Semisupervised Contrastive Learning for Bioactivity Prediction Using Cell Painting Image Data

This repo contains the code to reproduce results from our paper Semisupervised Contrastive Learning for Bioactivity Prediction Using Cell Painting Image Data. In this work, we demonstrate the application of SemiSupCon models for extracting meaningful features from Cell Painting data, to facilitate acurate bioactivity prediction. We used the loss functions from the Supervised Contrastive learning paper.

Baseline comparisson

Data

SemiSupCon(BBBC022) Embeddings

Baseline embeddings for the BBBC022 dataset

SemiSupCon(BBBC036) Embeddings

BBBC022 images processed with ImageJ