Tunable, mixed-resolution modeling using library-based Monte Carlo and graphics processing units.

摘要: Building on our recently introduced library-based Monte Carlo (LBMC) approach, we describe a flexible protocol for mixed coarse-grained (CG)/all-atom (AA) simulation of proteins and ligands. In the present implementation LBMC, protein side chain configurations are pre-calculated stored in libraries, while bonded interactions along backbone treated explicitly. Because AA coordinates maintained at minimal run-time cost, arbitrary sites interaction terms can be turned to create mixed-resolution models. For example, an region interest such as binding site coupled CG model rest protein. We have additionally developed hybrid generalized Born/surface area (GBSA) implicit solvent suitable models, which turn was ported graphics processing unit (GPU) faster calculation. The new software applied study two systems: (i) behavior spin labels B1 domain G (GB1) (ii) docking randomly initialized estradiol ligand estrogen receptor (ERα). performance GPU version code also benchmarked number additional systems.