Predicting changes to I_Na from missense mutations in human SCN5A

Area of Science:

• Genetics
• Cardiology
• Computational Biology

Background:

• SCN5A gene mutations disrupt cardiac sodium current (I_Na), increasing risk for Brugada and long-QT syndromes.
• Predicting clinical severity of SCN5A variants is difficult due to complex genotype-phenotype relationships.

Purpose of the Study:

• To investigate the predictability of I_Na changes caused by SCN5A mutations using machine learning.
• To assess the potential of in-silico methods to predict functional consequences of SCN5A variants.

Main Methods:

• Compiled a dataset of nonsynonymous missense SCN5A mutations and their effects on I_Na.
• Applied machine-learning algorithms to predict I_Na alterations based on mutation data.

Main Results:

• Machine learning predicted I_Na changes with higher sensitivity and specificity than existing clinical predictors.
• Mutation location on the protein was a key predictor, unlike residue conservation or physicochemical properties.
• Prediction accuracy was insufficient for direct use in mechanistic studies.

Conclusions:

• In-silico prediction of I_Na changes from SCN5A mutations is challenging, partly explaining poor clinical severity prediction.
• Functional studies of I_Na remain crucial for accurate cardiac risk assessment in SCN5A variant carriers.