Protein coarse-grained (CG) force fields (FF) are a promising alternative to conventional atomistic protein force fields. By subsuming groups of atoms into single interaction sites, and reducing the number of degrees of freedom, at the cost of reduced accuracy, much faster calculations can be realized. This represents one of the most important approximations for the construction and simulation of larger systems. One example is the SIRAH force field, which has support for proteins (and water, DNA and lipid models) and does not use elastic networks to overcome the problem of secondary structure stability. The use of a higher resolution backbone representation produces hydrogen bond-like interactions, which stabilize the secondary structure. Moreover, SIRAH models long-range electrostatic interactions using the particle mesh Ewald method (PME). New optimisation processes to parameterise these models could improve their quality and range of applicability. We present here an automated approach for the optimisation of the SIRAH CG protein FF. Parametrising force fields is often ad-hoc and time consuming. As such, automated procedures are particularly valuable. ForceBalance represents one such procedure where FF parameters are automatically optimised to reproduce experimental and/or simulation data. Here we apply ForceBalance to the optimisation of a protein coarse-grained force field. A full optimisation of the SIRAH water model WT4 was performed using the ForceBalance method, based on experimental water properties. We implemented hydration free energy gradients as a new target for force field optimisation in ForceBalance and applied it successfully to optimize the uncharged and charged side-chains, and backbone of SIRAH, closely reproducing the hydration free energies of atomistic models and improving agreement with experiment. Simulations of folded proteins and peptides in water result in improved protein stabilities for the new model.

Automated protein coarse-grained force field optimisation using free energy simulations

2019 Fall National Meeting

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