(Though the hybridization field is empty, I noticed just now.)
Wednesday, March 30, 2011
Reporting missing atom types...
About a month ago, I blogged about how to report new atom types the CDK has no knowledge about yet. Earlier this week I had a great overview from Julio of a view phosphorus atom types, a PDF looking like this:
(Though the hybridization field is empty, I noticed just now.)
(Though the hybridization field is empty, I noticed just now.)
Sure I have a list of few failures. As my chemistry is not my strength, nonetheless I will give it a shot.
ReplyDeletehttps://gist.github.com/1016328
I will follow the protocol for the bug report.
there are some molecules which i doubt about structures in the KEGG itself like C01793.mol Chlorophyll -Mg has neutral charge in the compound in it!!!
ReplyDeleteAbsolutely! The list of positives in such a screening is indeed a mix of true positives (the CDK incorrectly not recognizing a known atom type) as well as false positives (the input really being wrong, like the uncharged metal ion).
ReplyDeleteActually, people are using the CDK atom typing for that purpose too: to validate the input, and it is getting pretty good at that too :)
ReplyDeleteall the neutral metals are giving false results there should be some default atomtype for neutral metal atoms!!!!
ReplyDeleteSure, nimishg. The challenge is to know how many formal neighbors, lone pairs, pi bonds, etc, such an atom has. Can you tell me?
ReplyDelete