UC San Diego scientists have unveiled an AI platform named POLYGON for simulating early drug discovery stages.

POLYGON leverages a polypharmacology approach, targeting multiple proteins linked to diseases, unlike the traditional model focused on single targets.

The AI was trained on over a million bioactive molecules, enabling it to create novel chemical formulas potentially effective as cancer treatments.

Reinforcement learning further empowers POLYGON to formulate compounds active against multiple specific disease proteins.

Early tests indicate synthesized drugs are significantly active against cancer-related proteins, marking them as promising therapeutic candidates.

POLYGON's technology is open source, broadening access to its potential benefits in drug discovery.

The research, backed by the National Institutes of Health and published in Nature Communications, indicates ongoing improvements and wider applications for polypharmacology.