Skip to main content
Fig. 1 | BMC Bioinformatics

Fig. 1

From: Estimating colocalization probability from limited summary statistics

Fig. 1

Illustration of the POEMColoc method compared to coloc. Under \({H}_{1}\) or \({H}_{2}\), only one of the two traits has any causal SNP in the region. Under \({H}_{3}\), the two traits have two different causal SNPs. Under \({H}_{4}\) (colocalization) the two traits share a causal SNP. The first row of the plot illustrates the coloc method using \(p\)-values. Input data is \(p\)-value at each SNP in the region for each trait, and the output is a posterior probability of each hypothesis. The posterior probability of colocalization (\({H}_{4}\)) is shown on the plot. The second row illustrates input to the POEMColoc method. Full summary statistics are available for one trait only, and for the second trait only the position and \(p\)-value of the top SNP is known. We also require LD from the top SNP to at least some of the trait 1 SNPs to be known from a reference panel (shown below in red). We use the LD to impute missing \(p\)-values for input to coloc. Imputed \(p\)-values are shown in the bottom row. POEMColoc consists of applying coloc to the imputed datasets and outputting posterior probabilities of colocalization (shown in the bottom row)

Back to article page