Raphaël La Rocca, , , Anthony Cioppa, , , Enrico Ferrarini, , , Monica Höfte, , , Marc Van Droogenbroeck, , , Edwin De Pauw, , , Gauthier Eppe, , and , Loïc Quinton*,
{"title":"Relational Graph Convolutional Network for Robust Mass Spectrum Classification","authors":"Raphaël La Rocca, , , Anthony Cioppa, , , Enrico Ferrarini, , , Monica Höfte, , , Marc Van Droogenbroeck, , , Edwin De Pauw, , , Gauthier Eppe, , and , Loïc Quinton*, ","doi":"10.1021/jasms.5c00055","DOIUrl":null,"url":null,"abstract":"<p >Supervised machine learning methods have shown impressive performance in interpreting mass signals and automatically segmenting spatially meaningful regions in Mass Spectrometry Imaging (MSI). Such segmentation generates maps that provide researchers with valuable insights into sample composition and serve as a foundation for downstream statistical analyses. However, these models often require data set-specific preprocessing and do not fully exploit the rich mass features available in high-resolution mass spectrometry (HRMS). Unlike low-resolution mass spectrometry, HRMS reveals additional features such as mass defects and repeated mass differences that carry important chemical information. In this work, we propose a novel deep learning architecture based on a Relational Graph Convolutional Network (R-GCN) that captures and leverages those HRMS mass features. Our model explicitly encodes structural features such as mass defects and known mass differences to represent each spectrum as a graph, enabling the learning of associations between chemically related ion families. To the best of our knowledge, no existing deep learning models for MSI classification incorporate this level of chemically informed mass structure. Most existing methods treat spectra as flat vectors or image-like inputs, thereby ignoring the underlying mass relationships. We evaluate our R-GCN approach against several conventional machine learning and deep learning baselines across diverse MSI data sets, demonstrating its robustness to common signal variations (e.g., mass shift, ion loss). Finally, we integrate Class Activation Mapping (CAM) to enhance model interpretability, enabling the identification of ion families that are relevant to specific biological or spatial regions.</p>","PeriodicalId":672,"journal":{"name":"Journal of the American Society for Mass Spectrometry","volume":"36 10","pages":"2036–2047"},"PeriodicalIF":2.7000,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of the American Society for Mass Spectrometry","FirstCategoryId":"92","ListUrlMain":"https://pubs.acs.org/doi/10.1021/jasms.5c00055","RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"BIOCHEMICAL RESEARCH METHODS","Score":null,"Total":0}
引用次数: 0
Abstract
Supervised machine learning methods have shown impressive performance in interpreting mass signals and automatically segmenting spatially meaningful regions in Mass Spectrometry Imaging (MSI). Such segmentation generates maps that provide researchers with valuable insights into sample composition and serve as a foundation for downstream statistical analyses. However, these models often require data set-specific preprocessing and do not fully exploit the rich mass features available in high-resolution mass spectrometry (HRMS). Unlike low-resolution mass spectrometry, HRMS reveals additional features such as mass defects and repeated mass differences that carry important chemical information. In this work, we propose a novel deep learning architecture based on a Relational Graph Convolutional Network (R-GCN) that captures and leverages those HRMS mass features. Our model explicitly encodes structural features such as mass defects and known mass differences to represent each spectrum as a graph, enabling the learning of associations between chemically related ion families. To the best of our knowledge, no existing deep learning models for MSI classification incorporate this level of chemically informed mass structure. Most existing methods treat spectra as flat vectors or image-like inputs, thereby ignoring the underlying mass relationships. We evaluate our R-GCN approach against several conventional machine learning and deep learning baselines across diverse MSI data sets, demonstrating its robustness to common signal variations (e.g., mass shift, ion loss). Finally, we integrate Class Activation Mapping (CAM) to enhance model interpretability, enabling the identification of ion families that are relevant to specific biological or spatial regions.
期刊介绍:
The Journal of the American Society for Mass Spectrometry presents research papers covering all aspects of mass spectrometry, incorporating coverage of fields of scientific inquiry in which mass spectrometry can play a role.
Comprehensive in scope, the journal publishes papers on both fundamentals and applications of mass spectrometry. Fundamental subjects include instrumentation principles, design, and demonstration, structures and chemical properties of gas-phase ions, studies of thermodynamic properties, ion spectroscopy, chemical kinetics, mechanisms of ionization, theories of ion fragmentation, cluster ions, and potential energy surfaces. In addition to full papers, the journal offers Communications, Application Notes, and Accounts and Perspectives