PreprintCC BYChemRxiv
Universal Model for Atoms (UMA) in Computational Asymmetric Catalysis: Comparing Enantioselectivity Predictions
Abstract
Universal model for atoms (UMA) was evaluated for predicting enantioselectivity through direct modeling of diastereomeric transition states via conformational ensemble averaging. Across two organocatalytic and two metal-catalyzed transformations, UMA produced experimental selectivity trend with accuracy within 1-2 kcal/mol identifying key structural features governing stereoinduction. Findings discussed here highlight the promise of UMA foundation models as efficient alternatives for expensive quantum chemical methods in mechanistic modeling while also revealing current scope and limitations.
Related
- Bridging Chemical Literature and Molecular Graphs: LLM-Augmented Enantioselectivity Prediction for Iridium-Catalyzed Asymmetric C-H Borylation.
- Chiral Pyrazolinylidene Complexes for Asymmetric Catalysis
- Transition-State Steric Fields Capture Geometric Signatures Associated with Electronic Perturbations in Asymmetric Catalysis
- Crystal structure, Hirshfeld surface analysis, computational study, molecular docking predictions and biological evaluation of a new Schiff base (E)-2-((2,4-dichlorobenzylidene)amino)aniline
- Gold(I) vinylidenes as intermediates in catalysis: experimental and computational evidence
BibTeX
@misc{das2026universal,
title = {{Universal Model for Atoms (UMA) in Computational Asymmetric Catalysis: Comparing Enantioselectivity Predictions}},
author = {Soumik Das and Cyndi Qixin He and Yu-hong Lam},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15009465/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009465/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Universal Model for Atoms (UMA) in Computational Asymmetric Catalysis: Comparing Enantioselectivity Predictions
AU - Soumik Das
AU - Cyndi Qixin He
AU - Yu-hong Lam
PY - 2026
DA - 2026/09/24
PB - ChemRxiv
DO - 10.26434/chemrxiv.15009465/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009465/v1
AB - Universal model for atoms (UMA) was evaluated for predicting enantioselectivity through direct modeling of diastereomeric transition states via conformational ensemble averaging. Across two organocatalytic and two metal-catalyzed transformations, UMA produced experimental selectivity trend with accuracy within 1-2 kcal/mol identifying key structural features governing stereoinduction. Findings discussed here highlight the promise of UMA foundation models as efficient alternatives for expensive quantum chemical methods in mechanistic modeling while also revealing current scope and limitations.
ER -