Quickly and accurately calculating the electrostatic free energy change that occurs when proteins bind is essential to the analysis of protein-protein interactions. This free energy change, quantified as ?Gelec, expresses how energetically favorable it is for two proteins to come together. Calculating ?Gelec for a protein complex usually involves finding a numerical solution to the computationally intensive Poisson equation. More approximate methods can save time and computational cost. Our approach involves training a computer via regression-based machine learning techniques to predict ?Gelec using simple, structural features of the protein complex itself. Preliminary work involving shape-simplified model systems showed that our model was able to predict one component of ?Gelec, the ligand desolvation penalty,to a high degree of accuracy. We are currently extending this analysis to protein-like systems, with the ultimate goal of efficiently and accurately predicting ?Gelec for a variety of protein complexes.

Estimating protein-protein electrostatic binding energetics: A feature-based approach

245th National Meeting (2013)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.