Dataset Information

Deep neural networks predict class I major histocompatibility complex epitope presentation and transfer learn neoepitope immunogenicity.

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, creating a needle-in-a-haystack problem. Here we address this challenge, presenting computational methods for predicting class I major histocompatibility complex (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 (area under the receiver operating characteristic = 0.9733; area under the precision-recall curve = 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.

SUBMITTER: Albert BA

PROVIDER: S-EPMC10569228 | biostudies-literature | 2023 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep neural networks predict class I major histocompatibility complex epitope presentation and transfer learn neoepitope immunogenicity.

Albert Benjamin Alexander BA Yang Yunxiao Y Shao Xiaoshan M XM Singh Dipika D Smit Kellie N KN Anagnostou Valsamo V Karchin Rachel R

Nature machine intelligence 20230720 8

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, creating a needle-in-a-haystack problem. Here we address this challenge, presenting computational methods for predicting class I major histocompatibility complex (MHC-I) epitopes and identifying immunogenic neoepitopes with improved ...[more]

PMID: 37829001

Similar Datasets

Project description:UnlabelledHelper T-cell epitope dominance in human immunodeficiency virus type 1 (HIV-1) envelope glycoprotein gp120 is not adequately explained by peptide binding to major histocompatibility complex (MHC) proteins. Antigen processing potentially influences epitope dominance, but few, if any, studies have attempted to reconcile the influences of antigen processing and MHC protein binding for all helper T-cell epitopes of an antigen. Epitopes of gp120 identified in both humans and mice occur on the C-terminal flanks of flexible segments that are likely to be proteolytic cleavage sites. In this study, the influence of gp120 conformation on the dominance pattern in gp120 from HIV strain 89.6 was examined in CBA mice, whose MHC class II protein has one of the most well defined peptide-binding preferences. Only one of six dominant epitopes contained the most conserved element of the I-Ak binding motif, an aspartic acid. Destabilization of the gp120 conformation by deletion of single disulfide bonds preferentially enhanced responses to the cryptic I-Ak motif-containing sequences, as reported by T-cell proliferation or cytokine secretion. Conversely, inclusion of CpG in the adjuvant with gp120 enhanced responses to the dominant CD4+ T-cell epitopes. The gp120 destabilization affected secretion of some cytokines more than others, suggesting that antigen conformation could modulate T-cell functions through mechanisms of antigen processing.ImportanceCD4+ helper T cells play an essential role in protection against HIV and other pathogens. Thus, the sites of helper T-cell recognition, the dominant epitopes, are targets for vaccine design; and the corresponding T cells may provide markers for monitoring infection and immunity. However, T-cell epitopes are difficult to identify and predict. It is also unclear whether CD4+ T cells specific for one epitope are more protective than T cells specific for other epitopes. This work shows that the three-dimensional (3D) structure of an HIV protein partially determines which epitopes are dominant, most likely by controlling the breakdown of HIV into peptides. Moreover, some types of signals from CD4+ T cells are affected by the HIV protein 3D structure; and thus the protectiveness of a particular peptide vaccine could be related to its location in the 3D structure.

Dataset Information

Deep neural networks predict class I major histocompatibility complex epitope presentation and transfer learn neoepitope immunogenicity.

Publications

Deep neural networks predict class I major histocompatibility complex epitope presentation and transfer learn neoepitope immunogenicity.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets