DiffSpectra: Molecular Structure Elucidation from Spectra using Diffusion Models

2025/07/09 by Rong Yu, Wang, Liang, Tingyang Xu +16 · 3 citations
Computer Science · Materials Science · #Advanced Graph Neural Networks #Artificial Intelligence (cs.AI) #Chemical Physics (physics.chem-ph) #Computational Drug Discovery Methods #Computational Engineering #FOS: Biological sciences #FOS: Computer and information sciences #FOS: Physical sciences #Finance #Machine Learning (cs.LG) #Machine Learning in Materials Science #Molecular Networks (q-bio.MN) #and Science (cs.CE)

paper · pdf · doi:10.48550/arxiv.2507.06853

openalex publication_date 2025/07/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Molecular structure elucidation from spectra is a fundamental challenge in molecular science. Conventional approaches rely heavily on expert interpretation and lack scalability, while retrieval-based machine learning approaches remain constrained by limited reference libraries. Generative models offer a promising alternative, yet most adopt autoregressive architectures that overlook 3D geometry and struggle to integrate diverse spectral modalities. In this work, we present DiffSpectra, a generative framework that formulates molecular structure elucidation as a conditional generation process, directly inferring 2D and 3D molecular structures from multi-modal spectra using diffusion models. Its denoising network is parameterized by the Diffusion Molecule Transformer, an SE(3)-equivariant architecture for geometric modeling, conditioned by SpecFormer, a Transformer-based spectral encoder capturing multi-modal spectral dependencies. Extensive experiments demonstrate that DiffSpectra accurately elucidates molecular structures, achieving 40.76% top-1 and 99.49% top-10 accuracy. Its performance benefits substantially from 3D geometric modeling, SpecFormer pre-training, and multi-modal conditioning. To our knowledge, DiffSpectra is the first framework that unifies multi-modal spectral reasoning and joint 2D/3D generative modeling for de novo molecular structure elucidation.

Citations

• Materials Generation in the Era of Artificial Intelligence: A Comprehensive Survey
• A Survey of Graph Transformers: Architectures, Theories and Applications
• MolSpectra: Pre-training 3D Molecular Representation with Multi-modal Energy Spectra
• NExT-Mol: 3D Diffusion Meets 1D Language Modeling for 3D Molecule Generation
• DiffMS: Diffusion Generation of Molecules Conditioned on Mass Spectra
• Diffusion Models for Molecules: A Survey of Methods and Tasks
• MassSpecGym: A benchmark for the discovery and identification of molecules
• Accurate and efficient structure elucidation from routine one-dimensional NMR spectra using multitask machine learning
• Unraveling Molecular Structure: A Multimodal Spectroscopic Dataset for Chemistry
• Rethinking Graph Masked Autoencoders through Alignment and Uniformity
• GSLB: The Graph Structure Learning Benchmark
• Mass Spectra Prediction with Structural Motif-based Graph Neural Networks
• Geometric Latent Diffusion Models for 3D Molecule Generation
• Prefix-Tree Decoding for Predicting Mass Spectra from Molecules
• Conditional Diffusion Based on Discrete Graph Structures for Molecular Graph Generation
• Scalable Diffusion Models with Transformers
• A Time Series is Worth 64 Words: Long-term Forecasting with Transformers
• Structure-based Drug Design with Equivariant Diffusion Models
• Equivariant 3D-Conditional Diffusion Models for Molecular Linker Design
• Equivariant Energy-Guided SDE for Inverse Molecular Design
• MDM: Molecular Diffusion Model for 3D Molecule Generation
• Blurring Diffusion Models
• Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioning
• Pre-training via Denoising for Molecular Property Prediction
• GraphMAE: Self-Supervised Masked Graph Autoencoders
• Equivariant Diffusion for Molecule Generation in 3D
• Masked Autoencoders Are Scalable Vision Learners
• Variational Diffusion Models
• E(n) Equivariant Normalizing Flows
• Equivariant message passing for the prediction of tensorial properties and molecular spectra
• Score-Based Generative Modeling through Stochastic Differential Equations
• Denoising Diffusion Probabilistic Models
• Generative Modeling by Estimating Gradients of the Data Distribution
• Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules
• Representation Learning with Contrastive Predictive Coding
• FiLM: Visual Reasoning with a General Conditioning Layer
• Deep Unsupervised Learning using Nonequilibrium Thermodynamics
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Cited by

• NMIRacle: Multi-modal Generative Molecular Elucidation from IR and NMR Spectra
• MolSpectLLM: A Molecular Foundation Model Bridging Spectroscopy, Molecule Elucidation, and 3D Structure Generation
• SpectrumWorld: Artificial Intelligence Foundation for Spectroscopy

Related

• DiSE: A diffusion probabilistic model for automatic structure elucidation of organic compounds
• DiffMS: Diffusion Generation of Molecules Conditioned on Mass Spectra
• DiffNMR: Diffusion Models for Nuclear Magnetic Resonance Spectra Elucidation
• Sampling 3D Molecular Conformers with Diffusion Transformers
• Diffusion-Driven Generative Framework for Molecular Conformation Prediction