Figures
Incorporating physics helps overcome data scarcity in predicting mutational effects on BK channel voltage gating
Multi-decade efforts have yielded voltage gating voltage (∆V1/2) data on almost 500 single mutants of the big potassium (BK) channel. The left panel shows all residues of the BK channel where at least one mutation has been characterized experimentally. While representing enormous efforts, these data alone remain too scarce for deriving quantitative models for predicting the mutational effects using data-driven machine learning approaches. By incorporating physical features derived from molecular modeling and simulations, we show that a random forest model trained using 80% of the experimental shift in gating voltage is able to predict unseen experimental data with 0.79 correlation. The model is also able to capture a central role of hydrophobic gating in BK channels as well as new experimental results on several novel mutations. Nordquist et al 2023
Image Credit: Erik Nordquist, enordquist@umass.edu
Citation: (2023) PLoS Computational Biology Issue Image | Vol. 19(9) October 2023. PLoS Comput Biol 19(9): ev19.i09. https://doi.org/10.1371/image.pcbi.v19.i09
Published: October 2, 2023
Copyright: © 2023 . This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Multi-decade efforts have yielded voltage gating voltage (∆V1/2) data on almost 500 single mutants of the big potassium (BK) channel. The left panel shows all residues of the BK channel where at least one mutation has been characterized experimentally. While representing enormous efforts, these data alone remain too scarce for deriving quantitative models for predicting the mutational effects using data-driven machine learning approaches. By incorporating physical features derived from molecular modeling and simulations, we show that a random forest model trained using 80% of the experimental shift in gating voltage is able to predict unseen experimental data with 0.79 correlation. The model is also able to capture a central role of hydrophobic gating in BK channels as well as new experimental results on several novel mutations. Nordquist et al 2023
Image Credit: Erik Nordquist, enordquist@umass.edu