Phylogenetic and Physicochemical Analyses Enhance the Classification of Rare Nonsynonymous Single Nucleotide Variants in Type 1 and 2 Long-QT Syndrome

Abstract

Background— Hundreds of nonsynonymous single nucleotide variants (nsSNVs) have been identified in the 2 most common long-QT syndrome-susceptibility genes ( KCNQ1 and KCNH2 ). Unfortunately, an ≈3% background rate of rare KCNQ1 and KCNH2 nsSNVs amongst healthy individuals complicates the ability to distinguish rare pathogenic mutations from similarly rare yet presumably innocuous variants. Methods and Results— In this study, 4 tools [(1) conservation across species, (2) Grantham values, (3) sorting intolerant from tolerant, and (4) polymorphism phenotyping] were used to predict pathogenic or benign status for nsSNVs identified across 388 clinically definite long-QT syndrome cases and 1344 ostensibly healthy controls. From these data, estimated predictive values were determined for each tool independently, in concert with previously published protein topology–derived estimated predictive values, and synergistically when ≥3 tools were in agreement. Overall, all 4 tools displayed a statistically significant ability to distinguish between case-derived and control-derived nsSNVs in KCNQ1 , whereas each tool, except Grantham values, displayed a similar ability to differentiate KCNH2 nsSNVs. Collectively, when at least 3 of the 4 tools agreed on the pathogenic status of C-terminal nsSNVs located outside the KCNH2/Kv11.1 cyclic nucleotide−binding domain, the topology-specific estimated predictive value improved from 56% to 91%. Conclusions— Although in silico prediction tools should not be used to predict independently the pathogenicity of a novel, rare nSNV, our results support the potential clinical use of the synergistic utility of these tools to enhance the classification of nsSNVs, particularly for Kv11.1’s difficult to interpret C-terminal region.