Related Experiment Video
Updated: Jun 29, 2025
Immunopeptidomics: Isolation of Mouse and Human MHC Class I- and II-Associated Peptides for Mass Spectrometry Analysis
Published on: October 15, 2021
Deep Learning-Assisted Analysis of Immunopeptidomics Data
Wassim Gabriel1, Mario Picciani1, Matthew The2
1Computational Mass Spectrometry, TUM School of Life Sciences, Technical University of Munich, Freising, Germany.
Insights
Deep learning models enhance mass spectrometry analysis for human leukocyte antigen (HLA) peptides, improving identification accuracy. This approach aids in discovering disease-specific peptides and neo-epitopes, overcoming computational challenges in immunopeptidomics.
Area of Science:
- Mass Spectrometry
- Immunopeptidomics
- Computational Biology
Background:
- Liquid chromatography-coupled mass spectrometry (LC-MS/MS) is crucial for identifying human leukocyte antigen (HLA) peptides.
- Analyzing HLA peptides presents unique computational and statistical challenges compared to standard proteomics.
- Fragment ion intensity-based scores significantly improve peptide identification, especially for non-tryptic peptides.
Purpose of the Study:
- To detail procedures for applying deep learning models in mass spectrometry-based immunopeptidomics.
- To demonstrate how to analyze and validate spectral data using state-of-the-art deep learning tools.
- To showcase the benefits of deep learning for HLA peptide identification and neo-epitope discovery.
Main Methods:
- Utilizing deep learning frameworks like Prosit for fragment ion intensity and retention time prediction.
- Applying tools such as Universal Spectrum Explorer (USE) and Oktoberfest (online/offline) for spectral analysis.
- Leveraging intensity-based scoring for enhanced peptide matching in mass spectrometry data.
Main Results:
- Deep learning-assisted analysis increases the number of confidently identified HLA peptides.
- Facilitates the discovery of confidently identified neo-epitopes.
- Assists in the assessment of cryptic peptides, including spliced peptides.
Conclusions:
- Deep learning models offer powerful solutions to computational challenges in HLA peptide analysis.
- These methods enhance the accuracy and scope of immunopeptidomics studies.
- The described procedures provide a framework for advancing personalized medicine through improved peptide identification.
Abstract:
Liquid chromatography-coupled mass spectrometry (LC-MS/MS) is the primary method to obtain direct evidence for the presentation of disease- or patient-specific human leukocyte antigen (HLA). However, compared to the analysis of tryptic peptides in proteomics, the analysis of HLA peptides still poses computational and statistical challenges. Recently, fragment ion intensity-based matching scores assessing the similarity between predicted and observed spectra were shown to substantially increase the number of confidently identified peptides, particularly in use cases where non-tryptic peptides are analyzed. In this chapter, we describe in detail three procedures on how to benefit from state-of-the-art deep learning models to analyze and validate single spectra, single measurements, and multiple measurements in mass spectrometry-based immunopeptidomics. For this, we explain how to use the Universal Spectrum Explorer (USE), online Oktoberfest, and offline Oktoberfest. For intensity-based scoring, Oktoberfest uses fragment ion intensity and retention time predictions from the deep learning framework Prosit, a deep neural network trained on a very large number of synthetic peptides and tandem mass spectra generated within the ProteomeTools project. The examples shown highlight how deep learning-assisted analysis can increase the number of identified HLA peptides, facilitate the discovery of confidently identified neo-epitopes, or provide assistance in the assessment of the presence of cryptic peptides, such as spliced peptides.
