In silico relationship between configurational entropy and soft degrees of freedom in proteins and peptides

Configurational entropies from temperature-dependent free energy calculations of proteins and peptides are compared with entropies determined as the sum of the Shannon entropies of a kernel density estimate of the dihedral angle probability distributions. We find that the entropy changes can be separated into independent contributions from hard and soft degrees of freedom. Furthermore, because the effects of motional correlations between dihedral angles on entropy differences are small, the functional role of configurational entropy changes can be interpreted in a spatially resolved manner.