Aroop Sircar - A generalized approach to sampling backbone conformations with RosettaDock for CAPRI rounds 13-19

Document created by Aroop Sircar on Aug 22, 2014
Version 1Show Document
  • View in full screen mode

  Publication Details (including relevant citation   information):

  Aroop Sircar, Sidhartha Chaudhury, Krishna Praneeth Kilambi,   Monica Berrondo & Jeffrey J. Gray

  Proteins, 78(15), 3115-3123, May 18, 2010


  In CAPRI rounds 13–19, the most native-like structure predicted   by  RosettaDock resulted in two high, one medium, and one   acceptable  accuracy model out of 13 targets. The current   rounds of CAPRI were  especially challenging with many   unbound and homology modeled starting  structures. Novel   docking methods, including EnsembleDock and SnugDock,    allowed backbone conformational sampling during docking and   enabled the  creation of more accurate models. For Target   32, a-amylase/subtilisin  inhibitor-subtilisin savinase, we   sampled different backbone  conformations at an interfacial   loop to produce five high-quality models  including the most   accurate structure submitted in the challenge (2.1 A  ligand   rmsd, 0.52 A interface rmsd). For Target 41,   colicin-immunity  protein, we used EnsembleDock to sample   the ensemble of nuclear magnetic  resonance (NMR) models of   the immunity protein to generate a medium  accuracy   structure. Experimental data identifying the catalytic   residues  at the binding interface for Target 40   (trypsin-inhibitor) were used to  filter RosettaDock global   rigid body docking decoys to determine high  accuracy   predictions for the two distinct binding sites in which the    inhibitor interacts with trypsin. We discuss our generalized   approach to  selecting appropriate methods for different   types of docking problems.  The current toolset provides   some robustness to errors in homology  models,but   significant challenges remain in accommodating larger    backbone uncertainties and in sampling adequately for global   searches.

  Address (URL):