Analysis of Results(progress report)
We are working our way through a test set of 181 protein-ligand structures from the GOLD test set (Jones, et al, J. Mol. Biol., 267, 727 (1997)) and a set of structures used by H. Bohm (J. Comp. Aid. Mol. Design, 12, 309 (1998)).
These have all been run with the same set of parameters and run blind with the following exceptions:
a) We have tried to use the appropriate multimeric structure for the protein.
b) We have checked that the covalent bond orders have been assigned correctly by the bondage algorithm and corrected them where necessary.
c) In 1swd (biotin/strepavidin), we adjusted a parameter in sublime to prevent the spheres from filling binding sites from two adjacent monomers of the tetramer.
Two reasons for running the tests blind:
a) We wanted to see how well Dockit did automatically.
b) We wanted to avoid the temptation to hand tweak the runs too much.
125 calculated so far. These have been calculated with hydrogens and metals included.
Analysis is ongoing.