Decomposition of the free energy of a system in terms of specific interactions. Implications for theoretical and experimental studies.

Recently, a number of methods have been proposed that are designed to extract contributions to the change in free energy associated with a given perturbation or mutation of a protein originating from specific residue-residue or atom-atom interactions, both based on theoretical calculations and on experimental data. We caution here that detailed analysis based on these methods is unreliable. It is demonstrated, both from first principles using statistical mechanics and by way of example, that in a general case a meaningful decomposition of the free energy in terms of specific residue-residue or atom-atom interactions is not possible.