CCL:RE: clustering database molecules



 Hello Joey,
 I'm working with MOE and have used MOE:Similarity to cluster similar
 conformations. Our molecules are semi-peptidic macrocycles (about 40-50 atoms)
 and thus should be somewhat similar to your system. I've had success using Typed
 Atom Triangles (3D fingerprints) with a Tanimoto coefficient of 90%.
 Unfortunately, MOE:Similarity does not really cluster and/or does not make the
 results in a transferable data format such as to obtain the results without
 repetitively manually creating new files. I therefore proceeded in creating a
 script that uses MOE:Similarity to report the number of different families, the
 number of similar conformers within a family, export the separated families into
 new databases, and can even pipe the whole thing to population analyses. I only
 came up with a Tanimoto coefficient of 90% by trial and error (my modified
 scripts helped of course) obviously dependant on my definition of a cluster and
 what defines a new family. So, I suggest you do the same thing ...
 Say HI to Ghislain Deslongchamps, Rod Cooper, and Larry Calhoun for me ... ;O)
 Axel
 -----Original Message-----
 From: Joey Harriman [mailto:s808o)at(unb.ca]
 Sent: 6 novembre, 2003 10:14
 To: CCL posting
 Subject: CCL:clustering database molecules
 Hi,
 I am trying to cluster a database of a tetrapeptide with different geometries.
 I am
 currently using MOE and have tried varying fingerprints (at the 99%) threshold
 with no
 success.  Does anyone have a good system for clustering different geometries of
 the
 same molecule from a database?  Thanks in advance for any help you can offer.
 Joey
 ----------------------------------
 Joey Harriman
 MSc Computational Chemistry
 University of New Brunswick
 Room 228
 Toole Hall
 s808o)at(unb.ca
 476-6543
 ----------------------------------
 -= This is automatically added to each message by the mailing script =-
 To send e-mail to subscribers of CCL put the string CCL: on your Subject: line
 and send your message to:  CHEMISTRY)at(ccl.net
 Send your subscription/unsubscription requests to: CHEMISTRY-REQUEST)at(ccl.net
 HOME Page: http://www.ccl.net
 | Jobs Page: http://www.ccl.net/jobs
 If your mail is bouncing from CCL.NET domain send it to the maintainer:
 Jan Labanowski,  jkl)at(ccl.net (read about it on CCL Home Page)
 -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+