MeTAL enables multiparametric risk prediction for human KCNH2 variants

Abstract

BackgroundClinical interpretation of missense variants in the hERG potassium channel encoded by the KCNH2 gene remains a major challenge in inherited arrhythmia syndromes. Functional studies often rely on a minimal set of channel properties, mainly current amplitude measurements, which do not capture the multidimensional nature of channel gating and its impact on ventricular repolarization. We developed a multiscale computational framework to quantitatively link multiparametric channel dysfunction to ECG phenotypes. MethodsWe generated multiparametric electrophysiological profiles for KCNH2 variants using high-throughput patch-clamp, quantifying nine biophysical properties including conductance, voltage dependence, and gating kinetics. These parameters were incorporated into a modified formulation of IKr embedded in a human ventricular electrophysiology model. The resulting framework, termed MeTAL (Multiscale-enriched Transformation and Analysis for Long-QT), produces physiologically calibrated pseudo-ECGs enabling quantitative evaluation of QT dynamics. We systematically analyzed the contribution of individual and combined gating parameters and applied the model to 41 KCNH2 variants across ACMG classes, comparing simulated QTc values with clinical data. ResultsMultiparametric profiling revealed complex functional signatures in most variants, with concurrent gain- and loss-of-function effects affecting distinct gating processes. Simulations demonstrated that ventricular repolarization, though strongly determined by current amplitude, is substantially influenced by inactivation-related parameters, particularly the slope and voltage dependence of inactivation. Interaction analyses showed nonlinear relationships between gating parameters, explaining why variants with similar current density can produce divergent QT phenotypes. In heterozygous simulations, MeTAL reproduced clinically observed QTc distributions across variant classes and accurately predicted the repolarization regime (normal, long-QT, or short-QT) in most cases. ConclusionsMultiparametric integration of ion-channel function within a multiscale electrophysiological model enables mechanistic prediction of QT behavior beyond conductance-based metrics. This approach provides a scalable framework for interpretation of KCNH2 variants and solves the issue of risk stratification in inherited arrhythmia syndromes while offering new opportunities for variant-specific and pharmacological modeling of repolarization.