- Upload a CSV (like
activity_CHEMBL202_ERG.csv). - The CSV file can be generated from any original file containing these type of columns (ID [e.g. CHEMBLID, CID], SMILES, any activity in form of pchembl or 0/1 class labels).
- Keeps
pchembl_value(if present) and only feature columns starting withfr_orERG.... - Drops zero-variance feature columns.
- Creates
classwherepchembl_value >= 6.5→ 1 else 0. - Sorts compounds descending by
pchembl_value(fallback:class). - Renders a normalized (0..1) heatmap for the feature matrix.
- Shows the activity column as a separate strip on the left.
From this folder:
python -m pip install -r requirements.txt
streamlit run app.py