A model of antigen processing improves prediction of MHC I-presented peptides

O’Donnell, Timothy J.; Rubinsteyn, Alex; Laserson, Uri

doi:10.1101/2020.03.28.013714

Cited by 4 publications

(4 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other models derived from it, such as MHCflurry [O'Donnell et al, 2020] and DeepHLApan [Wu et al, 2019], have similar methods.…”

Section: Potential Application Suggested From Research Resultsmentioning

confidence: 99%

GraphMHC: neoantigen prediction model applying the graph neural network to molecular structure

Jeong,

Cho,

Gim

et al. 2023

Preprint

View full text Add to dashboard Cite

Neoantigens are biomarkers that can predict the prognosis associated with immune checkpoint inhibition by estimating the binding potential of candidate peptides to somatic mutation and major histocompatibility complex (MHC) proteins. Although deep neural networks have been primarily used for these prediction models, it is difficult to consider the models reported thus far as accurately representing the interactions between biomolecules. In this study, we propose the GraphMHC model, which utilizes a graph neural network model through molecular structure to simulate the binding between MHC proteins and peptide sequences. Amino acid sequences sourced from the immune epitope database (IEDB) undergo conversion into molecular structures. Subsequently, atomic intrinsic informations and inter-atomic connections are extracted and structured as a graph representation. Bindings are classified by feeding them into the GraphMHC network, comprising stacked graph attention and convolution layers. The prediction results from the test set using the GraphMHC model showed a high performance with an area under the receiver operating characteristic curve of 92.2% (91.9-92.5%), surpassing the baseline model. Moreover, by applying the GraphMHC model to melanoma patient data from the Cancer Genome Atlas project, we found a borderline difference in overall survival and a significant difference in stromal score between the high and low neoantigen load groups. This distinction was not present in the baseline model. This study presents the first feature-intrinsic method based on biochemical molecular structure for modeling the binding between MHC protein sequences and neoantigen candidate peptide sequences. The model can provide highly accurate suitability information for cancer patients who want to apply immune checkpoint inhibitors.Author summary

show abstract

“…Other models derived from it, such as MHCflurry [O'Donnell et al, 2020] and DeepHLApan [Wu et al, 2019], have similar methods.…”

Section: Potential Application Suggested From Research Resultsmentioning

confidence: 99%

GraphMHC: neoantigen prediction model applying the graph neural network to molecular structure

Jeong,

Cho,

Gim

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…For MHC class II design, we use NetMHCIIpan-4.0 ( Reynisson et al., 2020b ). For evaluation, we use our ensemble estimate of binding (MHC class I), as well as use binding predictions from a wide range of prediction algorithms (MHC class I: NetMHCpan-4.0 ( Jurtz et al., 2017 ), NetMHCpan-4.1 ( Reynisson et al., 2020a ), MHCflurry 1.6.0 ( O’Donnell et al., 2020 ), PUFFIN ( Zeng and Gifford, 2019 ); MHC class II: NetMHCIIpan-3.2 ( Jensen et al., 2018 ), NetMHCIIpan-4.0 ( Reynisson et al., 2020b ), PUFFIN ( Zeng and Gifford, 2019 )) to ensure that all methods agree that we have a good peptide vaccine. We validate our computational models on a dataset of SARS-CoV-2 peptides evaluated for stability ( Prachar et al., 2020 ).…”

Section: Methodsmentioning

confidence: 99%

“…Computational Peptide-HLA Prediction Models Computational Models For MHC class I design, we use an ensemble that outputs the mean predicted binding affinity of NetMHCpan-4.0 (Jurtz et al, 2017) and MHCflurry 1.6.0 (O'Donnell et al, 2020(O'Donnell et al, , 2018. We find this ensemble increases the precision of binding affinity estimates over the individual models on available SARS-CoV-2 experimental data (Table S1).…”

Section: Optivaxmentioning

confidence: 99%

Computationally Optimized SARS-CoV-2 MHC Class I and II Vaccine Formulations Predicted to Target Human Haplotype Distributions

Liu¹,

Carter²,

Bricken

et al. 2020

Cell Systems

View full text Add to dashboard Cite

Summary We present a combinatorial machine learning method to evaluate and optimize peptide vaccine formulations for SARS-CoV-2. Our approach optimizes the presentation likelihood of a diverse set of vaccine peptides conditioned on a target human-population HLA haplotype distribution and expected epitope drift. Our proposed SARS-CoV-2 MHC class I vaccine formulations provide 93.21% predicted population coverage with at least five vaccine peptide-HLA average hits per person (≥ 1 peptide: 99.91%) with all vaccine peptides perfectly conserved across 4,690 geographically sampled SARS-CoV-2 genomes. Our proposed MHC class II vaccine formulations provide 97.21% predicted coverage with at least five vaccine peptide-HLA average hits per person with all peptides having an observed mutation probability of ≤ 0.001. We provide an open-source implementation of our design methods (OptiVax), vaccine evaluation tool (EvalVax), as well as the data used in our design efforts here: https://github.com/gifford-lab/optivax .

show abstract

“…Candidate peptides are scored by their predicted binding affinity (IC50) to MHC molecules. For MHC class I, we use an ensemble that outputs the mean predicted binding affinity of NetMHCpan-4.0 (Jurtz et al, 2017) and MHCflurry 1.6.0 (O'Donnell et al, 2020(O'Donnell et al, , 2018. For MHC class II, we use NetMHCIIpan-4.0 (Reynisson et al, 2020b).…”

Section: Sars-cov-2 Proteome and Candidate Peptidesmentioning

confidence: 99%

Predicted Cellular Immunity Population Coverage Gaps for SARS-CoV-2 Subunit Vaccines and their Augmentation by Compact Peptide Sets

Liu¹,

Carter²,

Gifford

2020

Preprint

View full text Add to dashboard Cite

Subunit vaccines induce immunity to a pathogen by presenting a component of the pathogen and thus inherently limit the representation of pathogen peptides for cellular immunity based memory. We find that SARS-CoV-2 subunit peptides may not be robustly displayed by the Major Histocompatibility Complex (MHC) molecules in certain individuals. We introduce an augmentation strategy for subunit vaccines that adds a small number of peptides to a vaccine to improve the population coverage of pathogen peptide display. We augment a subunit vaccine by selecting additional pathogen peptides to maximize the total number of vaccine peptide hits against the distribution of MHC haplotypes in a population. For each subunit we design independent MHC class I and MHC class II peptide sets for augmentation, and alternatively design a combined set of peptides for MHC class I and class II display. We evaluate the population coverage of 9 different subunits of SARS-CoV-2, including 5 functional domains and 4 full proteins, and augment each of them to fill a predicted coverage gap. We predict that a SARS-CoV-2 receptor binding domain subunit vaccine will have fewer than six peptide-HLA hits with ≤ 50 nM binding affinity per individual in 51.31% (class I) and 32.99% (class II) of the population, and with augmentation, the uncovered population is predicted to be reduced to 0.54% (class I) and 1.46% (class II). We find that a joint set of pathogen peptides for MHC class I and class II display is predicted to produce a more compact vaccine design than using independent sets for MHC class I and class II. We provide an open source implementation of our design methods (OptiVax), vaccine evaluation tool (EvalVax), as well as the data used in our design efforts here: https://github.com/gifford-lab/optivax/tree/master/augmentation.

show abstract

A model of antigen processing improves prediction of MHC I-presented peptides

Cited by 4 publications

References 47 publications

GraphMHC: neoantigen prediction model applying the graph neural network to molecular structure

GraphMHC: neoantigen prediction model applying the graph neural network to molecular structure

Computationally Optimized SARS-CoV-2 MHC Class I and II Vaccine Formulations Predicted to Target Human Haplotype Distributions

Predicted Cellular Immunity Population Coverage Gaps for SARS-CoV-2 Subunit Vaccines and their Augmentation by Compact Peptide Sets

Contact Info

Product

Resources

About