Improved estimation of cancer dependencies from large-scale RNAi screens using model-based normalization and data...

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Genome-scale RNAi viability screens offer insights into cancer vulnerabilities.
• Analyzing these screens is challenging due to batch effects, variable quality, and absolute dependency assessment difficulties.

Purpose of the Study:

• To improve the analysis of RNAi screening data for cancer vulnerability identification.
• To create a unified resource of cancer cell line genetic dependencies by integrating multiple large-scale datasets.

Main Methods:

• Incorporated cell line screen-quality parameters into the DEMETER2 analytical framework.
• Utilized hierarchical Bayesian inference for improved gene dependency estimation.
• Integrated three large RNAi screening datasets.

Main Results:

• DEMETER2 significantly enhances gene dependency estimates.
• The model shows improved identification of essential genes.
• Results demonstrate strong agreement with CRISPR/Cas9-based viability screens.

Conclusions:

• DEMETER2 provides a robust framework for analyzing RNAi screening data.
• The integrated resource represents the most comprehensive compilation of cancer cell line genetic dependencies.
• This work advances the understanding of cancer vulnerabilities through integrated genomic analysis.