High‐throughput molecular dynamics simulations: toward a dynamic view of macromolecular structure

Abstract

AbstractMolecular dynamics (MD) simulation is nowadays the eligible theoretical technique to account for macromolecular flexibility. Molecular simulations, from ab initio to coarse‐grained representation levels, allow having a direct visualization of macromolecular dynamic behavior and its influence in molecular recognition. In the postgenomic era, where most bioinformatics studies should be performed genome‐wide, molecular simulations should not be an exception. However, moving MD simulations to the high‐throughput regime is not a trivial issue. High‐performance computing systems (highly parallel supercomputers or graphical processing units‐based systems) allow performing large and complex simulations at a significantly reduced time. Besides, data storage strategies have also been largely improved. However, system preparation and trajectory analysis are still performed almost manually and become highly limited by the need of human intervention. A number of projects, like Dynameomics, BioSimGrid, or MoDEL have addressed the issue of high‐throughput MD. Such initiatives have released tools that have meant a significant progress in the field. The extension of structural studies, mainly molecular simulations, to the high‐throughput regime, will allow matching genomic‐wide studies now performed.This article is categorized under: Molecular and Statistical Mechanics > Molecular Dynamics and Monte-Carlo Methods