A machine learning Automated Recommendation Tool for synthetic biology
Problem
Engineering microbes to produce valuable molecules (biofuels, drugs) is slow because traditional synthetic biology relies on ad-hoc trial-and-error practices.
Why it matters
Long development times delay production of renewable biofuels and anticancer drugs.
What the AI did
ART uses machine learning and probabilistic modeling to recommend which strains to build in the next engineering cycle, with probabilistic predictions of their production levels, without requiring full mechanistic understanding.
Results
not stated in source (abstract truncated; only says demonstrated on simulated data and real projects: biofuels, hoppy-flavored beer, fatty acids, tryptophan)
Impact
Demonstrated on real metabolic engineering projects, but no measured outcomes are given in the abstract.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2020 |
| Work type | article |
| Authors | Tijana Radivojević, Zak Costello, Kenneth Workman, Hector Garcia Martin |
| Citations | 277 (OpenAlex, at import time) |
| References | 72 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-020-18008-4 |
| Open access | gold |
| Keywords | machine learning, synthetic biology, metabolic engineering, strain recommendation, biofuels |
Importance rating
real but narrower advance
Abstract cites long development times for biofuels and anticancer drug production; ART demonstrated on real projects but no measured outcomes or scale stated.