Calculation of protein entropies from atomistic molecular dynamics trajectories

Biomolecular models are already an integral component of the drug-development pipeline. For example, docking studies are routinely used to select promising candidates from ligand libraries that can contain millions of small molecules for further computational and experimental analysis, and free energy perturbation methods have become increasingly common for rank ordering chemically similar compounds. However, biomolecular simulations do not always reliably predict binding free energies for ligand-protein interactions, and one potential reason for this is that they do not correctly account for changes in entropy when drugs bind. Here we are developing a theoretical framework and software suite for calculating absolute free energies of proteins and protein ligand complexes from atomistic molecular dynamics trajectories. The software, known as CodeEntropy, is being co-developed by CCPBioSim. The project also benefits from collaborations with experimentalists at Isis using neutron scattering to quantify entropy changes associated with the biotin-streptavadin binding. CodeEntropy has the potential to be very useful in computer aided drug design, because it provides a breakdown of the origins of entropy changes and so provides additional physical insight into the role of entropy in molecular recognition not accessible by free energy perturbation.