The quantitative prediction and qualitative understanding of noncovalent interactions is a critical challenge for quantum chemistry. Recently we have developed a new scheme to partition symmetry-adapted perturbation theory (SAPT) interaction energy contributions to the level of chemical functional groups. This functional group SAPT (F-SAPT) tool allows one to robustly determine the origins of complicated noncovalent interactions without needing to resort to chemical intuition. With an efficient density-fitted implementation, F-SAPT computations can be performed on systems with several hundred atoms. We have applied F-SAPT to several important problems in drug-protein interactions and find that these interactions may arise from surprising long-ranged contacts, which are not predicted by standard chemical intuition.
