Enhancing Protein-Drug Binding Prediction

Drugs typically exert their effects by interactions with and binding to proteins. Accurate methods are therefore needed to evaluate protein-drug binding and interaction. However, predicting their binding in an accurate way is typically difficult and computationally demanding. In this ASDI project we developed and implemented smart algorithms to accurately and automatically parameterize interactions involving a drug compound of interest, and we combined molecular dynamics (MD) simulation together with other computational approaches needed for protein-drug binding evaluation. By integrating these in the novel microservice-driven MDStudio platform and by developing a dedicated microservice to allow parallel execution of MD at heterogeneous (super)computing facilities, we have enabled and greatly facilitated efficient execution of complex MD-based workflows for protein-binding prediction. The platform, tools and algorithms developed during this project have been made available as open access software.