UCSF DOCK is a well-established program for finding and optimizing the structures of small molecules in protein or nucleic acid receptor sites. Recently, we have incorporated molecular dynamics-based scoring functions (using the Amber/GB implicit solvation models) into this code. This, for the first time in the DOCK program, allows one to consider receptor flexibility in estimating binding affinities. The new code also provides a "physics-based" alternative to the more traditional empirical, or "knowledge-based" scoring functions.
DOCK was originally developed in Tack Kuntz' group at UCSF, and continues to be maintained and extended by members of his group, those in Brian Shoichet's group (also at UCSF), and by others (principally Rob Rizzo at Stony Brook University and Dave Case at Rutgers). The latest version (6.8, April, 2017) can be obtained here: