[论文解读] De Novo Molecular Generation from Mass Spectra via Many-Body Enhanced Diffusion

Xichen Sun, Wentao Wei+3 more|arXiv (Cornell University)|Feb 2, 2026

Mass Spectrometry Techniques and Applications被引用 0

一句话总结

MBGen 引入一个边缘中心的扩散模型和多体注意力机制，从 MS/MS 谱图生成分子结构，在 NPLIB1 和 MassSpecGym 基准上达到最前沿的结果。

ABSTRACT

Molecular structure generation from mass spectrometry is fundamental for understanding cellular metabolism and discovering novel compounds. Although tandem mass spectrometry (MS/MS) enables the high-throughput acquisition of fragment fingerprints, these spectra often reflect higher-order interactions involving the concerted cleavage of multiple atoms and bonds-crucial for resolving complex isomers and non-local fragmentation mechanisms. However, most existing methods adopt atom-centric and pairwise interaction modeling, overlooking higher-order edge interactions and lacking the capacity to systematically capture essential many-body characteristics for structure generation. To overcome these limitations, we present MBGen, a Many-Body enhanced diffusion framework for de novo molecular structure Generation from mass spectra. By integrating a many-body attention mechanism and higher-order edge modeling, MBGen comprehensively leverages the rich structural information encoded in MS/MS spectra, enabling accurate de novo generation and isomer differentiation for novel molecules. Experimental results on the NPLIB1 and MassSpecGym benchmarks demonstrate that MBGen achieves superior performance, with improvements of up to 230% over state-of-the-art methods, highlighting the scientific value and practical utility of many-body modeling for mass spectrometry-based molecular generation. Further analysis and ablation studies show that our approach effectively captures higher-order interactions and exhibits enhanced sensitivity to complex isomeric and non-local fragmentation information.

研究动机与目标

• 推动从串联 MS 谱图准确去新生成分子结构的研究
• 通过建模多体边缘相互作用来解决原子中心和成对模型的局限性
• 开发一个以谱指纹为条件的边缘中心扩散框架
• 进行谱编码器的预训练与图解码器的预训练，随后进行端到端微调
• 展示在同分异构体辨识与化学合理性方面相较基线的改进

提出的方法

• 将分子表示为以边为中心的图，包含节点与边的嵌入
• 使用谱编码器（MIST Formula Transformer）从 MS/MS 数据生成全局指纹 y
• 通过一个多体增强扩散过程对边特征进行去噪，条件化地结合 y 与 H
• 引入多体注意力模块，通过三元组交互(i, j, k)更新边嵌入
• 应用 FiLM 条件化将全局谱信息注入到边表示中
• 采用三阶段训练策略：谱编码器的预训练、指纹-结构对的解码器预训练，以及端到端微调

Figure 1: The mass spectrum comprises fragments from complex bond cleavages, where isomers within the same peak each contribute distinct yet essential insights for accurate molecular generation.

实验结果

研究问题

• RQ1 通过多体注意力建模高阶边缘相互作用，是否能提升基于 MS/MS 谱图的去新分子生成性能？
• RQ2 以谱指纹为条件的边缘中心扩散是否在结构生成与同分异构体辨识方面优于原子中心基线？
• RQ3 对编码器/解码器和多体模块进行预训练对生成准确性与化学有效性有何影响？
• RQ4 MBGen 在具有相似碎裂模式的同分异构体辨识方面的区分能力如何？

主要发现

• MBGen 在 NPLIB1 与 MassSpecGym 基准上达到最先进的 Top-1 与 Top-10 精度
• 在 NPLIB1 上，MBGen 的 Top-1 精度为 12.20%，Top-10 为 22.29%，显著优于 DiffMS
• 在 MassSpecGym 上，MBGen 的 Top-1 精度为 7.58%，Top-10 为 12.54%，优于 DiffMS
• MBGen 能获得更高的 MCES 与 Tanimoto 相似度，表明子结构与总体分子相似性更好
• 消融研究表明多体模块与全面预训练共同推动性能提升
• 案例研究显示 MBGen 能捕捉到峰内同分异构信息，并比基线更准确地重建复杂分子

Figure 2: An overview of MBGen framework.