15 October 2025 in Molecular & Cellular Proteomics
Glycopeptide tandem mass spectra typically contain numerous glycan-specific fragments that can inform several features of glycan modifications, including glycan class, composition, and structure. While these fragment ions are often straightforward to observe by eye, few tools exist to systemically explore these common glycopeptide spectral features or explore their relationships with each other. Instead, most studies rely on manual inspection to understand glycan-informative ion content in their data, or they are restricted to evaluating the presence of these ions only in the small fraction of spectra that are identified by glycopeptide search algorithms. Here we introduce GlyCounter as a freely available, open-source tool to rapidly extract oxonium, Y-type, and custom ion information from raw data files. We highlight GlyCounter's utility by evaluating glycan-specific fragments in a diverse selection of publicly available datasets to demonstrate how others in the field can make immediate use of this software. In several cases, we show how conclusions drawn in these publications are evident simply through GlyCounter's extracted ion information without requiring database searches or experiment-specific programs. Although one of our goals is to decouple spectral evaluation from glycopeptide identification, we also show that evaluating oxonium ion content with GlyCounter can supplement a database search as valuable spectral evidence to validate conclusions. In all, we present GlyCounter as a user-friendly platform that can be easily incorporated into most glycoproteomic workflows to refine sample preparation, data acquisition, and post-acquisition identification methods through straightforward evaluation of the glycan content of glycoproteomic data. Software and instructions are available at https://github.com/riley-research/GlyCounter.
12 November 2024 in Journal of Proteome Research
7 October 2026 in Molecular & Cellular Proteomics
Mass spectrometry is the dominant technology in the field of proteomics, enabling high-throughput analysis of the protein content of complex biological samples. Due to the complexity of the instrumentation and resulting data, sophisticated computational methods are required for the processing and interpretation of acquired mass spectra. Machine learning has shown great promise to improve the analysis of mass spectrometry data, with numerous purpose-built methods for improving specific steps in the data acquisition and analysis pipeline reaching widespread adoption. Each of these tasks rely on a shared fundamental understanding of the information present in a mass spectrum. Here, we propose a transfer learning framework, CasanovoTL, which utilizes pretrained spectrum representations, derived from a de novo sequencing model trained on a massive proteomics dataset, rather than training a spectrum encoder from scratch for each task. We show that using these pre-trained spectrum representations significantly improves our performance on the four downstream tasks of spectrum quality prediction, chimericity prediction, phosphorylation prediction, and glycosylation status prediction. Finally, we perform end-to-end finetuning, as well as multi-task training, and find that finetuned spectrum representations further improve performance on each individual task. Overall, our work demonstrates that a pretrained spectrum encoder for tandem mass spectrometry proteomics trained on de novo sequencing learns generalizable representations of spectra, improves performance on downstream tasks where training data is limited, and can ultimately enhance data acquisition and analysis in proteomics experiments.
26 August 2024 in ChemRxiv
Tandem mass spectrometry (MS/MS) is the gold standard for intact glycopeptide identification, enabling peptide sequence elucidation and site-specific localization of glycan compositions. Beam-type collisional activation is generally sufficient for N-glycopeptides, while electron-driven dissociation is crucial for site localization in O-glycopeptides. Modern glycoproteomic methods often employ multiple dissociation techniques within a single LC-MS/MS analysis, but this approach frequently sacrifices sensitivity when analyzing multiple glycopeptide classes simultaneously. Here we explore the utility of intelligent data acquisition for glycoproteomics through real-time library searching (RTLS) to match oxonium ion patterns for on-the-fly selection of the appropriate dissociation method. By matching dissociation method with glycopeptide class, this autonomous dissociation-type selection (ADS) generates equivalent numbers of N-glycopeptide identifications relative to traditional beam-type collisional activation methods while also yielding comparable numbers of site-localized O-glycopeptide identifications relative to conventional electron transfer dissociation-based methods. The ADS approach represents a step forward in glycoproteomics throughput by enabling site-specific characterization of both N-and O-glycopeptides within the same LC-MS/MS acquisition.