BODO-CHEMINF alignment

[Egon Willighagen]

Hi all,

now that Cristian finished the alignment of the CDK descriptors which were using the BODO before with CHEMINF instead, I started looking at the CDK code to finally make this transition.

Replacing the URIs in the CDK for the descriptor specifications is easy enough, but the tricky thing is the DescriptorEngine...

That is, it seems that CHEMINF does not provide all the classification info that is used by the engine.

Some things seem fine. For example, in the BODO the acidGroupCount desciptor is:

        <isClassifiedAs rdf:resource="#molecularDescriptor"/>

        <isClassifiedAs rdf:resource="#constitutionalDescriptor"/>

And in CHEMINF it is a subclass of "constitutional descriptor" too.

However, the "topological PSA" descriptor is in BODO classified as a topological descriptor, but not in CHEMINF.

I will first patch the DescriptorEngine to use the CHEMINF in the first place, and then try to figure out what inconsistencies exist, and then update the CHEMINF to fix things.

Comments, ideas, thoughts, and help is very much appreciated!

Egon

[Gang Fu]

Hi Egon,

I looked at CHEMINF, and found the TPSA is a subclass of "surface area descriptor", which is a subclass of "dimensional extend descriptor", which is a subclass of "geometric descriptor".

Do you mean to move "TPSA" from (subclass*3) of "geometric descriptor" to "topological descriptor"?

Best,

Gang

--
You received this message because you are subscribed to the Google Groups "cheminf-ontology" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cheminf-ontolo...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.