Skip to content

pharmbio/phospholipidosis_project

Repository files navigation

Phospholipidosis Project

License: MIT

A codebase for training and evaluating of predicting phospholipidosis readout from Cell Painting data and molecular descriptors:

  • CellProfiler & DeepProfiler morphological feature pipelines
  • Compound embedding & classification workflows
  • Multimodal fusion of image + chemical data
  • Conformal prediction for uncertainty estimation

This project is under development.


Repository Layout

General scripts

  • classification.py — train single‐cell classifiers
  • generate_crops.py — extract and load image crops for CP pipeline
  • run_training.nf — Nextflow wrapper for classification

/CP – CellProfiler (Petter)

  • classification.py
  • generate_crops.py
  • Grit_check.ipynb
  • run_training.nf

/Compounds – Chemical embeddings & models (Petter + Benjamin)

  • compound_classification.py
  • compounds_embedding.ipynb

/DP – DeepProfiler (Petter + Benjamin)

  • DP_exploration.ipynb
  • DP_exploration_conformal.ipynb
  • DP_exploration_conformal_site.ipynb
  • DP_regression.ipynb

/multimodal – Image + compound fusion (Benjamin)

  • dataset.py
  • model.py, model_v2.py
  • train.py
  • optuna_tuning.py & sweep_config.yaml
  • wandb_sweep.py
  • utils.py

About

Data analysis and models for predicting phospholipidosis readouts from cell painting

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •  

Languages