Cloud Computing Based Immunopeptidomics Utilizing Community Curated Variant Libraries Simplifies and Improves...

Area of Science:

• Immunology
• Oncology
• Bioinformatics
• Personalized Medicine

Background:

• Unique peptide neo-antigens on cell surfaces are crucial targets in personalized medicine for therapies and diagnostics.
• Current pipelines often combine RNA/ribosomal sequencing with tandem mass spectrometry for neo-antigen identification.
• Existing proteogenomic approaches may miss a significant number of potential neo-antigen targets.

Purpose of the Study:

• To present an alternative, cloud-computing-powered approach for neo-antigen discovery.
• To reanalyze existing data from metastatic melanoma studies using an expanded sequence variant database.
• To identify novel non-canonical peptide neo-antigens and assess their potential for targeted therapies.

Main Methods:

• Utilized cloud computing to search for neo-antigens against a database of over 7 million human sequence variants.
• Reanalyzed proteogenomic data from two previously published metastatic melanoma studies.
• Validated identified neo-antigens using secondary analyses like predicted binding affinity (NetMHC).

Main Results:

• Identified 79% of non-canonical peptides previously reported and discovered 18-fold more than prior analyses.
• Reported 738 non-canonical peptides shared across at least five patient samples and 3258 shared between the two studies.
• Made a comprehensive list of shared peptides, annotations, and MS/MS spectra available via a web portal.

Conclusions:

• The cloud-based approach significantly enhances the power and depth of neo-antigen discovery compared to traditional methods.
• The identified novel neo-antigens represent promising targets for developing personalized cancer therapies and diagnostics.
• The publicly available dataset facilitates further community research into neo-antigen identification and utilization.