From owner-chemistry@ccl.net Tue Sep 20 14:03:01 2016 From: "Francois BERENGER francois.berenger:+:inria.fr" To: CCL Subject: CCL: substructure search Message-Id: <-52383-160920092619-9273-DBsI8WNQQN3+DfoQ711M8A[*]server.ccl.net> X-Original-From: Francois BERENGER Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=windows-1252 Date: Tue, 20 Sep 2016 15:26:12 +0200 MIME-Version: 1.0 Sent to CCL by: Francois BERENGER [francois.berenger===inria.fr] Maybe the fmcsR package for R allows to do what you need. Cf. http://bioinformatics.oxfordjournals.org/content/29/21/2792.long On 09/20/2016 11:26 AM, Yury Minenkov yury.minenkov ~~ gmail.com wrote: > Sent to CCL by: "Yury Minenkov" [yury.minenkov=-=gmail.com] > Dear colleagues, > > I would like to ask a (basic) question which is perhaps on the border between the fields of Computational chemistry, chemoinformatics and drug design, in particular I am interested in the sub-stricture search algorithms. > > At the beginning I have the two things: > > a) The small XYZ molecular fragment for which I know the connectivity (I know in which way the atoms are connected, but I do not know the bond orders this is not that important for me at the beginning) > > b) Many XYZ coordinates of different molecules for which I also know the connectives (again, not the bond orders) > > I want to search for the given fragment in each of the XYZ molecular file I have. Something similar is organized in the CSD Cambridge structural database. > > I am quite certain that this is a general problem and there are should be many ready solutions available. Do we have any available libraries (preferably open-source & free with C API) in which such substructure search is implemented? Perhaps any codes? Or easy to implement algorithms? I don't think those algorithms are easy (maximum common substructure search). They might be terribly inefficient also. > I tried once few Python-based implementations based on SMARTS/SMILES but these are failed for the Transition metal complexes. That is why I believe 3d search would be better. > > Thank you in advance! Sorry if this too general question. > > With kind regards, > Yury> > -- Regards, Francois. "When in doubt, use more types"