Generative structural elucidation from mass spectra as an iterative optimization problem

2026-02-07

AI Summary

This study addresses the challenge of annotating chemical structures in liquid chromatography–tandem mass spectrometry (LC-MS/MS) data when reference spectra or known structures are unavailable for the majority of spectral features. To this end, the authors formalize structure elucidation as an iterative optimization task and introduce a scalable generative paradigm. Their approach integrates a molecular formula-constrained graph-based genetic algorithm with in silico mass spectral simulation to iteratively generate and refine candidate structures. This framework can be deployed independently or used to augment existing inverse prediction models. Experimental results on the NIST’20 and MassSpecGym datasets demonstrate that the proposed method substantially enhances structural annotation accuracy under reference-free conditions.

Technology Category

• Machine Learning: Structured Learning
• Natural Language Processing: Generation
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Liquid chromatography tandem mass spectrometry (LC-MS/MS) is a critical analytical technique for molecular identification across metabolomics, environmental chemistry, and chemical forensics. A variety of computational methods have emerged for structural annotation of spectral features of interest, but many of these features cannot be confidently annotated with reference structures or spectra. Here, we introduce FOAM (Formula-constrained Optimization for Annotating Metabolites), a computational workflow that poses structure elucidation from LC-MS/MS as an iterative optimization problem. FOAM couples a formula-constrained graph genetic algorithm with spectral simulation to explore candidate annotations given an experimental spectrum. We demonstrate FOAM's performance on the NIST'20 and MassSpecGym datasets as both a standalone elucidation pipeline and as a complement to existing inverse models. This work establishes iterative optimization as an effective and extensible paradigm for structural elucidation.

Problem

• structural elucidation
• mass spectrometry
• molecular identification
• spectral annotation
• metabolomics

Innovation

• iterative optimization
• graph genetic algorithm
• formula-constrained
• spectral simulation
• structural elucidation