
+    "        edge_attr = torch.tensor(np.array(edge_feat), dtype=torch.long)\n",
+    "        \n",
+    "        return Data(x=x, edge_index=edge_index, edge_attr=edge_attr)\n",
+    "    \n",
+    "    def process_csv(self, csv_path, smiles_column='SMILES'):\n",
+    "        \"\"\"处理CSV文件中的SMILES字符串\"\"\"\n",
+    "        embeddings_dict = {}  # 使用字典存储结果，键为SMILES，值为embedding\n",
+    "        failed_smiles = []\n",
+    "        \n",
+    "        with open(csv_path, 'r') as f:\n",
+    "            reader = csv.DictReader(f)\n",
+    "            for i, row in enumerate(reader):\n",
+    "                smiles = row[smiles_column]\n",
+    "                try:\n",
+    "                    # 转换为图结构\n",
+    "                    graph = self.smiles_to_graph(smiles)\n",
+    "                    if graph is None:\n",
+    "                        failed_smiles.append(smiles)\n",
+    "                        continue\n",
+    "                        \n",
+    "                    # 通过模型获取嵌入\n",
+    "                    graph = graph.to(self.device)\n",
+    "                    with torch.no_grad():\n",
+    "                        output = self.model(graph)\n",
+    "                        # print(output.shape)\n",
+    "                    # break\n",
+    "                    \n",
+    "                    # 处理模型输出（可能是元组）\n",
+    "                    if isinstance(output, tuple):\n",
+    "                        # 假设我们只需要第一个元素作为embedding\n",
+    "                        embedding = output[0]\n",
+    "                    else:\n",
+    "                        embedding = output\n",
+    "                    \n",
+    "                    # 将结果存入字典\n",
+    "                    embeddings_dict[smiles] = embedding.cpu().numpy()\n",
+    "                    \n",
+    "                except Exception as e:\n",
+    "                    print(f\"Error processing SMILES {smiles}: {str(e)}\")\n",
+    "                    failed_smiles.append(smiles)\n",
+    "                \n",
+    "                if (i+1) % 1000 == 0:\n",
+    "                    print(f\"Processed {i+1} molecules...\")\n",
+    "        \n",
+    "        print(f\"\\nCompleted. Success: {len(embeddings_dict)}, Failed: {len(failed_smiles)}\")\n",
+    "        return embeddings_dict, failed_smiles\n",
+    "\n",
+    "# 使用示例\n",
+    "if __name__ == \"__main__\":\n",
+    "    # 替换为你的模型路径和CSV文件路径\n",
+    "    processor = SMILESToEmbedding(model_path='model_gin/Mole-BERT.pth')\n",
+    "    import time\n",
+    "    \n",
+    "    # ----- 计时开始 -----\n",
+    "    t0 = time.perf_counter()\n",
+    "    \n",
+    "    # 处理CSV文件\n",
+    "    embeddings_dict, failed = processor.process_csv(\n",
+    "        csv_path='/home/bob/boom/AIVC/VOBench/data/canonical_smiles_list.csv',\n",
+    "        smiles_column='SMILES'  # 替换为你的CSV中SMILES列的名称\n",
+    "    )\n",
+    "    print(len(embeddings_dict))\n",
+    "    \n",
+    "    t1 = time.perf_counter()\n",
+    "    # ----- 计时结束 -----    \n",
+    "    print(len(embeddings_dict))\n",
+    "    \n",
+    "    # 基本统计\n",
+    "    total_mols = len(embeddings_dict)\n",
+    "    avg_time_ms = (t1 - t0) / total_mols * 1000\n",
+    "\n",
+    "    print(f\"总分子数 : {total_mols}\")\n",
+    "    print(f\"成功嵌入 : {len(embeddings_dict)}\")\n",
+    "    print(f\"总耗时   : {t1 - t0:.2f} s\")\n",
+    "    print(f\"平均推理 : {avg_time_ms:.2f} ms / molecule\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 /data/boom/Virtual_Organoid/Molecule_encoder/Mole_BERT_emb300.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "# 保存字典为pickle文件\n",
+    "pickle_path = '/data/boom/Virtual_Organoid/Molecule_encoder/Mole_BERT_emb300.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(embeddings_dict, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

Molecular_representations/Get_embeddings/SMILES_Encoder.ipynb

@@ -0,0 +1,1653 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# look data_example"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "with open('./LINCS2020/idx2smi.pickle', 'rb') as f:\n",
+    "    idx2smi = pickle.load(f)\\\n",
+    "        \n",
+    "smiles_list = list(idx2smi.values())  # 假设idx2smi是一个字典，其值为SMILES字符串\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 BrC1C(Br)C(Br)C(Br)C(Br)C1Br\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(idx2smi), idx2smi[0])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'BrC1C(Br)C(Br)C(Br)C(Br)C1Br'"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "idx2smi[0]  # to path_to_input_smiles.csv"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "SMILES已成功保存到LINCS2020_smiles.csv文件中。\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "import csv\n",
+    "\n",
+    "# 将SMILES字符串写入CSV文件\n",
+    "with open('./LINCS2020/LINCS2020_smiles.csv', 'w', newline='') as csvfile:\n",
+    "    writer = csv.writer(csvfile)\n",
+    "    # 写入标题行（可选）\n",
+    "    writer.writerow(['SMILES'])\n",
+    "    # 遍历字典，写入SMILES\n",
+    "    for idx, smi in idx2smi.items():\n",
+    "        writer.writerow([smi])\n",
+    "\n",
+    "print(\"SMILES已成功保存到LINCS2020_smiles.csv文件中。\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate ECFP4 embedding for all smiles\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from rdkit import Chem\n",
+    "from rdkit.Chem import AllChem\n",
+    "import numpy as np\n",
+    "import pickle\n",
+    "\n",
+    "def smiles_to_ecfp4(smiles_list, radius=2, n_bits=2048):\n",
+    "    \"\"\"\n",
+    "    将SMILES列表转换为ECFP4特征向量。\n",
+    "\n",
+    "    参数:\n",
+    "    smiles_list (list): SMILES字符串列表。\n",
+    "    radius (int): ECFP的半径，默认为2（即ECFP4）。\n",
+    "    n_bits (int): 特征向量的长度，默认为2048。\n",
+    "\n",
+    "    返回:\n",
+    "    dict: 一个字典，键为SMILES字符串，值为对应的ECFP4特征向量。\n",
+    "    \"\"\"\n",
+    "    ecfp4_dict = {}\n",
+    "    \n",
+    "    for smiles in smiles_list:\n",
+    "        mol = Chem.MolFromSmiles(smiles)\n",
+    "        if mol is not None:\n",
+    "            # 替换编码器\n",
+    "            ecfp4 = AllChem.GetMorganFingerprintAsBitVect(mol, radius=radius, nBits=n_bits)\n",
+    "            ecfp4_dict[smiles] = np.array(ecfp4, dtype=np.float32)\n",
+    "        else:\n",
+    "            print(\"Error in \", smiles)\n",
+    "            # 如果SMILES无法解析为分子，则填充一个全零的向量\n",
+    "            ecfp4_dict[smiles] = np.zeros(n_bits, dtype=np.float32)\n",
+    "    \n",
+    "    return ecfp4_dict\n",
+    "\n",
+    "# 从字典中提取SMILES列表\n",
+    "smiles_list = list(idx2smi.values())  # 假设idx2smi是一个字典，其值为SMILES字符串\n",
+    "\n",
+    "# 计算ECFP4特征\n",
+    "ecfp4_features_dict = smiles_to_ecfp4(smiles_list)\n",
+    "\n",
+    "# ecfp4_features_dict\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 44,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 8316\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(smiles_list), len(ecfp4_features_dict))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ecfp4_features_dict\n",
+    "\n",
+    "# np.str_('BrC1C(Br)C(Br)C(Br)C(Br)C1Br'): array([0., 0., 0., ..., 0., 0., 0.], dtype=float32),"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 保存为 pickle 文件\n",
+    "with open('./LINCS2020/ECFP4_emb2048.pickle', 'wb') as f:\n",
+    "    pickle.dump(ecfp4_features_dict, f)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 50,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(2048,)"
+      ]
+     },
+     "execution_count": 50,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "ecfp4_features_dict['CN1CCN(CCCN2c3ccccc3Sc3ccc(cc23)C(F)(F)F)CC1'].shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# import pickle, h5py\n",
+    "# import numpy as np\n",
+    "# import pandas as pd\n",
+    "\n",
+    "# with open('./embeddings/ChemBERTa2_emb384.pickle', 'rb') as f:\n",
+    "#     KPGT_emb2304 = pickle.load(f)\n",
+    "# KPGT_emb2304"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(300,)"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# KPGT_emb2304['BrC1C(Br)C(Br)C(Br)C(Br)C1Br'].shape"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate KGPT embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "with open('./embeddings/KPGT_emb2304.pickle', 'rb') as f:\n",
+    "    KPGT_emb2304 = pickle.load(f)\n",
+    "KPGT_emb2304"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 111,755,936\n",
+      "111.8 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/KPGT/base.pth', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate InfoAlign embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# pip install infoalign\n",
+    "\n",
+    "# infoalign_predict --input LINCS2020/LINCS2020_smiles.csv \\\n",
+    "#                   --output InfoAlign_output.npy \\\n",
+    "#                   --output-to-input-column  # Adds the representation as an additional column in the input CSV\n",
+    "\n",
+    "# 是否存在数据泄露，因为这个是基于表征学习的"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 13,808,553\n",
+      "13.8 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/1463163983.py:2: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/infoalign_model/pretrain.pt', map_location='cpu')\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/infoalign_model/pretrain.pt', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')      # → 21.2 M"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import csv\n",
+    "import numpy as np\n",
+    "import pickle\n",
+    "\n",
+    "# CSV文件路径\n",
+    "csv_path = 'infoalign_model/LINCS2020_smiles.csv'\n",
+    "# NPY文件路径\n",
+    "npy_path = 'infoalign_model/InfoAlign_output.npy'\n",
+    "\n",
+    "# 初始化一个空字典来存储SMILES和对应的表征\n",
+    "smiles_to_representation = {}\n",
+    "\n",
+    "# 读取NPY文件\n",
+    "info_align_output = np.load(npy_path)\n",
+    "\n",
+    "# 读取CSV文件并同时替换字典中的值\n",
+    "with open(csv_path, mode='r', newline='', encoding='utf-8') as csvfile:\n",
+    "    reader = csv.DictReader(csvfile)\n",
+    "    for idx, row in enumerate(reader):\n",
+    "        smiles = row['SMILES']\n",
+    "        # 直接从info_align_output中获取对应的数值\n",
+    "        representation = info_align_output[idx]\n",
+    "        smiles_to_representation[smiles] = representation\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "8316"
+      ]
+     },
+     "execution_count": 34,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(smiles_to_representation)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 embeddings/InfoAlign_emb300.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/InfoAlign_emb300.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(smiles_to_representation, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate ChemBERTa-2 embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
+     ]
+    }
+   ],
+   "source": [
+    "from transformers import AutoModel, AutoTokenizer\n",
+    "model = AutoModel.from_pretrained(\"DeepChem/ChemBERTa-77M-MLM\")   # 77M SMILES（完整PubChem数据集)\n",
+    "tokenizer = AutoTokenizer.from_pretrained(\"DeepChem/ChemBERTa-77M-MLM\")\n",
+    "inputs = tokenizer(\"CCO\", return_tensors=\"pt\")\n",
+    "outputs = model(**inputs)  # 嵌入位于 outputs.last_hidden_state"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "torch.Size([1, 5, 384])"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "outputs.last_hidden_state.shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "ChemBERTa-77M 可训练参数量: 3,427,440\n",
+      "3.4 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "# from transformers import AutoModel\n",
+    "\n",
+    "# model = AutoModel.from_pretrained(\"DeepChem/ChemBERTa-77M-MLM\")\n",
+    "\n",
+    "# trainable = sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
+    "# print(f\"ChemBERTa-77M 可训练参数量: {trainable:,}\")\n",
+    "# print(f\"{trainable/1e6:.1f} M\")   # → 85.7 M"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# from transformers import AutoModel, AutoTokenizer\n",
+    "# import torch, os, shutil\n",
+    "\n",
+    "# repo_id = \"DeepChem/ChemBERTa-77M-MLM\"\n",
+    "\n",
+    "# # 1. 磁盘大小：下载后统计整个 snapshot 目录\n",
+    "# tmp_dir = \"./tmp_snapshot\"          # 临时目录，用完即删\n",
+    "# model = AutoModel.from_pretrained(repo_id, cache_dir=tmp_dir)\n",
+    "# tokenizer = AutoTokenizer.from_pretrained(repo_id, cache_dir=tmp_dir)\n",
+    "\n",
+    "# disk_mb = sum(\n",
+    "#     os.path.getsize(os.path.join(root, f))\n",
+    "#     for root, _, files in os.walk(tmp_dir)\n",
+    "#     for f in files\n",
+    "# ) / 1024 ** 2\n",
+    "\n",
+    "# print(f\"权重文件磁盘占用：{disk_mb:.1f} MB\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Some weights of RobertaModel were not initialized from the model checkpoint at DeepChem/ChemBERTa-77M-MLM and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']\n",
+      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n",
+      "100%|██████████| 260/260 [00:01<00:00, 168.25it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "总分子数 : 8316\n",
+      "总耗时   : 2.82 s\n",
+      "平均推理 : 0.34 ms / molecule\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "import time\n",
+    "from transformers import AutoModel, AutoTokenizer\n",
+    "import torch\n",
+    "import numpy as np\n",
+    "from rdkit import Chem\n",
+    "from tqdm import tqdm  # 可选：用于进度条\n",
+    "\n",
+    "def smiles_to_chemberta2(smiles_list, model_name=\"DeepChem/ChemBERTa-77M-MLM\", batch_size=32):\n",
+    "    \"\"\"\n",
+    "    将SMILES列表转换为ChemBERTa-2的嵌入向量（CLS token的隐藏状态）。\n",
+    "    \n",
+    "    参数:\n",
+    "        smiles_list (list): SMILES字符串列表\n",
+    "        model_name (str): Hugging Face模型ID，默认为ChemBERTa-2的MLM版本\n",
+    "        batch_size (int): 批处理大小，根据GPU内存调整\n",
+    "        \n",
+    "    返回:\n",
+    "        dict: 键为SMILES，值为768维嵌入向量（numpy数组）\n",
+    "    \"\"\"\n",
+    "    # 加载模型和tokenizer\n",
+    "    tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
+    "    model = AutoModel.from_pretrained(model_name)\n",
+    "    model.eval()  # 切换到推理模式\n",
+    "    \n",
+    "    # 检查CUDA可用性\n",
+    "    device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
+    "    model.to(device)\n",
+    "    \n",
+    "    chemberta2_dict = {}\n",
+    "    \n",
+    "    # 分批处理SMILES\n",
+    "    for i in tqdm(range(0, len(smiles_list), batch_size)):\n",
+    "        batch_smiles = smiles_list[i:i+batch_size]\n",
+    "        valid_indices = []\n",
+    "        batch_inputs = []\n",
+    "        \n",
+    "        # 过滤无效SMILES并生成输入\n",
+    "        for idx, smiles in enumerate(batch_smiles):\n",
+    "            mol = Chem.MolFromSmiles(smiles)\n",
+    "            if mol is not None:\n",
+    "                valid_indices.append(idx)\n",
+    "                batch_inputs.append(smiles)\n",
+    "            else:\n",
+    "                print(f\"Invalid SMILES: {smiles}\")\n",
+    "        \n",
+    "        if not batch_inputs:\n",
+    "            continue\n",
+    "            \n",
+    "        # Tokenize并移动到设备\n",
+    "        inputs = tokenizer(\n",
+    "            batch_inputs, \n",
+    "            return_tensors=\"pt\", \n",
+    "            padding=True, \n",
+    "            truncation=True, \n",
+    "            max_length=512\n",
+    "        ).to(device)\n",
+    "        \n",
+    "        # 生成嵌入\n",
+    "        with torch.no_grad():\n",
+    "            outputs = model(**inputs)\n",
+    "            embeddings = outputs.last_hidden_state[:, 0, :]  # 取CLS token的嵌入\n",
+    "        \n",
+    "        # 将有效结果存入字典\n",
+    "        for j, idx in enumerate(valid_indices):\n",
+    "            chemberta2_dict[batch_smiles[idx]] = embeddings[j].cpu().numpy()\n",
+    "        \n",
+    "        # 处理无效SMILES（填充零向量）\n",
+    "        for idx, smiles in enumerate(batch_smiles):\n",
+    "            if idx not in valid_indices:\n",
+    "                chemberta2_dict[smiles] = np.zeros(768, dtype=np.float32)\n",
+    "    \n",
+    "    return chemberta2_dict\n",
+    "\n",
+    "# 使用示例\n",
+    "import time\n",
+    "# ----- 计时开始 -----\n",
+    "t0 = time.perf_counter()\n",
+    "smiles_list = list(idx2smi.values())  # 假设idx2smi是SMILES字典\n",
+    "chemberta2_features_dict = smiles_to_chemberta2(smiles_list)\n",
+    "\n",
+    "t1 = time.perf_counter()\n",
+    "# ----- 计时结束 -----\n",
+    "\n",
+    "# 基本统计\n",
+    "total_mols = len(chemberta2_features_dict)\n",
+    "avg_time_ms = (t1 - t0) / total_mols * 1000\n",
+    "\n",
+    "print(f\"总分子数 : {total_mols}\")\n",
+    "print(f\"总耗时   : {t1 - t0:.2f} s\")\n",
+    "print(f\"平均推理 : {avg_time_ms:.2f} ms / molecule\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 384\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(chemberta2_features_dict), len(chemberta2_features_dict['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "chemberta2_features_dict['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 embeddings/ChemBERTa2_emb384.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/ChemBERTa2_emb384.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(chemberta2_features_dict, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate MolT5 embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "torch.Size([1, 768])\n"
+     ]
+    }
+   ],
+   "source": [
+    "from transformers import T5ForConditionalGeneration, AutoTokenizer\n",
+    "model = T5ForConditionalGeneration.from_pretrained(\"laituan245/molt5-base\")\n",
+    "tokenizer = AutoTokenizer.from_pretrained(\"laituan245/molt5-base\")\n",
+    "inputs = tokenizer(\"CCO\", return_tensors=\"pt\")\n",
+    "embeddings = model.encoder(**inputs).last_hidden_state.mean(dim=1)  # 平均池化\n",
+    "print(embeddings.shape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "MolT5-base 可训练参数量: 247.6 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "trainable = sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
+    "print(f'MolT5-base 可训练参数量: {trainable/1e6:.1f} M')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "已缓存权重磁盘占用：3784.1 MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "# molt5_disk_usage.py\n",
+    "from transformers import AutoModel, AutoTokenizer\n",
+    "import os, pathlib\n",
+    "\n",
+    "repo_id = \"laituan245/molt5-base\"\n",
+    "\n",
+    "# 让 transformers 只走本地缓存，不会触发下载\n",
+    "model = AutoModel.from_pretrained(repo_id, local_files_only=True)\n",
+    "tokenizer = AutoTokenizer.from_pretrained(repo_id, local_files_only=True)\n",
+    "\n",
+    "# 找到缓存目录\n",
+    "cache_root = pathlib.Path.home() / \".cache/huggingface/hub\"\n",
+    "# 仓库目录名格式：models--{user}--{repo}\n",
+    "repo_dir = cache_root / f\"models--{repo_id.replace('/', '--')}\"\n",
+    "\n",
+    "disk_bytes = sum(f.stat().st_size for f in repo_dir.rglob(\"*\") if f.is_file())\n",
+    "disk_mb = disk_bytes / 1024 ** 2\n",
+    "\n",
+    "print(f\"已缓存权重磁盘占用：{disk_mb:.1f} MB\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|██████████| 130/130 [00:09<00:00, 13.90it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "总分子数 : 8316\n",
+      "总耗时   : 11.55 s\n",
+      "平均推理 : 1.39 ms / molecule\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "import pickle\n",
+    "from transformers import T5ForConditionalGeneration, AutoTokenizer\n",
+    "import torch\n",
+    "import numpy as np\n",
+    "from rdkit import Chem\n",
+    "from tqdm import tqdm\n",
+    "\n",
+    "def smiles_to_molt5(smiles_list, model_name=\"laituan245/molt5-base\", batch_size=32, pooling=\"mean\"):\n",
+    "    \"\"\"\n",
+    "    使用 MolT5 生成 SMILES 的嵌入向量（基于 encoder 的隐藏状态）\n",
+    "    \n",
+    "    参数:\n",
+    "        smiles_list (list): SMILES 字符串列表\n",
+    "        model_name (str): Hugging Face 模型 ID，默认为 MolT5-base\n",
+    "        batch_size (int): 批处理大小，根据 GPU 内存调整\n",
+    "        pooling (str): 池化方法，可选 \"mean\"（平均池化）或 \"cls\"（取 CLS token）\n",
+    "        \n",
+    "    返回:\n",
+    "        dict: 键为 SMILES，值为嵌入向量（numpy 数组）\n",
+    "    \"\"\"\n",
+    "    # 加载 MolT5 的 tokenizer 和模型\n",
+    "    model = T5ForConditionalGeneration.from_pretrained(model_name)\n",
+    "    tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
+    "    model.eval()\n",
+    "    \n",
+    "    # 设备设置\n",
+    "    device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
+    "    model.to(device)\n",
+    "    \n",
+    "    molt5_dict = {}\n",
+    "    \n",
+    "    # 分批处理\n",
+    "    for i in tqdm(range(0, len(smiles_list), batch_size)):\n",
+    "        batch_smiles = smiles_list[i:i+batch_size]\n",
+    "        valid_indices = []\n",
+    "        batch_inputs = []\n",
+    "        \n",
+    "        # 过滤无效 SMILES\n",
+    "        for idx, smiles in enumerate(batch_smiles):\n",
+    "            mol = Chem.MolFromSmiles(smiles)\n",
+    "            if mol is not None:\n",
+    "                valid_indices.append(idx)\n",
+    "                batch_inputs.append(smiles)\n",
+    "            else:\n",
+    "                print(f\"Invalid SMILES: {smiles}\")\n",
+    "                molt5_dict[smiles] = np.zeros(model.config.d_model, dtype=np.float32)  # 填充零向量\n",
+    "        \n",
+    "        if not batch_inputs:\n",
+    "            continue\n",
+    "            \n",
+    "        # Tokenize 并生成嵌入\n",
+    "        inputs = tokenizer(\n",
+    "            batch_inputs,\n",
+    "            return_tensors=\"pt\",\n",
+    "            padding=True,\n",
+    "            truncation=True,\n",
+    "            max_length=512\n",
+    "        ).to(device)\n",
+    "        \n",
+    "        with torch.no_grad():\n",
+    "            # 获取 encoder 的隐藏状态（形状: [batch_size, seq_len, hidden_dim]）\n",
+    "            encoder_outputs = model.encoder(**inputs)\n",
+    "            hidden_states = encoder_outputs.last_hidden_state\n",
+    "            \n",
+    "            # 池化策略\n",
+    "            if pooling == \"mean\":\n",
+    "                embeddings = hidden_states.mean(dim=1)  # 平均池化\n",
+    "            elif pooling == \"cls\":\n",
+    "                embeddings = hidden_states[:, 0, :]     # CLS token\n",
+    "            else:\n",
+    "                raise ValueError(\"pooling 必须是 'mean' 或 'cls'\")\n",
+    "        \n",
+    "        # 存储结果\n",
+    "        for j, idx in enumerate(valid_indices):\n",
+    "            molt5_dict[batch_smiles[idx]] = embeddings[j].cpu().numpy()\n",
+    "    \n",
+    "    return molt5_dict\n",
+    "\n",
+    "\n",
+    "import time\n",
+    "# ----- 计时开始 -----\n",
+    "smiles_list = list(idx2smi.values())  # 假设 idx2smi 是 SMILES 字典\n",
+    "\n",
+    "t0 = time.perf_counter()\n",
+    "molt5_embeddings = smiles_to_molt5(smiles_list, batch_size=64, pooling=\"mean\")\n",
+    "\n",
+    "t1 = time.perf_counter()\n",
+    "# ----- 计时结束 -----\n",
+    "\n",
+    "# 基本统计\n",
+    "total_mols = len(molt5_embeddings)\n",
+    "avg_time_ms = (t1 - t0) / total_mols * 1000\n",
+    "\n",
+    "print(f\"总分子数 : {total_mols}\")\n",
+    "print(f\"总耗时   : {t1 - t0:.2f} s\")\n",
+    "print(f\"平均推理 : {avg_time_ms:.2f} ms / molecule\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 768\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(molt5_embeddings), len(molt5_embeddings['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "molt5_embeddings['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "字典已成功保存到 embeddings/molt5_emb768.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/molt5_emb768.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(molt5_embeddings, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate Chemprop embedding for all smiles"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# git clone https://github.com/chemprop/chemprop.git\n",
+    "# https://chemprop.readthedocs.io/en/main/installation.html\n",
+    "# https://chemprop.readthedocs.io/en/main/tutorial/cli/fingerprint.html#fingerprint \n",
+    "\n",
+    "# chemprop fingerprint --test-path ./embeddings/LINCS2020_smiles.csv --model-path chemprop/tests/data/example_model_v2_regression_mol.ckpt --output embeddings/fps.csv --ffn-block-index -2"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 319,505\n",
+      "0.3 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/3140832406.py:3: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  ckpt = torch.load('/home/bob/boom/VCBench/Molecule_encoder/chemprop/tests/data/example_model_v2_regression_mol.ckpt',\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "\n",
+    "ckpt = torch.load('/home/bob/boom/VCBench/Molecule_encoder/chemprop/tests/data/example_model_v2_regression_mol.ckpt',\n",
+    "                  map_location='cpu')\n",
+    "\n",
+    "# 取出模型参数部分\n",
+    "model_state = ckpt['state_dict']\n",
+    "\n",
+    "total = sum(v.numel() for v in model_state.values() if isinstance(v, torch.Tensor))\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "8316"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import csv\n",
+    "import numpy as np\n",
+    "\n",
+    "# CSV文件路径\n",
+    "csv1_path = 'infoalign_model/LINCS2020_smiles.csv'\n",
+    "csv2_path = 'embeddings/fps_0.csv'\n",
+    "\n",
+    "# 初始化一个空字典来存储SMILES和对应的表征\n",
+    "smiles_to_representation = {}\n",
+    "\n",
+    "# 读取第一个CSV文件（SMILES信息）\n",
+    "smiles_list = []\n",
+    "with open(csv1_path, mode='r', newline='', encoding='utf-8') as csvfile:\n",
+    "    reader = csv.DictReader(csvfile)\n",
+    "    for row in reader:\n",
+    "        smiles = row['SMILES']  # 假设CSV文件中列名为'SMILES'\n",
+    "        smiles_list.append(smiles)\n",
+    "\n",
+    "# 读取第二个CSV文件（表征信息），跳过第一行\n",
+    "representation_list = []\n",
+    "with open(csv2_path, mode='r', newline='', encoding='utf-8') as csvfile:\n",
+    "    reader = csv.reader(csvfile)\n",
+    "    next(reader)  # 跳过第一行\n",
+    "    for row in reader:\n",
+    "        # 将字符串转换为浮点数\n",
+    "        representation = [np.float32(x)  for x in row]  # 假设表征信息是整行数据\n",
+    "        representation_list.append(representation)\n",
+    "\n",
+    "# 将列表转换为numpy数组\n",
+    "representation_array = np.array(representation_list)\n",
+    "\n",
+    "# 将SMILES和对应的表征合并到字典中\n",
+    "for smiles, representation in zip(smiles_list, representation_array):\n",
+    "    smiles_to_representation[smiles] = representation\n",
+    "\n",
+    "len(smiles_to_representation)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "300 <class 'numpy.ndarray'>\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "dtype('float32')"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "print(len(smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']), type(smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))\n",
+    "\n",
+    "smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br'].dtype"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "300\n",
+      "字典已成功保存到 embeddings/Chemprop_emb300.pickle\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(smiles_to_representation['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))\n",
+    "\n",
+    "import pickle\n",
+    "# 保存字典为pickle文件\n",
+    "pickle_path = 'embeddings/Chemprop_emb300.pickle'\n",
+    "with open(pickle_path, 'wb') as pickle_file:\n",
+    "    pickle.dump(smiles_to_representation, pickle_file)\n",
+    "\n",
+    "print(f\"字典已成功保存到 {pickle_path}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate MolCLR embedding for all smiles\n",
+    "\n",
+    "## see /home/bob/boom/DrugIM/TranSiGen/data/MolCLR/Get_emd.ipynb\n",
+    "\n",
+    "### We use GIN model, which has a better performance in their paper"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 2,407,201\n",
+      "2.4 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/3871607093.py:2: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/MolCLR/ckpt/pretrained_gin/checkpoints/model.pth', map_location='cpu')\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/MolCLR/ckpt/pretrained_gin/checkpoints/model.pth', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')      # → 21.2 M"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Generate Mole-BERT embedding for all smiles\n",
+    "\n",
+    "## see /home/bob/boom/DrugIM/TranSiGen/data/Mole-BERT/Get_emd.ipynb\n",
+    "\n",
+    "### We use GNN_graphpred model, which has a better performance in their paper"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "预训练权重总参数量: 1,860,905\n",
+      "1.9 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/tmp/ipykernel_1948954/2103728332.py:2: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
+      "  state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/Mole-BERT/model_gin/Mole-BERT.pth', map_location='cpu')\n"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "state_dict = torch.load('/home/bob/boom/VCBench/Molecule_encoder/Mole-BERT/model_gin/Mole-BERT.pth', map_location='cpu')\n",
+    "total = sum(v.numel() for v in state_dict.values())\n",
+    "print(f'预训练权重总参数量: {total:,}')\n",
+    "print(f'{total/1e6:.1f} M')      # → 21.2 M"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 3D  UniMol"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "8316"
+      ]
+     },
+     "execution_count": 1,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import pickle, h5py\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "with open('./LINCS2020/idx2smi.pickle', 'rb') as f:\n",
+    "    idx2smi = pickle.load(f)\\\n",
+    "\n",
+    "unique_smiles = list(idx2smi.values())  # 假设idx2smi是一个字典，其值为SMILES字符串\n",
+    "len(unique_smiles)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2025-09-15 11:17:31 | unimol_tools/models/unimolv2.py | 161 | INFO | Uni-Mol Tools | Loading pretrained weights from /home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/unimol_tools/weights/modelzoo/310M/checkpoint.pt\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "开始生成32个SMILES的嵌入表示...\n",
+      "开始\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "处理批次:   0%|          | 0/1 [00:00<?, ?it/s]2025-09-15 11:17:32 | unimol_tools/data/conformer.py | 437 | INFO | Uni-Mol Tools | Start generating conformers...\n",
+      "32it [00:00, 82.76it/s]\n",
+      "2025-09-15 11:17:33 | unimol_tools/data/conformer.py | 452 | INFO | Uni-Mol Tools | Succeeded in generating conformers for 100.00% of molecules.\n",
+      "2025-09-15 11:17:33 | unimol_tools/data/conformer.py | 469 | INFO | Uni-Mol Tools | Succeeded in generating 3d conformers for 100.00% of molecules.\n",
+      "2025-09-15 11:17:33 | unimol_tools/tasks/trainer.py | 78 | INFO | Uni-Mol Tools | Number of GPUs available: 1\n",
+      "2025-09-15 11:17:33 | unimol_tools/tasks/trainer.py | 98 | INFO | Uni-Mol Tools | Using single GPU.\n",
+      "100%|██████████| 1/1 [00:00<00:00,  2.32it/s]\n",
+      "处理批次: 100%|██████████| 1/1 [00:01<00:00,  1.25s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "嵌入生成完成！共处理 32 个SMILES，32 个成功，0 个失败。\n",
+      "嵌入结果已保存至 embeddings/UniMolV2_emb1024.pkl\n",
+      "总分子数 : 32\n",
+      "总耗时   : 3.50 s\n",
+      "平均推理 : 109.37 ms / molecule\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import pickle\n",
+    "import numpy as np\n",
+    "from tqdm import tqdm\n",
+    "from unimol_tools import UniMolRepr\n",
+    "import os, sys, contextlib\n",
+    "\n",
+    "@contextlib.contextmanager\n",
+    "def suppress_stdout_stderr():\n",
+    "    \"\"\"with 块内所有 print / tqdm / warning 都不会显示\"\"\"\n",
+    "    with open(os.devnull, 'w') as devnull:\n",
+    "        old_out, old_err = sys.stdout, sys.stderr\n",
+    "        sys.stdout, sys.stderr = devnull, devnull\n",
+    "        try:\n",
+    "            yield\n",
+    "        finally:\n",
+    "            sys.stdout, sys.stderr = old_out, old_err\n",
+    "\n",
+    "def get_unimol_embeddings(smiles_list, output_file=\"UniMol_emb512.pkl\", \n",
+    "                         model_name='unimolv1', model_size='84m', \n",
+    "                         remove_hs=False, batch_size=32):\n",
+    "    \"\"\"\n",
+    "    使用Uni-Mol模型为SMILES列表生成分子嵌入，并保存为pickle文件\n",
+    "    \n",
+    "    参数:\n",
+    "    smiles_list (list): SMILES字符串列表\n",
+    "    output_file (str): 输出pickle文件路径\n",
+    "    model_name (str): 模型名称，可选'unimolv1'或'unimolv2'\n",
+    "    model_size (str): 模型大小，仅在使用unimolv2时有效\n",
+    "    remove_hs (bool): 是否移除氢原子\n",
+    "    batch_size (int): 批处理大小\n",
+    "    \n",
+    "    返回:\n",
+    "    dict: 包含SMILES及其对应嵌入的字典\n",
+    "    \"\"\"\n",
+    "    # 初始化模型\n",
+    "    clf = UniMolRepr(\n",
+    "        data_type='molecule',\n",
+    "        remove_hs=remove_hs,\n",
+    "        model_name=model_name,\n",
+    "        model_size=model_size\n",
+    "    )\n",
+    "    \n",
+    "    # 用于存储结果的字典\n",
+    "    embeddings_dict = {}\n",
+    "    error_smiles = []\n",
+    "    \n",
+    "    # 批处理SMILES\n",
+    "    total_batches = (len(smiles_list) + batch_size - 1) // batch_size\n",
+    "    \n",
+    "    print(f\"开始生成{len(smiles_list)}个SMILES的嵌入表示...\")\n",
+    "    \n",
+    "    print(\"开始\")\n",
+    "    # with suppress_stdout_stderr():\n",
+    "    for i in tqdm(range(total_batches), desc=\"处理批次\"):\n",
+    "        batch = smiles_list[i*batch_size : (i+1)*batch_size]\n",
+    "        \n",
+    "        try:\n",
+    "            # 获取嵌入表示\n",
+    "            batch_repr = clf.get_repr(batch, return_atomic_reprs=False)\n",
+    "            \n",
+    "            # 将结果存入字典 (使用CLS token作为分子表示)\n",
+    "            for idx, smiles in enumerate(batch):\n",
+    "                embeddings_dict[smiles] = batch_repr['cls_repr'][idx]\n",
+    "                \n",
+    "        except Exception as e:\n",
+    "            print(f\"处理批次 {i+1}/{total_batches} 时发生错误: {str(e)}\")\n",
+    "            # 记录处理失败的SMILES\n",
+    "            error_smiles.extend(batch)\n",
+    "        \n",
+    "    # 保存嵌入结果\n",
+    "    with open(output_file, 'wb') as f:\n",
+    "        pickle.dump(embeddings_dict, f)\n",
+    "    \n",
+    "    print(f\"嵌入生成完成！共处理 {len(smiles_list)} 个SMILES，\"\n",
+    "          f\"{len(smiles_list) - len(error_smiles)} 个成功，\"\n",
+    "          f\"{len(error_smiles)} 个失败。\")\n",
+    "    print(f\"嵌入结果已保存至 {output_file}\")\n",
+    "    \n",
+    "    if error_smiles:\n",
+    "        print(f\"处理失败的SMILES已记录。\")\n",
+    "    \n",
+    "    return embeddings_dict\n",
+    "\n",
+    "# 使用示例\n",
+    "if __name__ == \"__main__\":\n",
+    "    \n",
+    "    # 假设unique_smiles是你的SMILES列表\n",
+    "    unique_smiles = [\n",
+    "    \"CC(=O)OC1=CC=CC=C1C(=O)O\",  # Aspirin\n",
+    "    \"CN1C=NC2=C1C(=O)N(C(=O)N2C)C\" # Caffeine\n",
+    "    ]  # 示例SMILES列表\n",
+    "    unique_smiles = pd.read_csv('./LINCS2020/LINCS2020_smiles.csv')['SMILES'].tolist()\n",
+    "\n",
+    "    # UniMol V1\n",
+    "    # embeddings = get_unimol_embeddings(\n",
+    "    #     smiles_list=unique_smiles,\n",
+    "    #     output_file=\"embeddings/UniMol_emb512.pkl\",\n",
+    "    #     model_name='unimolv1',\n",
+    "    #     model_size='84m',\n",
+    "    #     remove_hs=False,\n",
+    "    #     batch_size=32\n",
+    "    # )\n",
+    "    \n",
+    "    # UniMol V2\n",
+    "    embeddings = get_unimol_embeddings(\n",
+    "    smiles_list=unique_smiles,\n",
+    "    output_file=\"embeddings/UniMolV2_emb1024.pkl\", # save path\n",
+    "    model_name='unimolv2',      # 修改为 v2 版本\n",
+    "    model_size='310m',          # 指定 v2 模型大小（根据需要选择）\n",
+    "    remove_hs=False,\n",
+    "    batch_size=32\n",
+    "    )\n",
+    "\n",
+    "    # 打印样例嵌入\n",
+    "    sample_smiles = unique_smiles[0]\n",
+    "    print(f\"SMILES: {sample_smiles}\")\n",
+    "    print(f\"嵌入向量: {embeddings[sample_smiles]}\")\n",
+    "    print(f\"嵌入维度: {len(embeddings[sample_smiles])}\")\n",
+    "    "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2025-10-16 21:29:18 | unimol_tools/models/unimol.py | 136 | INFO | Uni-Mol Tools | Loading pretrained weights from /home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/unimol_tools/weights/mol_pre_all_h_220816.pt\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "UniMol-unimolv1-84m 可训练参数量: 47.3 M\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2025-10-16 21:29:20 | unimol_tools/models/unimolv2.py | 161 | INFO | Uni-Mol Tools | Loading pretrained weights from /home/bob/anaconda3/envs/boom/lib/python3.11/site-packages/unimol_tools/weights/modelzoo/310M/checkpoint.pt\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "UniMol-unimolv2-310m 可训练参数量: 308.9 M\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import pickle\n",
+    "import numpy as np\n",
+    "from tqdm import tqdm\n",
+    "from unimol_tools import UniMolRepr\n",
+    "import os, sys, contextlib\n",
+    "\n",
+    "def get_unimol_embeddings(smiles_list, output_file=\"UniMol_emb512.pkl\", \n",
+    "                         model_name='unimolv1', model_size='84m', \n",
+    "                         remove_hs=False, batch_size=32):\n",
+    "    \"\"\"\n",
+    "    使用Uni-Mol模型为SMILES列表生成分子嵌入，并保存为pickle文件\n",
+    "    \n",
+    "    参数:\n",
+    "    smiles_list (list): SMILES字符串列表\n",
+    "    output_file (str): 输出pickle文件路径\n",
+    "    model_name (str): 模型名称，可选'unimolv1'或'unimolv2'\n",
+    "    model_size (str): 模型大小，仅在使用unimolv2时有效\n",
+    "    remove_hs (bool): 是否移除氢原子\n",
+    "    batch_size (int): 批处理大小\n",
+    "    \n",
+    "    返回:\n",
+    "    dict: 包含SMILES及其对应嵌入的字典\n",
+    "    \"\"\"\n",
+    "    # 初始化模型\n",
+    "    # ===== 在 clf 初始化完成后立即统计 =====\n",
+    "    clf = UniMolRepr(\n",
+    "            data_type='molecule',\n",
+    "            remove_hs=remove_hs,\n",
+    "            model_name=model_name,\n",
+    "            model_size=model_size)\n",
+    "\n",
+    "    # 取出真正的网络\n",
+    "    real_model = clf.model          # <== 关键属性\n",
+    "\n",
+    "    total = sum(p.numel() for p in real_model.parameters() if p.requires_grad)\n",
+    "    print(f'UniMol-{model_name}-{model_size} 可训练参数量: {total/1e6:.1f} M')\n",
+    "  \n",
+    "        \n",
+    "# 使用示例\n",
+    "if __name__ == \"__main__\":\n",
+    "    # 假设unique_smiles是你的SMILES列表\n",
+    "    # 生成并保存嵌入\n",
+    "    import time\n",
+    "    # ----- 计时开始 -----\n",
+    "    unique_smiles = list(idx2smi.values())  # 假设 idx2smi 是 SMILES 字典\n",
+    "\n",
+    "    t0 = time.perf_counter()\n",
+    "    \n",
+    "    embeddings = get_unimol_embeddings(\n",
+    "        smiles_list=unique_smiles[:32],\n",
+    "        output_file=\"embeddings/UniMol_emb512.pkl\",\n",
+    "        model_name='unimolv1',\n",
+    "        model_size='84m',\n",
+    "        remove_hs=False,\n",
+    "        batch_size=32\n",
+    "    )\n",
+    "    # UniMol 2\n",
+    "    embeddings = get_unimol_embeddings(\n",
+    "    smiles_list=unique_smiles[:32],\n",
+    "    output_file=\"embeddings/UniMolV2_emb1024.pkl\",\n",
+    "    model_name='unimolv2',      # 修改为 v2 版本\n",
+    "    model_size='310m',          # 指定 v2 模型大小（根据需要选择）\n",
+    "    remove_hs=False,\n",
+    "    batch_size=32\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# output_file=\"embeddings/UniMolV2_emb1024.pkl\"\n",
+    "# with open(output_file, 'wb') as f:\n",
+    "#     pickle.dump(embeddings, f)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "path = './embeddings/UniMolV2_emb1024.pkl'\n",
+    "import pickle\n",
+    "with open(path, 'rb') as f:\n",
+    "    unimol_feat = pickle.load(f)\n",
+    "\n",
+    "unimol_feat"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "8316 512\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(len(unimol_feat), len(unimol_feat['BrC1C(Br)C(Br)C(Br)C(Br)C1Br']))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "处理前维度: (1, 512)\n",
+      "处理后维度: (512,)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# import pickle\n",
+    "# import numpy as np\n",
+    "\n",
+    "# # 加载原始特征\n",
+    "# path = './embeddings/UniMol_emb512.pkl'\n",
+    "# with open(path, 'rb') as f:\n",
+    "#     unimol_feat = pickle.load(f)\n",
+    "\n",
+    "# # 处理特征维度：从(1,512)变为(512,)\n",
+    "# unimol_feat_processed = {k: v.squeeze() for k, v in unimol_feat.items()}\n",
+    "\n",
+    "# # 验证处理结果\n",
+    "# test_molecule = 'BrC1C(Br)C(Br)C(Br)C(Br)C1Br'\n",
+    "# if test_molecule in unimol_feat_processed:\n",
+    "#     print(f\"处理前维度: {unimol_feat[test_molecule].shape}\")\n",
+    "#     print(f\"处理后维度: {unimol_feat_processed[test_molecule].shape}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# # 保存处理后的特征\n",
+    "# # 选项1：覆盖原始文件\n",
+    "# with open(path, 'wb') as f:\n",
+    "#     pickle.dump(unimol_feat_processed, f)\n",
+    "\n",
+    "# 选项2：保存到新文件（推荐，避免覆盖原始数据）\n",
+    "# processed_path = './embeddings/UniMol_emb512_processed.pkl'\n",
+    "# with open(processed_path, 'wb') as f:\n",
+    "#     pickle.dump(unimol_feat_processed, f)\n",
+    "\n",
+    "# print(f\"处理后的特征已保存至: {processed_path}\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "boom",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

Tahoe-100M/lincs_in_scgpt_indices.csv

@@ -0,0 +1,966 @@
+scgpt_index
+106
+122
+135
+138
+172
+199
+200
+205
+225
+238
+240
+292
+313
+331
+346
+360
+377
+385
+387
+398
+403
+413
+425
+442
+458
+494
+499
+549
+553
+581
+629
+633
+637
+642
+650
+723
+728
+743
+769
+830
+832
+837
+839
+883
+889
+893
+903
+904
+939
+961
+964
+966
+975
+983
+996
+997
+1011
+1024
+1041
+1068
+1083
+1151
+1165
+1178
+1179
+1183
+1195
+1200
+1206
+1225
+1232
+1289
+1314
+1376
+1385
+1398
+1407
+1408
+1449
+1456
+1459
+1464
+1465
+1523
+1536
+1537
+1604
+1610
+1665
+1680
+1693
+1724
+1758
+1792
+1819
+1862
+1872
+1902
+1913
+1918
+1933
+2009
+2053
+2078
+2132
+2146
+2152
+2160
+2173
+2178
+2184
+2196
+2249
+2250
+2269
+2279
+2291
+2330
+2343
+2349
+2359
+2379
+2390
+2475
+2537
+2552
+2566
+2573
+2576
+2616
+2634
+2651
+2675
+2746
+2753
+2787
+2838
+2851
+2898
+2900
+2904
+2913
+2921
+2957
+2974
+2978
+3002
+3019
+3038
+3075
+3087
+3108
+3109
+3110
+3116
+3121
+3129
+3147
+3150
+3181
+3183
+3185
+3195
+3201
+3211
+3234
+3254
+3275
+3382
+3401
+3538
+3544
+3572
+3596
+3602
+3648
+3654
+3667
+3668
+3671
+3677
+3687
+3694
+3702
+3710
+3715
+3733
+3747
+3776
+3817
+3829
+3838
+3873
+3882
+3891
+3895
+3906
+3912
+3969
+4000
+4002
+4019
+4035
+4036
+4058
+4064
+4068
+4075
+4083
+4095
+4096
+4104
+4115
+4136
+4168
+4172
+4199
+4217
+4230
+4253
+4300
+4305
+4359
+4385
+4389
+4405
+4412
+4429
+4460
+4465
+4503
+4521
+4555
+4592
+4607
+4648
+4673
+4678
+4689
+4697
+4703
+4718
+4731
+4749
+4752
+4782
+4789
+4900
+4919
+4922
+4931
+4938
+4959
+4991
+4993
+5029
+5037
+5050
+5076
+5120
+5137
+5164
+5198
+5251
+5268
+5310
+5317
+5325
+5337
+5345
+5350
+5394
+5438
+5445
+5517
+5541
+5566
+5623
+5628
+5690
+5725
+5768
+5770
+5787
+5811
+5843
+5861
+5867
+5881
+5907
+5944
+5953
+5964
+5969
+6037
+6088
+6105
+6148
+6183
+6202
+6213
+6232
+6265
+6281
+6284
+6287
+6297
+6303
+6313
+6318
+6327
+6348
+6357
+6364
+6366
+6379
+6389
+6391
+6409
+6412
+6414
+6416
+6423
+6424
+6428
+6495
+6553
+6603
+6623
+6625
+6654
+6660
+6661
+6686
+6693
+6704
+6735
+6748
+6771
+6774
+6796
+6821
+6833
+6864
+6866
+6872
+6910
+6933
+6944
+6954
+6970
+7007
+7009
+7049
+7099
+7104
+7106
+7121
+7147
+7148
+7237
+7257
+7293
+7324
+7372
+7388
+7397
+7413
+7433
+7472
+7480
+7510
+7511
+7544
+7555
+7591
+7614
+7626
+7654
+7676
+7677
+7691
+7753
+7755
+7771
+7776
+7782
+7805
+7822
+7839
+7856
+7857
+7886
+7912
+7920
+7923
+7951
+7971
+8022
+8023
+8095
+8108
+8111
+8126
+8144
+8157
+8179
+8230
+8268
+8269
+8300
+8306
+8330
+8342
+8344
+8358
+8383
+8409
+8454
+8493
+8496
+8507
+8522
+8537
+8541
+8558
+8571
+8572
+8593
+8602
+8676
+8697
+8704
+8712
+8727
+8737
+8821
+8825
+8836
+8868
+8885
+8889
+8901
+8909
+8910
+8917
+8928
+8930
+8962
+8978
+8995
+8998
+9005
+9053
+9065
+9109
+9115
+9165
+9214
+9223
+9234
+9241
+9261
+9284
+9327
+9335
+9336
+9372
+9376
+9389
+9408
+9440
+9443
+9479
+9511
+9534
+9551
+9561
+9570
+9620
+9636
+9648
+9667
+9672
+9679
+9714
+9744
+9790
+9792
+9804
+9825
+9829
+9848
+9887
+9923
+9927
+9946
+9971
+9975
+10006
+10134
+10161
+10199
+10212
+10228
+10233
+10249
+10260
+10271
+10288
+10296
+10353
+10391
+10393
+10402
+10407
+10452
+10467
+10479
+10528
+10535
+10536
+10582
+10587
+10608
+10658
+10715
+10730
+10757
+10768
+10805
+10824
+10847
+10904
+10919
+10924
+10925
+10927
+10943
+10954
+10955
+10982
+11063
+11064
+11071
+11072
+11081
+11084
+11096
+11154
+11171
+11204
+11221
+11235
+11237
+11259
+11276
+11297
+11303
+11312
+11368
+11371
+11376
+11378
+11383
+11400
+11404
+11432
+11438
+11444
+11451
+11481
+11493
+11593
+11606
+11607
+11615
+11638
+11659
+11667
+11669
+11673
+11678
+11679
+11689
+11708
+11710
+11719
+11753
+11772
+11793
+11809
+11826
+11827
+11867
+11887
+11902
+11924
+11943
+11978
+11981
+12017
+12018
+12031
+12045
+12114
+12132
+12170
+12182
+12194
+12209
+12214
+12232
+12238
+12274
+12281
+12286
+12301
+12336
+12348
+12359
+12376
+12389
+12391
+12395
+12425
+12464
+12491
+12511
+12532
+12549
+12550
+12574
+12627
+12636
+12643
+12674
+12696
+12707
+12744
+12820
+12859
+12938
+12947
+12952
+12956
+12966
+12977
+12980
+12981
+12989
+13009
+13011
+13015
+13021
+13034
+13050
+13051
+13071
+13072
+13082
+13093
+13099
+13110
+13120
+13132
+13144
+13154
+13198
+13201
+13217
+13227
+13229
+13230
+13234
+13243
+13255
+13281
+13293
+13339
+13351
+13379
+13392
+13417
+13449
+13484
+13491
+13527
+13542
+13543
+13544
+13549
+13567
+13610
+13625
+13633
+13639
+13653
+13661
+13680
+13681
+13684
+13692
+13732
+13774
+13781
+13830
+13838
+13948
+13990
+14013
+14014
+14032
+14049
+14100
+14122
+14123
+14130
+14139
+14148
+14170
+14171
+14203
+14226
+14229
+14262
+14263
+14285
+14310
+14311
+14321
+14336
+14378
+14421
+14482
+14488
+14520
+14536
+14537
+14562
+14593
+14602
+14635
+14646
+14658
+14660
+14667
+14676
+14703
+14710
+14715
+14726
+14734
+14755
+14766
+14846
+14854
+14856
+14925
+14935
+14939
+14943
+14959
+14980
+14983
+15014
+15044
+15064
+15071
+15083
+15096
+15124
+15146
+15151
+15166
+15169
+15224
+15245
+15264
+15283
+15307
+15314
+15316
+15339
+15366
+15380
+15383
+15393
+15417
+15425
+15489
+15513
+15533
+15541
+15549
+15576
+15589
+15607
+15621
+15623
+15626
+15681
+15690
+15734
+15829
+15832
+15862
+15866
+15885
+15917
+15928
+15943
+15972
+15978
+16001
+16017
+16032
+16054
+16070
+16074
+16095
+16153
+16183
+16223
+16226
+16242
+16252
+16321
+16367
+16378
+16396
+16399
+16406
+16417
+16477
+16483
+16492
+16513
+16533
+16583
+16586
+16591
+16626
+16650
+16653
+16654
+16690
+16759
+16777
+16779
+16793
+16807
+16816
+16831
+16843
+16858
+16861
+16889
+16908
+16915
+16917
+16935
+16956
+16961
+16963
+16970
+16975
+16983
+17006
+17048
+17081
+17082
+17088
+17093
+17131
+17172
+17188
+17197
+17210
+17218
+17239
+17316
+17327
+17351
+17410
+17412
+17440
+17446
+17450
+17474
+17498
+17504
+17519
+17547
+17552
+17556
+17589
+17693
+17696
+17737
+17759
+17764
+17765
+17784
+17826
+17830
+17836
+17888
+17904
+17917
+17946
+17952
+17974
+17996
+18004
+18011
+18063
+18069
+18090
+18094
+18107
+18111
+18112
+18114
+18147
+18180
+18236
+18281
+18282
+18308
+18319
+18338
+18398
+18402
+18418
+18420
+18448
+18496
+18539
+18552
+18565
+18571
+18598
+18608
+18627
+18656
+18673
+18677
+18738
+18772
+18805
+18821
+18829
+18857
+18859
+18868
+18880
+18910
+18959
+18966
+18967
+18969
+18979
+18992
+19020

Tahoe-100M/unique_smiles.txt

@@ -0,0 +1,363 @@
+CC(C)CC(C(=O)O)NC(=O)C(C(CC1=CC=CC=C1)N)O.Cl
+C1=CC=C(C(=C1)C2=NC(=NO2)C3=CC(=CC=C3)C(=O)O)F
+C1CN=C(N1)NC2=C(C3=NC=CN=C3C=C2)Br
+C1CC1C(=O)NC2=NN3C(=N2)C=CC=C3C4=CC=C(C=C4)CN5CCS(=O)(=O)CC5
+CC(C)NCC(COC1=CC=C(C=C1)CCC(=O)OC)O.Cl
+CC1=CN=C(C(=C1OC)C)CS(=O)C2=NC3=C(N2)C=C(C=C3)OC
+CC(CS(=O)(=O)C1=CC=C(C=C1)F)(C(=O)NC2=CC(=C(C=C2)C#N)C(F)(F)F)O
+CC12CCC(=O)C=C1C3CC3C4C2CCC5(C4C6CC6C57CCC(=O)O7)C
+C1=CC=C(C=C1)NS(=O)(=O)C2=CC=CC(=C2)C=CC(=O)NO
+CC1=C(C=C(C=C1)C(=O)NC2=CC(=C(C=C2)CN3CCN(CC3)C)C(F)(F)F)C#CC4=CN=C5N4N=CC=C5
+COC1=CC2=C(C=CN=C2C=C1OC)OC3=CC=C(C=C3)NC(=O)C4(CC4)C(=O)NC5=CC=C(C=C5)F.C(C(C(=O)O)O)C(=O)O
+CN1C2CCC1CC(C2)OC(C3=CC=CC=C3)C4=CC=CC=C4.CS(=O)(=O)O
+CCC1=C(C=C2CC(CC2=C1)NCC(C3=C4C=CC(=O)NC4=C(C=C3)O)O)CC.C(=CC(=O)O)C(=O)O
+CC#CC(=O)N1CCC(C1)N2C3=NC=NC(=C3N(C2=O)C4=CC=C(C=C4)OC5=CC=CC=C5)N.Cl
+CN(C)C(=O)C(CCN1CCC(CC1)(C2=CC=C(C=C2)Cl)O)(C3=CC=CC=C3)C4=CC=CC=C4.Cl
+C1=CC(=CC=C1COC(CN2C=CN=C2)C3=C(C=C(C=C3)Cl)Cl)Cl
+CNCC(C1=CC(=CC=C1)O)O.Cl
+CN1CCN(CC1)C2=CC(=C(C=C2)C(=O)NC3=NNC4=C3C=C(C=C4)CC5=CC(=CC(=C5)F)F)NC6CCOCC6
+CC1=CC(=C(C=C1)SC2=CC=CC=C2N3CCNCC3)C
+CC1=CN=C(C(=C1OC)C)CS(=O)C2=NC3=C([N-]2)C=C(C=C3)OC.[Na+]
+CC(C)CN1C=NC2=C1C3=CC=CC=C3N=C2N.Cl
+CCCCCCCCC1=CC=C(C=C1)CCC(CO)(CO)N.Cl
+CC1C(C(CC(O1)OC2C(OC(CC2O)OC3C(OC(CC3O)OC4CCC5(C(C4)CCC6C5CCC7(C6(CCC7C8=CC(=O)OC8)O)C)C)C)C)O)O
+CC1=C(C2=CC=CC=C2C(=C1)OC(=O)C)OC(=O)C
+CC(C1=CC(=CC(=C1)C(F)(F)F)C(F)(F)F)OC2C(N(CCO2)CC3=NNC(=O)N3)C4=CC=C(C=C4)F
+C1=CC=C(C=C1)C2=C(OC(=N2)CCC(=O)O)C3=CC=CC=C3
+CC12CCC3C(C1CCC2=O)CCC4=C3C=CC(=C4)OS(=O)(=O)[O-].[K+]
+C(CC(C(F)F)(C(=O)O)N)CN.Cl
+CCP(CC)CC.CC(=O)OCC1C(C(C(C(O1)[S-])OC(=O)C)OC(=O)C)OC(=O)C.[Au+]
+C1=CC=C2C(=C1)C=CC3=CC=CC=C3N2C(=O)N
+CC(C)C(=O)NC1=CC(=C(C=C1)[N+](=O)[O-])C(F)(F)F
+CC(C1=CC2=CC=CC=C2S1)N(C(=O)N)O
+CCN(CC)CCNC(=O)C1=C(NC(=C1C)C=C2C3=C(C=CC(=C3)F)NC2=O)C
+CCC1C(C(C(C(=O)C(CC(C(C(C(C(C(=O)O1)C)OC2CC(C(C(O2)C)O)(C)OC)C)OC3C(C(CC(O3)C)N(C)C)O)(C)O)C)C)O)(C)O
+CC(=O)OC1CCC2(C3CCC4(C(C3CC=C2C1)CC=C4C5=CN=CC=C5)C)C
+C1CN(CC1O)C2=C(C=C(C=N2)C(=O)NC3=CC=C(C=C3)OC(F)(F)Cl)C4=CC=NN4
+COC(=O)C(C1=CC=CC=C1Cl)N2CCC3=C(C2)C=CS3
+COC(=O)CCC(=O)CN.Cl
+CC(CS)C(=O)N1CCCC1C(=O)O
+CC(C1=NC=NC=C1F)C(CN2C=NC=N2)(C3=C(C=C(C=C3)F)F)O
+CC1=C(C(=CC=C1)C)OCC(=O)NC(CC2=CC=CC=C2)C(CC(CC3=CC=CC=C3)NC(=O)C(C(C)C)N4CCCNC4=O)O
+CC1CC2C3CCC4=CC(=O)C=CC4(C3(C(CC2(C1(C(=O)CO)O)C)O)F)C
+CC1=C(C(=CC=C1)C)NC(=O)CN2CCN(CC2)CC(COC3=CC=CC=C3OC)O
+CCCN(CCC)C(=O)C(CCC(=O)O)NC(=O)C1=CC=CC=C1
+CCOC1=C(C=C2C(=C1)N=CC(=C2NC3=CC(=C(C=C3)OCC4=CC=CC=N4)Cl)C#N)NC(=O)C=CCN(C)C
+CC1=CN=C(S1)NC(=O)C2=C(C3=CC=CC=C3S(=O)(=O)N2C)O
+CN1C(=C(C2=CC=CC=C2S1(=O)=O)O)C(=O)NC3=CC=CC=N3
+CS(=O)(=O)O.CS(=O)(=O)O.C1=CC(=CC=C1C(=O)OC2=CC3=C(C=C2)C=C(C=C3)C(=N)N)N=C(N)N
+C1C(C(OC1N2C=C(C(=O)NC2=O)I)CO)O
+CC(C)(C)C(C(=O)NC(CC1=CC=CC=C1)C(CN(CC2=CC=C(C=C2)C3=CC=CC=N3)NC(=O)C(C(C)(C)C)NC(=O)OC)O)NC(=O)OC.OS(=O)(=O)O
+CC(C)OC(=O)C(=C1SC=CS1)C(=O)OC(C)C
+CCC1=CN=C(C=C1)CCOC2=CC=C(C=C2)CC3C(=O)NC(=O)S3
+CC1=CC2=C(C=C1C(=C)C3=CC=C(C=C3)C(=O)O)C(CCC2(C)C)(C)C
+CCOC(=O)C1=CN=C(C=C1)C#CC2=CC3=C(C=C2)SCCC3(C)C
+CN1CCC(CC1)COC2=C(C=C3C(=C2)N=CN=C3NC4=C(C=C(C=C4)Br)F)OC
+CC1(CN2C=C(N=C2O1)[N+](=O)[O-])COC3=CC=C(C=C3)N4CCC(CC4)OC5=CC=C(C=C5)OC(F)(F)F
+CCC(C)C1C(CCC2(O1)CC3CC(O2)CC=C(C(C(C=CC=C4COC5C4(C(C=C(C5O)C)C(=O)O3)O)C)OC6CC(C(C(O6)C)OC7CC(C(C(O7)C)O)OC)OC)C)C
+CN(C)CCCC1(C2=C(CO1)C=C(C=C2)C#N)C3=CC=C(C=C3)F.Br
+CC(=O)NC1=NN=C(S1)S(=O)(=O)N
+CC12CCC3C(C1CCC2O)C(CC4=C3C=CC(=C4)O)CCCCCCCCCS(=O)CCCC(C(F)(F)F)(F)F
+[OH-].[OH-].[O-][Si]12O[Si]3(O[Si](O1)(O[Si](O2)(O3)[O-])[O-])[O-].[Mg+2].[Mg+2].[Mg+2]
+CC(=O)CCCCN1C(=O)C2=C(N=CN2C)N(C1=O)C
+CC1CCC2C(C(OC3C24C1CCC(O3)(OO4)C)OC)C
+CCC1C(=O)N(CC(=O)N(C(C(=O)NC(C(=O)N(C(C(=O)NC(C(=O)NC(C(=O)N(C(C(=O)N(C(C(=O)N(C(C(=O)N(C(C(=O)N1)C(C(C)CC=CC)O)C)C(C)C)C)CC(C)C)C)CC(C)C)C)C)C)CC(C)C)C)C(C)C)CC(C)C)C)C
+CC(C1=C(C=CC(=C1Cl)F)Cl)OC2=C(N=CC(=C2)C3=CN(N=C3)C4CCNCC4)N.Cl
+CC1=C2C(=C(N(C1=O)C)NC3=C(C=C(C=C3)I)F)C(=O)N(C(=O)N2C4=CC=CC(=C4)NC(=O)C)C5CC5
+CC1=CN=C(N=C1NC2=CC(=CC=C2)S(=O)(=O)NC(C)(C)C)NC3=CC=C(C=C3)OCCN4CCCC4.O.Cl.Cl
+CC(CN1C=CC(=N1)C2=CC(=C(C=C2)C#N)Cl)NC(=O)C3=NNC(=C3)C(C)O
+CCC(=O)OC1(C(CC2C1(CC(C3(C2CCC4=CC(=O)C=CC43C)F)O)C)C)C(=O)CCl
+CCCC1OC2CC3C4CCC5=CC(=O)C=CC5(C4C(CC3(C2(O1)C(=O)CO)C)O)C
+CC(C1=CC=CC2=CC=CC=C21)NCCCC3=CC(=CC=C3)C(F)(F)F
+CC1=C(C(=NO1)C2=CC=CC=C2)C3=CC=C(C=C3)S(=O)(=O)N
+C1CN(CCC1N2C3=CC=CC=C3NC2=O)CCCC(C4=CC=C(C=C4)F)C5=CC=C(C=C5)F
+CN(CC1=CN=C2C(=N1)C(=NC(=N2)N)N)C3=CC=C(C=C3)C(=O)NC(CCC(=O)O)C(=O)O
+C1=C(C(=O)NC(=O)N1)F
+B1(OC(=O)C(O1)(CC(=O)O)CC(=O)O)C(CC(C)C)NC(=O)CNC(=O)C2=C(C=CC(=C2)Cl)Cl
+C1=NC(=NC(=O)N1C2C(C(C(O2)CO)O)O)N
+CC(=O)OC1=CC=CC=C1C(=O)O
+CS(=O)(=O)C1=CC(=C(C=C1)C(=O)NC2=CC(=C(C=C2)Cl)C3=CC=CC=N3)Cl
+CC1CCC2C(C(OC3C24C1CCC(O3)(OO4)C)O)C
+C.CC(C(=O)[O-])O.[Ca+2]
+CC1C(C(C(C(O1)OC2CC(C3(C4C(CCC3(C2)O)C5(CCC(C5(CC4O)C)C6=CC(=O)OC6)O)CO)O)O)O)O.O.O.O.O.O.O.O.O
+CC1=CC(=O)NC(=S)N1
+CN1C=NC2=C1C=C(C(=C2F)NC3=C(C=C(C=C3)Br)F)C(=O)NOCCO
+CCCC(C(=O)OCC)NC(C)C(=O)N1C2CCCCC2CC1C(=O)O.CC(C)(C)N
+CCOC1=CC=C(C=C1)CC2=C(C=CC(=C2)C3C(C(C(C(O3)CO)O)O)O)Cl
+CC1=C(C2=CC=CC=C2N1)CCNCC3=CC=C(C=C3)C=CC(=O)NO
+C1=CN(C(=O)NC1=O)C2C(C(C(O2)CO)O)O
+CN1CCCC1COC2=NC3=C(CCN(C3)C4=CC=CC5=C4C(=CC=C5)Cl)C(=N2)N6CCN(C(C6)CC#N)C(=O)C(=C)F
+CCN1CCN(CC1)CC2=CN=C(C=C2)NC3=NC=C(C(=N3)C4=CC5=C(C(=C4)F)N=C(N5C(C)C)C)F
+CN1C=C(C=N1)C2=NC=CC(=C2)OC3=C(C=C(C(=C3)F)NC(=O)C4(CC4)C(=O)NC5=CC=CC=C5)F
+CC(C)(C)CN1C2=C(C=CC(=N2)C3=C(N=C(N3)C(C)(C)C)C4=CC=C(C=C4)F)N=C1N.CS(=O)(=O)O.CS(=O)(=O)O
+C1=CC(=CC=C1C2=C(N=CC(=C2)C3=CN=C(C=C3)N)N)Cl
+CC1CN(CC(O1)C)C2=NC=C(C=C2)NC(=O)C3=CC=CC(=C3C)C4=CC=C(C=C4)OC(F)(F)F
+C1CNCCN(C1)S(=O)(=O)C2=CC=CC3=C2C=CNC3=O
+CC1=CC(=NN1)NC2=NC(=NC(=C2)C)C3CCC(CC3)(C(=O)NC(C)C4=CN=C(C=C4)N5C=C(C=N5)F)OC
+C1CC1C(=O)N2CCN(CC2)C(=O)C3=C(C=CC(=C3)CC4=NNC(=O)C5=CC=CC=C54)F
+CC(=O)NC(CC1=CC=C(C=C1)O)C(=O)NC(CCCN=C(N)N)C(=O)NC(CC2=CC=CC=C2)C(=O)N
+CC1CC2C3CCC(C3(CC(C2C4(C1=CC(=O)C=C4)C)O)C)(C(=O)COC(=O)CCC(=O)O)O
+CC(=O)O.CC(=O)O.C1=CC(=CC=C1NC(=NC(=NCCCCCCN=C(N)N=C(N)NC2=CC=C(C=C2)Cl)N)N)Cl
+CN1CCC2=CC(=C3C=C2C1CC4=CC=C(C=C4)OC5=C(C=CC(=C5)CC6C7=C(O3)C(=C(C=C7CCN6C)OC)OC)O)OC
+CC(C)(C)C1=CC=C(C=C1)C(CCCN2CCC(CC2)C(C3=CC=CC=C3)(C4=CC=CC=C4)O)O
+C1=CC(=CC=C1C(=N)N)OCCCCCOC2=CC=C(C=C2)C(=N)N.C(CS(=O)(=O)O)O.C(CS(=O)(=O)O)O
+CC1=C(ON=C1C)NS(=O)(=O)C2=CC=C(C=C2)N
+CS(=O)(=O)NC1=C(C=C(C=C1)[N+](=O)[O-])OC2=CC=CC=C2
+CN1CCC2=CC(=C3C=C2C1CC4=CC=C(C=C4)OC5=C(C=CC(=C5)CC6C7=C(O3)C(=C(C=C7CCN6C)OC)OC)O)OC.Cl.Cl
+C1=CC(=CC=C1C2=COC3=C(C2=O)C=CC(=C3)OC4C(C(C(C(O4)CO)O)O)O)O
+C(=CC(=O)O)C(=O)O
+CCC(C)C(=O)OC1CC(C=C2C1C(C(C=C2)C)CCC(CC(CC(=O)[O-])O)O)O.[Na+]
+CCC1=C2CN3C(=CC4=C(C3=O)COC(=O)C4(CC)O)C2=NC5=C1C=C(C=C5)OC(=O)N6CCC(CC6)N7CCCCC7.Cl
+CC(C)NC(=O)COC1=CC=CC(=C1)C2=NC3=CC=CC=C3C(=N2)NC4=CC5=C(C=C4)NN=C5
+CC1=C(C=C(C=N1)NC(=O)C2=CC(=C(C=C2)CN3CCN(CC3)C)C(F)(F)F)NC4=NC=CC(=N4)C5=CN=CC=C5.CS(=O)(=O)O
+C1=CN(C(=O)N=C1N)C2C(C(C(O2)CO)O)O
+CN(C)CC=CC(=O)NC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC(=C(C=C3)F)Cl)OC4CCOC4
+CCC1(C2=C(COC1=O)C(=O)N3CC4=CC5=C(C=CC(=C5CN(C)C)O)N=C4C3=C2)O.Cl
+CC(C1=CC2=C(C=C1)C=C(C=C2)OC)C(=O)O
+CCCCCCCCCC(=O)NC(CC1=CNC2=CC=CC=C21)C(=O)NC(CC(=O)N)C(=O)NC(CC(=O)O)C(=O)NC3C(OC(=O)C(NC(=O)C(NC(=O)C(NC(=O)CNC(=O)C(NC(=O)C(NC(=O)C(NC(=O)C(NC(=O)CNC3=O)CCCN)CC(=O)O)C)CC(=O)O)CO)C(C)CC(=O)O)CC(=O)C4=CC=CC=C4N)C
+CN1CCC(CC1C2=NC3=CC=CC=C3N2)NC(=O)NC4=CC=C(C=C4)C#N
+CNC(=O)C1=NC=CC(=C1)OC2=CC(=C(C=C2)NC(=O)NC3=CC(=C(C=C3)Cl)C(F)(F)F)F
+CC1(C(=O)N(C2=C(O1)C=CC(=N2)NC3=NC(=NC=C3F)NC4=CC(=C(C(=C4)OC)OC)OC)COP(=O)([O-])[O-])C.O.O.O.O.O.O.[Na+].[Na+]
+CN(CCCN1CCC2=CC(=C(C=C2CC1=O)OC)OC)CC3CC4=CC(=C(C=C34)OC)OC.Cl
+CC1C(C(CC(O1)OC2CC(OC(C2O)C)OC3=CC4=CC5=C(C(=O)C(C(C5)C(C(=O)C(C(C)O)O)OC)OC6CC(C(C(O6)C)O)OC7CC(C(C(O7)C)O)OC8CC(C(C(O8)C)O)(C)O)C(=C4C(=C3C)O)O)O)O
+CC1=NC=C(N1CC(CCl)O)[N+](=O)[O-]
+C1CC(N(C1)C2=NC3=C(C=NN3C=C2)NC(=O)N4CCC(C4)O)C5=C(C=CC(=C5)F)F.OS(=O)(=O)O
+CC1=CC=C(C=C1)S(=O)(=O)O.CN1CCC2=C(C1)SC(=N2)C(=O)NC3CC(CCC3NC(=O)C(=O)NC4=NC=C(C=C4)Cl)C(=O)N(C)C.O
+CC1=C(N(N=C1C(=O)NN2CCCCC2)C3=C(C=C(C=C3)Cl)Cl)C4=CC=C(C=C4)Cl.Cl
+CC(CC1=CC(=C(C=C1)O)O)(C(=O)O)NN
+CN(C)C(=O)C1=CC2=CN=C(N=C2N1C3CCCC3)NC4=NC=C(C=C4)N5CCNCC5
+C1=CC2=C(C(=C1)O)N=CC=C2
+C1C(C(OC1N2C=C(C(=O)NC2=O)C(F)(F)F)CO)O
+CC1=C(C(CCC1)(C)C)C=CC(=CC=CC(=CC(=O)O)C)C
+C(CS)N.Cl
+CCC1(CC2CC(C3=C(CCN(C2)C1)C4=CC=CC=C4N3)(C5=C(C=C6C(=C5)C78CCN9C7C(C=CC9)(C(C(C8N6C)(C(=O)OC)O)OC(=O)C)CC)OC)C(=O)OC)O.OS(=O)(=O)O
+CC(=O)O.C1C(CN2C1C(=O)NC(C(=O)NC(C(=O)NC(C(=O)NC(C(=O)NC(C2=O)CC3=CC=CC=C3)CC4=CC=C(C=C4)OCC5=CC=CC=C5)CCCCN)CC6=CNC7=CC=CC=C76)C8=CC=CC=C8)OC(=O)NCCN
+C1=CC2=C(C=C1OC(F)(F)F)SC(=N2)N.Cl
+COC1=C(C2=C[N+]3=C(C=C2C=C1)C4=CC(=C(C=C4CC3)OC)OC)OC.[Cl-]
+CN1CCC2=CC(=C(C3=C2C1CC4=C3C(=C(C=C4)OC)O)OC)OC
+C1=CC=C(C=C1)C(C2=CC=CC=C2)(C3=CC=CC=C3Cl)N4C=CN=C4
+C1=CC=C(C(=C1)C(=O)[O-])O.[Na+]
+C1=C(C=C(C(=C1O)O)O)C(=O)O.O
+CC(C)C(CCCN(C)CCC1=CC(=C(C=C1)OC)OC)(C#N)C2=CC(=C(C=C2)OC)OC.Cl
+CC1=NC=C(C(=C1O)CO)CO
+CC(=O)C1=CC=C(C=C1)S(=O)(=O)NC(=O)NC2CCCCC2
+CC1CCC2CC(C(=CC=CC=CC(CC(C(=O)C(C(C(=CC(C(=O)CC(OC(=O)C3CCCCN3C(=O)C(=O)C1(O2)O)C(C)CC4CCC(C(C4)OC)OC(=O)C(C)(CO)CO)C)C)O)OC)C)C)C)OC
+CC1=CC2C(CCC3(C2CCC3(C(=O)C)O)C)C4(C1=CC(=O)CC4)C
+CC1(C2CCC3(C(C2(CCC1OC4C(C(C(C(O4)C(=O)O)O)O)OC5C(C(C(C(O5)C(=O)O)O)O)O)C)C(=O)C=C6C3(CCC7(C6CC(CC7)(C)C(=O)O)C)C)C)C.N.N
+C1=C(C=C(C(=C1I)OC2=CC(=C(C(=C2)I)O)I)I)CC(C(=O)[O-])N.O.O.O.O.O.[Na+]
+CCCN1CCCCC1C(=O)NC2=C(C=CC=C2C)C.O.Cl
+C1=CC=C(C(=C1)C(=O)O)O
+CC12CCC3C(C1CCC2(C#C)O)CCC4=C3C=CC(=C4)OC5CCCC5
+CC1CCN(CC1N(C)C2=NC=NC3=C2C=CN3)C(=O)CC#N
+CC1=C(C=CC2=C1C(=NC(=N2)N)N)CNC3=CC(=C(C(=C3)OC)OC)OC
+CC1=CC(=C(C=C1)C)OCCCC(C)(C)C(=O)O
+CC1=CC=C(C=C1)C(=O)C2=CC(=C(C(=C2)O)O)[N+](=O)[O-]
+CC(C)C(CC1=CC(=C(C=C1)OC)OCCCOC)CC(C(CC(C(C)C)C(=O)NCC(C)(C)C(=O)N)O)N
+COCCCCC(=NOCCN)C1=CC=C(C=C1)C(F)(F)F.C(=CC(=O)O)C(=O)O
+CC1=CC=C(C=C1)S(=O)(=O)O.C[S+](CCC(C(=O)O)N)CC1C(C(C(O1)N2C=NC3=C(N=CN=C32)N)O)O.OS(=O)(=O)O.OS(=O)(=O)[O-]
+CC1C(C(CC(O1)OC2CC(CC3=C2C(=C4C(=C3O)C(=O)C5=CC=CC=C5C4=O)O)(C(=O)C)O)N)O.Cl
+CC1=C(C(=O)N(C(=O)N1CC2=C(C=CC=C2F)C(F)(F)F)CC(C3=CC=CC=C3)NCCCC(=O)[O-])C4=C(C(=CC=C4)OC)F.[Na+]
+C1=CC(=CC=C1C=CC2=CC(=CC(=C2)O)O)O
+CC1CCC2CC(C(=CC=CC=CC(CC(C(=O)C(C(C(=CC(C(=O)CC(OC(=O)C3CCCCN3C(=O)C(=O)C1(O2)O)C(C)CC4CCC(C(C4)OC)O)C)C)O)OC)C)C)C)OC
+CC(C)(C)NC(=O)C1CC2CCCCC2CN1CC(C(CC3=CC=CC=C3)NC(=O)C(CC(=O)N)NC(=O)C4=NC5=CC=CC=C5C=C4)O
+CC1CCN(CC1N(C)C2=NC=NC3=C2C=CN3)C(=O)CC#N.C(C(=O)O)C(CC(=O)O)(C(=O)O)O
+CC1=CC(=C(C=C1)C(C)C)O
+CN(C)CCN(C)C1=CC(=C(C=C1NC(=O)C=C)NC2=NC=CC(=N2)C3=CN(C4=CC=CC=C43)C5CC5)OC.Cl
+C[As](C)SCC(C(=O)NCC(=O)O)NC(=O)CCC(C(=O)O)N
+C1=CC(=C(C(=C1)Cl)COCCOCCCCCCNCC(C2=CC(=C(C=C2)O)CO)O)Cl
+C1CC(=O)NC(=O)C1N2CC3=C(C2=O)C=CC=C3N.C1CC(=O)NC(=O)C1N2CC3=C(C2=O)C=CC=C3N.O
+CCC1C(C(C(N(CC(CC(C(C(C(C(C(=O)O1)C)OC2CC(C(C(O2)C)O)(C)OC)C)OC3C(C(CC(O3)C)N(C)C)O)(C)O)C)C)C)O)(C)O.O.O
+CC(C)C1=NC(=CS1)CN(C)C(=O)NC(C(C)C)C(=O)NC(CC2=CC=CC=C2)CC(C(CC3=CC=CC=C3)NC(=O)OCC4=CN=CS4)O
+C1=NC(=C2C(=N1)N(C=N2)C3C(C(C(O3)CO)O)O)N
+C1=CC(=C(C=C1C(CN)O)O)O.Cl
+CC1=CC(=O)N(C(=C1)C2CCCCC2)O
+CCN(CC)CC(=O)NC1=C(C=CC=C1C)C.Cl
+CN(C)C1C2CC3C(C4=C(C=CC(=C4C(=C3C(=O)C2(C(=C(C1=O)C(=O)N)O)O)O)O)Cl)O
+C1=CC(=C2C(=C1NCCNCCO)C(=O)C3=C(C=CC(=C3C2=O)O)O)NCCNCCO.Cl.Cl
+C1CC(N(C1)C2=NC3=C(C=NN3C=C2)NC(=O)N4CCC(C4)O)C5=C(C=CC(=C5)F)F
+CS(=O)(=O)OCCCCOS(=O)(=O)C
+CN1CCC2=CC3=C(C4=C2C1CC5=CC=C(C=C5)OC6=C(C=CC(=C6)CC7C8=CC(=C(C=C8CCN7C)OC)O4)OC)OCO3
+CC1=C(C=C(C=C1)C(=O)NC2=CC(=CC(=C2)C(F)(F)F)N3C=C(N=C3)C)NC4=NC=CC(=N4)C5=NC=CN=C5
+CC1=C2C(C(=O)C3(C(CC4C(C3C(C(C2(C)C)(CC1OC(=O)C(C(C5=CC=CC=C5)NC(=O)OC(C)(C)C)O)O)OC(=O)C6=CC=CC=C6)(CO4)OC(=O)C)O)C)O
+C1=CC(=C(C=C1Cl)O)OC2=C(C=C(C=C2)Cl)Cl
+C1=CC(=CN=C1)C=CC(=O)NCC2=CC=C(C=C2)C(=O)NC3=C(C=C(C=C3)F)N
+COC1=C(C=C2C(=C1O)C3C(C(C(C(O3)CO)O)O)OC2=O)O
+C1(C(=O)NC(=O)N1)NC(=O)N
+CCNC(=O)CCCC=CCC1C(CC(C1C=CC(CCC2=CC=CC=C2)O)O)O
+CC12CC(C3(C(C1CC(C2(C(=O)CO)O)O)CCC4=CC(=O)C=CC43C)F)O
+CC1=CC2=C(S1)NC3=CC=CC=C3N=C2N4CCN(CC4)C
+CC1=C(N(N=C1C(=O)NN2CCCCC2)C3=C(C=C(C=C3)Cl)Cl)C4=CC=C(C=C4)Cl
+CC(C)(C)OC(=O)NCCC(=O)NC(CC1=CNC2=CC=CC=C21)C(=O)NC(CCSC)C(=O)NC(CC(=O)O)C(=O)NC(CC3=CC=CC=C3)C(=O)N
+CC1=CC=C(C=C1)S(=O)(=O)O.CN1CCN2C3CCN(CC3C4=C2C1=CC=C4)CCCC(=O)C5=CC=C(C=C5)F
+CN(C)CC1=C(SC2=C1C(=O)N(C(=O)N2CC3=C(C=CC=C3F)F)C4=NN=C(C=C4)OC)C5=CC=C(C=C5)NC(=O)NOC
+CC(C)(CCCC(CC(=O)OC)(C(=O)OC1C2C3=CC4=C(C=C3CCN5C2(CCC5)C=C1OC)OCO4)O)O
+CCCCCCC1=C(C=C(C=C1)O)O
+C1=CC=C(C=C1)C2=CC(=O)C3=C(C(=C(C=C3O2)OC4C(C(C(C(O4)C(=O)O)O)O)O)O)O
+COC1=C(C2=C[N+]3=C(C=C2C=C1)C4=CC5=C(C=C4CC3)OCO5)OC.O.[Cl-]
+CC1CC2C(CCC3(C2CCC3(C(=O)C)OC(=O)C)C)C4(C1=CC(=O)CC4)C
+CCOC1=C(C=C2C(=C1)N=CC(=C2NC3=CC(=C(C=C3)OCC4=CC=CC=N4)Cl)C#N)NC(=O)C=CCN(C)C.C(=CC(=O)O)C(=O)O
+CN1CCC23CC(=O)C(=CC2C1CC4=C3C(=C(C=C4)OC)O)OC
+CC(C)CC(C(=O)NC(CCCN=C(N)N)C(=O)N1CCCC1C(=O)NNC(=O)N)NC(=O)C(COC(C)(C)C)NC(=O)C(CC2=CC=C(C=C2)O)NC(=O)C(CO)NC(=O)C(CC3=CNC4=CC=CC=C43)NC(=O)C(CC5=CN=CN5)NC(=O)C6CCC(=O)N6.CC(=O)O
+CC1C2CCC3(C(C2(CCC1O)C)C(CC4C3(CC(C4=C(CCC=C(C)C)C(=O)O)OC(=O)C)C)O)C
+C1=CC(=CC=C1C(=O)NCCC(=O)[O-])N=NC2=CC(=C(C=C2)[O-])C(=O)O.O.O.[Na+].[Na+]
+CC1=C(C(=NO1)C2=CC=CC=C2Cl)C(=O)NC3C4N(C3=O)C(C(S4)(C)C)C(=O)[O-].[Na+]
+CC1CCC2C(C(OC3C24C1CCC(O3)(OO4)C)OC(=O)CCC(=O)O)C
+CC1=NC=C(C(=C1O)CO)CO.Cl
+C1=CC(=CC=C1C(=O)NC(CCC(=O)O)C(=O)O)NCC2=CN=C3C(=N2)C(=O)NC(=N3)N
+C1=CN=C(C=N1)NNC(=O)C=CN2C=NC(=N2)C3=CC(=CC(=C3)C(F)(F)F)C(F)(F)F
+C1=NC2=NC=NC(=C2N1)N
+C1=C(C=C(C(=C1O)O)O)C(=O)O
+C1CSSC1CCCCC(=O)O
+COC1=C(C=CC(=C1)NS(=O)(=O)C)NC2=C3C=CC=CC3=NC4=CC=CC=C42
+COC1=CC=C(C=C1)C2=CC(=S)SS2
+C1CN=C(N1)NC2=C(C=CC=C2Cl)Cl.Cl
+CCN1CCN(CC1)C2=CC=C(C=C2)NC3=CC(=NC=N3)N(C)C(=O)NC4=C(C(=CC(=C4Cl)OC)OC)Cl
+CC(C)NCCN(C1=CC2=NC(=CN=C2C=C1)C3=CN(N=C3)C)C4=CC(=CC(=C4)OC)OC
+CC1CCC2CC(C(=CC=CC=CC(CC(C(=O)C(C(C(=CC(C(=O)CC(OC(=O)C3CCCCN3C(=O)C(=O)C1(O2)O)C(C)CC4CCC(C(C4)OC)OCCO)C)C)O)OC)C)C)C)OC
+CCN1C2=C3C=C(NC3=NC=C2CN(C1=O)C4=C(C(=CC(=C4F)OC)OC)F)CN5CCOCC5
+C1CN(CCC1(C(=O)NC(CCO)C2=CC=C(C=C2)Cl)N)C3=NC=NC4=C3C=CN4
+C1=CN(C(=O)N=C1N)C2C(C(C(O2)CO)O)(F)F
+CCC1=C2CN3C(=CC4=C(C3=O)COC(=O)C4(CC)O)C2=NC5=C1C=C(C=C5)OC(=O)N6CCC(CC6)N7CCCCC7
+C1CCC(C(C1)[NH-])[NH-].C(=O)(C(=O)O)O.[Pt+2]
+CC1=C(SC(=N1)NC(=O)N2CCCC2C(=O)N)C3=CC(=NC=C3)C(C)(C)C(F)(F)F
+CC1=CC=C(C=C1)S(=O)(=O)O.CC1=CC=C(C=C1)S(=O)(=O)O.CS(=O)(=O)CCNCC1=CC=C(O1)C2=CC3=C(C=C2)N=CN=C3NC4=CC(=C(C=C4)OCC5=CC(=CC=C5)F)Cl
+CC1=C(C(=O)N(C2=NC(=NC=C12)NC3=NC=C(C=C3)N4CCNCC4)C5CCCC5)C(=O)C
+C1CN(CCC1CC(=O)N2CCC(CC2)C3C4=C(CCC5=C3N=CC(=C5)Br)C=C(C=C4Br)Cl)C(=O)N
+CCC1C(=O)N(C2=CN=C(N=C2N1C(C)C)NC3=C(C=C(C=C3)C(=O)NC4CCC(CC4)N5CCN(CC5)CC6CC6)OC)C
+CC1(CCC(=C(C1)C2=CC=C(C=C2)Cl)CN3CCN(CC3)C4=CC(=C(C=C4)C(=O)NS(=O)(=O)C5=CC(=C(C=C5)NCC6CCOCC6)[N+](=O)[O-])OC7=CN=C8C(=C7)C=CN8)C
+CNC(=O)C1=C(C=C(C=C1)C2=NN3C(=CN=C3N=C2)CC4=CC5=C(C=C4)N=CC=C5)F
+CC(C)OC1=NC=C(C=C1)C2=NNC3=C2C=C(C=C3)NC(=O)C4(CCN(C4)CC(=O)N5CCC(=CC5)C6=CC=C(C=C6)C7=NN(C=N7)C)SC
+CC1=C(C=CC(=C1)NC2=NC=NC3=C2C=C(C=C3)NC4=NC(CO4)(C)C)OC5=CC6=NC=NN6C=C5
+CCCS(=O)(=O)NC1=C(C(=C(C=C1)F)C(=O)C2=CNC3=C2C=C(C=N3)C4=CC=C(C=C4)Cl)F
+CC1=NC2=CC(=C(C=C2C(=N1)NC(C)C3=CC(=CC(=C3)N)C(F)(F)F)OC4CCOC4)OC
+CC1COCCN1C2=NC(=NC3=C2C=CC(=N3)C4=CC(=C(C=C4)OC)CO)N5CCOCC5C
+CC1(C2=C(C=C(S2)C3=NC(=NC=C3)NC4=CC=NN4C)C(=O)N1CCN5CCOCC5)C
+CN(C1CCS(=O)(=O)CC1)C(=O)CNC(=O)C=CC2=CC(=CC=C2)Cl
+C1COC2=C(O1)C=CC(=C2)N3C(=O)NN=C3SC4=NC=C(S4)[N+](=O)[O-]
+CC1=C2C(=O)NC3(N2C(=O)C(=C1)NC4=NC=NC(=C4)N)CCCCC3
+C1C(C(OC1N2C=C(C(=O)NC2=O)C=CBr)CO)O
+C1COCCN1CC2=CC=C(C=C2)COC3=NC=CC(=N3)C(C#N)C4=NC5=CC=CC=C5S4
+CC(C)(C)C1=NC(=C(S1)C2=NC(=NC=C2)N)C3=C(C(=CC=C3)NS(=O)(=O)C4=C(C=CC=C4F)F)F
+CC(C)N1C=C(C(=N1)C2=C(C(=CC(=C2)Cl)NS(=O)(=O)C)F)C3=NC(=NC=C3)NCC(C)NC(=O)OC
+CC1=C2C(C(=O)C3(C(CC4C(C3C(C(C2(C)C)(CC1OC(=O)C(C(C5=CC=CC=C5)NC(=O)C6=CC=CC=C6)O)O)OC(=O)C7=CC=CC=C7)(CO4)OC(=O)C)O)C)OC(=O)C
+CCN1C2=C3C=C(C=C2)C4=CSC(=N4)CC(C(=O)N5CCCC(N5)C(=O)OCC(CC3=C1C6=C(N=CC(=C6)N7CCN(CC7)C)C(C)OC)(C)C)NC(=O)C8CC8C
+CC1=C(C=C(C=C1)C(=O)NC2=CC=CC(=C2)C(F)(F)F)NC3=C4C=NN(C4=NC(=N3)C5=CN=CC=C5)C
+C1=CC(=C(C(=C1)Br)C(=O)C2=CNC(=C2)C(=O)NC3=CN=CC=C3)F
+CC1=C(C=C(C=C1)C(=O)NC)N2C(=CC(=C(C2=O)Br)OCC3=C(C=C(C=C3)F)F)C
+CCS(=O)(=O)C1=CC=CC(=C1)C2=CC(=C(C3=C2C4=C(N3)N=CC(=C4)C)C)C(=O)NC5CCN(CC5)C
+CC1CC(C2=C1C(=NC=N2)N3CCN(CC3)C(=O)C(CNC(C)C)C4=CC=C(C=C4)Cl)O
+CC(C)N1C2=NC=NC(=C2C(=N1)C3=CC4=C(C=C3)OC(=N4)N)N
+C1CC(C1)(C2=CC=C(C=C2)C3=C(N4C(=N3)C=CC(=N4)C(=O)N)C5=CC=CC=C5)N
+CN1C=C(C2=CC3=C(C=C21)OCO3)C4=C(C(=O)NC4=O)C5=COC6=C5C=C(C=C6)F
+C1=CC(=CC=C1COC2=C(C=C(C=C2)CNCCO)OCC3=CC=C(C=C3)F)F.Cl
+C1COCCN1C2=NC(=NC(=N2)C3=CN=C(C=C3C(F)(F)F)N)N4CCOCC4
+C1=CC2=NC=CC(=C2C=C1C=C3C(=O)NC(=O)S3)C4=CC=NC=C4
+CN1CCN(CC1)S(=O)(=O)C2=CC=C(C=C2)C3=CN=C(C(=N3)C(=O)NC4=CN=CC=C4)N
+C1=CC=C(C=C1)C2=C(C(=O)NC3=C2C=C(C=C3)Cl)C(=O)C=CC4=CC(=CN=C4)F
+C1COCCN1C2=CC=C(C=C2)C3=C(C=NC=C3)C4=CC(=C(C(=C4)F)O)F
+CC1=CC=C(C=C1)S(=O)(=O)O.CN1C=C(N=C1C2CCN(CC2)C3=NC=NC4=C3C=NN4)C5=CC(=C(C=C5)F)C(F)(F)F
+CC1=C(C=C(C=C1)OC2=NC3=C(N2)C=C(C(=C3)C4=CC5=C(C=C4)N(C=C5)C)Cl)C(=O)O
+CC1=C(C=C(C=C1)OC2=NC3=C(N2)C=C(C(=C3)C4=CC=C(C=C4)C5=CC=CC=C5)Cl)C(=O)O
+C1CC1C(C(F)(F)F)NC(=O)C2=CC(=CS2)C3=C4N=CC(=CN4N=C3)C5=CC=C(C=C5)F
+CC(=O)C1=CC=C(C=C1)NN=C(C2=NC(CC3=CC=CC=C32)(C)C)C(=O)N
+CN1C=C(C2=CC=CC=C21)C3=C(C(=O)NC3=O)C4=CC(=CC=C4)N
+CC(C)NC(=O)COC1=CC=CC(=C1)C2=NC3=CC=CC=C3C(=N2)NC4=CC5=C(C=C4)NN=C5.CS(=O)(=O)O
+CS(=O)(=O)C1=C2C(C(C(C2=C(C=C1)OC3=CC(=CC(=C3)C#N)F)F)F)O
+COC1=CC(=CC(=C1)C#CC2=NN(C3=NC=NC(=C23)N)C4CCN(C4)C(=O)C=C)OC
+CN1C2=C(C(=C(C1=O)F)NC3=C(C=C(C=C3)I)F)C(=O)N(C=N2)CC(CO)O
+CC1COCCN1C2=NC3=C(C=CN=C3C4=CC=NN4)C(=C2)C5=CC=NN5C.Cl
+CC(C)N1C2=C(C=NC3=CC(=C(C=C32)C4=CN=C(C=C4)OCCCN5CCCCC5)F)N(C1=O)C
+COC1=CC=C(C=C1)C(=O)N2C3=CC=CC=C3SC4=C2C=C(C=C4)Cl
+CC(C)N1C2=NC=NC(=C2C(=N1)C3=CC4=C(N3)C=CC(=C4)O)N
+C1=CC=C(C(=C1)CC(=NNC2=NC(=CS2)C3=CC(=C(C=C3)Cl)Cl)C(=O)O)[N+](=O)[O-]
+CCC1=C(C=CC(=C1)C(=O)N)N2C3=NC=CC(=C3C(=N2)C(C)C)N4C=C(N=C4)C5=CN(N=C5)C
+B(C(CC(C)C)NC(=O)CNC(=O)C1=C(C=CC(=C1)Cl)Cl)(O)O
+CC1=CC(=C(C2=C1NC(=O)C3=C2C=CS3)C4=CC=C(C=C4)C(C)CN)O
+CCN1CCN(CC1)CC2=C(C=C(C=C2)NC(=O)C3=CC(=C(C=C3)C)OC4=C5C=CNC5=NC=C4)C(F)(F)F
+C1CC(C(C(C1)N2C3=CC=CC=C3OC4=CC=CC=C42)O)NS(=O)(=O)C5=CC=C(C=C5)OC(F)(F)F
+CC1=CC2=NC=NN2C=C1NC3=NC=C4C(=N3)N(C(=O)N4C)C5CCOCC5
+CC1=CC=C(C=C1)C2=CC(=NN2C3=CC=C(C=C3)S(=O)(=O)N)C(F)(F)F
+CCC1(CC2CC(C3=C(CCN(C2)C1)C4=CC=CC=C4N3)(C5=C(C=C6C(=C5)C78CCN9C7C(C=CC9)(C(C(C8N6C=O)(C(=O)OC)O)OC(=O)C)CC)OC)C(=O)OC)O
+C1=CC=C2C(=C1)N=C3C(=C4C=C(C=CN4C3=N2)NC(=O)C5=CC=CS5)C#N
+CC1=C(C2=C(N1)C=CC(=C2)F)C3=NC4=C5C=CC=NC5=C6C(=C4N3)C=CC=N6
+CN1CCN(CC1)C(=O)C2C3CCC(C2C(=O)O)O3
+C1CNCCC1NC(=O)C2=C(C=NN2)NC(=O)C3=C(C=CC=C3Cl)Cl
+C1COCCN1C(=O)C2=CC=C(C=C2)C3=CN4C(=NC=C4C5=CC=C(C=C5)C#N)C=C3
+C1CCN(CC1)C(=O)N2CCN3C=C(C4=CC(=CC(=C43)C2)F)C5=C(C(=O)NC5=O)C6=CN=C7N6C=CC=C7
+CCCCCCCCC=CCCCCCCCC(=O)O
+C1=CC2=NC=C(N2C=C1)CC(O)(P(=O)(O)O)P(=O)(O)O
+CC(C)(C#N)C1=CC(=CC(=C1)CN2C=NC=N2)C(C)(C)C#N
+CC(C)(C)C(=O)OC1=CC=C(C=C1)S(=O)(=O)NC2=CC=CC=C2C(=O)NCC(=O)[O-].O.O.O.O.[Na+]
+CCC(C(C)O)N1C(=O)N(C=N1)C2=CC=C(C=C2)N3CCN(CC3)C4=CC=C(C=C4)OCC5CC(OC5)(CN6C=NC=N6)C7=C(C=C(C=C7)F)F
+CC1(C2CC3C(C(=O)C(=C(C3(C(=O)C2=C(C4=C1C=CC=C4O)O)O)O)C(=O)N)N(C)C)O.Cl
+C=C1C(CC(C1CO)O)N2C=NC3=C2N=C(NC3=O)N.O
+CCCC1=NN(C2=C1N=C(NC2=O)C3=C(C=CC(=C3)S(=O)(=O)N4CCN(CC4)C)OCC)C
+COC(=O)C=CC(=O)OC
+CNCC1=CC=C(C=C1)C2=C3CCNC(=O)C4=C3C(=CC(=C4)F)N2.OP(=O)(O)O
+CC(C1=C(C=CC(=C1Cl)F)Cl)OC2=C(N=CC(=C2)C3=CN(N=C3)C4CCNCC4)N
+CCC1=C2N=C(C=C(N2N=C1)NCC3=C[N+](=CC=C3)[O-])N4CCCCC4CCO
+CC(C)(C)C1=CC=C(C=C1)S(=O)(=O)NC2=C(C(=NC(=N2)C3=NC=CC=N3)OCCO)OC4=CC=CC=C4OC.O
+CC(C)CC(C(=O)O)NC(=O)C(C(CC1=CC=CC=C1)N)O
+CN1CC(=O)N2C(C1=O)CC3=C(C2C4=CC5=C(C=C4)OCO5)NC6=CC=CC=C36
+CCN(CC)CCNC1=C2C(=C(C=C1)C)SC3=CC=CC=C3C2=O
+CC(COC1=CC=CC=C1CC2=CC=CC=C2)N3CCCCC3.OP(=O)(O)O
+COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C)OCCOC
+CC(C)OC(=O)C(C)(C)OC1=CC=C(C=C1)C(=O)C2=CC=C(C=C2)Cl
+CC1C(C(C(C(O1)OCC2C(C(C(C(O2)OC3=C(OC4=CC(=CC(=C4C3=O)O)O)C5=CC(=C(C=C5)O)O)O)O)O)O)O)O.O.O.O
+C1=CC(=NC=C1CC2=CNC3=C2C=C(C=N3)Cl)NCC4=CN=C(C=C4)C(F)(F)F.Cl
+CC#CC(=O)N1CCC(C1)N2C3=NC=NC(=C3N(C2=O)C4=CC=C(C=C4)OC5=CC=CC=C5)N
+CN(C)CCOC(C1=CC=CC=C1)C2=CC=CC=C2
+CC(C)(CCC(CC(=O)OC)(C(=O)OC1C2C3=CC4=C(C=C3CCN5C2(CCC5)C=C1OC)OCO4)O)O
+CC1C(C(CC(O1)OC2CC(CC3=C2C(=C4C(=C3O)C(=O)C5=C(C4=O)C(=CC=C5)OC)O)(C(=O)CO)O)N)O.Cl
+C1=CC=C(C=C1)C2(C(=O)NC(=N2)[O-])C3=CC=CC=C3.[Na+]
+CCCCCCCCCCCC(CC1C(C(=O)O1)CCCCCC)OC(=O)C(CC(C)C)NC=O
+CCCCC(=O)N(CC1=CC=C(C=C1)C2=CC=CC=C2C3=NN=N[N-]3)C(C(C)C)C(=O)[O-].CCCCC(=O)N(CC1=CC=C(C=C1)C2=CC=CC=C2C3=NN=N[N-]3)C(C(C)C)C(=O)[O-].CCOC(=O)C(C)CC(CC1=CC=C(C=C1)C2=CC=CC=C2)NC(=O)CCC(=O)[O-].CCOC(=O)C(C)CC(CC1=CC=C(C=C1)C2=CC=CC=C2)NC(=O)CCC(=O)[O-].O.O.O.O.O.[Na+].[Na+].[Na+].[Na+].[Na+].[Na+]
+CC1C(=O)OC2CCN3C2C(=CC3)COC(=O)C(C1(C)O)(C)O
+CC12CCC3C(C1CCC2C(=O)NC(C)(C)C)CCC4C3(C=CC(=O)N4)C
+C(=O)(N)NO
+CC1C2C(C3C(C(=O)C(=C(C3(C(=O)C2=C(C4=C1C=CC=C4O)O)O)O)C(=O)N)N(C)C)O.O
+C1CCNC(C1)C2(CN(C2)C(=O)C3=C(C(=C(C=C3)F)F)NC4=C(C=C(C=C4)I)F)O
+COC1=C(C=C2C(=C1)N=CN=C2NC3=CC(=C(C=C3)F)Cl)OCCN4CC5(CC5)C6(C4)OCCO6
+CCN(CC)CCNC(=O)C1=CC=C(C=C1)N.Cl
+CC(CC1=CC2=C(C(=C1)C(=O)N)N(CC2)CCCO)NCCOC3=CC=CC=C3OCC(F)(F)F
+CC1=C2C(=NC=C1)N(C3=C(C=CC=N3)C(=O)N2)C4CC4
+C1CCC(CC1)N2C(=NN=N2)CCCCOC3=CC4=C(C=C3)NC(=O)CC4
+CCNC1CC(S(=O)(=O)C2=C1C=C(S2)S(=O)(=O)N)C.Cl
+CC1=C2C=C(C=CC2=C(C(=N1)C(=O)NCC(=O)O)O)OC3=CC=CC=C3
+CC1=C(C=C(C=C1)C2C(C(C(C(O2)CO)O)O)O)CC3=CC=C(S3)C4=CC=C(C=C4)F
+C1C(C(OC1N2C=NC(=NC2=O)N)CO)O
+CC1=C(N=C(C(=N1)C)C)C
+CC(C)C(C(=O)NC(CC1=CC=C(C=C1)O)C(=O)O)NC(=O)C(CC(=O)O)NC(=O)C(CCCCN)NC(=O)C(CCCN=C(N)N)N
+CC1C(C(C(C(O1)OCC2C(C(C(C(O2)OC3=CC(=C4C(=O)CC(OC4=C3)C5=CC(=C(C=C5)OC)O)O)O)O)O)O)O)O
+COC(=O)NC1=NC2=C(N1)C=C(C=C2)C(=O)C3=CC=CC=C3
+CCC(=O)OCC(=O)C1(C(CC2C1(CC(C3(C2CCC4=CC(=O)C=CC43C)F)O)C)C)OC(=O)CC
+CCCCN1CCCCC1C(=O)NC2=C(C=CC=C2C)C.Cl
+CN(CCOC1=CC=C(C=C1)CC2C(=O)NC(=O)S2)C3=CC=CC=N3
+CC(C)C(CCCN(C)CCC1=CC(=C(C=C1)OC)OC)(C#N)C2=CC(=C(C=C2)OC)OC
+C.CC(C)NCC(COC1=CC=C(C=C1)COCCOC(C)C)O.C(=CC(=O)O)C(=O)O
+C1CN(CCC1(C2=CC(=C(C=C2)Cl)C(F)(F)F)O)CCCC(C3=CC=C(C=C3)F)C4=CC=C(C=C4)F
+COC1=C(C=CC(=C1)C=CC(=O)O)O
+C1=CC(=CC=C1CCC2=CNC3=C2C(=O)NC(=N3)N)C(=O)NC(CCC(=O)O)C(=O)O
+COCCCCC(=NOCCN)C1=CC=C(C=C1)C(F)(F)F
+C1=CN(C(=O)N=C1N)C2C(C(C(O2)CO)O)O.Cl
+CN1C2=C(C=C(C=C2)N(CCCl)CCCl)N=C1CCCC(=O)O
+CN(C)C1=NC(=NC(=N1)N(C)C)N(C)C
+CC1C=CC=C(C(=O)NC2=C(C3=C(C4=C(C(=C3O)C)OC(C4=O)(OC=CC(C(C(C(C(C(C1O)C)O)C)OC(=O)C)C)OC)C)C5=C2N6C=CC(=CC6=N5)C)O)C
+CC1=C(C(=CC=C1)C(C)C2=CN=CN2)C
+CC1=CC(=O)C2=CC=CC=C2C1=O
+CC1(C2CCC3(C(C2(CCC1OC4C(C(C(C(O4)C(=O)O)O)O)OC5C(C(C(C(O5)C(=O)O)O)O)O)C)C(=O)C=C6C3(CCC7(C6CC(CC7)(C)C(=O)O)C)C)C)C
+CCOC1=CC=C(C=C1)CC2=C(C=CC(=C2)C3C(C(C(C(O3)CO)O)O)O)Cl.CC(CO)O.O
+CC(C)CN1C=NC2=C1C3=CC=CC=C3N=C2N.C(=CC(=O)O)C(=O)O
+C1=NC2=C(N=C(N=C2N1C3C(C(C(O3)CO)O)F)Cl)N
+CC1=CC=CC=C1C(=O)NC2=CC=C(C=C2)C(=O)N3CCCC(C4=CC=CC=C43)N(C)C
+CC1=CC2=C(N1)C=CC(=C2)OC3=NC(=NC=C3)NC4=CC=CC(=C4)CS(=O)(=O)NCCN(C)C
+C1=CC(=CC=C1O)OC2C(C(C(C(O2)CO)O)O)O
+CC1=C2C(C(=O)C3(C(CC4C(C3C(C(C2(C)C)(CC1OC(=O)C(C(C5=CC=CC=C5)NC(=O)OC(C)(C)C)O)O)OC(=O)C6=CC=CC=C6)(CO4)OC(=O)C)O)C)O.O.O.O
+COC1=C(C=C(C=C1)C=CC(=O)NC2=CC=CC=C2C(=O)O)OC
+CCCNS(=O)(=O)NC1=C(C(=NC=N1)OCCOC2=NC=C(C=N2)Br)C3=CC=C(C=C3)Br
+CC12CCC(=O)C=C1CC(C3C24C(O4)CC5(C3CCC56CCC(=O)O6)C)C(=O)OC
+CC#CC1(CCC2C1(CC(C3=C4CCC(=O)C=C4CCC23)C5=CC=C(C=C5)N(C)C)C)O
+CNC(=O)C1=C(C=C(C=C1)N2C(=S)N(C(=O)C23CCC3)C4=CC(=C(N=C4)C#N)C(F)(F)F)F
+C1=CC(=C(C=C1[N+](=O)[O-])Cl)NC(=O)C2=C(C=CC(=C2)Cl)O.C(CO)N
+CC(=CCCC(=CCCC(=CC=CC(=CC(=O)O)C)C)C)C
+CC1=CC=C(C=C1)C(=O)C2=CC=C(N2C)CC(=O)O

UniMol2_embedding.py

@@ -0,0 +1,124 @@
+import os
+import pickle
+import numpy as np
+import pandas as pd
+import torch
+import torch.nn as nn
+from tqdm import tqdm
+from unimol_tools import UniMolRepr
+import os, sys, contextlib
+
+@contextlib.contextmanager
+def suppress_stdout_stderr():
+    """with 块内所有 print / tqdm / warning 都不会显示"""
+    with open(os.devnull, 'w') as devnull:
+        old_out, old_err = sys.stdout, sys.stderr
+        sys.stdout, sys.stderr = devnull, devnull
+        try:
+            yield
+        finally:
+            sys.stdout, sys.stderr = old_out, old_err
+
+def get_unimol_embeddings(smiles_list, output_file="UniMol_emb512.pkl", 
+                         model_name='unimolv1', model_size='84m', 
+                         remove_hs=False, batch_size=32):
+    """
+    使用Uni-Mol模型为SMILES列表生成分子嵌入，并保存为pickle文件
+    
+    参数:
+    smiles_list (list): SMILES字符串列表
+    output_file (str): 输出pickle文件路径
+    model_name (str): 模型名称，可选'unimolv1'或'unimolv2'
+    model_size (str): 模型大小，仅在使用unimolv2时有效
+    remove_hs (bool): 是否移除氢原子
+    batch_size (int): 批处理大小
+    
+    返回:
+    dict: 包含SMILES及其对应嵌入的字典
+    """
+    # 初始化模型
+    clf = UniMolRepr(
+        data_type='molecule',
+        remove_hs=remove_hs,
+        model_name=model_name,
+        model_size=model_size
+    )
+    
+    # 用于存储结果的字典
+    embeddings_dict = {}
+    error_smiles = []
+    
+    # 批处理SMILES
+    total_batches = (len(smiles_list) + batch_size - 1) // batch_size
+    
+    print(f"开始生成{len(smiles_list)}个SMILES的嵌入表示...")
+    
+    print("开始")
+    # with suppress_stdout_stderr():
+    for i in tqdm(range(total_batches), desc="处理批次"):
+        batch = smiles_list[i*batch_size : (i+1)*batch_size]
+        
+        try:
+            # 获取嵌入表示
+            batch_repr = clf.get_repr(batch, return_atomic_reprs=False)
+            
+            # 将结果存入字典 (使用CLS token作为分子表示)
+            for idx, smiles in enumerate(batch):
+                embeddings_dict[smiles] = batch_repr['cls_repr'][idx]
+                
+        except Exception as e:
+            print(f"处理批次 {i+1}/{total_batches} 时发生错误: {str(e)}")
+            # 记录处理失败的SMILES
+            error_smiles.extend(batch)
+        
+    # 保存嵌入结果
+    with open(output_file, 'wb') as f:
+        pickle.dump(embeddings_dict, f)
+    
+    print(f"嵌入生成完成！共处理 {len(smiles_list)} 个SMILES，"
+          f"{len(smiles_list) - len(error_smiles)} 个成功，"
+          f"{len(error_smiles)} 个失败。")
+    print(f"嵌入结果已保存至 {output_file}")
+    
+    if error_smiles:
+        print(f"处理失败的SMILES已记录。")
+    
+    return embeddings_dict
+
+# 使用示例
+if __name__ == "__main__":
+    
+    # 假设unique_smiles是你的SMILES列表
+    unique_smiles = [
+    "CC(=O)OC1=CC=CC=C1C(=O)O",  # Aspirin
+    "CN1C=NC2=C1C(=O)N(C(=O)N2C)C" # Caffeine
+    ]  # 示例SMILES列表
+    unique_smiles = pd.read_csv('./LINCS2020/LINCS2020_smiles.csv')['SMILES'].tolist()
+
+    # UniMol V1
+    # embeddings = get_unimol_embeddings(
+    #     smiles_list=unique_smiles,
+    #     output_file="embeddings/UniMol_emb512.pkl",
+    #     model_name='unimolv1',
+    #     model_size='84m',
+    #     remove_hs=False,
+    #     batch_size=32
+    # )
+    
+    # UniMol V2
+    embeddings = get_unimol_embeddings(
+    smiles_list=unique_smiles,
+    output_file="embeddings/UniMolV2_emb1024.pkl", # save path
+    model_name='unimolv2',      # 修改为 v2 版本
+    model_size='310m',          # 指定 v2 模型大小（根据需要选择）
+    remove_hs=False,
+    batch_size=32
+    )
+
+    # 打印样例嵌入
+    sample_smiles = unique_smiles[0]
+    print(f"SMILES: {sample_smiles}")
+    print(f"嵌入向量: {embeddings[sample_smiles]}")
+    print(f"嵌入维度: {len(embeddings[sample_smiles])}")
+
+

upload.py

@@ -0,0 +1,21 @@
+from huggingface_hub import HfApi
+
+print("Initializing upload...")
+
+api = HfApi()
+
+# 你的配置
+src_folder = "/data/boom/MVCBench/Github/data"  # 本地源文件夹
+repo_id = "Boom5426/MVCBench"
+
+try:
+    api.upload_large_folder(
+        folder_path=src_folder,
+        repo_id=repo_id,
+        repo_type="dataset",
+        # 注意：这里去掉了 path_in_repo
+        # upload_large_folder 会默认把 src_folder 里的内容直接铺在仓库根目录下
+    )
+    print("Upload completed successfully!")
+except Exception as e:
+    print(f"An error occurred: {e}")