Comprehensive gene heritability estimation reveals the genetic architecture of rare coding variants underlying...

Area of Science:

• Genetics and Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Estimating heritability from rare coding variants using whole-exome sequencing (WES) data is challenging due to linkage disequilibrium (LD) and data sparsity.
• Existing methods often struggle to accurately capture the contribution of rare and ultra-rare variants to complex trait variation.

Purpose of the Study:

• To introduce FLEX (Fast, LD-aware Estimation of eXome-wide and gene-level heritability), a novel framework for estimating and partitioning heritability from WES data.
• To develop a scalable and flexible method that integrates all coding variants, from common to ultra-rare, while correcting for LD effects.

Main Methods:

• FLEX utilizes a unified model to incorporate all coding variants (common to ultra-rare) from WES data.
• The framework corrects for linkage disequilibrium (LD) effects to enhance heritability estimation accuracy.
• FLEX supports both individual-level and summary statistic data, ensuring computational efficiency for large-scale biobank datasets.

Main Results:

• Simulations demonstrated that FLEX is well-calibrated and provides accurate heritability estimates.
• Application to UK Biobank WES data (N=153,351) identified 64 significant gene-trait pairs, with rare coding variants contributing 38% of gene-level heritability on average.
• Incorporating rare and ultra-rare coding variants increased heritability estimates by 24.8% compared to imputed SNPs, with effect sizes approximately 18x larger.

Conclusions:

• FLEX offers an adaptable and accurate approach for quantifying gene-level heritability, advancing the understanding of complex trait genetic architecture.
• The study highlights the substantial role of rare and ultra-rare coding variants in complex traits, often missed by traditional methods.
• FLEX facilitates the discovery of trait-relevant genes by providing precise heritability estimates at the gene level.