openmolecules.org

 
Home » DataWarrior » Bug Reports » Stereo issue? (Conversion from Smiles to MOL: missing stereo center)
Stereo issue? [message #65] Mon, 01 June 2015 01:49 Go to next message
bruno.b is currently offline  bruno.b
Messages: 10
Registered: October 2014
Location: Germaniy
Junior Member
Hi Thomas,

I wrote a small Java program that uses the open source Datawarrior library. This program converts the smiles of the drug benazepril into a MOL file. I used the smiles from Pubchem. The generated MOL has only one stereo center instead of two. I have attached the MOL output and the java code to this post.

I am not sure that the Java program I wrote is correct.

Thank you for providing the library,


Bruno
Re: Stereo issue? [message #66 is a reply to message #65] Thu, 18 June 2015 20:50 Go to previous messageGo to next message
thomas is currently offline  thomas
Messages: 119
Registered: June 2014
Senior Member
Hi Bruno,

I am sorry for answering so late. My intention is to answer within 24 hours after being aware of a question. If you don't get a quick answer, you may assume, that I simply didn't see your question (as in this case). Then please send me an e-mail...

I checked your source code, which seem perfectly correct. With the current SmilesParser it also works fine. The problem is
that Smiles were not used at Actelion and therefore I didn't notice that the stereo recognition was not working well.
Recently, we published a paper and a website with structure search on all Wikipedia structures. Since this requires parsing
the Smiles strings from the Wikipedia database, I enhanced the SmilesParser and fixed lots of stereochemistry related and other
issues. You can find the up-to-date version on GitHub in the OpenChemLib project. It will also arrive as part of the
source code of the upcoming new DataWarrior release in hopefully early June.

Thomas
Re: Stereo issue? [message #116 is a reply to message #66] Thu, 05 November 2015 10:23 Go to previous messageGo to next message
bruno.b is currently offline  bruno.b
Messages: 10
Registered: October 2014
Location: Germaniy
Junior Member
It is indeed improved.


I tested further the SMILES to structure generator, Here is a SMILES that does not give good 2D coordinates: c1ccc(C23C(N(c4ccc(cc4)C)C(C2C2c4c(C3c3c2cccc3)cccc4)=O)=O)c c1


This is not an easy case.


Best regards,

Bruno
Re: Stereo issue? [message #119 is a reply to message #116] Tue, 10 November 2015 15:54 Go to previous message
thomas is currently offline  thomas
Messages: 119
Registered: June 2014
Senior Member
This is indeed a little tricky. But it is a problem of the CoordinateInventor and not of the SmilesParser.
Generating coordinates is a never ending story. Paul Labute has once written a nice paper about it...

Best wishes,

Thomas
Previous Topic: copy/paste from ChemBioDraw 14 to latest DW causes crash
Next Topic: Macros
Goto Forum:
  


Current Time: Tue Jan 23 12:38:28 CET 2018

Total time taken to generate the page: 0.00275 seconds