Large Scale Simulation of Protein Mechanics and Function

Abstract

The chapter illustrates the state of art in large-scale biomolecular modeling. Such modeling has become feasible only through parallel computing utilizing hundreds and soon thousands of processors. Fortunately, the necessary computers are available to many researchers and the needed new generation of molecular dynamics and molecular visualization programs has been developed and these resources are widely shared. Harnessing them effectively requires great effort, to which purpose the authors' group has developed the molecular dynamics program Not (just) Another Molecular Dynamics program (NAMD). For the graphical analysis of the ensuing gigabytes and terabytes of data the group has developed the program Visual Molecular Dynamics (VMD). These programs are widely used today since the increasing availability of protein structures has led to most biomedical researchers using structural information for the design and analysis of their experiments. The chapter describes NAMD and VMD and illustrates three exemplary simulations of large-scale systems that are presently the subject of intense research. The functions of all three proteins investigated are mainly mechanical. NAMD is primarily designed to work with Chemistry at HARvard Molecular Mechanics (CHARMM) force field parameters. Input files for NAMD can be generated using CHARMM, X-PLOR, or VMD. NAMD has also been extended to read Assisted Model Building with Energy Refinement (AMBER) and GROningen MAchine for Chemical Simulations (GROMACS) input file formats. NAMD running on 768 processors has enabled a record-breaking two million atom 5 ns simulation of the ribosome using an AMBER force field. To relate the observed mechanics to the architecture of proteins is the domain of so-called steered molecular dynamics simulations, contributing to the founding of the new field of mechanobiology, which studies the role of forces in cellular processes.