Agreement between two large pan-cancer CRISPR-Cas9 gene dependency data sets

Area of Science:

• Genomics
• Cancer Biology
• CRISPR-Cas9 Technology

Background:

• Genome-scale CRISPR-Cas9 screens systematically identify cancer dependencies and therapeutic targets.
• Assessing the reproducibility of these large-scale screens is crucial as more data becomes available.

Purpose of the Study:

• To formally assess the reproducibility of pan-cancer CRISPR-Cas9 viability screens conducted at the Broad and Sanger Institutes.
• To identify factors contributing to batch effects in CRISPR-Cas9 screening data.

Main Methods:

• Analysis of data from pan-cancer CRISPR-Cas9 viability screens from Broad and Sanger Institutes.
• Comparison of screen results across multiple metrics.
• Replication experiments to investigate batch effects.

Main Results:

• High concordance between Broad and Sanger Institute screen results, identifying both common and specific gene dependencies.
• Recovery of robust gene dependency biomarkers across datasets.
• Identification of reagent library and assay length as principal drivers of batch effects.

Conclusions:

• CRISPR-Cas9 viability screens from the Broad and Sanger Institutes yield robust and reproducible findings.
• Understanding and controlling experimental parameters like reagent library and assay length is key for reliable results.
• Reproducibility supports the use of these screens for identifying cancer dependencies and therapeutic targets.