Depth-First: Extending InChI Stereochemistry

Extending InChI Stereochemistry

Rich Apodaca — Wed, 09 Jul 2008 10:18:00 +0000

As covered by Reuters and many other wire services, ArtusLabs and Boston University's CMLD have teamed up to extend InChI's stereochemistry support:

DURHAM, N.C.--(Business Wire)-- ArtusLabs, Inc., a leading provider of life science software tools and data management solutions, has entered into a partnership with Boston University's Center for Chemical Methodology and Library Development (CMLD) to develop a way to standardize and expand the way in which stereochemistry, and ultimately a three-dimensional structures, are represented in the International Chemical Identifier (InChI(TM)).

With the increasing use of molecules containing axial chirality , planar chirality and other forms of non-tetrahedral stereogenicity in chemistry, the move by ArtusLabs and CMLD could be significant.

Put simply, the ability of cheminformatics to represent certain kinds of compounds has fallen way behind the ability of chemistry to make them. While molecules once considered mere oddities 30 years ago continue to pour into corporate compound collections, laboratory notebooks, and product catalogs, cheminformatics has been stuck with a form of molecular representation that hasn't changed significantly in several decades.

InChI isn't alone. All three of the most widely-used molecular representation systems now in use (Molfile, SMILES, and CML) suffer from fundamental limitations in representing axial chirality, planar chirality, and multicenter bonding.

The kind of work being undertaken by ArtusLabs and CMLD is essential if cheminformatics is to continue to keep pace with new developments in chemistry.

"Extending InChI Stereochemistry" by ChemSpiderMan

Wed, 20 Aug 2008 02:19:38 +0000

My question is why the InChI development team ave not been consulted on this work. SInce InChI is "blessed" by IUPAC it doesn't make good sense to do this without the participation of the team who developed InChI in the first place. The InChI should not fork but be kept consistent and developed in collaboration with IUPAC...it just makes sense to do it that way.

"Extending InChI Stereochemistry" by CLYDE DAVIES

Tue, 19 Aug 2008 18:27:12 +0000

I'd like to see INChi extended to handle generics properly. Did you know that if you try to generate INChIs for the structures in the Available Chemicals Directory, about 7% of them fail? This is because they comtain are weird atoms like 'Pol' for a polymer substrate.

"Extending InChI Stereochemistry" by Rich Apodaca

Fri, 11 Jul 2008 01:51:51 +0000

Keith, sounds interesting. Where can I find some examples of this molfile extension?

"Extending InChI Stereochemistry" by keith.taylor@symyx.com

Thu, 10 Jul 2008 19:43:29 +0000

The molfile can support axial, planar, and multi-center bonding. You can draw the structures, save them, and read them back in. The problem is to get a search engine to differentiate the isomers. To date Symyx's database engine can recognize axial isomers but not planar isomers or multi-center bonded structures where hydrogen is the bridging species, chlorine bridges are recognised.

"Extending InChI Stereochemistry" by baoilleach

Thu, 10 Jul 2008 07:47:21 +0000

Rich - I sense your hand in this. Come on, own up - you've just gone and revolutionized cheminformatics with your blog posts.

"Extending InChI Stereochemistry" by Egon Willighagen

Thu, 10 Jul 2008 00:14:03 +0000

Yeah, it's weird that this got found via IT channels, instead of science channels...