A Massively Parallel Trafficking Assay Accurately Predicts Loss of Channel Function inKCNH2Variants

AbstractHigh throughput genomics has greatly facilitated identification of genetic variants. However, determining which variants contribute to disease causation is challenging with more than half of all missense variants now classified asvariants of uncertain significance(VUS). A VUS leaves patients and their clinicians unable to utilize the variant information in clinical decision-making. In long QT syndrome type 2,KCNH2channel function is directly associated with disease presentation. Therefore, functional phenotypingof KCNH2variants can provide direct evidence to aid variant classification. Here, we investigated the expression of all codon variants in exon 2 ofKCNH2using a massively parallel trafficking assay and for a subset of 458 single nucleotide variants compared the results with peak tail current density and gating using automated patch clamp electrophysiology. Trafficking could correctly classify loss of peak tail current density variants with an AUC reaching 0.94 compared to AUCs of 0.75 to 0.8 forin silicovariant classifiers. We suggest massively parallel trafficking assays can provide prospective and accurate functional assessment for all missense variants inKCNH2and most likely many other ion channels and membrane proteins.