broadinstitute / profiling-resistance-mechanisms

Predicting pharmacodynamic responses to cancer drugs using cell morphology
BSD 3-Clause "New" or "Revised" License
7 stars 5 forks source link
cancer carpenter-lab cell-painting machine-learning morphology pharmacodynamics resistance

DOI

Discovering Morphological Markers of Drug Resistance

In this repository we analyze Cell Painting data generated from multiple cell line clones that were resistant or sensitive to bortezomib.

Citation

Kelley ME, Berman AY, Stirling DR, Cimini BA, Han Y, Singh S, Carpenter AE, Kapoor TM, Way GP. High-content microscopy reveals a morphological signature of bortezomib resistance. (2023) eLife; 12:e91362. DOI: https://doi.org/10.7554/eLife.91362.

Data collection and processing

We cultured a colon cancer cell line (HCT116), treated with a proteosome inhibitor (Bortezomib), and selected two resistant clones. We applied Cell Painting to these cell lines (in triplicate) under four conditions (DMSO, 0.7nm, 7nm, and 70nm Bortezomib).

The Cell Painting assay captures several cellular morphology features (described in more detail here). Our hypothesis was that morphological features could distinguish wildtype from resistant clones.

We processed the cell painting data using CellProfiler. We use CellProfiler to test quality control, segment images to extract nuclei, and measure features captured by cell painting.

This repository contains all image analysis pipelines and image-based profiling pipelines (see 0.generate-profiles).

Pilot analyses

This repository ingests the processed Cell Painting data and performs several downstream analyses.

Using the triplicate measurements, and two batches, we perform the following pilot analyses:

UMAP Batch Analysis

UMAP

T-test to Determine Morphological Differences

ttest

Reproducibility

We use conda to manage package versions. After installing conda, obtain all required packages:

conda env create --force --file environment.yml

# Activate environment
conda activate resistance-mechansisms

Clone the github repository. First, generate and enable SSH Keys if you haven't already.

# Then clone and enter repo
git clone git@github.com:broadinstitute/profiling-resistance-mechanisms
cd profiling-resistance-mechanisms

All analyses are presented in analysis.sh. To reproduce, perform the following:

./analysis.sh

Bug Reporting

Please file an issue with any questions or bug reports.

Internal documents

GDrive folder