RDKit Functional Group Filter node bug


Ive been using the RDkit Functional Group Filter node and believe there is a bug.

If I use the KNIME Public Server workflow "Complex SAR" which I uploaded, and take this dataset of compounds which contains aryl rings with a NMe Piperazine group coming off the aromatic ring(amongst others), I then use the Functional Group Filter node with >0 set for Amine, then I get no matches.

If I choose any of the more specific amines, i.e. aliphatic amine or aromatic amine with >0 set, these also return no hits.

I checked the node is working by choosing ARomatic Halogen >0, and this did work pulling back the desired cpds. So I believe there is a bug with the amine filter list.



Apologies if this is a stupid question, chemical nomenclature is not my thing, but aren't both nitrogens in aryl NMe piperazine's tertiary?

The current patterns used in the functional group filters, which were originally intended for use in building-block selection, require that an amine have at least one H attached.



Ah okay.

You are right, both the Nitrogens in the example I mention are tertiary.

I didnt realise that the "amine" header mentioned in the node, is referring to secondary or primary amines only.

I was trying to use it to get only "basic" compounds out from a list of cpds.



If the tertiary amines were charged and protonated (the ultimated proof of basicity), the node would do what you want, but as it stands... not really.

Note: if this is something you regularly want to do, you could either use your own customized functional-group definition file or use one of the substructure nodes with the appropriate patterns.