Iambic uses AI predictions and AI agents in closed-loop framework to guide wet-lab experiments.
“These three elements can work independently, but complement one another rather than act as substitutes: each performs a distinct function, and we use them together in our discovery programs. All three elements operate together as a closed-loop DMT framework, in which our team uses AI predictions and AI agents to guide wet-lab experiments, with resulting data continuously fed back to retrain our models. Within each cycle, the output of one element can inform the others, and one element may use data generated by another. As an example, a therapeutic molecule generally must attach to a specific biological target, often a disease-associated protein, to have a therapeutic effect. NeuralPLexer is used to predict structures and protein-ligand affinities, meaning how a molecule would fit and bind that target, and to enable generation of hypotheses. Its predicted structures and binding characteristics may inform which molecules Enchant evaluates for preclinical and clinical properties. Enchant's predictions may in turn guide which molecules our robotic workflows synthesize and test; and the resulting laboratory data are then used to fine-tune both Enchant and NeuralPLexer for the next cycle. We use the platform both to advance our pipeline of wholly owned drug candidates and to power collaborations with technology, biotechnology, and pharmaceutical development partners.”