How Genomic Language Models Could Accelerate Biomanufacturing

A unique team of specialists in data generation and artificial intelligence are co-developing an AI genomic model capable of predicting genomic edits that improve commercial-scale protein production. This collaboration between Triplebar and UC Berkeley—with funding from the U.S. Department of Defense and the National Science Foundation through a BioMADE grant—aims to accelerate strain optimization for more resilient, cost-effective protein manufacturing across therapeutic and industrial applications.

In this episode of Off Script, we spoke with Shawn Manchester, CEO of Triplebar, about how the company's ultra-high-throughput microfluidics platform is being paired with UC Berkeley's genomic language model to transform strain engineering. Manchester discussed the motivation behind the collaboration, explained how this genomic language model differs from traditional AI approaches, and explored how they could improve protein expression, shorten development timelines, and reduce manufacturing costs. He also shared his vision for how AI-driven strain optimization could help enable the next generation of protein therapeutics and other biomanufactured products.