Augmenting BMKGs With Cheminformatics And CADD
Biomedical knowledge graphs (BMKG) contain chemical compounds such as drugs, toxins, metabolites, cofactors or signaling molecules. These entities and some of their relations can be richly augmented with qualitative & quantitative properties by methods from cheminformatics, computer-aided drug design (CADD) and related fields. This enables numerical queries and many analyses such as filtering, clustering, embedding, similarity/outlier detection, QSAR modeling, ML, etc. The aim of this project is to make existing but scattered methods available in a Python package with a unified functional API, expose it to OpenCog Hyperon, and apply it in a PoC study to annotate and analyze Hetionet in MORK.
