The abstract is read at the publisher; the summary is JClub's.
A large language model–driven pipeline trained on 29 006 metabolic engineering entries predicts cell-factory engineering targets and, in yeast, enabled geraniol overproduction with several novel targets.
- Why it matters: Metabolic complexity makes it hard to predict effective engineering targets, so designs often rely on trial and error; extracting strategies from prior work could improve future cell factory development.
- What they did: The team built a database of 29 006 entries across 1210 products and 751 organisms, then trained a deep learning model to predict engineering targets from this knowledge.
- The result: The model outperformed traditional algorithms, generalized to unseen products and multigene combinations, and experimental validation in Saccharomyces cerevisiae for geraniol identified several novel targets.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.