Deep learning boosts sensitivity of mass spectrometry-based immunopeptidomics

Abstract

Characterizing the human leukocyte antigen (HLA) bound ligandome by mass spectrometry (MS) holds great promise for developing vaccines and drugs for immune-oncology. Still, the identification of non-tryptic peptides presents substantial computational challenges. To address these, we synthesized and analyzed >300,000 peptides by multi-modal LC-MS/MS within the ProteomeTools project representing HLA class I & II ligands and products of the proteases AspN and LysN. The resulting data enabled training of a single model using the deep learning framework Prosit, allowing the accurate prediction of fragment ion spectra for tryptic and non-tryptic peptides. Applying Prosit demonstrates that the identification of HLA peptides can be improved up to 7-fold, that 87% of the proposed proteasomally spliced HLA peptides may be incorrect and that dozens of additional immunogenic neo-epitopes can be identified from patient tumors in published data. Together, the provided peptides, spectra and computational tools substantially expand the analytical depth of immunopeptidomics workflows.

RHybridFinder: An R package to process immunopeptidomic data for putative hybrid peptide discovery

“…More recently, MS-based immunopeptidomics has been used to expedite the identification of PSPs in a systematic manner, including cis- and trans-spliced peptides ( Berkers et al., 2015 ; Faridi et al., 2018 ; Liepe et al., 2010 , 2016 ; Rolfs et al., 2019 ; Specht et al., 2020 ). However, MS-based studies using different computational approaches have led to a debate around the proportion of those PSPs in the MHC class I immunopeptidome ( Lichti, 2021 ; Mylonas et al., 2018 ; Wilhelm et al., 2021 ).…”

Is the Immunopeptidome Getting Darker?: A Commentary on the Discussion around Mishto et al., 2019

“…However, it is clear that there are still informatic challenges in the peptide-centric data analysis pipelines (12,13) with a much lower proportion of spectra confidently assigned from immunopeptidomics data compared with more conventional samples such as tryptic digests of cell lysates. The search to explain these dark elements of the immunopeptidome has driven differences in interpretation and peptide sequence assignments to these MS/MS spectra (12,(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25).…”

“…It has certainly shone a spotlight on the uncomfortably large proportion of peptides in the immunopeptidome that cannot be readily assigned using reference proteomes. It has also driven developments in the informatics used to search immunopeptidomics data with improvements in assignments of post-translational modifications ( 34 , 35 ), incorporation of proteogenomics databases ( 36 ) and a series of alternative workflows to address the presence of spliced peptides ( 16 , 18 , 20 – 22 , 24 , 25 , 37 – 42 ). However, no consensus has been reached with somewhere between 0 and 45% of peptides in the immunopeptidome believed to be spliced.…”

“…It is also time as a field that we decide how we debate these issues and in my opinion time to stop subjectively re-searching others data but to design and implement our own experiments to more definitely corroborate or repudiate various sources of antigenic peptides in the immunopeptidome. Our field has made great strides and can look back at great community initiatives: the human immunopeptidome project ( 46 ) that strives to advance the documentation of peptides available for immune recognition in a variety of healthy and diseased states; minimal information requirements for an immunopeptidomics experiment ( 47 ) and HLA peptide atlases ( 48 , 49 ) and data repositories ( 50 ) that have driven improvements in predictive algorithms and will provide the data required for the next generation of AI-based informatic tools ( 25 ). Let’s use this same spirit to explore the immunopeptidome and illuminate its dark side as a community united in this purpose.…”

The HLA Ligandome Comprises a Limited Repertoire of O-GlcNAcylated Antigens Preferentially Associated With HLA-B*07:02

“…A major source of therapeutic intervention for cancer patients comes from neoantigens in the HLA Class I ligandome, introduced by cancer induced mutations ( 4 , 5 ). Mass spectrometry (MS) based immunopeptidomics provides means to identify disease or patient specific HLA antigen presentation, and chart the HLA bound peptide ligandome ( 6 , 7 ). Next to the peptide sequence and protein of origin, mass spectrometry can also be used to identity post-translation modifications (PTMs) on the antigens.…”