Learning Effective Representations for Small Molecules

Speaker(s): Mukundh Murthy
Topic: Bioactivity Descriptors for Uncharacterized Chemical Compounds

Quantitative structure-activity modeling (QSAR) in computational chemistry is a task that involves predicting the binding affinity of a small molecule to a protein target given solely its molecular structure. Now, however, we are also interested in predicting more downstream properties including toxicity, side effects, and effects on gene expression – properties that concern both the biological and chemical properties of a molecule. This talk discussed the paper “Bioactivity Descriptors for uncharacterized chemical compounds,” which revolves around learning a generalizable and multi-modal representation for small molecules that can be applied across a large array of drug-discovery related tasks through integration of 25 small molecule datasets and a triplet network training task.

You can find a recording of this talk here (UM only).

Supplemental Resources

Bioactivity Descriptors for Uncharacterized Chemical Compounds
Computational Biochemistry Primer by Mukundh Murthy and Michael Trinh
ML for Molecular Property Prediction
MoleculeNet