Is there a simple way to count functional groups?

Groups Cheminformatics
1

Hello!

I'm having a list of SMILES and I would like to receive columns which contain the numbers of each functional group for a molecule.

 

I've noticed that there is an RDKit Functional Group Filter Node which lists many functional groups. Since there is a node to filter by functional groups I assumed that there might also be one that counts and lists those functional groups.

 

Help is much appreciated!

Kind regards

mtest

2

The SMARTS patterns for the functional groups are viewable if you press the 'i' button in the Functional Group filter node - the best you could do probably at the moment is to use those as the basis for a second input table for the RDKit Substructure Counter node, although I would assume that it wouldn't be a huge leap for the RDKit developers to create a counter rather than filter version of the FG node

Steve

1 Like
3

Hey,

that helped me a lot already - thank you very much!

Also I'd really appreciate such a node for future...

 

mtest

4

The functional group definitions that the RDKit uses for the Functional Group Filter node happen to all be in github:

https://github.com/rdkit/rdkit/blob/master/Data/Functional_Group_Hierarchy.txt

Thanks to the fact that KNIME's file reader node can read from a URL, you can use these directly in a workflow after a bit of post-processing.

The attached workflow does just that.

Since the RDKit repo has another set of functional groups (these were collected for a different purpose and are broader, but perhaps not as carefully defined):

https://github.com/rdkit/rdkit/blob/master/Data/FunctionalGroups.txt

I created a second metanode that reads those in and does the required post-processing. 

And, finally, since there is also a set of SMARTS for "Topliss descriptors":

https://github.com/rdkit/rdkit/blob/master/Data/FragmentDescriptors.csv

I created a metanode for those as well.

Hopefully the general pattern I use here, as well as the workflow itself, is useful.

-greg

 

2 Likes
5

Hey greg,

thanks this workflow is very helpful!!