From: Stefan K. <ste...@un...> - 2005-11-23 22:53:06
|
Hi all, we found that the daylight fingerprints do not work well (i. e. give a bi= g=20 result set) when searching for a common substructure, e. g. naphtalene. W= e=20 had the idea to extend the fingerprint with bits for fragments. I suggest= the=20 keep these fragments in a config file as smarts. The extended fingerprint= er=20 itself would then be easy to write. Any comments on this? Stefan --=20 Stefan Kuhn M. A. Cologne University BioInformatics Center (http://www.cubic.uni-koeln.de) Z=C3=BClpicher Str. 47, 50674 Cologne Tel: +49(0)221-470-7428 Fax: +49 (0) 221-470-7786 My public PGP key is available at http://pgp.mit.edu |
From: Egon W. <e.w...@sc...> - 2005-11-24 07:57:28
|
On Thursday 24 November 2005 12:23 am, Stefan Kuhn wrote: > we found that the daylight fingerprints do not work well (i. e. give a big > result set) when searching for a common substructure, e. g. naphtalene. We > had the idea to extend the fingerprint with bits for fragments. I suggest > the keep these fragments in a config file as smarts. MQL in the end. > The extended fingerprinter itself would then be easy to write. Any comments > on this? Biggest problem here is to define useful substructures. But some of-the-mind set would already help a lot. To do it properly, you would need to optimize it for a large data set, and find the set of substructures that is most discriminating and diverse. Egon |
From: Egon W. <e.w...@sc...> - 2005-11-27 13:33:20
|
On Thursday 24 November 2005 12:23 am, Stefan Kuhn wrote: > I suggest the keep these fragments in a config file as smarts. The extended > fingerprinter itself would then be easy to write. Any comments on this? OpenBabel 2.0 has just been released, and contains a substructure list based FingerPrinter. I highly recommend to consult with Geoff on the list of substructures. Maybe another BO activity... Egon |
From: Geoffrey H. <ge...@ge...> - 2005-11-27 20:14:02
|
On Nov 27, 2005, at 8:33 AM, Egon Willighagen wrote: > On Thursday 24 November 2005 12:23 am, Stefan Kuhn wrote: >> I suggest the keep these fragments in a config file as smarts. The >> extended >> fingerprinter itself would then be easy to write. Any comments on >> this? > > OpenBabel 2.0 has just been released, and contains a substructure > list based > FingerPrinter. I highly recommend to consult with Geoff on the list of > substructures. Maybe another BO activity... We actually now have two SMARTS based fingerprints as well as a pretty flexible fingerprint class. Chris Morley, who contributed the fingerprint code provided the SMARTS patterns more as examples than comprehensive lists. So yes, this would be a great BO activity to sync the fingerprint patterns. http://cvs.sourceforge.net/viewcvs.py/openbabel/openbabel/data/ patterns.txt?rev=1.2&view=log http://cvs.sourceforge.net/viewcvs.py/openbabel/openbabel/data/ SMARTS_InteLigand_051110.txt?rev=1.2&view=log Cheers, -Geoff |
From: Stefan K. <ste...@un...> - 2005-11-28 15:19:29
|
Hi Geoff, I would like to use your files. Is the SMARTS_InteLigand_051110.txt in so= me=20 standard format? It would be easier if it would be the same format as=20 patterns.txt. Stefan Am Sunday 27 November 2005 21:13 schrieb Geoffrey Hutchison: > On Nov 27, 2005, at 8:33 AM, Egon Willighagen wrote: > > On Thursday 24 November 2005 12:23 am, Stefan Kuhn wrote: > >> I suggest the keep these fragments in a config file as smarts. The > >> extended > >> fingerprinter itself would then be easy to write. Any comments on > >> this? > > > > OpenBabel 2.0 has just been released, and contains a substructure > > list based > > FingerPrinter. I highly recommend to consult with Geoff on the list o= f > > substructures. Maybe another BO activity... > > We actually now have two SMARTS based fingerprints as well as a > pretty flexible fingerprint class. Chris Morley, who contributed the > fingerprint code provided the SMARTS patterns more as examples than > comprehensive lists. > > So yes, this would be a great BO activity to sync the fingerprint > patterns. > > http://cvs.sourceforge.net/viewcvs.py/openbabel/openbabel/data/ > patterns.txt?rev=3D1.2&view=3Dlog > http://cvs.sourceforge.net/viewcvs.py/openbabel/openbabel/data/ > SMARTS_InteLigand_051110.txt?rev=3D1.2&view=3Dlog > > Cheers, > -Geoff > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files for problems? Stop! Download the new AJAX search engine that ma= kes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick > _______________________________________________ > Cdk-devel mailing list > Cdk...@li... > https://lists.sourceforge.net/lists/listinfo/cdk-devel --=20 Stefan Kuhn M. A. Cologne University BioInformatics Center (http://www.cubic.uni-koeln.de) Z=FClpicher Str. 47, 50674 Cologne Tel: +49(0)221-470-7428 Fax: +49 (0) 221-470-7786 My public PGP key is available at http://pgp.mit.edu |
From: Geoffrey H. <ge...@ge...> - 2005-11-28 19:08:45
|
Hi Stefan, I should note that the file has been renamed SMARTS_InteLigand.txt (the last bit was a datestamp). There are two pretty simple formats to the files. From src/ fingerprints/finger3.cpp: bool ReadPatternFile(const string& filename, vector<string>& lines) { //Reads two types of file: SMARTS + comments and vice versa //depending on whether the first line is #Comments after SMARTS So to answer your question, yes it's in a standard format. It contains a comment, whitespace, and a SMARTS pattern. Currently, the comment must not contain whitespace. Cheers, -Geoff On Nov 28, 2005, at 10:49 AM, Stefan Kuhn wrote: > Hi Geoff, > I would like to use your files. Is the SMARTS_InteLigand_051110.txt > in some > standard format? It would be easier if it would be the same format as > patterns.txt. > Stefan |
From: Christoph S. <c.s...@un...> - 2005-11-27 21:43:26
|
Stefan Kuhn wrote: > Hi all, > we found that the daylight fingerprints do not work well (i. e. give a = big=20 > result set) when searching for a common substructure, e. g. naphtalene.= We=20 > had the idea to extend the fingerprint with bits for fragments. I sugge= st the=20 > keep these fragments in a config file as smarts. The extended fingerpri= nter=20 > itself would then be easy to write. Any comments on this? > Stefan Just a minor remark. We do *not* have Daylight fingerprint!!!! This mailing list is publicly searchable and we have to be careful about=20 spreading strings like "the daylight fingerprints do not work well". We h= ave no=20 idea whatsoever what the Daylight fingerprints are exactly like. I doubt = that we=20 are anywhere close their implementation. So, the thing is: The CDK fingerprints do not work well. Full Stop. Stefan, we have also spoken about a solution. A number of additional bits= , for=20 example coding for rings and ring sizes, will already help a lot. Cheers, Chris --=20 Priv. Doz. Dr. Christoph Steinbeck (c.s...@un...) Head of the Research Group for Molecular Informatics Cologne University BioInformatics Center (http://almost.cubic.uni-koeln.d= e) Z=C3=BClpicher Str. 47, 50674 Cologne Tel: +49(0)221-470-7426 Fax: +49 (0) 221-470-7786 What is man but that lofty spirit - that sense of enterprise. ... Kirk, "I, Mudd," stardate 4513.3.. |
From: Stefan K. <ste...@un...> - 2005-11-28 09:43:42
|
Am Sunday 27 November 2005 22:43 schrieb Christoph Steinbeck: > Stefan Kuhn wrote: > > Hi all, > > we found that the daylight fingerprints do not work well (i. e. give = a > > big result set) when searching for a common substructure, e. g. > > naphtalene. We had the idea to extend the fingerprint with bits for > > fragments. I suggest the keep these fragments in a config file as sma= rts. > > The extended fingerprinter itself would then be easy to write. Any > > comments on this? > > Stefan > > Just a minor remark. We do *not* have Daylight fingerprint!!!! > This mailing list is publicly searchable and we have to be careful abou= t > spreading strings like "the daylight fingerprints do not work well". We > have no idea whatsoever what the Daylight fingerprints are exactly like= . I > doubt that we are anywhere close their implementation. > > So, the thing is: The CDK fingerprints do not work well. Full Stop. Thanks for clarification > > Stefan, we have also spoken about a solution. A number of additional bi= ts, > for example coding for rings and ring sizes, will already help a lot. Indeed. So my suggestion was to keep these fragments in a smarts file to = keep=20 the fingerprinter easily extensible. I just wanted to make sure this is a= =20 good idea. Stefan > > Cheers, > > Chris --=20 Stefan Kuhn M. A. Cologne University BioInformatics Center (http://www.cubic.uni-koeln.de) Z=C3=BClpicher Str. 47, 50674 Cologne Tel: +49(0)221-470-7428 Fax: +49 (0) 221-470-7786 My public PGP key is available at http://pgp.mit.edu |
From: Rajarshi G. <rx...@ps...> - 2005-11-28 17:05:10
|
On Mon, 2005-11-28 at 11:13 +0100, Stefan Kuhn wrote: > Am Sunday 27 November 2005 22:43 schrieb Christoph Steinbeck: > > Stefan Kuhn wrote: > > > > > Stefan, we have also spoken about a solution. A number of additional bits, > > for example coding for rings and ring sizes, will already help a lot. > Indeed. So my suggestion was to keep these fragments in a smarts file to keep > the fingerprinter easily extensible. I just wanted to make sure this is a > good idea. Has the CDK SMARTS parser been updated? ------------------------------------------------------------------- Rajarshi Guha <rx...@ps...> <http://jijo.cjb.net> GPG Fingerprint: 0CCA 8EE2 2EEB 25E2 AB04 06F7 1BB9 E634 9B87 56EE ------------------------------------------------------------------- COBOL is for morons. -- E.W. Dijkstra |
From: Egon W. <eg...@us...> - 2005-11-28 19:08:45
|
On Monday 28 November 2005 14:35, Rajarshi Guha wrote: > On Mon, 2005-11-28 at 11:13 +0100, Stefan Kuhn wrote: > > Am Sunday 27 November 2005 22:43 schrieb Christoph Steinbeck: > > > Stefan Kuhn wrote: > > > > > > Stefan, we have also spoken about a solution. A number of additional > > > bits, for example coding for rings and ring sizes, will already help a > > > lot. > > > > Indeed. So my suggestion was to keep these fragments in a smarts file to > > keep the fingerprinter easily extensible. I just wanted to make sure this > > is a good idea. > > Has the CDK SMARTS parser been updated? No, but adding stuff is really easy. Let me know is something is missing, then I will explain how to add it. Egon -- eg...@us... Blog: http://chem-bla-ics.blogspot.com/ GPG: 1024D/D6336BA6 |