Created on 10th June 2017
This paper has been published in GigaScience.
Summary: Identifying phenotypic correlations between complex traits and diseases can provide useful etiological insights. Restricted access to individual-level phenotype data makes it difficult to estimate large-scale phenotypic correlation across the human phenome. Here, we present a novel method, PhenoSpD, to estimate phenotypic correlations using genome-wide association study (GWAS) summary statistics from the same sample and utilizes the correlations to inform correction of multiple testing for human GWAS studies. In a case study using GWAS summary results, PhenoSpD suggested 324.4 independent tests among 452 metabolites, which is close to the 296 independent tests estimated using true phenotypic correlation. We then estimated 120,713 pair-wise phenotypic correlations among 24 categories of human traits and diseases (total 862 traits) and further corrected multiple testing for these traits using PhenoSpD. The atlas of phenotypic correlations provides novel insights into the relationships between traits, while the PhenoSpD multiple testing correction function provides a simple and conservative way to reduce dimensionality for GWAS of complex molecular traits. Availability: R codes and Documentation for PhenoSpD V1.0.0 is available online (https://github.com/MRCIEU/PhenoSpD).Show more
|Donghyung Lee||Completed||29 Aug 2017||View review|
|Aaron Day-Williams||Completed||8 Sep 2017||View review|
|3||Completed||21 May 2018||View review|