Hi Taka, I've been playing with the UFFAddTorsionConstraint suggestion but I can't make it work for my systems (although it is great to know about as I can think of lots of other uses). The problem is that the Embedded geometries are too far from the geometry that I want to enforce on the torsions so that if I actually apply a large enough force constant to get the desired torsions I blow up the molecule and if I reduce the force constant the torsions don't go anywhere near where I need them.
I think that for now I might just go with some other conformer generator like Balloon, generate a silly number conformers and filter them for 'reasonable' ones using RDKit and go from there. Not ideal, but better than throwing my computer out the window because I can't get the RDKit embedding to do what I want :D Thanks for you suggestions. I know they'll come in handy for other things. Cheers, Bruce On Sun, 26 Apr 2020 at 13:28, Taka Seri <serit...@gmail.com> wrote: > Hi Bruce, > > Thanks for your reply. > I searched rdkit documents and found that RDKit has some constraint method. > https://www.rdkit.org/docs/source/rdkit.ForceField.rdForceField.html > And I tried to use UFFAddTorsionConstraint for an example. > https://gist.github.com/iwatobipen/42c325f2471166ff4084051381d0ef78 > By using these method I think you can keep the torsions. > Thanks. > > Taka > > 2020年4月26日(日) 19:35 Bruce Milne <bfmi...@gmail.com>: > >> Hi Taka, >> >> Thanks for the suggestion. The ConstrainedEmbed method is certainly >> useful but I need to only freeze some selected torsions and allow the rest >> of the macrocycle to remain flexible. If this was only a single part of the >> macrocycle the ConstrainedEmbed method would work but I need to >> simultaneously restrain several parts of the ring that are connected by >> flexible regions. >> >> Cheers, >> Bruce >> >> On Sun, 26 Apr 2020 at 02:00, Taka Seri <serit...@gmail.com> wrote: >> >>> Dear Bruce, >>> >>> How about to use the defined core which has your desired torsions? >>> Here are examples to generate restricted conformers with user defined >>> cores. >>> >>> https://www.discngine.com/blog/2019/6/6/tethered-minimization-of-small-molecules-with-rdkit-towards-tethered-docking-on-proteins-with-rdock >>> http://rdkit.blogspot.com/2019/01/more-on-constrained-embedding.html >>> Thanks, >>> >>> Taka >>> >>> >>> >>> 2020年4月26日(日) 1:11 Bruce Milne <bfmi...@gmail.com>: >>> >>>> Hi, >>>> >>>> Is it possible to add torsional restraints whilst >>>> using AllChem.EmbedMultipleConfs() to generate conformers? I did look at >>>> this some time back but ended up moving on to another project and never >>>> found out if it could be done. I asked a similar thing in this group at the >>>> time (about avoiding the generation of cis-peptide bonds) but I don't think >>>> it got an answer. >>>> >>>> I suspect that it may not be possible with AllChem.EmbedMultipleConfs() >>>> but if anyone knows of an alternative for generating partially constrained >>>> structures it would be good to know. >>>> >>>> Cheers, >>>> Bruce >>>> >>>> -- >>>> Prof. Dr. Bruce F. Milne >>>> CFisUC >>>> Department of Physics >>>> University of Coimbra >>>> Rua Larga >>>> 3004 - 516 Coimbra >>>> Portugal >>>> >>>> https://orcid.org/0000-0002-5522-4808 >>>> https://publons.com/researcher/2905148/bruce-milne/ >>>> https://www.linkedin.com/in/brucemilne >>>> >>>> A/h = S/k >>>> _______________________________________________ >>>> Rdkit-discuss mailing list >>>> Rdkit-discuss@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/rdkit-discuss >>>> >>> >> >> -- >> Prof. Dr. Bruce F. Milne >> CFisUC >> Department of Physics >> University of Coimbra >> Rua Larga >> 3004 - 516 Coimbra >> Portugal >> >> https://orcid.org/0000-0002-5522-4808 >> https://publons.com/researcher/2905148/bruce-milne/ >> https://www.linkedin.com/in/brucemilne >> >> A/h = S/k >> > -- Prof. Dr. Bruce F. Milne CFisUC Department of Physics University of Coimbra Rua Larga 3004 - 516 Coimbra Portugal https://orcid.org/0000-0002-5522-4808 https://publons.com/researcher/2905148/bruce-milne/ https://www.linkedin.com/in/brucemilne A/h = S/k
_______________________________________________ Rdkit-discuss mailing list Rdkit-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
