Phenome-wide association study of loci harboring de novo tandem repeat mutations in UK Biobank exomes

BackgroundTandem repeats (TRs) are a major source of variation in the human genome under-investigated by large-scale genetic studies. When present in coding regions, TRs may have large effects on protein structure and function contributing to health and disease. MethodsIn a family-based design of 39 European ancestry trios from the UK Biobank (UKB), the GangSTR and MonSTR methods were used to identify de novo TRs in whole-exome sequences. TRs were annotated for association with gene expression and gene set enrichment. The loci harboring de novo TRs were investigated in a phenome-wide association study in up to 148,607 unrelated UKB participants of European descent. Linear and logistic regression included age, sex, sexxage, age2, sexxage2, and ten within-ancestry principal components as covariates. TR loci were fine-mapped to identify likely causal associations. ResultsThere were 427 mutated TRs with a trend towards expansions versus contractions ({chi}2=5.46, df=1, P=0.019). These TRs were enriched for targets of the tumor suppressor microRNA-184 (21.1-fold, P=4.30x10-5). There were 123 TR-phenotype associations with posterior probabilities>0.95. These were related to body structure, cognition, and cardiovascular, metabolic, psychiatric, and respiratory outcomes. The most significant was between NCOA6-[GT]N and "ease of skin tanning" (beta=0.069, se=0.003, P=1.51x10-155). There were several loci with large likely causal effects on tissue microstructure, including the association of FAN1-[TG]N with carotid intima-media thickness (mean thickness: beta=5.22, se=1.08, P=1.22x10-6; maximum thickness: beta=6.44, se=1.32, P=1.12x10-6. ConclusionsCombined with the TR de novo mutational background characterized herein, TR-phenotype associations contribute clear and testable hypotheses of dose-dependent TR implications linking genetic variation and protein structure with health and disease outcomes.