> Specifically, I was wondering if there is an example of how to use > OBConformerSearch from Python (and is there a way how to use this directly > from the command line?).
Sorry the documentation isn't online -- we should definitely fix that. It's easy to use from the command-line: obabel -L conformer One of the ops conformer Conformer Searching (not displayed in GUI) Typical usage: obabel infile.xxx -O outfile.yy --conformer --nconf options: description --log output a log of the energies (default = no log) --nconf # number of conformers to generate forcefield based methods for finding stable conformers: --systematic systematically generate all conformers --random randomly generate conformers --weighted weighted rotor search for lowest energy conformer --ff # select a forcefield (default = MMFF94) genetic algorithm based methods (default): --children # number of children to generate for each parent (default = 5) --mutability # mutation frequency (default = 5) --converge # number of identical generations before convergence is reached --score # scoring function [rmsd|energy] (default = rmsd) For example: # generate 50 conformers, scoring with MMFF94 energies, but default children, etc. obabel EtOT5D.cml -O EtOT5D0.xyz --conformer --nconf 50 --score energy As far as Python, I haven't tried, but the code is "wrapped" so it should be available. If you want some example code, I can probably whip something up based on the C++ (e.g., src/ops/conformer.cpp). > Also, do the conformers undergo some form of energy minimization in this GA > approach? As of yet, no. That could be done new scoring function (e.g., minenergy) which determines the energy after some minimization. It's a good idea. > Also, I can see I can generate a fixed number of conformers (e.g. 50) using > Weighted Rotor Score, what is a decent parameter for geomSteps in > WeightedRotorSearch (unsigned int conformers, unsigned int geomSteps) ? Why > do the other structure methods have a default for 2500 and this one has not? > Is this an oversight? I usually use a few hundred geomSteps, maybe 250 or 500. IMHO, the point isn't to produce a fully converged structure, but to quickly minimize steric collisions generated from a particular rotamer. Once I have some low-energy conformers, I'll minimize the low-energy set to find more converged energies and geometries. Hope that helps, -Geoff ------------------------------------------------------------------------------ Cloud Services Checklist: Pricing and Packaging Optimization This white paper is intended to serve as a reference, checklist and point of discussion for anyone considering optimizing the pricing and packaging model of a cloud services business. Read Now! http://www.accelacomm.com/jaw/sfnl/114/51491232/ _______________________________________________ OpenBabel-discuss mailing list OpenBabel-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
