Predicting the functional effects of voltage-gated potassium channel missense variants with multi-task learning

Abstract

PurposeVariants in genes encoding voltage-gated potassium channels are associated with a broad spectrum of neurological diseases including epilepsy, ataxia, and intellectual disability. Knowledge of the resulting functional changes, characterized as overall ion channel gain- or loss-of-function, is essential to guide clinical management including precision medicine therapies. However, for an increasing number of variants, little to no experimental data is available. New tools are needed to evaluate variant functional effects.MethodsWe catalogued a comprehensive dataset of 959 functional experiments across 19 voltage-gated potassium channels, leveraging data from 782 unique disease-associated and synthetic variants. We used these data to train a taxonomy-based multi-task learning support vector machine (MTL-SVM), and compared performance to a baseline of standard SVMs.ResultsMTL-SVM maintains channel family structure during model training, improving overall predictive performance (mean balanced accuracy 0.729 ± 0.029, AU-ROC 0.757 ± 0.039) over baseline (mean balanced accuracy 0.645 ± 0.041, AU-ROC 0.710 ± 0.074). We can obtain meaningful predictions even for channels with few known variants (KCNC1, KCNQ5).ConclusionOur model enables functional variant prediction for voltage-gated potassium channels. It may assist in tailoring current and future precision therapies for the increasing number of patients with ion channel disorders.

Learning with phenotypic similarity improves the prediction of functional effects of missense variants in voltage-gated sodium channels

“…Each variant was identified by its channel, sequence position, and amino acid substitution. Sequence- and structure-based features were extracted and pre-processed as previously described (28). These 62 features included amino acid physicochemical properties, radicality of substitution, paralog and ortholog conservation, secondary and tertiary structure, residue accessibility, disordered regions, and protein-protein interaction or binding sites.…”

“…the hierarchical taxonomy of voltage-gated sodium channels across the channel superfamily, is represented as an alignment-based tree distance measure and enables multi-task (parallel transfer) learning. Both methods have been previously described (28). We can consider each of these three data input types of our variants (channel, sequence/structure, and phenotype) as different measurement modalities with individual kernel matrices for each modality.…”

“…Multi-task learning (MTL) SVM embeds explicit prior knowledge on the structural and evolutionary pairwise similarity between ion channels in the feature space induced by our kernel function, which enables the model to assign more weight when learning from variants of similar channels (28). MTL-SVM outperformed the baseline models across all metrics (Table 1a).…”

“…They used known genotype-phenotype correlations to infer likely functional labels for supervised learning, which may not make full use of available clinical information and risks introducing bias and uncertainty to the training data set. In this study, we extend our own previous work on kernel-based multi-task learning methods applied to potassium channel variants (28) by designing a framework capable of integrating functional, structural and clinical data while leveraging prior knowledge on channel family structure.…”

Current practice in diagnostic genetic testing of the epilepsies

“…If available, results of functional data on the identified variant should be mentioned in the report. If no functional data is available, it may be possible to predict the variant effect with the growing number of gene-family specific in silico prediction tools [49,50]. This is especially relevant if both loss-and gain-offunction variants are described for the gene of interest, as they often require different precision medicine approaches.…”