Lab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio
DNA encoded librariessmall molecule drug discoverymachine learning modelsprotein-ligand bindingcombinatorial chemistryKaggle competitionHermes modelchemist biasactivity cliffsprotein mutationslanguage modelslab-in-the-loopdrug discovery scalingADMET propertiesexperimental rigorreinforcement learningbinding predictionstructural biologychemical library design
Ian Quigley, co-founder and CEO of Leash Bio, discusses the challenges and innovations in small molecule drug discovery using DNA encoded libraries (DELs) and machine learning. The episode covers the technical details of DELs, lessons from a large Kaggle competition on binding prediction, the development of the Hermes model, and the use of language models for lab-in-the-loop reasoning to accelerate drug discovery.