Multivariate Bayesian variable selection for multi-trait genetic fine mapping

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) identify numerous single-nucleotide polymorphisms (SNPs) linked to complex traits.
• Pinpointing causal variants from correlated SNPs in GWAS-enriched regions is a significant challenge.
• Current multi-trait fine mapping methods often fail to account for trait heterogeneity.

Purpose of the Study:

• To develop a novel multivariate Bayesian variable selection method for multi-trait fine mapping.
• To address the challenge of identifying causal variants for multiple correlated and potentially heterogeneous traits.
• To incorporate prior biological knowledge and identify trait-specific variant targets.

Main Methods:

• Developed a multivariate Bayesian variable selection approach for multi-trait fine mapping.
• Implemented multi-level selection and integrated prior biological information.
• Identified the optimal subset of traits targeted by specific variants.

Main Results:

• The novel method demonstrated superior performance compared to existing approaches in simulations.
• Comprehensive simulations mimicked realistic fine-mapping scenarios.
• Applied to UK Biobank data, the method identified critical causal variants for addictive behaviors and risk factors.

Conclusions:

• The proposed Bayesian method offers an advanced solution for multi-trait fine mapping.
• It effectively identifies causal variants in the presence of trait heterogeneity and group structures.
• The approach has significant implications for understanding the genetic architecture of complex, multi-faceted traits.