Speaker: Dr. Garrett Roell
Abstract:
Non-model microbial species offer unique potential for biomanufacturing. However, the development of novel microbial hosts requires holistic metabolic knowledge and modeling tools for strain engineering and fermentation optimization. I will present computational methods for unravelling the metabolic regulations of two non-model microbes: Rhodococcus opacus and Clostridium carboxidivorans. R. opacus is an aromatic-tolerant bacteria that can produce lipid-based biofuels from lignocellulose. C. carboxidivorans is a syngas-consuming bacteria that can produce alcohol-based biofuels. Among tested machine learning methods, random forests and support vector machines performed best, and these trained models were able to offer guidance for fermentation optimizations and model predictive control of bioreactor operations.
By 2027, the annual market for the products of synthetic biology are projected to be over $30 billion in the USA alone. However, strain development suffers both high failing rates and high experimental costs. To overcome this challenge, knowledge engineering and AI methods are essential approaches. I will also discuss my current work focusing on the development of a ChatGPT-enhanced biomanufacturing database (ImpactDB; impact-database.com) and AI methods for guiding synthetic biologists and bioprocess engineers to enhance microbial cell factory performance.
August 18, 2023 recording (click here)
May 6, 2022 Transcript (download here)
May 6, 2022 Recording (click here)