An exercise in the form of a Jupyter Notebook with describes how to use Python, pandas, and rdkit to enumerate isosteric replacements of a given drug-like chemical structure.
There are two versions of the notebook
- the "normal" one,
isosteric-replacements, that should be the first one to be tried out. - the "expanded" version,
isosteric-replacements-expanded, which include cross-joining operations on the isosteric database, increasing greatly the number of isosteric replacements the notebook can suggest.
The logic done is quite simple. If chemical groups A and B are both isosteres of a chemical group C, that means that A and B should (or must) be isosteres of each other. This logic was implemented using the free software KNIME to create the extended table. The workflow is available on the KNIME public hub.
It is required that you have the Python modules pandas and rdkit (and their dependences) configured in your conda environment.
(c) Cedric Graebin (Drug Discovery Unit, University of Dundee, Scotland). This notebook is distributed under the terms of the GPL v3 license.