Comprehensive allele genotyping in critical pharmacogenes reduces residual clinical risk in diverse populations.
Genomic-guided pharmaceutical prescribing is increasingly recognized as an important clinical application of genetics. Accurate genotyping of pharmacogenomic (PGx) genes can be difficult, owing to their complex genetic architecture involving combinations of SNPs and structural variation. Here we introduce the Helix PGx database, an open-source star allele, genotype, and resulting metabolic phenotype frequency database for CYP2C9, CYP2C19, CYP2D6, and CYP4F2, based on short-read sequencing of >86,000 unrelated individuals enrolled in the Helix DNA Discovery Project. The database is annotated using a pipeline that is clinically validated against a broad range of alleles and designed to call CYP2D6 structural variants with high (98%) accuracy. We find that CYP2D6 has greater allelic diversity than the other genes, manifest in both a long tail of low-frequency star alleles as well as a disproportionate fraction (36%) of all novel predicted loss-of-function variants identified. Across genes, we observe that many rare alleles (<0.1% frequency) in the overall cohort have 10x higher frequency in one or more subgroups with non-European genetic ancestry. Extending these PGx genotypes to predicted metabolic phenotypes, we demonstrate that >90% of the cohort harbors a high risk variant in one of the four pharmacogenes. Based on the recorded prescriptions for >30,000 individuals in the Healthy Nevada Project, combined with predicted PGx metabolic phenotypes, we anticipate that standard-of-care screening of these four pharmacogenes could impact nearly half of the general population.
Authors: Shishi Luo, Ruomu Jiang, Joseph Grzymski, William Lee, James Lu, Nicole L Washington