CAREER: Learning mechanistic models with automated experiments

While the microbiome revolution revealed thousands of new species, it also presents a challenge as new species are identified at a faster pace than can be characterized by microbiologists using conventional methods. Without knowing what these new species do, understanding how a microbiome contributes to the behavior of biological systems is difficult. To address this challenge, this proposal develops autonomous systems that combine artificial intelligence and automated wet experiments to build mechanistic models of these understudied microbes. In addition, the proposal seeks to train students at rural-serving institutions to use these autonomous systems and to enable these students to continue this research upon return to their home institution. The proposal builds on the PIs previous work in developing BacterAI, an AI-driven robotic laboratory that answers scientific questions in a closed loop of designing, executing, and interpreting wet-lab experiments – all without human intervention. BacterAI will be enhanced in three ways. First, automated experiments will be combined with metabolic models to uncover the genome-scale regulatory networks for understudied bacteria. Second, a recommender system will be built to select experiments for efficient mapping genotype-to-phenotype across thousands of environments. Finally, access to this robotic laboratory will be extended to students at rural-serving institutions to train a new generation of biologists that leverage AI and automated science. All three of these goals will accelerate microbiology and illuminate the numerous understudied microbes that live on and around us. This project is supported by the Systems and Synthetic Biology Cluster of the Division of Molecular and Cellular Biosciences. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.