Small molecule structure elucidation from NMR spectra with generative and reasoning models
- 1Princeton University
Abstract
Elucidating the structures of unknown small molecules remains a major challenge in chemistry, pharmacology, and drug discovery. While nuclear magnetic resonance (NMR) spectroscopy is the gold standard for structure elucidation, interpreting spectra remains labor-intensive and requires expert knowledge to navigate the vast space of possible structures. Here, we present an AI system for rapid, automated de novo structure elucidation that combines a molecular generative model with a scientific reasoning agent. Its generative component, ChefNMR2, is an atomic flow-matching model that predicts small-molecule structures from molecular formulae and one-dimensional 1H and 13C NMR spectra. ChefNMR2-Agent augments our generative model with large language models (LLMs) to reason over and refine candidate structures. Across diverse benchmarks, ChefNMR2 substantially outperforms existing tools, offering calibrated confidence estimates and two to three orders of magnitude faster inference than frontier LLMs. We use ChefNMR2-Agent to elucidate the structures of four novel natural products, matching human-chemist accuracy in a fraction of the time. Our work demonstrates how specialized generative models and general-purpose reasoning agents can unite to solve challenging scientific inverse problems and accelerate NMR structure elucidation, thus opening new possibilities for AI-driven small-molecule discovery across diverse disciplines.
Related
- Advancing herbicide safeners: Integrating structure-guided design, pathway mechanism elucidation, and smart delivery
- Recent advances on structure and dynamics of G protein–coupled receptors using single-molecule Förster resonance energy transfer
- SMARTFlexDB: a paired apo-holo RNA structure database and analysis platform for RNA conformational flexibility and small-molecule recognition
- CausalOdyssey: Multi-Agent Local LLM Reasoning with Per-Patient Causal Structure for Rare Disease Retrieval
- scFLAME: a unified generative model for interpretable clustering, hierarchical structure discovery and marker-gene identification in single-cell RNA-seq data
BibTeX
@misc{xiong2026small,
title = {{Small molecule structure elucidation from NMR spectra with generative and reasoning models}},
author = {Z. Xiong and V. Y. Ying and J. Z. Huang and Z. Koch and J. D. Clark and R. C. Heeter and J. S. An and K. Matanachai and S. G. Rodriques and A. D. White and M. R. Seyedsayamdost and E. Zhong},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.29.762711},
url = {https://www.biorxiv.org/content/10.64898/2026.10.29.762711v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Small molecule structure elucidation from NMR spectra with generative and reasoning models
AU - Z. Xiong
AU - V. Y. Ying
AU - J. Z. Huang
AU - Z. Koch
AU - J. D. Clark
AU - R. C. Heeter
AU - J. S. An
AU - K. Matanachai
AU - S. G. Rodriques
AU - A. D. White
AU - M. R. Seyedsayamdost
AU - E. Zhong
PY - 2026
DA - 2026/10/29
PB - bioRxiv
DO - 10.64898/2026.10.29.762711
UR - https://www.biorxiv.org/content/10.64898/2026.10.29.762711v1
AB - Elucidating the structures of unknown small molecules remains a major challenge in chemistry, pharmacology, and drug discovery. While nuclear magnetic resonance (NMR) spectroscopy is the gold standard for structure elucidation, interpreting spectra remains labor-intensive and requires expert knowledge to navigate the vast space of possible structures. Here, we present an AI system for rapid, automated de novo structure elucidation that combines a molecular generative model with a scientific reasoning agent. Its generative component, ChefNMR2, is an atomic flow-matching model that predicts small-molecule structures from molecular formulae and one-dimensional 1H and 13C NMR spectra. ChefNMR2-Agent augments our generative model with large language models (LLMs) to reason over and refine candidate structures. Across diverse benchmarks, ChefNMR2 substantially outperforms existing tools, offering calibrated confidence estimates and two to three orders of magnitude faster inference than frontier LLMs. We use ChefNMR2-Agent to elucidate the structures of four novel natural products, matching human-chemist accuracy in a fraction of the time. Our work demonstrates how specialized generative models and general-purpose reasoning agents can unite to solve challenging scientific inverse problems and accelerate NMR structure elucidation, thus opening new possibilities for AI-driven small-molecule discovery across diverse disciplines.
ER -