This repository contains Python code to reproduce the core analysis described in the manuscript: "A stability-weighted combined map for Cell Painting across regular and prolonged incubation".

1. Loads a single CSV file that contains both exposure windows.
2. Keeps only compounds with at least N independent long-incubation iterations (optional filter).
3. Splits metadata vs numeric features.
4. Standardises features and runs PCA.
5. Builds a kNN graph in PCA space and performs Leiden clustering (Scanpy).
6. Computes within-window replicate stability (mean within-compound replicate distance in PCA space).
7. Builds a Combined dataset by selecting, for each compound, the exposure window with higher stability.
8. Exports cluster tables and (optionally) UMAP plots for Short / Long / Combined.

Outputs are saved to the folder specified in out_dir.

The input CSV should contain:

• metadata columns starting with Metadata_ (e.g., Metadata_Name, Metadata_Source, Metadata_MoA)
• numeric feature columns (CellProfiler features)

By default, the script uses Metadata_Source to split Short vs Long using string patterns:

• short: contains sh
• long: contains l

If your naming differs, update the patterns in the config.

• UMAP is used for 2D visualisation.
• The code sets a random seed for reproducibility.

• A stability-weighted combined map for Cell Painting.ipynb
• env/siriusrequirements.txt: full pip freeze shared by the supervisor (may include non-Windows packages).
• env/requirements_windows.txt: Windows-friendly requirements (same as above but with non-Windows lines removed).
• environment.yml: conda environment file that installs Python and then installs env/requirements_windows.txt.