Joint genotypic and phenotypic outcome modeling improves base editing variant effect quantification.

Area of Science:

• Genetics and Genomics
• Molecular Biology
• Bioinformatics

Background:

• CRISPR base editing screens are valuable for studying disease variants but face challenges with variable editing efficiency and precision.
• Accurate assessment of variant-induced phenotypic effects is confounded by inconsistencies in base editing perturbations.

Approach:

• Developed an integrated pipeline for high-throughput ABE8e-SpRY base editing screens using an integrated reporter construct.
• Introduced BEAN, a Bayesian network, to account for per-guide editing outcomes and chromatin accessibility for improved variant impact estimation.
• Validated the pipeline's superior performance in variant classification and effect size quantification compared to existing tools.

Key Points:

• The pipeline accurately measures editing efficiency and outcomes alongside phenotypic consequences.
• BEAN improves variant impact estimation by integrating editing data and chromatin accessibility.
• Identified common variants affecting LDL uptake and implicated novel genes.
• Enabled quantitative prediction of missense variant effects on LDL-C levels via saturation base editing of LDLR.

Conclusions:

• This integrated pipeline significantly enhances the power of base editor screens for characterizing disease-associated variants.
• The approach provides a widely applicable method for improving variant impact assessment in genetic screens.
• Findings advance understanding of genetic factors influencing LDL metabolism and disease pathogenicity.