Multi-Agent Closed-Loop Reasoning for Organic Structure Elucidation from Multimodal Spectra

📅 2026-08-12
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
This study addresses the limited generalizability of conventional spectral interpretation to unseen samples by proposing MACROS, a multi-agent system that simulates expert iterative hypothesis testing and closed-loop reasoning. By integrating multimodal spectral data with large-scale training, the model develops emergent chemical intuition and autonomously learns underlying spectral correlations rather than merely memorizing patterns. Experimental results demonstrate that MACROS achieves precise zero-shot structural elucidation of organic molecules. Furthermore, in human-AI collaborative scenarios, the system enhances efficiency sixfold and improves accuracy by 40%, significantly overcoming the generalization bottleneck in automated structural analysis.
📝 Abstract
Following the molecular discovery and synthesis revolutions, scalable automated structure elucidation from routine spectroscopic data remains an outstanding challenge. Despite decades of computational efforts, no existing system achieved reliable reasoning over unseen spectra. Here, we propose MACROS, a multi-agent system automating structure elucidation by emulating expert iterative hypothesis-testing. Trained on 100M simulated and 1.6M experimental spectra-molecule pairs, it natively supports arbitrary combinations of routine spectroscopic techniques. It achieves unprecedented zero-shot generalization to diverse real-world samples, correctly identifying synthetic compounds, natural products and metabolites above 500 Da with 1D NMR. Remarkably, MACROS spontaneously recovers textbook spectroscopic correlations from unassigned data and exhibits emergent chemical intuition such as a ring-first parsing preference, learning fundamental chemical principles rather than memorizing database patterns. MACROS augments chemists via collaboration to deliver sixfold faster, 40% more accurate elucidation. MACROS establishes a scalable foundation for fully automated structure elucidation, and catalyzes accelerated molecular discovery toward autonomous laboratories.
Problem

Research questions and friction points this paper is trying to address.

Automated Structure Elucidation
Multimodal Spectra
Zero-shot Generalization
Organic Chemistry
Innovation

Methods, ideas, or system contributions that make the work stand out.

Multi-Agent System
Closed-Loop Reasoning
Zero-Shot Generalization
Emergent Chemical Intuition
Organic Structure Elucidation
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
No related jobs found.
B
Bingsen Xue
Institute of Medical Robotics, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Z
Zhuojun Jiang
CAS Key Laboratory of Green Process and Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China
J
Jianhao Zhang
Institute of Medical Robotics, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
M
Mingcheng Gu
Institute of Medical Robotics, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Y
Yizhe Yuan
Institute of Medical Robotics, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Y
Yongtai Zhuo
Institute of Medical Robotics, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Y
Yifan Zhang
CAS Key Laboratory of Green Process and Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China
Li Wang
Li Wang
Fudan University
Computer Vision
Ya Su
Ya Su
Department of Neurology, National Centre for Neurological Disorders, National Clinical Research Centre for Aging and Medicine, Huashan Hospital, Fudan University, Shanghai 200240, China
Y
Yue Yuan
Department of Microbiology Laboratory, Minhang District Center for Disease Control, Shanghai 200240, China
J
Jiang Liu
Department of Electrical and Computer Engineering, Johns Hopkins University, Baltimore, USA
X
Xueqian Kong
School of Chemistry and Chemical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Cheng Jin
Cheng Jin
Ph.D. Student, School of Computer Science and Engineering, HKUST
Knowledge DistillationComputational PathologyAI for Science