MultiXcan

MultiXcan integrates GWAS and eQTL data to perform multivariate gene-level association analysis that prioritizes candidate target genes and elucidates biological mechanisms underlying GWAS loci.

Key Features:

• Statistical Methodology: Employs multivariate regression that leverages the sharing of eQTLs across tissues and accounts for the correlation structure among predictors.
• Enhanced Detection Capability: Synthesizes evidence across multiple eQTL panels to detect a larger set of significantly associated genes than analysis of each panel separately.
• S-MultiXcan Extension: Provides a summary-statistic-based extension (S-MultiXcan) that yields results concordant with the individual-level multivariate approach when linkage disequilibrium (LD) is well matched and can be calibrated against individual-level outputs.

Scientific Applications:

• Gene-Level Association Studies: Prioritizes candidate genes at GWAS loci to aid interpretation of the genetic basis of complex traits.
• Therapeutic Target Identification: Integrates diverse eQTL data across tissues to improve identification of putative therapeutic targets and implicated biological pathways.

Methodology:

Uses multivariate regression to integrate multiple eQTL panels, applies a multivariate model-based framework that can be calibrated with individual-level results, operates via a summary-statistic extension (S-MultiXcan) when LD is well matched, and has been applied to simulated and real traits including analyses using UK Biobank data.