The genetic architecture of lipoprotein(a): simplicity, complexity, and mutability
The genetic architecture of lipoprotein(a): simplicity, complexity, and mutability
Mukamel, R. E.; Rubinacci, S.; Kamitaki, N.; Handsaker, R. E.; Hujoel, M. L. A.; Tang, D.; McCarroll, S. A.; Loh, P.-R.
AbstractHigh levels of lipoprotein(a) (Lp(a)) increase risk of cardiovascular diseases. Lp(a) concentration is primarily determined (>80%) by genetic variation at LPA, particularly by its large, size-polymorphic kringle IV-2 (KIV-2) repeat. However, because variation of the KIV-2 repeat is difficult to measure, the precise ways that genetic variants at LPA and other loci determine Lp(a) levels have been unclear, especially in individuals of non-European ancestry. Here, we developed new computational techniques to accurately genotype KIV-2 variation in whole-genome-sequenced population cohorts. Applying these methods to UK Biobank (n=490,414) and All of Us (n=245,394) provided insights into the genetic determinants of Lp(a) and the mutational dynamics of the KIV-2 repeat, which we found expands or contracts intra-allelically once every [~]500 generations. Variation in Lp(a) levels among people of all genetic ancestries was accurately predicted by a single, transferable genetic model capturing cis-epistatic effects of KIV-2 length and protein-altering polymorphisms in LPA. Controlling for predicted effects of LPA alleles revealed an additional, highly polygenic contribution to Lp(a) variation: LPA-adjusted Lp(a) levels associated with genetic variation at over 300 loci (P<5x10-8), much like other blood lipids. These results elucidate the genetic architecture and mutational origins of Lp(a) variation, showing how detailed characterization of a largely monogenic trait can provide surprising insights for complex trait genetics.