Published September 26, 2022 | Version 1.0
Software Open

CellPainting-CNN model

Authors/Creators

  • 1. Biological Research Center, Szeged, Hungary
  • 2. Broad Institute of MIT and Harvard
  • 3. Carnegie Mellon University
  • 4. Harvard College
  • 5. Meta

Description

The supplementary model for the publication "Learning representations for image-based profiling of perturbations", if you use this model, please cite our publication.

It was trained on images of single-cells obtained with Cell Painting assay (488 treatments and 2 negative controls from 5 source datasets). The architechture used is EfficientNet. 

You can find instructions specifically related on profiling with this model in our DeepProfiler handbook. 

We recommend to use this model with DeepProfiler software. It is possible to use this model separately, but your Python environment should have the following package installed: 

efficientnet==1.1.1

 

Files

Files (19.2 MB)

Name Size Download all
md5:b36679578418ec7f856390e857213214
19.2 MB Download

Additional details

Related works

Is supplement to
Preprint: 10.1101/2022.08.12.503783 (DOI)