4-19-2024 Release Notes for minor update v5.189 (vs 5.186): Tools: Improved recognition of certain types of ring chirality that are not detectable by topology (affects the regen3d command). Added finer control of conformer pool size for ForceGen (-nfinal) to allow for conformer pool size control following full conformer search. Docking: Improved RMSD calculations for evaluation of docking quality (rms_list and rms_fam commands). Added a method (bbox command) to automatically define bounding boxes given collections of aligned ligands. Implemented a multi-core approach to pose family generation when using eSim for exploiting known poses. 2-14-2024 Release Notes for minor update v5.186 (vs 5.173): General: We have added a new chapter to the manual (and associated data to distribution examples) to cover Advanced Applications, which include usage of the newly released xGen PyMOL interface, docking of conformationally restrained PROTACs, and restraint-based docking and analysis of large macrocyclic peptides. If a user has set OMP_THREAD_LIMIT, that number of threads will be used by default. Added a user-settable environment variable SURFLEX_ROOT to aid in license file location and automatic finding of data such as the verifypdb.smi file. Tools: Added termination line for ForceGen and FGen3D log output to indication successful completion and number of mols processed. The command combine_sfdb will accept input arg of a single SFDB file to apply energy threshold. Added +3dfast option for faster 3D structure generation. Added commands to aid in the determination of ligand strain (see bound_energy and unbound_energy) for xGen ensembles and pose families from docking or other operations. For macrocycles, additional focus on identifying and considering different triplet H-bond sets. Improved make_sfdb to recalculate conformer energies when needed. Similarity: Minor changes in memory management for mult_esim. More careful checking of user-specified target molecule for esim_list. PDBGrind: Added option for faster grinding when the ligand details matter more than the whole complex (+pdbquick). This incorporates a simple estimate of tautomer strain in choosing which is best for a binding site. Docking: Minor changes to protomol core voxel method to eliminate disconnected areas. Added new methods to define binding sites during protomol generation. 8-22-2023 Release Notes for minor update v5.173 (vs 5.164): One substantive change: multiple ligand alignment has been improved with respect to cross-platform behavior using a normalization within pose-clique scoring. To turn off this new default behavior, in the Similarity module use -me_norm and in the QuanSA module use the -clnorm option. One bug fix: the docking pose family generation implementation has been patched to use the correct score values. This affects pose family results especially in cases where prior ligand knowledge is not used. ---------------------------------------------------------------------------------------------------------------------- 5-6-2023 Release Notes for minor update v5.164 (vs 5.162): 1) eSim: Fixed memory leak when using -poscon 2) Tools: Implemented new behavior for torsional restraints. Given multiple fragments that match to separate parts of a molecule to be searched, the individual fragments create positional restraints so that a linker between the two parts will keep them in geometrically correct positions. Adherence to the geometry restraint is done through the -pospen and -pwiggle parameters. 3) Tools: Provide the ability to turn off inclusion of torsional restraints in SFDB conformer files. Useful if one wants to restrain conformer search but not restrain deviations in downstream calculations like docking and similarity optimization. ---------------------------------------------------------------------------------------------------------------------- 3-6-2023 Release Notes for minor update v5.162 (vs 5.142): Tools Module: 1) Improvement in enforcement of torsional restraints during fgen3d (and regen3d) procedures. 2) Fixed issue with parsing variant SDF files with non-standard tags 3) Added tautomer recognition for imide/amide proton shift 4) Increased max SDF tag line width to 2000 characters Docking Module: 1) Changed default PDB ligand parsing size to 100 heavy atoms 2) Fixed bug in reporting of polar component of docking scores QuanSA Module: 1) Fixed issue with reading named molecules where mols are repeated ESim and xGen Modules: No significant user-facing changes. ---------------------------------------------------------------------------------------------------------------------- 5-16-2022 Release Notes for minor update v5.142 (vs v5.137): Sim and Dock Modules: 1) Positional constraints: Specification of a positional constraint argument (-poscon) will now cause filtering of input to similarity and docking operations. Only those input molecules that contain the given positional constraint substructure will be processed. Behavior can be suppressed with -skipnonmatch. 2) SDF output: Specification of +sdf will produce a tagged SDF file as output to similarity and docking operations, where the information present in the log file (corresponding to the mol2 output) is offered in the standard MDL tag/value syntax. 3) Similarity display: esim_disp will now produce a more refined display of similarity between two molecules. Load the -disp.pml file in PyMol to visualize. 4) Molecular imprinting: molecular imprinting has been updated with eSim similarity calculations. The commands imprint, choose_ref, and iscreen_list are described in the manual. 5) PSIM (see manual for details): increased thoroughness of pocket alignment, added a molecular overlap criterion to prevent marginally overlapping binding-site ligands from forming tree edges (controlled by -psim_overlap), removed brittle grid-caching behavior, added re-engineered psim_findcav command (experimental). 6) PDBGrind: Better handling of peptidic ligands by identifying them as non-connected components. 7) Docking fingerprints: added dock_fp command to produce protein interaction fingreprints (suitable as input to iscreen_list). Tools Module: 1) Changed behavior of pretty_sdf command to also flatten molecules. Suitable for vizualization of aligned sets of ligands in pseudo-2D as well as import into ChemDraw (chirality is preserved). 2) Added -molid option to specify the SDF tag name to use for molecule ID in SDF files. 3) Added parse_sdf command to grab tag/value data from SDF files and produce tab-delimited text suitable for input into Excel. 4) Profile command will not center molecules by default (can turn on centering with +pcenter). ---------------------------------------------------------------------------------------------------------------------- 1-17-2022 Release Notes for minor update v5.137 (vs. 5.125): All Modules: The default number of desired threads is 36, but the machine architecture may specify a different value, and the environment variable OMP_THREAD_LIMIT may specify a different one yet. The default thread utilization is now set to the minimum of these three values. If -nthreads is specified, that value will be used. However, if that value is higher than the machine architecture limit or the environmental thread limit, thread utilization will depend on operating system. Tools Module: 1) Addition of +qmin option for approximating r-6 averaging in NMR restraints on ambiguous protons 2) Minor change in macrocycle search in order to increase search breadth of twist motions. 3) The fgen3d command now allows for the -multiproc option. 4) Minor changes to core utilization strategy for complex peptidic macrocycles. 5) Addition of fgen_deep_list command and automatic file cleanup for fgen_deep. 6) Fixed case where nominally chiral protonated N was being counted as a chiral center to be enumerated if -enum_chiral was specified. Now, Tools correctly perceives such N-H cases as being treated properly with amine inversion mechanics. Docking Module: 1) Reduced memory footprint for grindpdblist and increased robustness. Windows multi-core now runs with high utilization. NOTE: with very large numbers of PDB files, the approximately 1% chance of a fatal error on any particular PDB file may manifiest in a parallel run. In such cases, use the serial grinding script. 2) Fixed odd case of incorrect bond order assignment to certain nitro groups in grindpdb functions. 3) Protomol generation now checks for proteins being assigned partial charges. Proteins automatically charged if detected as not charged in docking commands. Better to do this right ahead of time to avoid repeated time wastage. Similarity Module: 1) The -vrange option now scales the min volume against the smallest ligand of a multi-ligand query and the max volume against the largest. 2) Fixed bug with -pfast option leading to incorrect MMFF94sf calculation. 3) Added targprep command to validate/prepare molecular queries. QuanSA Module: 1) Minor changes in multiple ligand alignment when known poses are provided via -clknown to improve utilization of the information. 2) Default values for -clkthresh and -addthresh are now both 6.5 3) Fixed a bug with the add command when employing a new SFDB for the mols to be added xGen Module: No significant changes. ---------------------------------------------------------------------------------------------------------------------- 8-19-2021 Release Notes for minor update v5.125 (vs. 5.114): Tools Module: 1) Added fgen_deep command for extremely thorough conformer search, especially for macrocycles. 2) Addition of dihedral restraint type to NMR specification to automatically determine dihedral angle sign, rather than expecting the user to make an informed guess. 3) Added commands for conformer pool compression and clustering (comp_rms and comp_macrms) 4) Minor changes to profile command to improve characterization of macrocyclic ring systems. Similarity Module: 1) Exposed control of the joint similarity switch +-joint for multi-ligand targets. Default is +joint. If -joint is specified, then eSim will seek to optimally match a single ligand within a multi-ligand target rather than optimizing against the best matching parts of each ligand. 2) Minor changes in multiple ligand alignment when known poses are provided via -me_known to improve utilization of the information. QuanSA Module: 1) Minor changes in multiple ligand alignment when known poses are provided via -clknown to improve utilization of the information. 2) Default values for -clkthresh and -addthresh are now both 6.5 3) Fixed a bug with the add command when employing a new SFDB for the mols to be added Docking Module: No significant changes. xGen Module: No significant changes. ---------------------------------------------------------------------------------------------------------------------- 4-20-2021 Release Notes for minor update v5.114 (vs. 5.111): 1) Fixed bug in +findbeta macrocycle search that limited the number of conformations produced. A new command: comp_macrms allows for compression and clustering of macrocycle conformer pools with the clusters being driven by macrocyclic backbone geometry. 2) Fixed bug in output of two-way eSim values to the log file. Slight changes to calculation of two-way eSim values to increase accuracy in higher-throughput modes (with a time cost for the +two_way options). 3) Observer points are now placed around a canonically oriented target ligand prior to transformation back to the original coordinate frame. This reduces the relatively minor effects of coordinate frame changes on eSim calculations. 3) eSim reference ligands that have zero partial charges will be automatically charged. Note that reference ligands that have (non-SF) partial charges should be explicitly assigned SF charges (e.g. by using the "charge" command of sf-tools). Similarly, molecules provided as known poses will also be automatically charged if existing charges are zero. 4) The default value for -me_kthresh has been decreased to 4.0. This controls the level of similarity required for a molecule that is part of a mult_esim operation to be retained as part of the multiple alignment when making use of the -me_known option. The -me_known option allows a user to specify, for example, a set of crystallographic poses to drive pose generation. Final pose selection optimizes a multi-objective function that seeks high mutual similarity and low ligand strain. Molecules that are less similar than -me_kthresh to the -me_given mols are omitted from the output. Users can align them to the molecules that "survived" through an esim_list operation if desired.