Modeling Protein Conformational Transition Pathways Using Collective Motions and the LASSO Method

Many proteins can adopt multiple distinct conformational states which often play different functional roles. Previous studies have shown that the underlying global dynamics through which these states are accessed are, at least in part, encoded by the protein’s topology. In this work we present a method for generating transition pathways between states by perturbing the protein toward a target conformational state along thermally accessible collective motions calculated from the starting conformation. Specifically, the least absolute shrinkage and selection operator (LASSO) is used to identify the most parsimonious route along soft modes calculated using the anisotropic network model. In a survey of 436 conformational changes following protein–protein interaction, we show that such a path exists for most cases and that selected paths are low in energy. We discuss the implications for the atomic modeling of protein recognition and provide soft energy and parameter bounds which can be employed to efficiently...