Quantifying rare, deleterious variation in 12 human cytochrome P450 drug-metabolism genes in a large-scale exome dataset

Abstract

The study of genetic influences on drug response and efficacy ('pharmacogenetics') has existed for over 50 years. Yet, we still lack a complete picture of how genetic variation, both common and rare, affects each individual's responses to medications. Exome sequencing is a promising alternative method for pharmacogenetic discovery as it provides information on both common and rare variation in large numbers of individuals. Using exome data from 2203 AA and 4300 Caucasian individuals through the NHLBI Exome Sequencing Project, we conducted a survey of coding variation within 12 Cytochrome P450 (CYP) genes that are collectively responsible for catalyzing nearly 75% of all known Phase I drug oxidation reactions. In addition to identifying many polymorphisms with known pharmacogenetic effects, we discovered over 730 novel nonsynonymous alleles across the 12 CYP genes of interest. These alleles include many with diverse functional effects such as premature stop codons, aberrant splicesites and mutations at conserved active site residues. Our analysis considering both novel, predicted functional alleles as well as known, actionable CYP alleles reveals that rare, deleterious variation contributes markedly to the overall burden of pharmacogenetic alleles within the populations considered, and that the contribution of rare variation to this burden is over three times greater in AA individuals as compared with Caucasians. While most of these impactful alleles are individually rare, 7.6-11.7% of individuals interrogated in the study carry at least one newly described potentially deleterious alleles in a major drug-metabolizing CYP.

Figure 1.

Distribution of exonic variation across the 12 drug-metabolizing CYP genes separated by variant consequence. Variant types (missense, nonsense, synonymous, splicesite, frameshift) were determined using SeattleSeq annotation. For genes that produce more than one known transcript (CYP2D6, CYP2C8, CYP3A4), annotation was based on the primary transcript.

Figure 2.

MAF for novel and known variants across the CYP-12 in EAs and AAs.

Figure 3.

MAF for all CYP-12 variants as well as for only nonsynonymous variants (missense, nonsense, splicespite, indels).