NetMHCpan, a Method for Quantitative Predictions of Peptide Binding to Any HLA-A and -B Locus Protein of Known Sequence

By Morten Nielsen, Claus Lundegaard, Thomas H. Blicher, Kasper Lamberth, M. Harndahl, S. Justesen, G. Røder, Bjoern Peters, Alessandro Sette, Ole Lund and 1 more. 2007

693 citations recorded by the source corpus.

Background Binding of peptides to Major Histocompatibility Complex (MHC) molecules is the single most selective step in the recognition of pathogens by the cellular immune system. The human MHC class I system (HLA-I) is extremely polymorphic. The number of registered HLA-I molecules has now surpassed 1500. Characterizing the specificity of each separately would be a major undertaking. Principal Findings Here, we have drawn on a large database of known peptide-HLA-I interactions to develop a bioinformatics method, which takes both peptide and HLA sequence information into account, and generates quantitative predictions of the affinity of any peptide-HLA-I interaction. Prospective experimental validation of peptides predicted to bind to previously untested HLA-I molecules, cross-validation, and retrospective prediction of known HIV immune epitopes and endogenous presented peptides, all successfully validate this method. We further demonstrate that the method can be applied to perform a clustering analysis of MHC specificities and suggest using this clustering to select particularly informative novel MHC molecules for future biochemical and functional analysis. Conclusions Encompassing all HLA molecules, this high-throughput computational method lends itself to epitope searches that are not only genome- and pathogen-wide, but also HLA-wide. Thus, it offers a truly global analysis of immune responses supporting rational development of vaccines and immunotherapy. It also promises to provide new basic insights into HLA structure-function relationships. The method is available at http://www.cbs.dtu.dk/services/NetMHCpan.

Subfields: Immunology, Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Immunology

In Molecular Biology

Citations in the corpus

Simulation-assisted machine learning (2018, neutral): “, (Nielsen et al., 2007)) predict if a given peptide will bind to a given HLA allele using properties of the amino acids but without using 3D details of the chemical structure of the peptide or information on the structural binding of the peptide and HLA molecule.”
Computational Immunology Meets Bioinformatics: The Use of Prediction Tools for Molecular Binding in the Simulation of the Immune System (2010, methodology): “To this purpose, we use the binding motif matrices generated from the NN methods described in [5].”
The Immune Epitope Database and Analysis Resource in Epitope Discovery and Synthetic Vaccine Design (2017, methodology): “This process is then repeated four more times, with each subset being omitted from training and used to compare predictions (10).”
pVAC-Seq: A genome-guided in silico approach to identifying tumor neoantigens (2016, methodology): “Previous studies [38, 39] have shown that allele-specific epitope prediction software, such as NetMHC, perform slightly better when compared to pan-specific methods such as NetMHCpan [40–42] in case of wellcharacterized alleles due to availability of large amounts of training data.”
NetMHCpan, a method for MHC class I binding prediction beyond humans (2008, positive): “…algorithm to capture general features of the relationship between peptides and MHC sequences, and it is therefore apparent that the method should perform better when the query MHC molecule is represented by closely related MHC molecules with characterized binding specificity (Nielsen et al. 2007).”
Fundamentals and Methods for T- and B-Cell Epitope Prediction (2017, methodology): “To overcome this limitation, some researchers have developed pan-MHCspecific methods by training ANNs on input data combining MHC residues that contact the peptide with peptide-binding affinity that are capable of predicting peptide-binding affinities to uncharacterized HLA alleles [54, 55].”
Genomic Correlates of Immune-Cell Infiltrates in Colorectal Carcinoma (2016, methodology): “We predicted the binding affinities of all possible 9- and 10-mer mutant peptides to the corresponding POLYSOLVER-inferred HLA alleles using NetMHCpan (v.2.4; Nielsen et al., 2007).”
Evaluation of machine learning methods to predict peptide binding to MHC Class I proteins (2017, methodology): “Our results indicate that four methods: NetMHC, NetMHCpan, MHCflurry, and our new method MHCnuggets have best-in-class prediction performance, and of these methods MHCnuggets was substantially faster.”
A comprehensive review and performance evaluation of bioinformatics tools for HLA class I peptide-binding prediction (2020, methodology): “Specifically, two NN-based prediction tools and one PSSM-based tool have been included in this consensus method, namely (i) NetMHC 3.4 [94, 95], (ii) NetMHCpan 2.8 [86, 96] and (iii) Pickpocket 1.1 [24].”
Improved methods for predicting peptide binding affinity to MHC class II molecules (2018, methodology): “Earlier work has demonstrated that the prediction performance of both NetMHCII and NetMHCIIpan is dependent on the amount of peptide binding data (10),(11) and one would therefore expect the two methods to improve in performance if retrained on an extended peptide binding data set.”
Neoantigens: promising targets for cancer therapy (2023, methodology): “It is noteworthy that NetMHCpan, in contrast to state-of-the-art methods, improves the prediction performance of tumor neoantigens by combining information from binding affinity data with MS peptidome data to give a "panspecific" machine-learning strategy for MHC-I alleles.(230,264,267) Two recent s”
Predicting Peptide-MHC Binding Affinities with Imputed Training Data (2016, methodology): “Two of these predictors, NetMHC (7) and NetMHCpan (8), have emerged as the methods of choice across multiple fields of study within immunology, including virology (9), tumor immunology (10), and autoimmunity (11).”
NetMHCpan 4.0: Improved peptide-MHC class I interaction predictions integrating eluted ligand and peptide binding affinity data (2017, methodology): “The inputs to the neural networks consisted of the peptide and the MHC molecule in terms of a pseudo-sequence (8).”
Genomic and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic Melanoma (2016, neutral): “8 (Hoof et al., 2009; Nielsen et al., 2007) and NetMHCIIpan v3.”