A new review published in the American Journal of Biomedical Science & Research explores how artificial intelligence is revolutionizing preclinical drug development. The paper highlights the pharmaceutical industry’s struggle with declining productivity over the past 70 years, where FDA drug approvals have halved nearly every decade despite technological advances.
The authors detail how modern AI, particularly transformer-powered deep learning networks, is helping to address these challenges by enabling more sophisticated phenotypic drug discovery approaches. These AI systems can now process and integrate complex biological data from multiple sources, including genomics, proteomics, and metabolomics, to create more accurate models of human biology.
The review showcases how AI is being implemented across the entire drug discovery pipeline, from initial screening to lead optimization. Of particular note is the integration of transformer architecture, originally developed for natural language processing, which has proven remarkably effective at handling the complexity of biological systems and drug interactions.
The authors, including researchers from GATC Health and the University of California Irvine, present a comprehensive AI-powered pipeline that could potentially reverse the troubling trend of rising drug development costs and declining returns, known as “Eroom’s Law” (Moore’s Law spelled backward).
