The Se.2 atom type is defined in the CDKAtomTypeMatcher.java but its definition is missing in the ontology file cdk-atom-types.owl.
This patch fixes the missing atom type.
Few examples with Selenium combinations:
Asad, the result lists shows various atom types. Which one particular do you have in mind?
Need to change the default milestone
Excellent - will review tomorrow. This is a long standing issue - see related bugs
I think the actual matcher needs changing here... the Se.2 appears twice - once for 0 neighbours and once for 1 neighbour. This molecule perceives as Se.2: http://www.ebi.ac.uk/chebi/advancedSearchFT.do?searchString=32744&queryBean.stars=3&queryBean.stars=-1
Which would make sense, 2=2 formal neighbours. This also matches other atom type schemes (Hg.2 = 2 neighbours etc.). I wonder if this would still be broken though as that molecule has an implicit hydrogen which the typer doesn't considered in this case...
John, I will take it from here...
Gilleain's and my initial investigation showed a false negative, but then Steffen found a hit too...
A full fix for this requires various unit tests... I will look at this one today.
BTW, having 0 and 1 or even 2 neighbors makes perfect sense, because there may be missing hydrogens :)
BTW, H2Se should already be correctly perceived... Asad, it seems your email reply did not make it into the tracker yet, but thanx for these data links into ChEBI and PDBechem:
This patch matches half of the patch of Gilleain, and is for a single selenium atom without any neighbors.
I'll close it as a duplicate.
Updated patches pending.