The search for risky genes of ischemic stroke by the analysis of singular linkage disequilibrium matrix on imputed genotypic data

Abstract

In genome-wide association studies, single-marker and multi-marker approaches are distinguished. Single-marker methods analyze each single nucleotide polymorphism (SNP) individually, which allows for rapid identification of genetic markers associated with a phenotype. Multi-marker methods consider combinations of SNPs, which increases the power of tests and the overall informative value of the analysis, although it also increases the computational load. Previously, we proposed breaking down the linkage disequilibrium matrix into singular submatrices, using the matrix determinant as a measure of SNP correlation within a group. By applying this approach to the genotypes of individuals from Western Europe (4929 patients with ischemic stroke, 652 control individuals, 883,749 SNPs), we identified candidate genes for the disease that were not detected by single-marker methods. In this work, we reapplied this method to the same sample of individuals but using 10 times more SNPs obtained by enriching the original marker panels through imputation. The study results demonstrated the applicability of the determinant method for analyzing large SNP datasets, including those obtained via imputation, and suggested a potential role of neurogenesis in the development of ischemic stroke.