Premium
SOFTDOCK application to protein–protein interaction benchmark and CAPRI
Author(s) -
Li Nan,
Sun Zhonghua,
Jiang Fan
Publication year - 2007
Publication title -
proteins: structure, function, and bioinformatics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.699
H-Index - 191
eISSN - 1097-0134
pISSN - 0887-3585
DOI - 10.1002/prot.21728
Subject(s) - cluster analysis , docking (animal) , voronoi diagram , solvation , computer science , grid , data mining , macromolecular docking , biological system , algorithm , artificial intelligence , chemistry , molecular dynamics , computational chemistry , mathematics , molecule , geometry , biology , medicine , nursing , organic chemistry
The success of molecular docking requires cooperation of sampling and scoring of various conformations. The SOFTDOCK package uses a coarse-grained docking method to sample all possible conformations of complexes. SOFTDOCK uses a new Voronoi molecular surface and calculates several grid-based scores. It is shown by the leave-one-out test that three geometry scores and an FTDOCK-like electrostatics score contribute the most to the discrimination of near-native conformations. However, an atom-based solvation score is shown to be ineffective. It is also found that an increased Voronoi surface thickness greatly increases the accuracy of docking results. Finally, the clustering procedure is shown to improve the overall ranking, but leads to less accurate docking results. The application of SOFTDOCK in Critical Assessment of PRedicted Interactions involves four steps: (i) sampling with INTELEF; (ii) clustering; (iii) AMBER energy minimization; and (iv) manual inspection. Biological information from literature is used as filters in some of the sampling and manual inspection according to different targets. Two of our submissions have L_rmsd around 10 A. Although they are not classified as acceptable solutions, they are considered successful because they are comparable to the accuracy of our method.