Molecular: Efficient free energy calculations for biomolecular applications

Free energy calculations is an important task in molecular simulations, since it is probably the most obvious property that can be easily obtained directly in the lab, without several intermediate steps of interpreting raw data. Typical examples include solubility of small compounds in water, the binding energy of a drug to a protein, or how conformational transitions in macromolecules are affected by changes in the environment or ligand binding.

In practice, free energy calculation can be relatively difficult since it involves entropic contributions, and these can only be sampled accurately in a molecular dynamics simulation. Within SeRC, we are working on improving these methods by implementing Bennett Acceptance Ratio techniques in GROMACS. This approach is more efficient than previous alternatives, but one of the most important aspects is that it directly provides standard error estimates for the free energies, which has been quite difficult before. In particular, we working on methods to automatically optimize the number of and type of subcalculations that have to be performed to calculate a single free energy, and in the next stage we will automatically account for all the multiple poses that will contribute to an experimental binding energy. The results are of direct relevance for high-throughput free energy calculation e.g. in the pharmaceutical industry.

Responsible scientist: Erik Lindahl
Additional researchers: Berk Hess, Szilard Pall, Sander Pronk