CRISPhieRmix: a hierarchical mixture model for CRISPR pooled screens.

Area of Science:

• Genomics
• Bioinformatics
• Genetic screening

Background:

• Pooled CRISPR screens offer genome-wide interrogation of genetic causes for complex phenotypes.
• CRISPR interference (CRISPRi) and CRISPR activation (CRISPRa) screens face challenges with variable guide efficiency and rare off-target effects.
• Accurate analysis is crucial for reliable results in large-scale genetic screens.

Purpose of the Study:

• To introduce CRISPhieRmix, a novel statistical method for analyzing pooled CRISPRi/a screens.
• To resolve issues of guide efficiency variability and rare off-target effects in CRISPR screening data.
• To enhance the accuracy and statistical power of inferences from genome-wide CRISPR screens.

Main Methods:

• Development of CRISPhieRmix, a hierarchical mixture model.
• Incorporation of a broad-tailed null distribution to handle data characteristics.
• Application to large-scale pooled CRISPRi/a screening data.

Main Results:

• CRISPhieRmix demonstrates improved accuracy and power in pooled CRISPRi/a screens.
• The method effectively addresses challenges related to guide efficiency and off-target effects.
• Enhanced ability to identify genetic drivers of complex phenotypes.

Conclusions:

• CRISPhieRmix provides a robust solution for analyzing pooled CRISPRi/a screens.
• The method facilitates more reliable genome-wide discovery of genetic causes.
• Considerations for screen design, including guide number, are important for robust findings.