Finding Reaction Coordinates for Protein Folding and Functional Motion

Mehr zum Buch

To understand life at the molecular level, a grasp of protein function and dynamics is essential. Molecular dynamics simulations facilitate the study of biomolecular dynamics over femto- to milliseconds and beyond. However, the vast data generated necessitates methods to transform input coordinates into new collective coordinates, capturing relevant processes to form a multidimensional reaction coordinate. Several challenges arise in this transformation process, including the choice of input coordinates—whether Cartesian atomic positions, dihedral angles, or distances between atom groups—and its impact on results. Different transformations yield unique collective coordinates, complicating the interpretation of molecular motion. Additionally, selecting meaningful reaction coordinates remains unclear. This work emphasizes principal component analysis for generating collective coordinates, systematically using inter-residue distances as input. By comparing this method to traditional approaches, we find that the choice of input coordinates is crucially linked to the type of dynamics observed. For instance, in the folding system villin headpiece (HP35), backbone dihedral angles are most effective, while T4 lysozyme's large-scale domain motion is better described by inter-residue distances. We advocate evaluating simple parameters first to identify dominant motions before selecting appropriate coordinates. Furthermore, simulations e