Stereo Group in Reactor

Hi,





I am using the following code to react acid and amine.

Reactor r = new Reactor() ;


r.setMappingStyle(1) ;


r.setReactionString(args[0]) ;





MolImporter im = new MolImporter() ;


MolExporter xp = new MolExporter(System.err,"mol:V3ec") ;


Molecule[] reagents = new Molecule[args.length-1] ;


for(int i=1;i<args.length;i++) {


im.setFileName(args) ;


reagents = im.read() ;


}


r.setReactants(reagents) ;


Molecule[] p ;


int k = 0 ;


while( (p = r.react()) != null ) {


++k ;


for(int i=0;i<p.length;i++) {


p.clean(2,null) ;


p.setName(k + "." + i) ;


xp.write(p);


}


}

Reaction string is

([C;$(C=O):1][OH][H]).([N;!H0;!$(N[C,S]=O);!$(N[a]);!$(NC=N):3][H])>>([C:1][N:3])

args[1] is mol7.mol (a chiral acid), and args[2] is mol8.mol (a racemic amine) (both attached). The program spits out an amide, which I export to V3 mol file: mol9.mol. I am using evaluation JChem3.1.7 that I just downloaded a few days ago, JDK 1.5/WinXP and ISIS 2.5.





I am expecting my amide to have one AND stereo group (from the amine) and one ABS stereo center (from the acid), but I am seeing three ABS stereo centers in the product. If I re-export the racemic reagent as V3 mol file, I see one AND stereo group as expected.





I am very new to JChem (less than a week), so please go slow. Am I missing something?





Thanks,


dak

We are examining the problem and will answer you soon in details.

The conversion of the MDL chiral flag to enhanced stereo is not automatical. You explicitly request this if you convert the molecule with the "V3ec" export parameter. Therefore as long as your input is in "V2" and you use the chiral flag to denote absolute stereo, only this chiral flag will be set in your output if any of your chiral reagents has the chiral flag set. To set enhanced stereo in you output you must have enhanced stereo set in your input as well, e.g. by pre-converting your reagents with "V3ec".





The reason is that we do not want to set enhanced stereo in every chiral reagent - if we applied the "V3ec" conversion automatically then we should have convert every chiral molecule without a chiral flag to an "and1" group - even if the molecule does not come from an MDL mol file.

Nora wrote:

The conversion of the MDL chiral flag to enhanced stereo is not automatical. You explicitly request this if you convert the molecule with the "V3ec" export parameter. Therefore as long as your input is in "V2" and you use the chiral flag to denote absolute stereo, only this chiral flag will be set in your output if any of your chiral reagents has the chiral flag set. To set enhanced stereo in you output you must have enhanced stereo set in your input as well, e.g. by pre-converting your reagents with "V3ec". The reason is that we do not want to set enhanced stereo in every chiral reagent - if we applied the "V3ec" conversion automatically then we should have convert every chiral molecule without a chiral flag to an "and1" group - even if the molecule does not come from an MDL mol file.

Thank you for your reply.





It would be very convenient to have an import option that tells the Importer to automatically add enhanced stero information if structures are imported from a format that does not support enhanced stereochemisty. Native mol file writers (including ISIS, MDL/Draw, Cheshire and MDL/Direct) usually try to export mol files as V2 format, unless the molecule being exported contains some features that are not supported by V2 format, and I do not know of a way to change this behavior. As a result, I may be dealing with a mixture of V2 and V3 mol files for reagents, depending on which reagent require V3 representations, and this new import option would save me considerable code that (a) checks if I have v2 or v3 mol file and (b) exports structures to v3:ec and re-imports.





Having said that, I followed your suggestion and added the following code to generate stereo group information

Code:

private static Molecule makeStereoGroups(Molecule m) {       for(int i=0;i<m.getAtomCount();i++) {          MolAtom a = m.getAtom(i) ;          int p = a.getFlags() & MolAtom.PARITY_MASK ;          if( p == MolAtom.PARITY_ODD || p == MolAtom.PARITY_EVEN ) {             if( !m.isAbsStereo() ) {                a.setStereoGroupNumber(1) ;                a.setStereoGroupType(MolAtom.STGRP_AND) ;                            } else {                a.setStereoGroupType(MolAtom.STGRP_ABS) ;             }          }       }       return m ;    }

This gets me closer to where I want to be, but not quite. Here is another acid/amine pair: mol7a.mol (an acid with 1 AND stereo group), mol8a.mol (an amine with 1 AND stereo group), which result in the following product: mol9a.mol





I am expecting TWO AND stereo groups with two stereo centers each, but I only get ONE AND stereo group with four stereo centers. Exporting each reagent to V3 molfile (using MolExporter) and re-importing does not fix the problem either.





Thanks again,


dak

Instead of an import option, we are planning to add the MDL chiral flag -> enhanced stereo conversion to Standardizer in form of a new standardization task. But this will not be included in the next major JChem 3.2 release yet. I agree that it is useful to have this transformation available in some way.





mol7a, mol8a:


I have also discovered this bug in the enhanced stereo handling and fixed it for the next JChem 3.,2 release. A minor remark: with this correction, I get a slightly different molceule with your reaction SMARTS (mol9aec.mol). But the essence os that there will be two AND groups with 2 stereo centers each. Thanks for your comments.