Markov state models (MSMs) are increasingly used to combine information from many short molecular dynamics trajectories to characterize biomolecular dynamics. Based on the assumption that conformational dynamics are independent of the history of states, MSMs fully describe the thermodynamics and kinetics of the states. Given the rich information contained within MSMs, it would be advantageous to exploit them in binding free energy calculations. The statistical mechanics linking MSMs to binding free energy calculations is provided by implicit ligand theory (ILT) [2,5]. ILT specifies that noncovalent binding free energies between a flexible ligand and receptor may be determined through an exponential average of the binding potential of mean force (BPMF) - the binding free energy between a flexible ligand and rigid receptor. The rigid receptor conformations must be drawn from or reweighed to an apo [4] or holo [2] ensemble, as is possible in MSMs. To demonstrate this approach, we construct a MSM of T4 lysozyme based on 4 microseconds of molecular dynamics initiated from four different crystal structures representing open, intermediate, and closed conformations of the enzyme. Representative conformations are selected based on the Jaccard distance between occupancy fingerprints [1]. BPMFs for 141 small organic molecules are computed for the representative snapshots using our program AlGDock [4]. Free energy estimates were compared to previous results in which snapshots were selected from alchemical binding free energy calculations [3].<br/><br/>

From Markov state models to absolute binding free energies

2019 Fall National Meeting

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