| from dash import Dash, Input, Output, callback, dcc, html |
| import copy |
| import plotly.graph_objects as go |
| import pandas as pd |
|
|
|
|
| |
| file = open('./sankey_data_y145.pkl', 'rb') |
|
|
| |
| data = pd.read_pickle(file) |
|
|
| df = pd.read_csv("./chatgpt_explanation.csv") |
|
|
| external_script = ["https://tailwindcss.com/", {"src": "https://cdn.tailwindcss.com"}] |
|
|
| app = Dash( |
| __name__, |
| external_scripts=external_script, |
| ) |
| app.scripts.config.serve_locally = True |
|
|
| app.layout = html.Div( |
| [ |
| html.Div( |
| [ |
| html.Div( |
| [ |
| html.H1("BioXNet: a biologically inspired neural network for deciphering anti-cancer drug response in precision medicine", className="text-black text-4xl font-bold pb-8 pt-32"), |
| html.H2("https://doi.org/10.1101/2024.01.29.576766", className="text-black flex text-left text-lg font-semibold pb-8"), |
| html.H3("For an improved experience, it is recommended to use either Chrome or Edge.", className="text-black flex text-left text-lg font-semibold pb-8"), |
| html.Hr(className="border-t-2 border-slate-500") |
| ],className="max-w-6xl flex flex-col justify-center mx-auto px-8 mx-8 pb-4 bg" |
| ) |
| ],id="background-container" |
| ), |
| html.Div( |
| [ |
|
|
|
|
| html.Div( |
| [ |
| html.Div( |
| [ |
| html.H5("Drug Name", className="text-lg font-medium pb-2 text-slate-500"), |
| dcc.Dropdown( |
| id="drugname", |
| options=sorted(df["drug_name"].unique()), |
| value='bevacizumab', |
| className="w-80" |
| ), |
| ], className="flex flex-col" |
| ), |
|
|
| html.Div( |
| [ |
| html.H5("Cancer type in TCGA", className="text-lg font-medium pb-2 text-slate-500"), |
| dcc.Dropdown( |
| id="tcga", options=sorted(df["tcga_id"].unique()), |
| value='BRCA', |
| className="w-80" |
| ), |
| ], className="flex flex-col" |
| ), |
| ], className="flex flex-row max-md:flex-col gap-20 items-center mx-auto" |
| ), |
|
|
| ],className="max-w-6xl flex flex-col justify-center mx-auto mt-8 px-8 mx-8" |
| ), |
| html.H3("Interpretable paths", className="px-8 mx-8 flex mx-auto max-w-6xl font-extrabold text-transparent text-4xl bg-clip-text bg-gradient-to-r from-purple-400 to-pink-600 pb-8 text-left pt-12"), |
|
|
| html.Div( |
| [ |
| dcc.Graph(id="graph", style={'width': '60vw', 'height': '70vh','align':'justify-center'}), |
| ],className="max-w-6xl flex flex-col items-center mx-auto mt-8 px-8 mx-8" |
| ), |
| html.Div( |
| [ |
| html.H3("Readable interpretations generate by ChatGPT", className="text-left font-extrabold text-transparent text-4xl max-md:text-2xl bg-clip-text bg-gradient-to-r from-purple-400 to-pink-600 pb-12"), |
| html.P(id="explaination", className="text-lg max-md:text-base leading-10 text-justify") |
| ],className="max-w-6xl flex flex-col mx-auto mt-8 px-8 mx-8 pb-32" |
| ) |
| ] |
| ) |
|
|
|
|
| @callback( |
| Output("drugname", "options"), |
| Input("tcga", "value"), |
| ) |
| def chained_callback_drugname(tcga): |
|
|
| dff = copy.deepcopy(df) |
|
|
| if tcga is not None: |
| dff = dff.query("tcga_id == @tcga") |
|
|
| return sorted(dff["drug_name"].unique()) |
|
|
|
|
|
|
| @callback( |
| Output("tcga", "options"), |
| Input("drugname", "value"), |
| ) |
| def chained_callback_tcga(drugname): |
|
|
| dff = copy.deepcopy(df) |
|
|
| if drugname is not None: |
| dff = dff.query("drug_name == @drugname") |
|
|
| return sorted(dff["tcga_id"].unique()) |
|
|
|
|
| @callback( |
| Output("graph", "figure"), |
| Input("drugname", "value"), |
| Input("tcga", "value"), |
| ) |
| def graph_chart(drugname, tcga): |
|
|
| comb_name= tcga+'_'+drugname |
|
|
| node = data[comb_name]['node'] |
| |
| link = data[comb_name]['link'] |
|
|
| fig = go.Figure(go.Sankey(link=link, node=node,)) |
| |
| fig.update_layout( |
| margin=dict( l= 0, r=0, t=8,b= 0, autoexpand=False,), |
| font= { |
| "size": 12, |
| "family": "Helvetica" |
| }, |
| paper_bgcolor='rgba(0,0,0,0)', |
| plot_bgcolor='rgba(0,0,0,0)' |
| ) |
|
|
| return fig |
|
|
| @callback( |
| Output("explaination", "children"), |
| [Input("tcga", "value"), Input("drugname", "value")] |
| ) |
| def update_explanation(tcga, drugname): |
| dff = copy.deepcopy(df) |
|
|
| if tcga is not None: |
| dff = dff.query("tcga_id == @tcga") |
|
|
| if drugname is not None: |
| dff = dff.query("drug_name == @drugname") |
|
|
| explanation = dff["explanation"].values[0] if not dff.empty else "" |
|
|
| return explanation |
|
|
|
|
| if __name__ == "__main__": |
| app.run(debug=False, host="0.0.0.0", port=7860) |
