File size: 6,341 Bytes
bcd1195
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
149
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
188
189
190
191
192
193
194
195
196
197
198
199
import colorsys
import pathlib
import random
import textwrap

import numpy as np
import pandas as pd
import streamlit as st

BASE_DIR = pathlib.Path(__file__).parent
DATA_DIR = BASE_DIR / "data"


@st.cache_data
def latent_load_data():
    reactome_dims = DATA_DIR / "all_reactome_results.parquet"
    corum_dims = DATA_DIR / "all_corum_results.parquet"
    drug_dims = DATA_DIR / "all_drug_results.parquet"

    reactome_df = pd.read_parquet(reactome_dims)
    corum_df = pd.read_parquet(corum_dims)
    drug_df = pd.read_parquet(drug_dims)

    reactome_df["feature"] = reactome_df["reactome_pathway"]
    reactome_df = reactome_df.drop(
        columns=["model", "latent_dim_total", "init", "seed", "z", "reactome_pathway"]
    )

    corum_df["feature"] = corum_df["reactome_pathway"]
    corum_df = corum_df.drop(
        columns=["model", "latent_dim_total", "init", "seed", "z", "reactome_pathway"]
    )

    drug_df["feature"] = drug_df["name"]
    drug_df = drug_df.drop(
        columns=["model", "latent_dim_total", "init", "seed", "z", "name"]
    )

    meta_cols = ["ModelID", "OncotreePrimaryDisease"]
    reactome_meta = reactome_df[meta_cols].drop_duplicates()
    corum_meta = corum_df[meta_cols].drop_duplicates()
    drug_meta = drug_df[meta_cols].drop_duplicates()

    reactome_matrix = reactome_df.pivot(
        index="ModelID", columns="feature", values="latent_score"
    )
    corum_matrix = corum_df.pivot(
        index="ModelID", columns="feature", values="latent_score"
    )
    drug_matrix = drug_df.pivot(
        index="ModelID", columns="feature", values="latent_score"
    )

    reactome_matrix = reactome_matrix.merge(reactome_meta, on="ModelID", how="left")
    corum_matrix = corum_matrix.merge(corum_meta, on="ModelID", how="left")
    drug_matrix = drug_matrix.merge(drug_meta, on="ModelID", how="left")

    return reactome_matrix, corum_matrix, drug_matrix


@st.cache_data
def spider_load_data():
    files = {
        "Reactome Pathways": DATA_DIR / "all_reactome_results.parquet",
        "CORUM Complexes": DATA_DIR / "all_corum_results.parquet",
        "Drug Responses": DATA_DIR / "all_drug_results.parquet",
    }
    feature_colnames = {
        "Reactome Pathways": "reactome_pathway",
        "CORUM Complexes": "reactome_pathway",
        "Drug Responses": "name",
    }
    score_col = "pathway_score"

    global_max = 0
    dfs = {}
    for name, file in files.items():
        df = pd.read_parquet(file)
        dfs[name] = df
        if df[score_col].max() > global_max:
            global_max = df[score_col].max()

    return dfs, global_max, feature_colnames


def clean_label(x: str) -> str:
    if "Polymerase Switching" in x:
        return "Polymerase Switching"
    if "Negative Regulation Of MET" in x or "Negative Regulation of MET" in x:
        return "Negative MET Regulation"
    return truncate_label(x)


def truncate_label(
    label: str,
    max_chars_per_line: int = 22,
    max_total_words: int = 4,
    max_lines: int = 2,
) -> str:
    words = label.split()
    short = " ".join(words[:max_total_words])
    lines = textwrap.wrap(short, width=max_chars_per_line)
    if len(lines) > max_lines:
        lines = lines[:max_lines]
    wrapped = "\n".join(lines)
    if len(words) > max_total_words or (
        len(lines) >= max_lines
        and len(lines) < len(textwrap.wrap(short, width=max_chars_per_line))
    ):
        wrapped += "..."
    return wrapped


def place_labels_polar(ax, angles, labels, radius: float) -> None:
    for angle, label in zip(angles[:-1], labels):
        if 0 <= angle <= np.pi / 2 or 3 * np.pi / 2 <= angle <= 2 * np.pi:
            ha = "left"
        else:
            ha = "right"
        ax.text(
            angle, radius, label, fontsize=9, color="black", ha=ha, va="center", wrap=True
        )


def generate_random_palette(num_colors, seed=12):
    random.seed(seed)
    colors = []
    for _ in range(num_colors):
        h = random.random()
        lightness = random.uniform(0.2, 0.8)
        s = random.uniform(0.5, 1.0)
        colors.append(colorsys.hls_to_rgb(h, lightness, s))
    return colors


def generate_model_color_map(model_ids, seed=12):
    colors = generate_random_palette(len(model_ids), seed)
    return {model_id: color for model_id, color in zip(model_ids, colors)}


def make_radar(
    ax,
    df: pd.DataFrame,
    feature_col: str,
    title: str,
    global_max: float,
    score_col: str = "pathway_score",
    model_ids: list = ["ACH-000323", "ACH-002083", "ACH-002228"],
    num_circles: int = 5,
    max_processes: int = None,
) -> None:
    if model_ids == ["ACH-000323", "ACH-002083", "ACH-002228"]:
        model_colors = {
            "ACH-000323": "#00FFFF",
            "ACH-002083": "#FF69B4",
            "ACH-002228": "#800080",
        }
    else:
        model_colors = generate_model_color_map(model_ids, seed=12)

    top_by_model = (
        df.groupby("ModelID", group_keys=False)
        .apply(lambda x: x.nlargest(5, score_col))
        .reset_index(drop=True)
    )
    top_features = top_by_model[feature_col].drop_duplicates().tolist()
    if max_processes is not None:
        top_features = top_features[:max_processes]
    n_vars = len(top_features)
    labels = [clean_label(x) for x in top_features]

    angles = np.linspace(0, 2 * np.pi, n_vars, endpoint=False).tolist()
    angles += [angles[0]]

    circle_levels = np.linspace(0, global_max, num_circles + 1)[1:]
    for r in circle_levels:
        ax.plot(
            angles, [r] * (n_vars + 1), color="grey", linestyle="--", linewidth=1, zorder=0
        )

    for mid in model_ids:
        subset = df[df["ModelID"] == mid]
        merged = pd.DataFrame({feature_col: top_features}).merge(
            subset[[feature_col, score_col]], on=feature_col, how="left"
        )
        values = merged[score_col].fillna(0).tolist()
        values += [values[0]]
        ax.plot(angles, values, color=model_colors[mid], linewidth=3, label=mid)
        ax.scatter(
            angles, values, color=model_colors[mid], s=30, edgecolor="black", zorder=5
        )

    ax.set_ylim(0, global_max * 1.25)
    ax.set_xticks([])
    ax.set_yticks([])
    ax.set_aspect("equal")
    place_labels_polar(ax, angles, labels, global_max * 1.4)
    ax.set_title(title, fontsize=18, pad=30, fontweight="bold", y=1)