Hi all

I've been looking at this feature request as well and I have found this article:

J. Chem. Inf. Comput. Sci., 1996, 36 (1), 25-34; DOI: 10.1021/ci950061u

It contains an algorithm for the use of graph theory to calculate the CIP priorities of an e.g. quatrivalent carbon atom and detect stereoisomeric centres.  anyway it reports a non-factorial computational cost with increasing atom count...

I am still on chapter 1, page 1 of the book: "graph theory for dummies" (trying to understand the difference between backtracking and partitioning and relaxation)  so is taking me a while to read through it. So in the meantime, I hope it is useful for someone else.



Message Received: Apr 08 2009, 03:41 PM
From: "Kieron Taylor"
To: "Developers forum for discussion about the Chemistry Development Kit
Subject: Re: [Cdk-devel] [Cdk-user] Detect stereocenters of a molecule


I've been following the stereocenter discussions on the mailing list,
and discovering how daunting a task stereochemistry is to capture.

I found this: http://www.mdpi.org/molecules/papers/61100915/61100915.htm

CIP rules on the face of it appear to be rather difficult to handle in
computational terms, on account of the "view" required to decide
clockwise vs anticlockwise. If someone better with matrices than I has
anything to say about it, I would be delighted to hear.

If I were to try and implement some of the above, can Java handle the
matrix stuff adequately? I only assume the capability is there and worth

Would it be helpful if I try to enhance the Morgan number code first in
order to accommodate this?



This SF.net email is sponsored by:
High Quality Requirements in a Collaborative Environment.
Download a free trial of Rational Requirements Composer Now!
Cdk-devel mailing list