Prediction of transition state structures of gas-phase chemical reactions via machine learning
Problem
Predicting transition state structures of chemical reactions is challenging because constructing initial structures is difficult and computational costs are heavy.
Why it matters
Transition state structures are essential for understanding reaction mechanisms and exploring reaction networks.
What the AI did
A machine learning model derives the interatomic distances of a transition state from atomic pair features reflecting reactant, product, and linearly interpolated structures.
Results
93.8% success ratio for quantum chemical saddle point optimizations; 88.8% of optimization results had energy errors below 0.1 kcal/mol.
Impact
The model enables accurate transition state prediction for general organic reactions, aiding mechanism studies and reaction network exploration.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Sunghwan Choi |
| Citations | 58 (OpenAlex, at import time) |
| References | 50 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-023-36823-3 |
| Open access | gold |
| Keywords | transition states, machine learning, quantum chemistry, saddle point, reaction prediction |
Importance rating
significant — thousands to millions affected
Abstract states TS structures are essential for understanding reaction mechanisms and exploring reaction networks, benefiting chemistry broadly; no scale stated and impact is on computational method, not direct lives.