Similarity-guided Monte Carlo Tree Search for bio-retrosynthesis

Metabolic engineering aims to produce chemicals of interest from living organisms, to advance towards greener chemistry. However, the research and development process is still long and costly and efficient computational design tools are required to explore the chemical biosynthetic space. Here, we provide RetroPath 3.0, an open-source, modular, command line tool that explores the bio-retrosynthesis space using a similarity-guided Monte Carlo Tree Search. We validate it on a dataset of manually curated experimental pathways as well as on a larger dataset of successful metabolic engineering projects. Moreover, we provide a novel feature, that suggests potential media supplements to complement the enzymatic synthesis plan.