RGroup Composition


Is it possible to have a node which will take a query indigo molecule in one in port, depicted with Rx groups, and in another in-port to have a table of indigo molecule columns depicted with a point for connection, such as with an A atom (i.e. 4-A-Pyridine) (it would be good to be able to specify the atom you want to use for the connection, this way you can simply load in a set of benzylbromides for instance an specify Br as the connection point), and there is then the possibility to match up these multiple Indigo Molecule columns to the Rx points on the scaffold in the other in-port. The node then enumerates all possible combinations of groups at each Rx position and outputs the structures.

Basically an enumerator of sorts.




Actually it may be better to have an additional node (Functionality Clipper node) which will take any group of Indigo molecules and you specify which atom you want to be a connecting point such as A, Br, I, NH2 etc, and it clips this atom off and replaces it with a connecting point (as depicted with a squiggle in the RGroup Decomposition node output).

These molecules with the connecting point are then loaded into the R Group Composition node, along with a scaffold with Rx groups in a separate in-port. This then allows you to match up the molecules with connecting points to all the Rx positions and then enumerates all possibilities.

How about it,



Thank you again for another useful suggestion.

Yes, the R-Group composer node would be useful, and currently in the core Indigo library we have a reaction enumerator, that does somewhat similar thing with reactions instead of molecules. This functionality is currently not in KNIME because, well, we have not got around to reactions in KNIME yet.

Implementing a molecule R-Group enumerator sounds reasonable, and we will be trying to implement it this year in Indigo, and then add a corresponding node to KNIME.


Best regards,


Fantastic, I'm glad there are some reaction enumerators in the pipeline, that would be really handy.

An R-Group enumerator would be good too without the need of having to create a reaction template and just want to do some simple joining of scaffolds and R groups.