White matter hyperintensities (WMH) detected on MRI are commonly identified abnormalities in the adult brain, and are associated with a greater risk of stroke, dementia, and death. Genetic factors play a significant role in WMH etiology, yet, common genetic variants identified by GWAS explain little of the variance in WMH burden. Rare variants with larger effect on the phenotype may be identified from sequence analysis of the protein-coding region of the genome (exome). We sequenced the protein-coding regions of 16,541 genes in 2510 individuals of European or African ancestry from three NHLBI cohorts, and investigated whether putatively functional exomic variants were associated with WMH burden, either individually or in aggregate within a gene. Within each cohort, we used the SeqMeta R package to compute race-specific score statistics for each variant and genotypic covariance matrices within predefined gene regions. These were then combined by meta-analysis to generate single-variant and gene-based tests of association. Only missense, nonsense, and splice variants were included in the analyses. Analyses of 12,790 single variants with minor allele frequency (MAF)≥1% did not identify statistically significant associations based on a Bonferroni-corrected significance threshold. The most significantly associated variant was a common missense variant in the ELL gene (p=2.6x10-5). In the sequence kernel association test (SKAT), which included variants with MAF<5%, five genes were significantly associated with WMH burden (p<0.05/16,541). These have known function in phospholipid binding, transport, and signaling (3 genes); Abeta metabolism (1 gene); and have been previously implicated in microvascular complications of diabetes, hypertension, and obesity. Among 35 candidate genes mapping to five previously reported WMH GWAS loci, association of MRPL38 (chr17q25) was nominally significant by SKAT (p=0.011). This gene contained two missense variants (MAF~1%) also nominally significantly associated with WMH (p=0.040 and 0.017). This study suggests that rare and low frequency variants significantly influence WMH burden and that genes involved in cardiovascular health and disease play a role in WMH etiology.
Automatic clustering of white matter fibers in brain diffusion MRI with an application to genetics