Machine Learning to Predict Enzyme–Substrate Interactions in Elucidation of Synthesis Pathways: A Review
Problem
Experimentally determining enzyme-substrate interactions is slow and costly, and physics-based simulations are too slow for large search spaces.
Why it matters
These interactions are fundamental to elucidating synthesis pathways and synthetic biology, but experiments are slow and expensive.
What the AI did
Reviews AI methods (SVMs, neural networks, decision trees) that automate repetitive processes, extract patterns, and cut computation time while covering vast search spaces.
Results
No metrics; abstract only claims AI 'significantly reduces' computation time
Impact
Per this review, AI methods have sped up computational prediction of enzyme-substrate interactions.
Record
| Venue | Metabolites |
|---|---|
| Published | 2024 |
| Work type | article |
| Authors | Luis Fernando Salas Nuñez, Álvaro Barrera-Ocampo, Paola A. Caicedo, Natalie Cortés, Edison H. Osorio, María Francisca Villegas-Torres, Andrés Fernando González Barrios |
| Citations | 45 (OpenAlex, at import time) |
| References | 67 |
| Model or method | support vector machines, neural networks, decision trees (reviewed) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.3390/metabo14030154 |
| Open access | gold |
| Keywords | machine learning, enzyme-substrate interactions, synthesis pathways, synthetic biology, neural networks |
Importance rating
real but narrower advance
Enzyme-substrate interactions are called fundamental to synthesis pathways and synthetic biology, but this is a review with no metrics or scale; real but narrower computational advance.