PhenomeXcan: Mapping the genome to the phenome through the transcriptome

Abstract

Large-scale genomic and transcriptomic initiatives offer unprecedented ability to study the biology of complex traits and identify target genes for precision prevention or therapy. Translation to clinical contexts, however, has been slow and challenging due to lack of biological context for identified variant-level associations. Moreover, many translational researchers lack the computational or analytic infrastructures required to fully use these resources. We integrate genomewide association study (GWAS) summary statistics from multiple publicly available sources and data from Genotype-Tissue Expression (GTEx) v8 using PrediXcan and provide a user-friendly platform for translational researchers based on state-of-the-art algorithms. We develop a novel Bayesian colocalization method, fastENLOC, to prioritize the most likely causal gene-trait associations. Our resource, PhenomeXcan, synthesizes 8.87 million variants from GWAS on 4,091 traits with transcriptome regulation data from 49 tissues in GTEx v8 into an innovative, gene-based resource including 22,255 genes. Across the entire genome/phenome space, we find 65,603 significant associations (Bonferroni-corrected p-value of 5.5 x 10 -10 ), where 19,579 (29.8 percent) were colocalized (locus regional colocalization probability > 0.1). We successfully replicate associations from PheWAS Catalog (AUC=0.61) and OMIM (AUC=0.64). We provide examples of (a) finding novel and underreported genome-to-phenome associations, (b) exploring complex gene-trait clusters within PhenomeXcan, (c) studying phenome-to-phenome relationships between common and rare diseases via further integration of PhenomeXcan with ClinVar, and (d) evaluating potential therapeutic targets. PhenomeXcan (phenomexcan.org) broadens access to complex genomic and transcriptomic data and empowers translational researchers. Spain 41.

Probabilistic Colocalization of Genetic Variants from Complex and Molecular Traits: Promise and Limitations

“…and eQTL data, we analyze 4,091 complex trait datasets and the final release of the GTEx data (v8) from 49 tissues [18,6]. In total, we perform colocalization analysis on 200,459 trait-tissue pairs using fastENLOC.…”

“…In total, we perform colocalization analysis on 200,459 trait-tissue pairs using fastENLOC. The biological implications from the colocalization analysis, coupled with PrediXcan analysis [4], have been reported and discussed in [6]. In this section, we focus on the technical perspective of the colocalization analysis and provide a high-level summary of colocalization analysis over a wide range of complex traits with currently available GWAS and eQTL data sets.…”

“…Subsequently, utilizing the genetic association discoveries to explore the molecular mechanisms of complex disease etiology has become a standard practice in human genetics research. Various types of analytical approaches designed for the integrative analysis of data from expression quantitative trait loci (eQTL) mapping and genome-wide association analysis (GWAS) of complex traits have shown promise in implicating molecular pathways connecting genetic variations, molecular phenotype changes, and complex diseases [1,2,3,4,5,6].…”

“…We refer to this type as the locuslevel colocalization analysis. The second kind, illustrated by coloc [12], eCAVIAR [13], and ENLOC/fastENLOC [14,6], attempts to uncover colocalization signals at the single SNP resolution. We refer to this type as SNP-level colocalization analysis.…”

MendelVar: gene prioritization at GWAS loci using phenotypic enrichment of Mendelian disease genes

“…In general, Mendelian disease genes show enrichment among genes flanking the low p-value disease GWAS loci and their occurrence positively correlates with association strength (Chong et al, 2015). Widespread comorbidity has also been detected between Mendelian disease and complex disease (Blair et al, 2006) as well as cancer (Melamed, Emmett, Madubata, Rzhetsky, & Rabadan, 2015), which can potentially be driven by pleiotropy (Pividori et al, 2019). The established involvement of certain Mendelian disease genes in complex traits has started to become utilised in evaluating GWAS gene prioritisation algorithms Guo et al, 2019) and indeed, in gene prioritisation itself (Schlosser et al, 2020).…”