Key result
A machine learning model trained on sequence- and structure-based features predicted loss-of-function versus gain-of-function effects of missense variants in sodium and calcium channels with an ROC of 0.85.
Why the study?
Experimental studies of voltage-gated sodium and calcium channel function are important for understanding disease manifestations and drug response, but are laborious and typically focus on few variants at a time.
Effect estimate: ROC 0.85
A machine learning model can accurately predict gain- or loss-of-function effects of missense variants in voltage-gated sodium and calcium channels, potentially facilitating precision medicine.
May aid variant interpretation in channelopathies; leaves open prospective clinical validation.
Summary Malfunctions of voltage-gated sodium and calcium channels (SCN and CACNA1 genes) have been associated with severe neurologic, psychiatric, cardiac and other diseases. Altered channel activity is frequently grouped into gain or loss of ion channel function (GOF or LOF, respectively) which is not only corresponding to clinical disease manifestations, but also to differences in drug response. Experimental studies of channel function are therefore important, but laborious and usually focus only on a few variants at a time. Based on known gene-disease-mechanisms, we here infer LOF (518 variants) and GOF (309 variants) of likely pathogenic variants from disease phenotypes of variant carriers. We show regional clustering of inferred GOF and LOF variants, respectively, across the alignment of the entire gene family, suggesting shared pathomechanisms in the SCN/CACNA1 genes. By training a machine learning model on sequence- and structure-based features we predict LOF- or GOF- associated disease phenotypes (ROC = 0.85) of likely pathogenic missense variants. We then successfully validate the GOF versus LOF prediction on 87 functionally tested variants in SCN1/2/8A and CACNA1I (ROC = 0.73) and in exome-wide data from > 100.000 cases and controls. Ultimately, functional prediction of missense variants in clinically relevant genes will facilitate precision medicine in clinical practice.
No takes yet. Share an insight, caveat, or question.
Heyne et al. (2019) studied Channelopathies (SCN and CACNA1 gene variants). Machine learning model (FunNCion) vs. Known functional effects from electrophysiology or disease phenotypes was evaluated on Prediction of LOF- or GOF-associated disease phenotypes (ROC 0.85). A machine learning model trained on sequence- and structure-based features predicted loss-of-function versus gain-of-function effects of missense variants in sodium and calcium channels with an ROC of 0.85.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: