A fast and efficient colocalization algorithm for identifying shared genetic risk factors across multiple traits

Abstract

Genome-wide association studies (GWAS) have identified thousands of genomic regions affecting complex diseases. The next challenge is to elucidate the causal genes and mechanisms involved. One approach is to use statistical colocalization to assess shared genetic aetiology across multiple related traits (e.g. molecular traits, metabolic pathways and complex diseases) to identify causal pathways, prioritize causal variants and evaluate pleiotropy. We propose HyPrColoc (Hypothesis Prioritisation for multi-trait Colocalization), an efficient deterministic Bayesian algorithm using GWAS summary statistics that can detect colocalization across vast numbers of traits simultaneously (e.g. 100 traits can be jointly analysed in around 1 s). We perform a genome-wide multi-trait colocalization analysis of coronary heart disease (CHD) and fourteen related traits, identifying 43 regions in which CHD colocalized with ≥1 trait, including 5 previously unknown CHD loci. Across the 43 loci, we further integrate gene and protein expression quantitative trait loci to identify candidate causal genes.

Beyond factor H: The impact of genetic-risk variants for age-related macular degeneration on circulating factor-H-like 1 and factor-H-related protein concentrations

“… 40 Nevertheless, there are still questions that remain unanswered by the present study and could be addressed through the use of recently developed analytical tools to perform, for example, multivariate analyses of the FHR protein concentrations as well as conditional and genetic colocalization analyses. 41 , 42 , 43 , 44 , 45 At present, given the relatively moderate statistical power of our study, we did not fully disentangle the relative role of the concentration of each FHR protein on AMD risk and/or narrowing down the specific CFH genes and genetic variants that are likely to be involved in the causal cascade with AMD. There was limited pairwise correlation between the FHR protein concentrations in controls ( Figure 2 ), and the GWAS signals of the FHR protein concentrations are not in high LD with each other, except for the top signals of FHR-2 and FHR-4 (R 2 = 0.83, D′ = 0.95, Table S6 ) and for both the top signal of FHR-1 and the secondary signal of FHR-3 (R 2 = 0.96, D′ = 0.99).…”

A Bayesian model selection approach to mediation analysis

“…Emerging MR methods relax the assumption of no direct effect and may be useful in this context, but this field is rapidly developing and a full review of these methods is beyond the scope of this paper. A different, non-causal method for evaluating the relationship between M and Y is colocalization analysis [38,39,40,41,42], which evaluates if two variables (M and Y) share a common X, when only one of many independent variables are under consideration. This approach is typically motivated by fine-mapping genetic variants (i.e., evaluating candidate X, rather than candidate M), and it does not establish a causal relationship between M and Y in the way that mediation analysis can.…”

Investigating the genetic architecture of eye colour in a Canadian cohort

“…We performed colocalization analyses with hyprcoloc (38) of the eye colour meta-analysis with melanocyte gene expression and methylation cis-QTLs (eQTLs, meQTLs, respectively) to explore if there were shared causal signals (See Methods for details). Through the colocalization of GWAS with eQTLs, we identified a region overlapping OCA2 and AC090696.2 (the latter being a transcript which partially overlaps OCA2) (Table 2), in which the candidate marker is rs12913832.…”