POEMColoc

POEMColoc estimates colocalization probabilities between genetic variants and phenotypes using complete or limited summary statistics to evaluate whether the same causal variant underlies GWAS and eQTL signals.


Key Features:

  • Approximation of coloc: Implements an approximation of the coloc method to assess shared causality when full summary statistics are not available for both traits.
  • Imputation using LD from a reference panel: Imputes missing summary statistics by leveraging linkage disequilibrium (LD) structure from a reference panel.
  • Support for single top SNP data: Operates with datasets that contain either full summary statistics or only the single reported top SNP for a trait.
  • Application to GWAS and eQTL data: Designed to test GWAS Catalog hits against GTEx expression quantitative trait loci (eQTL) data for colocalization analysis.
  • Simulation-based validation: Performance has been evaluated via simulations showing comparable ability to identify shared causality to the coloc method.
  • Factors affecting accuracy: Accuracy is influenced by multiple independent causal variants in a region, imputation from a limited subset of typed variants, and modestly by mismatched ancestry in the reference panel.

Scientific Applications:

  • Colocalization of GWAS Catalog and GTEx eQTLs: Estimating colocalization between GWAS Catalog entries and GTEx eQTLs to link trait-associated loci to gene expression.
  • Understanding genetic mechanisms: Determining whether the same variant affects multiple phenotypes to strengthen evidence for shared genetic mechanisms beyond shared significance.
  • Prioritizing disease-relevant genes and drug targets: Identifying trait–gene pairs enriched in disease-relevant tissues and associations relevant to approved drug target–indication pairs.

Methodology:

Accepts full summary statistics or only a single top SNP; imputes missing summary statistics using LD structure from a reference panel; performs colocalization analysis between imputed statistics and complete data for a second trait; validation performed by simulation comparisons to coloc.

Topics

Details

License:
Apache-2.0
Programming Languages:
R
Added:
1/18/2021
Last Updated:
1/24/2021

Operations

Publications

King EA, Dunbar F, Davis JW, Degner JF. Estimating colocalization probability from limited summary statistics. Unknown Journal. 2020. doi:10.1101/2020.05.19.104927.