Researchers unveil PhenMol for drug discovery breakthroughs

A team of researchers from institutions including the University of Science and Technology of China and IBM Research has introduced **PhenMol**, a novel AI framework designed to revolutionize drug discovery by preserving molecular structural integrity while incorporating cellular phenotype data.

PhenMol tackles a longstanding challenge in multimodal representation learning for drug discovery: how to align molecular structures with cellular responses without distorting chemical space. The framework achieves this by disentangling molecular and cellular representations into shared and private components, allowing phenotype-guided alignment while maintaining chemical neighborhood organization. In experiments involving 30,400 molecule-cell morphology pairs, PhenMol demonstrated superior performance across 270 bioactivity tasks, molecule-phenotype retrieval, and clinical trial outcome prediction. The model’s ability to preserve molecular neighborhoods—validated through ECFP4-based structural analysis—reduced embedding distortion compared to existing multimodal alignment methods.

The work highlights a critical advancement for drug discovery pipelines, where accurate molecular representations are essential for predicting drug efficacy and safety. By bridging cellular phenotypes with chemical knowledge, PhenMol offers a more holistic approach to understanding how molecules interact with biological systems, potentially accelerating the discovery of new therapeutics.