mGPfusion: Predicting protein stability changes with Gaussian process kernel learning and data fusion

TL;DR

mGPfusion is a Gaussian process-based method that combines limited experimental data with molecular simulation data to accurately predict how mutations affect protein stability, aiding protein design.

Contribution

It introduces a Bayesian data fusion approach and a protein-specific Gaussian process model that effectively integrates diverse data sources for stability prediction.

Findings

Outperforms existing methods on 15 proteins

Incorporating simulation data improves accuracy

Effective with limited experimental data

Abstract

Proteins are commonly used by biochemical industry for numerous processes. Refining these proteins' properties via mutations causes stability effects as well. Accurate computational method to predict how mutations affect protein stability are necessary to facilitate efficient protein design. However, accuracy of predictive models is ultimately constrained by the limited availability of experimental data. We have developed mGPfusion, a novel Gaussian process (GP) method for predicting protein's stability changes upon single and multiple mutations. This method complements the limited experimental data with large amounts of molecular simulation data. We introduce a Bayesian data fusion model that re-calibrates the experimental and in silico data sources and then learns a predictive GP model from the combined data. Our protein-specific model requires experimental data only regarding the protein of interest and performs well even with few experimental measurements. The mGPfusion models proteins by contact maps and infers the stability effects caused by mutations with a mixture of graph kernels. Our results show that mGPfusion outperforms state-of-the-art methods in predicting protein stability on a dataset of 15 different proteins and that incorporating molecular simulation data improves the model learning and prediction accuracy.