Exome array analysis of ischaemic stroke: results from a southern Swedish study

Background and purpose Genome-wide association (GWA) studies have identified a few risk loci for ischaemic stroke, but these variants explain only a small part of the genetic contribution to the disease. Coding variants associated with amino acid substitutions or premature termination of protein synthesis could have a large effect on disease risk. We performed an exome array analysis for ischaemic stroke. Methods Patients with ischaemic stroke (n = 2385) and control subjects (n = 6077) from three Swedish studies were genotyped with the Illumina HumanOmniExpressExome BeadChip. Single-variant association analysis and gene-based tests were performed of exome variants with minor allele frequency of < 5%. A separate GWA analysis was also performed, based on 700 000 genotyped common markers and subsequent imputation. Results No exome variant or gene was significantly associated with all ischaemic stroke after Bonferroni correction (all P > 1.8 x 10(-6) for single-variant and >4.15 x 10(-6) for gene-based analysis). The strongest association in single-variant analysis was found for a missense variant in the DNAH11 gene (rs143362381; P = 5.01 x 10(-6)). In gene-based tests, the strongest association was for the ZBTB20 gene (P = 7.9 x 10(-5)). The GWA analysis showed that the sample was homogenous (median genomic inflation factor = 1.006). No genome-wide significant association with overall ischaemic stroke risk was found. However, previously reported associations for the PITX2 and ZFHX3 gene loci with cardioembolic stroke subtype were replicated (P = 7 x 10(-15) and 6 x 10(-3)). Conclusions This exome array analysis did not identify any single variants or genes reaching the pre-defined significance level for association with ischaemic stroke. Further studies on exome variants should be performed in even larger, well-defined and subtyped samples.