BCAM Alumni & Former Members Seminar: Guiding synthetic biology via machine learning and automation technologies

Title: Guiding synthetic biology via machine learning and automation technologies

Abstract:
Synthetic biology allows us to bioengineer cells to synthesize novel valuable molecules such as renewable biofuels, anticancer drugs, or develop novel therapies. However, traditional synthetic biology approaches involve ad-hoc
engineering practices, which lead to long development times. One of the most important challenges in bioengineering is effectively using -omics data to guide metabolic engineering towards higher production levels. In this talk, I will show our efforts in developing automation pipelines for data collection, their analysis through machine learning (ML), and the generation of recommendations, with the goal of accelerating the engineering cycle. I will focus on the Automated Recommendation Tool (ART), a tool that leverages ML and probabilistic modeling techniques to guide synthetic biology in a systematic fashion, without the need for a full mechanistic understanding of the biological system.
Furthermore, using sampling based optimization ART provides a set of actionable recommendations for the next engineering cycle, alongside probabilistic predictions of their outcomes. We demonstrate the capabilities of ART on simulated and real data sets and discuss possible difficulties in achieving satisfactory predictive power. The presented work was done in collaboration with Zak Costello, Mark Forrer, Matthew Incha, Vincent Blay Roger, Jose Manuel Marti, Hector Garcia Martin.