The DOI matches its Crossref record; 1 of the 1 references it cites were found in a registry or an index.
Deep Neural Networks Predict MHC-I Epitope Presentation and Transfer Learn Neoepitope Immunogenicity
Open accessPreprint — published as 10.1038/s42256-023-00694-6
Abstract
Identifying neoepitopes that elicit an adaptive immune response is a major bottleneck to developing personalized cancer vaccines. Experimental validation of candidate neoepitopes is extremely resource intensive, and the vast majority of candidates are non-immunogenic, making their identification a needle-in-a-haystack problem. To address this challenge, we present computational methods for predicting MHC-I epitopes and identifying immunogenic neoepitopes with improved precision. The BigMHC method comprises an ensemble of seven pan-allelic deep neural networks trained on peptide-MHC eluted ligand data from mass spectrometry assays and transfer learned on data from assays of antigen-specific immune response. Compared with four state-of-the-art classifiers, BigMHC significantly improves the prediction of epitope presentation on a test set of 45,409 MHC ligands among 900,592 random negatives (AUROC=0.9733, AUPRC=0.8779). After transfer learning on immunogenicity data, BigMHC yields significantly higher precision than seven state-of-the-art models in identifying immunogenic neoepitopes, making BigMHC effective in clinical settings. All data and code are freely available at https://github.com/KarchinLab/bigmhc .
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Highly accurate protein structure prediction with AlphaFold
Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03819-2 · Cited by 43,764
Flagged first, then by citations
Research works printed but not linked: 38
- 10.1158/2326-6066.CIR-19-0464 · DOI
- 2020, MHCflurry 2.0: Improved Pan-Allele Prediction of MHC Class I-Presented Peptides by Incorporating Antigen Processing, Cell Systems, 11, 42, 10.1016/j.cels.2020.06.010 · DOI · Paper page
- 2020, A large peptidome dataset improves HLA class I epitope prediction across most of the human population, Nature Biotechnology, 38, 199, 10.1038/s41587-019-0322-9 · DOI
- 10.1007/s00251-010-0441-4 · DOI
- 2016, NetMHCpan-3.0; improved prediction of binding to MHC class I molecules integrating information from multiple receptor and peptide length datasets, Genome Medicine, 8, 1
- 2020, NetMHCpan-4.1 and NetMHCIIpan-4.0: improved predictions of MHC antigen presentation by concurrent motif deconvolution and integration of MS MHC eluted ligand data, Nucleic Acids Research, W1, 48
- 10.1007/s00251-008-0341-z · DOI
- 10.1371/journal.pone.0000796 · DOI
- 2017, Deciphering HLA-I motifs across HLA peptidomes improves neo-antigen predictions and identifies allostery regulating HLA specificity, PLoS computational biology, 13, e1005725, 10.1371/journal.pcbi.1005725 · DOI
- 2018, The Length Distribution and Multiple Specificity of Naturally Presented HLA-I Ligands, The Journal of Immunology, 201, 3705, 10.4049/jimmunol.1800914 · DOI
- 2022, A transformer-based model to predict peptide–HLA class I binding and optimize mutated peptides for vaccine design, Nature Machine Intelligence, 4, 300, 10.1038/s42256-022-00459-7 · DOI
- 2018, MHCflurry: Open-Source Class I MHC Binding Affinity Prediction, Cell Systems, 7, 129, 10.1016/j.cels.2018.05.014 · DOI
- Gfeller, D. et al. Improved predictions of antigen presentation and TCR recognition with MixMHCpred2.2 and PRIME2.0 reveal potent SARS-CoV-2 CD8+ T-cell epitopes. Cell Systems (2023). https://doi.org:https://doi.org/10.1016/j.cels.2022.12.002 · DOI
- 2021, Deep learning-based prediction of the T cell receptor–antigen binding specificity, Nature Machine Intelligence, 3, 864, 10.1038/s42256-021-00383-2 · DOI · Paper page
- 2021, Prediction of neo-epitope immunogenicity reveals TCR recognition determinants and provides insight into immunoediting, Cell Reports Medicine, 2, 100194, 10.1016/j.xcrm.2021.100194 · DOI
- 2020, Key Parameters of Tumor Epitope Immunogenicity Revealed Through a Consortium Approach Improve Neoantigen Prediction, Cell, 183, 818, 10.1016/j.cell.2020.09.015 · DOI · Paper page
- Vaswani, A. et al. in Advances in neural information processing systems 30 (NeurIPS).
- Huang, G. , Liu, Z. , van der Maaten, L. & Weinberger, K. Q. in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 4700–4708.
- 10.5555/2627435.2670313 · DOI
- Gfeller, D. et al. Predictions of immunogenicity reveal potent SARS-CoV-2 CD8+ T-cell epitopes. bioRxiv, 2022.2005.2023.492800 (2022). https://doi.org:10.1101/2022.05.23.492800 · DOI
- 2021, NEPdb: a database of T-cell experimentally-validated neoantigens and pan-cancer predicted neoepitopes for cancer immunotherapy, Frontiers in Immunology, 12, 644637, 10.3389/fimmu.2021.644637 · DOI
- 10.1158/2326-6066.CIR-18-0129 · DOI
- 10.1158/2159-8290.CD-16-0828 · DOI
- 2021, Transcriptional programs of neoantigen-specific TIL in anti-PD-1-treated lung cancers, Nature, 596, 126, 10.1038/s41586-021-03752-4 · DOI
- 2020, Multimodal genomic features predict outcome of immune checkpoint blockade in non-small-cell lung cancer, Nature Cancer, 1, 99, 10.1038/s43018-019-0008-8 · DOI
- 10.1126/scitranslmed.aaa7161 · DOI
- 10.1007/s00251-010-0441-4 · DOI
- Paszke, A. et al. in Advances in Neural Information Processing Systems 32 (NeurIPS).
- Kingma, D. P. & Ba, J. L. in 3rd International Conference for Learning Representations (ICLR).
- Loshchilov, I. & Hutter, F. Decoupled Weight Decay Regularization. arXiv:1711.05101 (2017). <https://ui.adsabs.harvard.edu/abs/2017arXiv171105101L>. · arXiv
- 10.1093/nar/gkz950 · DOI
- 2016, IPD-MHC 2.0: an improved inter-species database for the study of the major histocompatibility complex, Nucleic Acids Research, 45, D860
- 2022, UniProt: the Universal Protein Knowledgebase in 2023, Nucleic Acids Research, 51, D523
- 10.1093/bioinformatics/12.2.95 · DOI
- 2001, What is the value added by human intervention in protein structure prediction?, Proteins: Structure, Function, and Bioinformatics, 45, 86, 10.1002/prot.10021 · DOI
- 10.1006/jmbi.1994.1104 · DOI
- 10.1186/1471-2105-10-394 · DOI
- 10.1093/bioinformatics/btu829 · DOI
Metadata
- Authors
- Albert, B. A., Yang, Y., Shao, X. M., Singh, D., Smith, K. N., Anagnostou, V., & Karchin, R.
- Affiliations
- Johns Hopkins University (US)
- Corresponding author
- Karchin, Rachel
- ORCID
- Venue
- Nature Machine Intelligence
- Publisher
- openRxiv
- Published
- 2022-08-29
- Volume
- 5
- Issue
- 8
- Pages
- 861-872
- ISSN
- 2522-5839 (ISSN-L)
- PMC
- PMC10569228
- OpenAlex
- W4384934897
- Cited by
- 108
- References
- 1
- Citations by year
- 2026: 37 · 2025: 40 · 2024: 29 · 2023: 2
- FWCI
- 9.62
- Type
- journal article
- Language
- en
- Open access
- green
- doi.org
- Locations
- Nature Machine Intelligence (publishedVersion)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion)
- Funders
- U.S. Department of Defense (CA190755)
- American College of Radiology Imaging Network
- ECOG-ACRIN Cancer Research Group
- National Institutes of Health (5r01ca121113-15, CA121113)
- National Cancer Institute (R01-CA121113, UG1CA233259)
- Congressionally Directed Medical Research Programs
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › vaccines and immunoinformatics approaches
- Life Sciences › Immunology and Microbiology › Immunology › Immunotherapy and Immune Responses
- Health Sciences › Medicine › Oncology › Cancer Immunotherapy and Biomarkers
- Keywords
- transfer learning, deep neural network, cancer vaccines, personalized cancer immunotherapy, ensemble learning, long short-term memory networks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
