NEW FEATURES IN DOCK 6.1 ###New Pruning Algorithm for Anchor & Grow### As a result of previous studies, we determined that the major sampling limitation for the flexible ligand algorithm was the pruning. After exploring several alternatives, we have added a new method. Here, the goal was to bias the sampling toward conformations that are close to the correct binding mode (as optimized using a test set of experimentally solved structures). Much as the older method, the algorithm ranks the generated poses and conformations. Then, all poses that violate a user-defined score cutoff are removed. To facilitate the speed of the calculation, the remaining list is additionally pared back to a user-defined length. In this technique, the sampling is driven toward molecules that sample closer to the experimentally determined binding site, but results in a significantly less diverse set of final poses. The original pruning algorithm is still available and remains the default setting. ###Ability to Perform Analysis with Primary Score Before Rescoring### In previous versions, the user was able to perform both ranking and clustering on the dock runs for each ligand using the primary score. However, if the user wanted to then apply a more advanced scoring function, the analyzed data needed to be save to a file and then rescored in a separate run. The primary scoring, analysis, and secondary scoring procedures can now all be performed in the same run. ###Finer Control Over Structural Output### In a typical run, the best-scoring pose of each ligand in the library is written to a file called XXX_scored.mol2. In addition, there are several more advanced output options: (1) Users can choose to write out anchor orientations to a file called XXX_orients.mol2. (2) Users can also write out conformers after the final level of optimization to a file called XXX_conformers.mol2. If clustering is enabled, all conformations will be written to the XXX_conformers.mol2 file while the clusterheads will be written to the XXX_scored.mol2 file. This seemingly redundant writing is to ensure that no portion of the analysis is lost if a failure occurs. In addition, both the number of ranked conformations and clusterheads can now be controlled by the user. (3) Users can write molecules ranked by score to a file called XXX_ranked.mol2. Once again, the user can control the number of ligands from a library are saved. If secondary scoring is enabled, all of the above files will be generated for both primary and secondary scoring. The secondary scoring files contain both the original primary score as well as the new secondary score. ###Amber Score Improvements### Two new movable regions have been added: distance and nothing. In the distance movable region receptor residues that are within a cutoff of the ligand are flexible. The nothing movable region is a mildly performance optimized method for freezing all atoms. The DOCK output has been reorganized; all scoring details are controlled by the -v verbose flag. The ligand_outfile_prefix_scored.mol2 file contains the ligand coordinates extracted from the final structure of the complex Final structure pdb files are written for the receptor, ligand, and complex as necessary. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ DEPRECATED FEATURES 1. For a variety of reasons still under active development, amber_score cannot effectively be used as a secondary_score. This function has been temporarily deprecated, and using input parameter amber_score_secondary causes program termination. The recommended protocol is to perform two DOCK runs with the second run specifying amber_score as the primary_score. 2. Due to the restructuring of the structure output files (see "Finer Control Over Structural Output," the scored_conformer_output_override and write_conformations options are no longer available. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ BUG FIXES Several critical bugs have been identified and fixed since the last release. These bugs include: For SHOWSPHERE: • Tabs and open statement keywords have been removed to enable compiling on various platforms, in particular gfortran 4.1.1 For DOCK: • The heavy atom RMSD reported after docking is now being properly calculated • The default values for missing input parameters are properly assigned in batch mode • If primary_score is called with minimization and secondary_score is called without minimization, secondary_score is now being calculated • Amber_score now properly writes the final pose of the complex and the ligand properly • When final_min is called, all user defined iterations are now performed. • The reported problems with the bump filter have been addressed. The behavior for the anchor filtering now reproduces DOCK 5. The behavior for the growth filtering has been tested more extensively.