Ian Quigley on Lab-in-the-Loop Systems at Compute Speed
Ian Quigley from Leash Bio discussed how automating DNA-encoded library assays generates high-quality data to train AI agents and identify flawed models in drug discovery. This approach addresses the critical need for reliable, high-throughput biological data required to train modern AI agents to think like chemists and accelerate drug development. Leash Bio successfully automated DNA-encoded library (DEL) assays, producing clean data that helps expose cheating or overfitted machine learning models.
## BACKGROUND
DNA-encoded libraries are collections of small molecule compounds used in drug discovery to screen for biological targets. Lab-in-the-loop systems integrate automated wet-lab experiments with computational AI models to continuously generate, analyze, and learn from biological data at high speeds.