Combining Molecular Dynamics and Machine Learning to Improve Protein Function Recognition

Abstract

As structural genomics efforts succeed in solving protein structures with novel folds, the number of proteins with known structures but unknown functions increases. Although experimental assays can determine the functions of some of these molecules, they can be expensive and time consuming. Computational approaches can assist in identifying potential functions of these molecules. Possible functions can be predicted based on sequence similarity, genomic context, expression patterns, structure similarity, and combinations of these. We investigated whether simulations of protein dynamics can expose functional sites that are not apparent to the structurebased function prediction methods in static crystal structures. Focusing on Ca 2+ binding, we coupled a machine learning tool that recognizes functional sites, FEATURE, with Molecular Dynamics (MD) simulations. Treating molecules as dynamic entities can improve the ability of structure-based function prediction methods to annotate possible functional sites.

Role of surgical outcome as prognostic factor in advanced epithelial ovarian cancer: A combined exploratory analysis of 3 prospectively randomized phase 3 multicenter trials

“…Computational methods to predict Ca 2+ ‐binding sites have been actively pursued using various approaches 4, 25–28. Most of the published structure‐based Ca 2+ ‐binding site‐prediction algorithms, including FEATURE,29 Fold‐X,30 and the approaches by Nayal et al 20.…”

Integration of Diverse Research Methods to Analyze and Engineer Ca2+- Binding Proteins: From Prediction to Production

“…An online server WebFEATURE is available at http://feature.stanford.edu/webfeature/. Besides being applied to static X-ray crystal structures, FEATURE has recently been applied to multiple conformations of protein parvalbumin β generated from molecular dynamics (MD) simulation [88]. For MD generated conformations of both holo and apo forms, FEATURE correctly identifies the calcium-binding sites and non-sites with some interesting fluctuations in score.…”

Efficient Algorithms to Explore Conformation Spaces of Flexible Protein Loops

“…If a sampled conformation is recognized by the method, not only does this indicate that the loop may be a possible calcium-binding site, but it also tells us what the holo conformation may look like. In fact, molecular dynamics simulation has already been used successfully to generate conformations starting with apo proteins in order to identify unrecognized calcium-binding sites in them [15]. …”