Analysis of 200,000 exome-sequenced UK Biobank subjects illustrates the contribution of rare genetic variants to hyperlipidaemia

Abstract A few genes have previously been identified in which very rare variants can have major effects on lipid levels. Weighted burden analysis of rare variants was applied to 200,000 exome sequenced UK Biobank subjects with hyperlipidaemia as the phenotype, with the strength of association characterised by the signed log 10 p value (SLP). With principal components included as covariates there was a tendency for genes on the X chromosome to produce strongly negative SLPs, and this was found to be due to the fact that rare X chromosome variants were identified less frequently in males than females. The test performed well when both principal components and sex were included as covariates and strongly implicated LDLR (SLP = 50.08) and PCSK9 (SLP = ?10.42) while also highlighting other genes previously found to be associated with lipid levels. Category-specific analyses of variants in these two genes revealed that, while there were loss of function variants with major effects, there were much larger numbers of protein-altering variants which had moderate effects on risk of hyperlipidaemia. Variants classified by SIFT as deleterious have on average a two-fold effect and their cumulative frequency is such that they are present in approximately 1.5% of the population. There was no evidence for association of HUWE1, which had produced statistically significant results in an earlier analysis of a subset of 50,000 exomes, and with hindsight this result had been caused by the excess of X gene variants in females. These analyses shed further light on the way that genetic variation contributes to risk of hyperlipidaemia and in particular that there are very many protein-altering variants which have on average moderate effects and whose effects can be detected when large samples of exome-sequenced subjects are available.