Full release of LINCS Cell Painting dataset

Summary

This release contains finalized profile and metadata information for the CellProfiler-derived features from the LINCS Cell Painting dataset.

We assayed two batches of data and applied an image analysis pipeline using CellProfiler to segment cells and extract morphology features.
Using the CellProfiler output, we applied an image-based profiling pipeline to process the morphology profiles into an analysis-ready form.

In this repository, we provide the full image-based profiling pipeline we used to process the data, as well as most of the intermediate data.

Data counts

• Two batches ("2016_04_01_a549_48hr_batch1" and "2017_12_05_Batch2")
• 159,717,488 unique cells total

2016_04_01_a549_48hr_batch1

• 52,223 profiles (batch corrected)
• 10,752 unique perturbations
• 1,571 unique compounds
• 1 time point (48 hour)
• 7 doses
• 1 cell line (A549)
• 110,012,425 unique cells

2017_12_05_Batch2

• 51,447 profiles (batch corrected)
• 10,368 unique perturbations
• 349 unique compounds
• 3 time points (6 hour, 24 hour, 48 hour)
• 6 doses
• 3 cell lines (A549, MCF7, U2OS)
• 49,705,063 unique cells

Initial Processed Data Release

This is the very first release of 136 plates (384 well) of Cell Painting (Gustafsdottir et al. 2013) data derived from 1,571 compounds from the Drug Repurposing Hub (Corsello et al. 2017).

This release includes processed data alongside the processing code that produced the data. Raw microscopy images will be provided elsewhere at a later date. We also include processed metadata for each Drug Repurposing Compound. Metadata were originally downloaded from the CLUE resource portal at https://clue.io/repurposing.

See https://github.com/broadinstitute/lincs-cell-painting for specific data and processing details.