Cross-reactive MHC Class I T Cell Epitopes May Dictate Heterologous Immune Responses Between Respiratory Viruses and Food Allergens

Abstract

Respiratory virus infections play a major role in asthma inception, persistence, and exacerbations. There is also a close correlation between asthma and food allergy, and we hypothesize that food-allergen-induced T cell-mediated heterologous immunity likely plays a role in inducing asthma symptoms in sensitized individuals. In this study, we used two independent in silico pipelines for the identification of cross-reactive virus- and food allergen- derived T cell epitopes, considering individual peptide sequence similarity, MHC binding affinity and immunogenicity. We assessed the proteomes of human rhinovirus (RV1b), respiratory syncytial virus (RSVA2) and influenza-strains contained in the seasonal quadrivalent influenza vaccine 2019/2020 (QIV 2019/2020), as well as SARS-CoV-2 for the most frequent human HLA alleles, in addition to more than 200 most common food allergen protein sequences. All resulting allergen-derived peptide candidates were subjected to an elaborate scoring system considering multiple criteria, including clinical relevance. In both bioinformatics approaches, we found that shortlisted peptide pairs that are potentially binding to MHC class II molecules scored up to 10x lower compared to MHC class I candidate epitopes. For MHC class I food allergen epitopes, several candidate peptides from shrimp, kiwi, apple, soy bean and chicken were identified. Such allergen sources contained potentially cross-reactive epitopes to the aforementioned viruses. The shortlisted set of peptide pairs may be implicated as heterologous virus-mediated immune response to food allergens. Our findings may be translated to peptide immunization strategies with immunomodulatory properties.

Full text

Page 1/13
Cross-reactive MHC Class I T Cell Epitopes May Dictate
Heterologous Immune Responses Between Respiratory
Viruses and Food Allergens
Kathrin Balz
Universities of Giessen and Marburg Lung Center (UGMLC), Philipps University Marburg
Abhinav Kaushik
Sean N. Parker Center for Allergy and Asthma Research at Stanford University
Franz Cemic
TH Mittelhessen, University of Applied Sciences Gießen
Vanitha Sampath
Sean N. Parker Center for Allergy and Asthma Research at Stanford University
Vanessa Heger
TH Mittelhessen, University of Applied Sciences Gießen
Harald Renz
Universities of Giessen and Marburg Lung Center (UGMLC), Philipps University Marburg
Kari Nadeau
Sean N. Parker Center for Allergy and Asthma Research at Stanford University
Chrysanthi Skevaki (  Chrysanthi.Skevaki@uk-gm.de )
Universities of Giessen and Marburg Lung Center (UGMLC), Philipps University Marburg
Article
Keywords: T cell-mediated heterologous immunity, Food allergy, Asthma, Allergen sensitization, Cross-reactive T cells
Posted Date: March 30th, 2023
DOI: https://doi.org/10.21203/rs.3.rs-2660592/v1
License:   This work is licensed under a Creative Commons Attribution 4.0 International License.  Read Full License

Page 2/13
Abstract
Respiratory virus infections play a major role in asthma inception, persistence, and exacerbations. There is also a close
correlation between asthma and food allergy, and we hypothesize that food-allergen-induced T cell-mediated heterologous
immunity likely plays a role in inducing asthma symptoms in sensitized individuals. In this study, we used two independent
in silico
pipelines for the identication of cross-reactive virus- and food allergen- derived T cell epitopes, considering
individual peptide sequence similarity, MHC binding anity and immunogenicity. We assessed the proteomes of human
rhinovirus (RV1b), respiratory syncytial virus (RSVA2) and inuenza-strains contained in the seasonal quadrivalent inuenza
vaccine 2019/2020 (QIV 2019/2020), as well as SARS-CoV-2 for the most frequent human HLA alleles, in addition to more
than 200 most common food allergen protein sequences. All resulting allergen-derived peptide candidates were subjected to
an elaborate scoring system considering multiple criteria, including clinical relevance. In both bioinformatics approaches, we
found that shortlisted peptide pairs that are potentially binding to MHC class II molecules scored up to 10x lower compared
to MHC class I candidate epitopes. For MHC class I food allergen epitopes, several candidate peptides from shrimp, kiwi,
apple, soy bean and chicken were identied. Such allergen sources contained potentially cross-reactive epitopes to the
aforementioned viruses. The shortlisted set of peptide pairs may be implicated as heterologous virus-mediated immune
response to food allergens. Our ndings may be translated to peptide immunization strategies with immunomodulatory
properties.
Introduction
During the last few decades, the prevalence of allergic diseases has dramatically increased in developed countries. The
incidence of asthma has increased fourfold since the 1950´s (1) and food allergy prevalence among children has increased
to 3.5-8 % (2) Food allergy is often classied into either IgE-mediated, non-IgE-mediated, or mixed IgE/non-IgE- mediated
allergic disease. Non-IgE-mediated food allergy is thought to be initiated by T cells, although the pathophysiological
mechanisms underlying these reactions are not yet fully understood (3, 4). Studies have observed that food allergy is
associated with subsequent increases in the development of allergic rhinitis and asthma (5). Both consumption and
inhalation of food allergens can cause allergic reactions in sensitized individuals. Inhalation of aerosolized wheat, lupin, and
other food allergens are thought to stimulate mast cells in the lung and have been associated with respiratory symptoms,
including wheeze (6–8). As wheezing in childhood is mostly associated with viral infections, an indirect correlation between
viral infections and food allergy has been suggested. Gastrointestinal viral infections are also more relevant in the context of
food allergy (9). Some mouse studies provide evidence for the subsequent development of food allergen specic IgE on
exposure to food allergens after gastrointestinal infection with certain RNA viruses, including murine norovirus type 1 or
reovirus (10, 11).
We have previously shown an inuenza virus-mediated protective effect over development of experimental asthma in
models of ovalbumin and house dust mite-induced asthma. This effect was mediated by cross-reactive T effector memory
cells (12). We hypothesized that cross-reactive T cell epitopes present in respiratory viruses and food allergens may provide
the missing link between allergy and wheeze.
Therefore, our aim was to identify potentially cross-reactive T cell epitope pairs among food allergens and clinically relevant
respiratory viruses. Such data can serve as the basis for investigating heterologous immune responses among patients with
a specic food allergy and a viral infection or even for following outcomes (benecial or harmful) of antiviral vaccination.
Recently, computational advances have signicantly improved our understanding of cross-reactive sites across different
allergens. These methods screen for antigenic peptides with similar epitope prole to bind a particular T cell receptor (TCR)
(13). However, most of these methods have their input dataset requirements (e.g., TCRseq or expression data or 3D structure
of proteins) and with limited or no options to customize the source of training dataset (e.g., training with experimentally
validated T-cell epitopes only) to achieve case-specic objectives (14–16). Therefore, in this study, we applied two
independent approaches developed in-house to predict MHC binding cross-reactive peptide sequences across viral and food

Page 3/13
allergen sequences that can potentially cause similar T-cell response. These tools also include prediction of MHC binding
anity, which is highly dependent on the so called “anchor residues” of the peptides. Such residues represent at least two
amino acids, which are exactly tting into the groove of the MHC molecule. Thus, anchor residues play an important role in
the denition of antigenic peptides (17).
Results
We used partially modied versions of our previously published
in-silico
pipelines for prediction of potentially cross-reactive
T cell epitope pairs between allergens and viruses, focusing on food allergens and clinically relevant respiratory viruses. The
pipeline-1 uses an agnostic approach to predict the antigenic peptides by means of T cell epitope prediction for both viruses
and allergens as well as for the sequence homology of the predicted epitopes, whereas pipeline-2 uses a supervised
approach, wherein IEDB T-cell epitope repertoire was used to predict the cross-reactive antigenic regions across viruses and
allergens.
Pipeline-1 predicted multiple potentially-cross-reactive T cell epitope pairs for each virus, which were further ranked based on
the calculated pair combined score and subsequently scored for the top 5 candidate pairs based on additional criteria.
Among the top 5 results for all viruses, we identied T cell epitope pairs with the allergenic epitopes deriving from kiwi
(
Actinidia deliciosa, Act d
), chicken (
gallus gallus, Gal d
) and apple (
malus domestica, Mal d
) (Fig.1). The top 5 candidates
for RSVA2, RV1b and inuenza strains of the seasonal quadrivalent inuenza vaccine 2019/2020 (QIV 19/20) on the
background of the most frequent human HLA class I alleles are depicted in Table1. The corresponding top 5 candidate pairs
for SARS-CoV-2 for pipeline-1 were published previously (21).

Page 4/13
Table 1
Top 5 candidate epitope pairs for RV1b, RSVA2 and Inuenza vaccine 2019/2020 on the background of the most frequent
human HLA alleles (Suppl. Table3), based on pipeline-1 and calculation of the nal score as described (see Material and
Methods)
A: RV1b       
 allergen
epitope allergen allergen
protein
family
viral epitope viral
protein MHC allele nal
score 
Nr.
1ASDVICQEY Gal d 5 Serum
albumin MHDSILVSY P2-A HLA-A*01:01 1258 
Nr.
2VSDDGLNIY Gal d 6 Lipoprotein LSCKFLPLY P2-C HLA-A*01:01 1032 
Nr.
3FADLTNEEY Act d 1 Papain-like
cysteine
protease
YIPETEDDY Protease HLA-A*01:01 610 
Nr.
4KVFRFSMFK Gal d 6 Lipoprotein GIFGENMYY VP2 HLA-A*11:01 508 
Nr.
5FLGDVIPPGI Gal d 6 Lipoprotein LLLAYTPPGI VP3 HLA-A*02:01 484 
B: RSVA2       
 allergen
epitope allergen allergen
protein
family
viral epitope viral
protein MHC allele nal
score 
Nr.
1FLGDKFYTV Gal d 3 Transferrin FLPDKISLT RNA-
directed
RNA
polymerase
L
HLA-A*02:01 1166 
Nr.
2VYMDLPHGI Mal d Unknown LYMNLPMLF RNA-
directed
RNA
polymerase
L
HLA-A*24:02 1148 
Nr.
3TTYKEFLGDK Gal d 3 Transferrin TTYNQFLTWK RNA-
directed
RNA
polymerase
L
HLA-A*11:01 642 
Nr.
4FADLTNEEY Act d 1 Papain-like
cysteine
protease
KSNRYNDNY RNA-
directed
RNA
polymerase
L
HLA-A*01:01 586 
Nr.
5YLLDLLPAA Gal d 6 Lipoprotein YLSELLNSL RNA-
directed
RNA
polymerase
L
HLA-A*02:01 516 
C: Inuenza vaccine 2019 /
2020 
 allergen
epitope allergen allergen
protein
family
viral epitope virus
protein virus strain MHC
allele nal
score

Page 5/13
A: RV1b       
Nr.
1SPARLYNAL Mal d 1 Bet v 1
family NPAMLYNKM PB2 A/Kansas/2017 HLA-
B*07:02 1784
Nr.
2YLLQNPEAYV Mal d 1 Bet v 1
family SLYQNADAYV HA A/Brisbane/2018 HLA-
A*02:01 636
Nr.
3FLGDKFYTV Gal d 3 Transferrin FMYSDFHFI PA A/Brisbane/2018,
A/Kansas/2017 HLA-
A*02:01 604
Nr.
4KPTSKRMAI Act d 1 Papain-like
cysteine
protease
EPESKRMSL NS1 B/Colorado/2017,
B/Phuket/2013 HLA-
B*07:02 556
Nr.
5YTQTYGVDY Mal d Unknown YTDTYHSYA NA B/Colorado/2017,
B/Phuket/2013 HLA-
A*01:01 432
Interestingly, sequences from apple only appeared in the QIV/19/20 top 5 candidate epitope sequences. Sequences from Gal
d 6 and Act d 1 were predicted in 3 out of the 4 virus analyses, suggesting an important role in cross-reactivity between
respiratory viruses and food allergens. With regards to the viral proteins, the predicted top 5 sequences for RSVA2 were
derived exclusively from the RNA-directed polymerase, whereas the viral protein source for the other viruses were more
diverse. Of note, when applying the pipeline for human HLA class II prediction, no candidate epitope pairs could be identied
for RV1b and QIV19/20. Epitope pairs for RSVA2 achieved similar scores in both HLA class I and II analyses (Table2).
Importantly, such epitopes derived from a variety of viral proteins and different allergenic sources compared with candidates
for HLA class I. Allergenic sources include buckwheat (
Fagopyrum esculentum, Fag e
), sesame (
Sesamum indicum, Ses i)
,
potato
(Solanum tuberosum, Sola t)
, and hazelnut
(Corylus avellana, Cor a)
, with almost all belonging to the protein family of
cupins.
Table 2
Top 5 candidate pairs for RSVA2 on the background of the most frequent human HLA class II alleles (Suppl.Table3) based
on pipeline-1 and calculation of the nal score as described (see Material and Methods)
allergen
epitope allergen allergen
protein
family
viral epitope viral protein MHC allele nal
score
Nr.
1Fag e 1 LPILEFLQLSAQHVV Cupin KGAFKYIKPQSQFIV Matrix protein HLA-
DRB1_01_01 1292
Nr.
2Ses i 3 VLFALLLASAVVASE Cupin MELLILKANAITTIL Fusion
glycoprotein
F0
HLA-
DRB1_01_01 666
Nr.
3Cor a 9 ARRLKYNRQETTLAR Cupin LRWLTYYKLNTYPSL RNA-directed
RNA
polymerase L
HLA-
DRB1_04_01 638
Nr.
4Sola t 1 FAKLLSDRKKLRANK Patatin
family TELNSDDIKKLRDNE Matrix M2-1 HLA-
DRB1_03_01 350
Nr.
5Fag e 1 LPILEFIQLSAQHVV Cupin KGAFKYIKPQSQFIV Matrix protein HLA-
DRB1_01_01 168
In order to validate our
in-silico
results we applied an independent pipeline that uses known T cell epitope features as they
appear on IEDB tools. The pipeline is an extension of a computational framework we published previously (21), wherein
similarity between identical
k
-mers with experimentally validated T cell epitopes was used to predict cross-reactive peptides.
In this analysis, we observed a limited number of cross-reactivity sites between food and viral antigens, wherein, the largest
number of cross-reactive peptides were predicted between
Glycine max
(soybean) and SARS-CoV-2 replicase polyprotein.
Besides the polyprotein, SARS-CoV-2 spike glycoprotein also shares potential cross-reactive amino acids with the

Page 6/13
sarcoplasmic ca-binding proteins of
Crangon crangon
(shrimp). In addition to SARS-CoV-2, we also observed that human
respiratory syncytial virus A proteins also share potential cross-reactive sites with food allergen sequences from
Crangon
crangon
and
Malus domestica
(apple). Overall, six different protein sequences from two viral species were found to share
cross-reactive amino acids with three food allergens. Wherein, we predicted 44 redundant (16 non-redundant) cross-reactive
peptide pairs between those six viral sequences with 16 different food allergen sequences (Table3; Supplementary Table4).
A total of 22 unique peptides from food and viral allergens were nally shortlisted that were predicted to bind 27 unique
class-I HLA alleles, that includes both HLA-A and HLA-B genes.

Page 7/13
Table 3
Cross-reactive peptide peptides predicted between viral and allergen sequences using pipeline-2
Virus
name Protein Viral Sequence Allergen Protein Allergen Sequence
SARS-
CoV2 Replicase_polyprotein RLQAGNATEVPANST
Glycine
max
Beta-conglycinin
protein LQSGDALRVPAGTTF
SARS-
CoV2 Spike_glycoprotein GINITRFQTLLALHR
Glycine
max
Beta-conglycinin
protein LQSGDALRVPAGTTY
SARS-
CoV2 Spike_glycoprotein LPIGINITRFQTLLA
Glycine
max
Beta-conglycinin
protein LQSGDALRVPAGTTY
SARS-
CoV2 Replicase_polyprotein DEGNCDTLKEILVTY
Crangon
crangon
Iosephosphate
isomerase PCIGEKLDERESNRT
SARS-
CoV2 Replicase_polyprotein LQAGNATEVPANSTV
Crangon
crangon
Iosephosphate
isomerase PCIGEKLDERESNRT
HRSV A
(strain
A2)
RNA-directed RNA
polymerase L RLMEGQTHAQADYLL
Malus
domestica
Predicted protein
(E4Z8N9) RLFARTRQVESLTAE
SARS-
CoV2 Replicase_polyprotein RLQAGNATEVPANST
Malus
domestica
Predicted protein
(E4Z8N9) RLFARTRQVESLTAE
SARS-
CoV2 Replicase_polyprotein RLQAGNATEVPANST
Malus
domestica
Putative COBL7
(COBRA-LIKE 7) RLFARTRQVESLAAE
HRSV A
(strain
A2)
Non-structural protein
2CIVRKLDERQATFTF
Crangon
crangon
Sarcoplasmic
calcium-binding
protein
AGGINIARYQELYAQ
SARS-
CoV2 Replicase_polyprotein DEGNCDTLKEILVTY
Crangon
crangon
sarcoplasmic
calcium-binding
protein
GINIARYQELYAQFI
HRSV A
(strain
A2)
Non-structural protein
2CIVRKLDERQATFTF
Crangon
crangon
sarcoplasmic
calcium-binding
protein
KAGGINIARYQELYA
SARS-
CoV2 Membrane protein RLFARTRSMWSFNPE
Crangon
crangon
Sarcoplasmic
calcium-binding
protein
AGGINIARYQELYAQ
SARS-
CoV2 Spike_glycoprotein PIGINITRFQTLLAL
Crangon
crangon
sarcoplasmic
calcium-binding
protein
GGINIARYQELYAQF
SARS-
CoV2 Membrane protein RLFARTRSMWSFNPE
Crangon
crangon
sarcoplasmic
calcium-binding
protein
GINIARYQELYAQFI
SARS-
CoV2 Replicase_polyprotein DEGNCDTLKEILVTY
Gallus
gallus
Serum albumin DHGEADFLKSILIRY
SARS-
CoV2 Spike_glycoprotein GINITRFQTLLALHR
Gallus
gallus
Serum albumin DHGEADFLKSILIRY
SARS-
CoV2 Spike_glycoprotein GINITRFQTLLALHR
Malus
domestica
Two-component
response regulator MKGVTHGACDYLIKP
Comparing the T cell epitope pairs predicted with both pipelines, no exact sequences were predicted commonly. However,
predicted allergenic sequences from
malus domestica
and
gallus gallus
were identied with both pipelines and sequences
from
gallus gallus
were even predicted to potentially cross-react with viral epitopes from SARS-CoV-2 by both approaches.

Page 8/13
Discussion
We identied several potentially cross-reactive T cell epitope pairs between food allergens and epidemiologically relevant
respiratory viruses, using two independent
in-silico
pipelines. To our knowledge, this is the rst study investigating
heterologous immunity between T cell epitopes among several clinically relevant RNA viruses and food allergens. Our study
revealed allergenic sequences from
malus domestica
to be important for cross-reactivity. Further, we found that the epitope
sequences from SARS-CoV-2 and RSVA2 were those predicted to most likely cross-react with food allergens, as T cell epitope
pairs were predicted by both pipelines. Finally, we observed more candidate epitope pairs for HLA class I compared to HLA
class II.
In pipeline-1, we made use of several T cell epitope prediction tools followed by alignments based on sequence homology
and amino acid properties. The latter is particularly important for detection of cross-reactivity, which is known to also exist
between structurally unrelated antigens with little sequence homology (33). In addition, it is known that the position of
specic amino acids in their non-anchor region is critical for T cell epitope immunogenicity (31). A limitation of pipeline-1 is
the lack of such tools for prediction of immunogenicity, which is compensated by pipeline-2, which takes immunogenicity
into consideration.
We applied pipeline-2 to identify only known T cell epitopes in combination with alignment against respiratory bacteria in
order to avoid prediction contaminations. Hence, this restriction led to the lack of any identied epitope pairs for RV1b and
inuenza strains containing the seasonal quadrivalent inuenza vaccine. The allergen sequences among the short-listed
epitope pairs for SARS-CoV-2 and RSVA2 were derived from different allergen sources, including apple and chicken.
Importantly, using pipeline-1, several candidate epitope pairs involving SARS-CoV-2 contained chicken allergens as published
previously (21). Thus, such epitope pairs require further investigation.
Of note, we also predicted epitope pairs for RSVA2 on the background of human HLA class II alleles. Interestingly, no
candidate pairs could be identied using RV1b or inuenza proteins. In addition, we were able to identify candidate pairs for
SARS-CoV-2 but with a 10 x lower score compared to HLA class I-restricted epitopes (21). However, no HLA class II
candidates were predicted with pipeline-2.
Aerosolized food allergens may trigger asthma symptoms via inhalation and subsequent inammatory response in the lung
(7). Such cases have been shown for sh allergens (33) and soy proteins (34). Importantly, we predicted T cell epitope
sequences from RSVA2 and SARS-CoV-2 which appeared to be cross-reactive with sequences from
Crangon crangon
, a
shrimp, as well as sequences from RSVA2 which showed cross-reactivity with sequences from soybean.
Our study provides
in-silico
data, which support a yet unexplored pathogenic mechanism for the connection between food
allergy and virus-associated asthma. Future studies using PBMCs from individuals with relevant food allergies may be the
rst step to validate cross-reactivity involving the predicted T cell epitope pairs. Results of such studies may advice peptide
immunization strategies for a favorable outcome in the context of allergy and infection.
Material And Methods
We used our previously published
in-silico
analysis for prediction of potentially cross-reactive T cell epitope pairs between
viral and allergenic proteins, applying two independent pipelines. We downloaded all available food allergens from Allergen
Online (10.09.2017) (18–20), as well as protein sequences from the most clinically relevant respiratory viruses SARS-CoV-2,
RSVA2, RV1b and inuenza strains of the seasonal quadrivalent inuenza vaccine 2019/2020 (QIV19/20) (Uniprot and
GISAID). Protein sequences are provided in Supplementary Tables1 and 2.
Pipeline-1 was performed as described previously (21), using the most relevant human HLA alleles (Supplementary Table3).
Briey, viral T cell epitopes were predicted using smm (22), ann (23) and consensus (24) algorithm tools for MHCI (IC50
threshhold < = 5000nm), and netMHCII (25) for MHCII. Allergenic proteins were aligned against predicted viral epitopes (NCBI

Page 9/13
protein blast platform) (26) and subsequently used for T cell epitope prediction by netMHC (27) and netMHCpan (28) for
MHCI, and netMHCII and netMHCIIpan (29) for MHC class II prediction. Predicted viral and allergen epitopes were pairwise
aligned with a Biopython module pairwise 2 (30). A nal pair combined score was calculated and allergenic epitopes of the
top 30 candidate pairs further underwent a comprehensive scoring system taking clinical relevance and conservation criteria
into consideration. Epitope pairs were nally ranked for the top 5 candidates for each virus based on the nal score. A
schematic overview of the pipeline is depicted in Supplementary Fig.1A. The scoring system is shown in Supplementary
Fig.2.
In addition to the above pipeline-1, we also performed an analysis using a modied version of the independent pipeline we
published previously (Supplementary Fig.1B). We used IEDB database, which hosts known epitope peptide sequences, to
predict peptides that are known to bind MHC molecules. To identify cross-reactive antigenic peptides between the given food
allergenic protein sequences and viral proteomes, we split each of a given protein sequence into a set of sequential
k-mers
or
peptides (length = 15). To avoid contaminations in the downstream analysis, we ltered out the sequential
k-mers
that
mapped with bacterial protein sequences, using blastp (e-value < 1 & identity > 70% and coverage > 70%). Here, eight bacterial
species (
Moraxella catarrhalis, Chlamydophila pneumoniae, Mycoplasma pneumoniae, Coxiella burnetii, Streptococcus
pneumoniae, Haemophilus inuenzae, Streptococcus pyogenes
and
Legionella pneumophila
) were selected that are known
to cause common respiratory infections. Subsequently, using the IEDB immunogenicity prediction tool (31), we identied
from the remaining sequential
k
-mers those that can potentially form peptide MHC (pMHC) complexes. This tool uses the
properties of amino acids within a given peptide to predict its potential to form pMHC complexes. The above steps ltered
out a large proportion of sequential
k
-mers to retain only 134,171 high condence peptide sequences. However, it is likely
that some of the peptides may not have a strong MHC-I binding anity despite the homology, and therefore may be less
likely to be presented as antigens by HLA molecules. Therefore, using IEDB tools (32), the homologous peptides were further
evaluated for their binding anity with human MHC-I molecules for a broad range of alleles (
n
= 54) that are known to bind
viral and allergenic proteins. The peptides with MHC binding anity rank > 50 (arbitrarily selected threshold to retain at most
only 50% of the total number of sequences) were further selected for homology analysis. Wherein, using the blastp (e-value <
1 & identity > 50% and coverage > 70%), we identied those food allergen and viral antigenic peptides that share homologous
peptide sequences and which were considered as cross-reactive MHC binding peptides.
We evaluated the statistical signicance of predicted peptides (from method-2) by comparing the cross-reactive peptides
with 1000 randomly generated sequences using BLASTp program. The random peptide sequences were generated using
RSAT webserver (http://rsat.sb-roscoff.fr/random-seq_form.cgi). We measured the statistical signicance, by calculating the
p-value, as the number of times each of the predicted cross-reactive peptide matches with randomly generated sequences,
using following equation,
Here,
n
= number of times a cross-reactive peptide shares sequence identity > 30% and e-value < 1.0 with the set of 1000
random peptide sequences. P-value < 0.05 was considered as statistically signicant, to support alternative hypothesis that
cross-reactive peptide sequence is signicantly different from random sequences.
Declarations
Author contributions
FC and VH performed the
in-silico
analysis of pipeline-1, KB further processed and evaluated the data. AK performed the
in-
silico
analysis of pipeline-2. CS planned the study. All authors wrote or carefully reviewed sections of the manuscript
Data Statement
p
=
n
/1000

Page 10/13
All data generated or analysed during this study are included in this published article and its supplementary information
les.
Competing interests
For CS: Consultancy and research funding, Hycor Biomedical, Bencard Allergie and Thermo Fisher Scientic, Research
Funding, Mead Johnson Nutrition (MJN).
For KC: Dr. Nadeau reports grants from National Institute of Allergy and Infectious Diseases (NIAID), National Heart, Lung,
and Blood Institute (NHLBI), National Institute of Environmental Health Sciences (NIEHS), and Food Allergy Research &
Education (FARE), stock options from IgGenix, Seed Health, ClostraBio, and ImmuneID, is Director of the World Allergy
Organization Center of Excellence for Stanford, Advisor at Cour Pharma, Consultant for Excellergy, Red tree ventures, Eli Lilly,
and Phylaxis, Co-founder of Before Brands, Alladapt, Latitude, and IgGenix, and National Scientic Committee member at
Immune Tolerance Network (ITN), and National Institutes of Health (NIH) clinical research centers, outside the submitted
work, patents include, “Mixed allergen composition and methods for using the same,” “Granulocyte-based methods for
detecting and monitoring immune system disorders,” and “Methods and Assays for Detecting and Quantifying Pure
Subpopulations of White Blood Cells in Immune System Disorders.”
Acknowledgements
CS is supported by the Universities Giessen and Marburg Lung Center (UGMLC), the German Center for Lung Research (DZL),
University Hospital Giessen and Marburg (UKGM) research funding according to article 2, section 3 cooperation agreement,
and the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)-Project-ID 197785619 – SFB 1021, KFO
309 (P10), and SK 317/1-1 (Project number 428518790) as well as by the Foundation for Pathobiochemistry and Molecular
Diagnostics. This work was also supported by the NIAID AADCRC U19AI057229 (PI Davis), U19AI104209 (PI Galli),
U01AI140498 (PI Nadeau), NIEHS R21 ES033049 (PI Nadeau), Sean N. Parker Center for Allergy and Asthma Research at
Stanford University and Crown Foundation.
We would like to thank L. England, P. Nelson, M. Bruhn for their help with the
in-silico
pipeline-1 and associated scoring
system.
References
1. Lambrecht BN, Hammad H. The immunology of asthma. Nat Immunol 2015, 16(1):45–56.
2. Foong R-X, Du Toit G, Fox AT. Asthma, Food Allergy, and How They Relate to Each Other. Front Pediatr 2017, 5:89.
3. Cianferoni A. Non-IgE Mediated Food Allergy. Curr Pediatr Rev 2020, 16(2):95–105.
4. Labrosse R, Graham F, Caubet J-C. Non-IgE-Mediated Gastrointestinal Food Allergies in Children: An Update. Nutrients
2020, 12(7).
5. Hill DA, Grundmeier RW, Ram G, Spergel JM. The epidemiologic characteristics of healthcare provider-diagnosed
eczema, asthma, allergic rhinitis, and food allergy in children: a retrospective cohort study. BMC Pediatr 2016, 16:133.
. Emons JAM, van Gerth Wijk R. Food Allergy and Asthma: Is There a Link? Curr Treat Options Allergy 2018, 5(4):436–44.
7. Foong R-X, Du Toit G, Fox AT. Mini Review - Asthma and Food Allergy. Curr Pediatr Rev 2018, 14(3):164–70.
. Heymann PW, Carper HT, Murphy DD, Platts-Mills TAE, Patrie J, McLaughlin AP et al. Viral infections in relation to age,
atopy, and season of admission among children hospitalized for wheezing. J Allergy Clin Immunol 2004, 114(2):239–
47.
9. Cheung DS, Grayson MH. Role of viruses in the development of atopic disease in pediatric patients. Curr Allergy Asthma
Rep 2012, 12(6):613–20.

Page 11/13
10. Fecek RJ, Marcondes Rezende M, Busch R, Hassing I, Pieters R, Cuff CF. Enteric reovirus infection stimulates peanut-
specic IgG2a responses in a mouse food allergy model. Immunobiology 2010, 215(12):941–8.
11. Chen X, Leach D, Hunter DA, Sanfelippo D, Buell EJ, Zemple SJ et al. Characterization of intestinal dendritic cells in
murine norovirus infection. Open Immunol J 2011, 4:22–30.
12. Skevaki C, Hudemann C, Matrosovich M, Möbs C, Paul S, Wachtendorf A et al. Inuenza-derived peptides cross-react
with allergens and provide asthma protection. J Allergy Clin Immunol 2018, 142(3):804–14.
13. Montemurro A, Schuster V, Povlsen HR, Bentzen AK, Jurtz V, Chronister WD et al. NetTCR-2.0 enables accurate prediction
of TCR-peptide binding by using paired TCRα and β sequence data. Commun Biol 2021, 4.
14. Maurer-Stroh S, Krutz NL, Kern PS, Gunalan V, Nguyen MN, Limviphuvadh V et al. AllerCatPro-prediction of protein
allergenicity potential from the protein sequence. Bioinformatics 2019, 35(17):3020–7.
15. Jaravine V, Mösch A, Raffegerst S, Schendel DJ, Frishman D. Expitope 2.0: a tool to assess immunotherapeutic antigens
for their potential cross-reactivity against naturally expressed proteins in human tissues. BMC Cancer 2017, 17(1):892.
1. Qiu T, Yang Y, Qiu J, Huang Y, Xu T, Xiao H et al. CE-BLAST makes it possible to compute antigenic similarity for newly
emerging pathogens. Nat Commun 2018, 9(1):1772.
17. Hobohm U, Meyerhans A. A pattern search method for putative anchor residues in T cell epitopes. Eur J Immunol 1993,
23(6):1271–6.
1. Mari A, Scala E. Allergome: a unifying platform. Arb Paul Ehrlich Inst Bundesamt Sera Impfstoffe Frankf A M 2006,
(95):29-39, discussion 39-40.
19. Mari A, Rasi C, Palazzo P, Scala E. Allergen databases: current status and perspectives. Curr Allergy Asthma Rep 2009,
9(5):376–83.
20. Mari A, Scala E, Palazzo P, Ridol S, Zennaro D, Carabella G. Bioinformatics applied to allergy: allergen databases, from
collecting sequence information to data integration. The Allergome platform as a model. Cell Immunol 2006, 244(2):97–
100.
21. Balz K, Kaushik A, Chen M, Cemic F, Heger V, Renz H et al. Homologies between SARS-CoV-2 and allergen proteins may
direct T cell-mediated heterologous immune responses. Sci Rep 2021, 11(1):4792.
22. Peters B, Sette A. Generating quantitative models describing the sequence specicity of biological processes with the
stabilized matrix method. BMC Bioinformatics 2005, 6:132.
23. Buus S, Lauemøller SL, Worning P, Kesmir C, Frimurer T, Corbet S et al. Sensitive quantitative predictions of peptide-MHC
binding by a 'Query by Committee' articial neural network approach. Tissue Antigens 2003, 62(5):378–84.
24. Moutaftsi M, Peters B, Pasquetto V, Tscharke DC, Sidney J, Bui H-H et al. A consensus epitope prediction approach
identies the breadth of murine T(CD8+)-cell responses to vaccinia virus. Nat Biotechnol 2006, 24(7):817–9.
25. Nielsen M, Lund O. NN-align. An articial neural network-based alignment algorithm for MHC class II peptide binding
prediction. BMC Bioinformatics 2009, 10:296.
2. Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K et al. BLAST+: architecture and applications. BMC
Bioinformatics 2009, 10:421.
27. Nielsen M, Lundegaard C, Worning P, Lauemøller SL, Lamberth K, Buus S et al. Reliable prediction of T-cell epitopes
using neural networks with novel sequence representations. Protein Sci 2003, 12(5):1007–17.
2. Hoof I, Peters B, Sidney J, Pedersen LE, Sette A, Lund O et al. NetMHCpan, a method for MHC class I binding prediction
beyond humans. Immunogenetics 2009, 61(1):1–13.
29. Andreatta M, Karosiene E, Rasmussen M, Stryhn A, Buus S, Nielsen M. Accurate pan-specic prediction of peptide-MHC
class II binding anity with improved binding core identication. Immunogenetics 2015, 67(11-12):641–50.
30. Cock PJA, Antao T, Chang JT, Chapman BA, Cox CJ, Dalke A et al. Biopython: freely available Python tools for
computational molecular biology and bioinformatics. Bioinformatics 2009, 25(11):1422–3.

Page 12/13
31. Calis JJA, Maybeno M, Greenbaum JA, Weiskopf D, Silva AD de, Sette A et al. Properties of MHC Class I Presented
Peptides That Enhance Immunogenicity. PLoS Comput Biol 2013, 9(10).
32. Reynisson B, Alvarez B, Paul S, Peters B, Nielsen M. 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
Res 2020, 48(W1):W449-W454.
33. Welsh RM, Che JW, Brehm MA, Selin LK. Heterologous immunity between viruses. Immunol Rev 2010, 235(1):244–66.
34. Antó JM, Sunyer J, Rodriguez-Roisin R, Suarez-Cervera M, Vazquez L. Community outbreaks of asthma associated with
inhalation of soybean dust. Toxicoepidemiological Committee. N Engl J Med 1989, 320(17):1097–102.
Figures
Figure 1
Venn diagram depicting the molecular allergen components containing the allergen counterpart of the top 5 predicted
epitope pairs for HLA class I prediction with pipeline-1. The top 5 candidates were predicted with pipeline-1 and ranked based
on the described scoring system. Molecular allergen components contained in the predicted top 5 allergen epitopes for each
virus are depicted in the Venn diagram.
Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.
SupplementaryTable1.xlsx
SupplementaryTable2.xlsx
SupplementaryTable4new.xlsx

Page 13/13
Supplementaryinformationnal.pdf