A CCS-centric HLA-specific trained de novo module for precise and accurate real-time immunopeptide identification on the Bruker ProteoScape platform

Importance of the Topic

The characterization of immunopeptides is critical for understanding antigen presentation mechanisms and developing targeted immunotherapies. Real-time de novo sequencing of HLA-bound peptides enhances the speed and depth of immunopeptidomics workflows, enabling rapid identification of tumor neoantigens and infection biomarkers. The integration of specialized algorithms into data acquisition platforms supports timely decision making in clinical and research settings.

Objectives and Study Overview

This study presents the development and evaluation of an HLA-specific de novo sequencing module integrated into the Bruker ProteoScape platform. The goals were to retrain the Rapid Novor algorithm using HLA-derived peptide spectra, implement the optimized MHC scoring model for timsTOF acquisitions, and compare performance against standard models on published immunopeptidomics datasets.

Methodology and Instrumentation

• Retraining dataset: 1.4 million PSMs mapping to over 150,000 MHCI and MHCII peptides, filtered to 1% PSM FDR using ProLuCID results.
• Sequencing engine: Modified Rapid Novor incorporated into Bruker ProteoScape software for real-time data streaming (Run & Done workflow).
• Computational platform: First-generation PaSER workstation with AMD Epyc 7302P (16 cores, 128 GB RAM).
• MS instrumentation: Bruker timsTOF series operated in dda-PASEF mode with 45 to 120 min LC gradients.

Main Results and Discussion

• Precision and accuracy: The MHC-optimized model improved amino acid correctness by ~2% and peptide correctness by ~1% compared to the pretrained Orbitrap HCD model in the Feola et al. dataset.
• Sequence coverage: In the melanoma cohort (Phulphagar et al.), the MHC model increased amino acid and peptide accuracy by ~6% and yielded an average 8.5% rise in identified 8–11mer sequences.
• Processing speed: Despite the enriched scoring complexity, the MHC model maintained high throughput (~1338±226 spectra/s) with negligible impact on overall run time relative to LC gradients.

Benefits and Practical Applications

• Immediate data availability: The Run & Done feature delivers processed results upon acquisition completion, accelerating downstream analysis.
• Enhanced immunopeptidomics: Improved de novo sequencing precision supports biomarker discovery, vaccine design, and personalized medicine research.
• Platform extensibility: Seamless integration into Bruker ProteoScape allows real-time customization and scalability across proteomics applications.

Future Trends and Applications

De novo algorithms will increasingly leverage machine learning and CCS information to refine peptide identification. Integration with cloud-based computing and artificial intelligence will enable distributed real-time analysis. Advancements in ion mobility and next-generation mass spectrometers will demand adaptive models for broader immunopeptidome coverage.

Conclusion

The HLA-specific MHC scoring model in Bruker ProteoScape Novor significantly enhances real-time de novo immunopeptide identification without compromising throughput. This innovation streamlines immunopeptidomics workflows and lays the foundation for future algorithmic and hardware integration in proteomics research.

Reference

1. Hoenisch Gravel N. et al. TOFIMS mass spectrometry-based immunopeptidomics refines tumor antigen identification. Nat Commun. 2023;14:7472.
2. Phulphagar KM. et al. Sensitive, high-throughput HLA-I and HLA-II immunopeptidomics using parallel accumulation-serial fragmentation mass spectrometry. Mol Cell Proteomics. 2023;22.
3. Feola S. et al. PeptiCHIP: A microfluidic platform for tumor antigen landscape identification. ACS Nano. 2021;15:15992–16010.