Machine learning predictions of MHC-II specificities reveal alternative binding mode of class II epitopes - Research summary

Aug 5, 2026

Upload by Anda

CD4+ T cells recognize peptides displayed on class II major histocompatibility complex (MHC-II) molecules, yet epitope prediction remains limited by the extreme polymorphism of MHC-II genes. The work curates a mass-spectrometry dataset containing 627,013 unique MHC-II ligands and uses it to define high-resolution binding motifs for 88 MHC-II alleles across multiple species. Motif analysis, supported by X-ray crystallography, uncovers a widespread reverse-binding mode for HLA-DP ligands. A machine-learning framework is then developed to predict binding specificities and ligands for any MHC-II allele, improving CD4+ T cell epitope predictions and enabling discovery of viral and bacterial epitopes.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages31

Reading Time1 h 18 min

AccessFree to download

Highlights

Curated MHC-II ligand dataset

Binding motifs for 88 alleles

Reverse-binding mode in HLA-DP

MixMHC2pred improves epitope prediction

Summary

Immunological context of CD4+ T cells and MHC-II polymorphism

Dataset size and cross-species coverage

Structural refinement and binding-mode discovery

Machine-learning framework for allele-agnostic prediction

Introduction

Role of CD4+ T cells in infection and cancer immunotherapy

MHC-II antigen presentation pathway and peptide loading

Canonical binding mode and known exceptions

Article Machine learning predictions of MHC-II speciﬁcities reveal alternative binding mode of class II epitopes Graphical abstract Highlights d Analysis of MHC-II peptidomics data identiﬁes >600,000 MHC-II ligands Authors Julien Racle, Philippe Guillaume, Julien Schmidt, ..., Michal Bassani-Sternberg, Alexandre Harari, David Gfeller Correspondence [julien.racle@unil.ch](julien.racle@unil.ch) (J.R.), [david.gfeller@unil.ch](david.gfeller@unil.ch) (D.G.) In brief CD4+ T cells recognize peptides displayed on MHC-II molecules. Racleet al. curate >600,000 MHC-II ligands and derive high-resolution motifs for 88 MHC-II alleles. These motifs reveal a widespread reverse-binding mode for HLA-DP ligands and improve predictions of CD4+ T cell epitopes with MixMHC2pred. d Motif deconvolution determines high-resolution binding motifs for 88 MHC-II alleles d Structural analysis uncovers alternative binding modes of MHC-II ligands d MixMHC2pred improves predictions of CD4+ T cell epitopes Racle et al., 2023, Immunity 56, 1359–1375 June 13, 2023 ª 2023 The Authors. Published by Elsevier Inc. [https://doi.org/10.1016/j.immuni.2023.03.009](https://doi.org/10.1016/j.immuni.2023.03.009) ll ll OPEN ACCESS Article Machine learning predictions of MHC-II speciﬁcities reveal alternative binding mode of class II epitopes Julien Racle,1,2,3,4,* Philippe Guillaume,1,4,5 Julien Schmidt,1,4,5 Justine Michaux,1,3,4,5,6 Am´ed´e Larabi,7 Kelvin Lau,7 Marta A.S. Perez,1,2,4 Giancarlo Croce,1,2,3,4 Rapha€el Genolet,1,4,5 George Coukos,1,3,4,5 Vincent Zoete,1,2,4 Florence Pojer,7 Michal Bassani-Sternberg,1,3,4,5,6 Alexandre Harari,1,3,4,5 and David Gfeller1,2,3,4,8,* 1Department of Oncology UNIL CHUV, Ludwig Institute for Cancer Research, University of Lausanne, Lausanne, Switzerland 2Swiss Institute of Bioinformatics (SIB), Lausanne, Switzerland 3Agora Cancer Research Centre, Lausanne, Switzerland 4Swiss Cancer Center Leman (SCCL), Lausanne, Switzerland 5Department of Oncology UNIL CHUV, Ludwig Institute for Cancer Research, University Hospital of Lausanne, Lausanne, Switzerland 6Center of Experimental Therapeutics, Department of Oncology, Centre Hospitalier Universitaire Vaudois (CHUV), Lausanne, Switzerland 7School of Life Sciences, Ecole Polytechnique F´ed´erale de Lausanne, Lausanne, Switzerland 8Lead contact *Correspondence: [julien.racle@unil.ch](julien.racle@unil.ch) (J.R.), [david.gfeller@unil.ch](david.gfeller@unil.ch) (D.G.) [https://doi.org/10.1016/j.immuni.2023.03.009](https://doi.org/10.1016/j.immuni.2023.03.009) SUMMARY CD4+ T cells orchestratethe adaptive immune response against pathogens and cancer by recognizing epitopes presented on class II major histocompatibility complex (MHC-II) molecules. The high polymorphism of MHC-II genes represents an important hurdle toward accurate prediction and identiﬁcation of CD4+ T cell epitopes. Here we collected and curated a dataset of 627,013 unique MHC-II ligands identiﬁed by mass spectrometry. This enabled us to precisely determine the binding motifs of 88 MHC-II alleles across humans, mice, cattle, and chickens. Analysis of these binding speciﬁcities combined with X-ray crystallography reﬁned our understanding of the molecular determinants of MHC-II motifs and revealed a widespread reverse-binding mode in HLA-DP ligands. We then developed a machine-learning framework to accurately predict binding speciﬁcities and ligands of any MHC-II allele. This tool improves and expands predictions of CD4+ T cell epitopesand enables us to discover viral and bacterial epitopes following the aforementioned reverse-binding mode. INTRODUCTION CD4+ T cells are key components of the adaptive immune system. They are implicated in priming and modulating the immune response to pathogens and cancer. CD4+ T cells also play an essential role in cancer immunotherapy,1 ,2 as demonstrated by CD4+ T cell responses following neoantigen-based cancer vaccines3–

FAQs

What dataset is curated to study MHC-II specificities?

A curated collection of 627,013 unique MHC-II ligands identified by mass spectrometry.

How many MHC-II alleles and what output are derived from the analysis?

High-resolution binding motifs are derived for 88 MHC-II alleles across humans, mice, cattle, and chickens.

What is the key binding-mode finding for HLA-DP ligands?

Structural and motif analyses reveal a widespread reverse-binding mode for HLA-DP ligands.

How does MixMHC2pred support CD4+ T cell epitope prediction?

It is a machine-learning framework that predicts binding specificities and ligands for any MHC-II allele, improving and expanding CD4+ T cell epitope predictions, including viral and bacterial epitopes.

Aug 5, 2026

Upload by Anda