[2S00a] 01. ASBA - Part 5

2:00 PM - 2:20 PM JST(5:00 AM - 5:20 AM UTC)

[2ASBA-07-KL] Accelerating Strain Engineering by combining AI Guidance and ME Wisdom

○Gregory Stephanopoulos (MIT)

Metabolic engineering (ME) emerged, 35 years ago, as the directed modification of cellular metabolic networks for the purpose of creating efficient biocatalysts for the production of fuels, chemicals and pharmaceutical products. As such, it has had a profound impact in the advancement of biomanufacturing and industrial biotechnology. To achieve its goal, Metabolic Engineering combined principles of chemical reaction engineering with stable isotope labeling and measurements of mass spectrometry to the analysis of metabolic reaction networks and used the results of this analysis to guide the application of state-of-the-art molecular biological methods to remove rate-limiting steps and streamline the function of metabolism towards efficient product synthesis. Numerous applications arose, in industrial biotechnology but also in human health, such as diabetes and insulin resistance and metabolism of cancer. Yet, this quest for correlations between metabolic function and cellular physiological data has been demanding, opaque and the domain of only a few skilled researchers. These limitations can now be overcome with the use of tools from Artificial Intelligence. AI methods bypass the need of rational approaches for pathway analysis and evaluation in favor of very efficient black box methods that make use of LLM-based correlations. These methods enable the analysis and evaluation of enormous amounts of data of very diverse genetic, spectroscopic, chemical and bioprocessing nature allowing extraction of rich information content that guides further strain improvement and bioprocess optimization. I will illustrate these trends with examples from my current research on natural product synthesis, and attempt to project lessons learned to charting the most impactful new directions to which AI technologies will direct industrial biotechnology in the foreseeable future.