Predicting functional effects of ion channel variants using new phenotypic machine learning methods

Missense variants in genes encoding ion channels are associated with a spectrum of severe diseases. Variant effects on biophysical function correlate with clinical features and can be categorized as gain- or loss-of-function. This information enables a timely diagnosis, facilitates precision therapy, and guides prognosis. Functional characterization presents a bottleneck in translational medicine. Machine learning models may be able to rapidly generate supporting evidence by predicting variant functional effects. Here, we describe a multi-task multi-kernel learning framework capable of harmonizing functional results and structural information with clinical phenotypes. This novel approach extends the human phenotype ontology towards kernel-based supervised machine learning. Our gain- or loss-of-function classifier achieves high performance (mean accuracy 0.853 SD 0.016, mean AU-ROC 0.912 SD 0.025), outperforming both conventional baseline and state-of-the-art methods. Performance is robust across different phenotypic similarity measures and largely insensitive to phenotypic noise or sparsity. Localized multi-kernel learning offered biological insight and interpretability by highlighting channels with implicit genotype-phenotype correlations or latent task similarity for downstream analysis.Author summary: Genetic variants can impact the function of voltage-gated ion channels, which are transmembrane channels that are important for the electrical signalling of neurons. Individuals affected by these variants may have severe neurological or cardiac disease. Knowing the functional effects of variants can help us improve the diagnosis and clinical care for these individuals, and may also enable precision medicine aimed at correcting channel function. Machine learning methods are capable of predicting these functional effects, but previous models have only included information from protein sequence and structure. Here, we have developed a novel algorithm that can integrate clinical information by using the Human Phenotype Ontology (HPO), a standardized vocabulary of phenotypic features encountered in human disease. These different sources of information are used to train a multi-kernel support-vector machine. Our results suggest that this model is robust, interpretable, and more accurate than previous methods. Variant functional effect prediction engines will be useful tools for clinical geneticists, neurologists and biophysicists.