Lab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio

Small molecule drug discovery faces significant hurdles in data quality and scalability, prompting Leash Bio to leverage DNA Encoded Libraries (DELs) for high-throughput measurement. By attaching DNA barcodes to millions of small molecules, researchers can physically screen vast chemical libraries against hundreds of protein targets, generating massive, reproducible datasets. Machine learning models often inadvertently learn the "sociology" of chemical creation—such as individual chemist biases or manufacturing shortcuts—rather than fundamental biophysics. To combat this, Leash Bio employs rigorous experimental controls, including competitive binding assays and site-directed mutagenesis, to force models to learn causal interactions rather than memorizing data fingerprints. Integrating these physically grounded datasets with large language models enables reasoning agents to navigate chemical space, potentially transforming early-stage drug discovery by replacing traditional, resource-intensive screening with predictive, lab-in-the-loop reinforcement learning.