Addressing Challenges to Genomic Medicine: Effects of Rare and Common Variation on Arrhythmia and Pharmacogenetic Phenotypes

Abstract

The promise of genomic medicine has yet to be fully realized. In the realm of rare disease, one of the challenges faced by genomic medicine is that of variant interpretation. A majority of variants in clinically actionable genes either lack the data needed to ascertain their pathogenicity or have conflicting clinical interpretations, and thus are called variants of uncertain significance (VUS). Furthermore, variant discovery far outpaces functional assessment, one of the strongest criteria for variant classification. Another challenge faced by genomic medicine is in the realm of common disease: precise and reliable genome markers to predict phenotypes are lacking. In fact, even the utility of such predictors is unknown for some phenotypes.

To assist with variant classification, we probed the function of variants in two cardiac ion-channel genes associated with cardiac arrhythmia disorders, SCN5A and KCNE1, using medium- and high-throughput approaches, respectively. We studied 179 SCN5A in-frame insertions/deletions and missense variants, including 94 VUS, by automated patch clamp. Functional data obtained from our studies assisted in the classification of 51/94 VUS to likely pathogenic or likely benign. We also used deep mutational scanning to assess cell surface expression of 1,886 KCNE1 variants, identifying 277 loss-of-trafficking and 183 gain-of-trafficking variants. Our data set the stage to classify hundreds of KCNE1 VUS and understand the structure/function relationship in the protein.

To assess the benefit of using polygenic predictors to predict inter-individual variability in drug response, we calculated the heritability of 7 pharmacodynamic and 5 pharmacokinetic phenotypes across 8 different drugs using a Bayesian Hierarchical Mixed Model. We found 8/12 phenotypes to have a heritability >25%, indicating the feasibility for a polygenic predictor. Small-effect size variants had the largest contribution to phenotype heritability, indicating that current clinical approaches focusing on one, or a few, large effect alleles do not capitalize on the full potential of the genome to reduce adverse drug reactions and increase drug efficacy.

Our work highlights the role of the genome in the delivery of safe, effective and efficient healthcare for both rare and common diseases.