Machine learning applied to enzyme turnover numbers reveals protein structural correlates and improves metabolic models
Problem
Experimental data on enzyme turnover numbers is sparse and noisy, limiting quantitative understanding of organism growth, proteome composition and physiology.
Why it matters
Turnover numbers are essential for understanding metabolism and the proteome at genome scale.
What the AI did
Machine learning predicted catalytic turnover numbers in E. coli from enzyme biochemistry, protein structure and network context, and the predictions parameterized two genome-scale proteome-limited metabolic models.
Results
Significantly higher accuracy in predicting quantitative proteome data than previous approaches; no numeric benchmarks stated.
Impact
Provides a tool for genome-scale understanding of metabolism and improved proteome prediction accuracy.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2018 |
| Work type | article |
| Authors | David Heckmann, Colton J. Lloyd, Nathan Mih, Yuanchi Ha, Daniel Craig Zielinski, Zachary B. Haiman, Abdelmoneim Amer Desouki, Martin J. Lercher et al. |
| Citations | 217 (OpenAlex, at import time) |
| References | 76 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-018-07652-6 |
| Open access | gold |
| Keywords | machine learning, enzyme turnover, metabolic modeling, proteome, E. coli |
Importance rating
significant — thousands to millions affected
Turnover numbers essential for understanding growth, proteome composition, and physiology at genome scale; broad foundational relevance to metabolism but no quantified impact scale given.