key: cord-353877-wzndpcq3
authors: Albagi, Sahar Obi Abd; Al-Nour, Mosab Yahya; Elhag, Mustafa; Abdelihalim, Asaad Tageldein Idris; Haroun, Esraa Musa; Essa, Mohammed Elmujtba Adam; Abubaker, Mustafa; Hassan, Mohammed A.
title: A Multiple Peptides Vaccine against nCOVID-19 Designed from the Nucleocapsid phosphoprotein (N) and Spike Glycoprotein (S) via the Immunoinformatics Approach
date: 2020-05-20
journal: bioRxiv
DOI: 10.1101/2020.05.20.106351
sha: 
doc_id: 353877
cord_uid: wzndpcq3

Due to the current COVID-19 pandemic, the rapid discovery of a safe and effective vaccine is an essential issue, consequently, this study aims to predict potential COVID-19 peptide-based vaccine utilizing the Nucleocapsid phosphoprotein (N) and Spike Glycoprotein (S) via the Immunoinformatics approach. To achieve this goal, several Immune Epitope Database (IEDB) tools, molecular docking, and safety prediction servers were used. According to the results, The Spike peptide peptides SQCVNLTTRTQLPPAYTNSFTRGVY is predicted to have the highest binding affinity to the B-Cells. The Spike peptide FTISVTTEI has the highest binding affinity to the MHC I HLA-B1503 allele. The Nucleocapsid peptides KTFPPTEPK and RWYFYYLGTGPEAGL have the highest binding affinity to the MHC I HLA-A0202 allele and the three MHC II alleles HLA-DPA1*01:03/DPB1*02:01, HLA-DQA1*01:02/DQB1- *06:02, HLA-DRB1, respectively. Furthermore, those peptides were predicted as non-toxic and non-allergen. Therefore, the combination of those peptides is predicted to stimulate better immunological responses with respectable safety.

Human coronaviruses (HCoVs) including the Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2, COVID-19) are enveloped, positive-sense, single-stranded polyadenylated RNA viruses belong to the Coronaviridae family. They cause systemic and respiratory zoonotic diseases [1] .

The SARS-CoV-2 is a novel strain detected firstly in the city of Wuhan, the Republic of China in December 2019 [2] . It causes fever, cough, dyspnea, bilateral infiltrates on chest imaging and may progress to Pneumonia [3] . The COVID-19 is characterized by rapid spreading; "As the 27 Feb, it is reported in 47 countries, causing over 82,294 infections with 2,804 deaths" [4] and till the 5th May, more than 3. 5 million positive cases and 0.25 million deaths have been identified globally [5] . Unfortunately, until now COVID-19 has no effective antiviral drug for the treatment or vaccine for the prevention, hence extensive researches should be conducted on the development of safe and effective vaccines and antiviral drugs [4] .

To develop a safe and effective COVID-19 vaccine rapidly, the WHO recommended that "we must test all candidate vaccines until they fail to ensure that all of them have the chance of being tested at the initial stage of development". Ensuing this point, recently, there are over 120 proposed vaccines. Six of them in the clinical evaluation and 70 in pre-clinical evaluation [6] . The vaccine development is achieved by multiple approaches including the Inactivated, Live-attenuated, Non-replicating viral vector, DNA, RNA, Recombinant proteins, and Peptide-based vaccines. "As of 8 April 2020, the global COVID-19 vaccine R&D landscape confirmed 78 active candidate vaccine" [7] .

Consistent with global efforts, this study aims to predict potential COVID-19 Peptide-based vaccine utilizing the Nucleocapsid phosphoprotein (N) and Spike Glycoprotein (S) via the Immunoinformatics approach. Due to the respectable antigenicity of the Nucleocapsid and Spike Glycoprotein, they are appropriate targets for vaccine design [8] . The peptide vaccines are sufficient to stimulate cellular and humoral immunity without allergic responses [9] . They are" safe, simply produced, stable, reproducible, cost-effective" [10] , and permits a broad spectrum of immunity [9] , consequently, they are the targets for this study as well as they utilized in multiple studies concerning COVID-19 vaccine [11] [12] [13] [14] .

A total of 100 COVID-19 Nucleocapsid phosphoprotein (N) and Spike Glycoprotein (S) were retrieved from the National Center for Biotechnology Information (NCBI) database [15] as FASTA format in March 2020. The sequences with their accession number are listed in the Supplementary file S1.

The protein sequences of MHC I alleles HLA-A*02:01, HLA-B15:03, HLA-C*12:03, and MHC II alleles HLA-DPA1*01:03/DPB1*02:01, HLA-DQA1*01:02/DQB1*06:02, HLA-DRB1 were obtained from the Immuno Polymorphism Database (IPD-IMGT/HLA) [16] .

The retrieved COVID-19 Nucleocapsid phosphoprotein (N) and Spike Glycoprotein (S) sequences were aligned using the ClustalW algorithm [17] on the BioEdit software version 7.2.5 [18] to identify the conserved regions between sequences.

The B-Cells peptides were predicted from the conserved regions using the linear Epitope Prediction tool "BepiPred-test" on the Immune Epitope Database (IEDB) [19] at the default threshold value -0.500. To predict the epitopes accurately, a combination between the hidden (Parker and Levitt) method and the Markov model (HMM) [20] was used.

The Surface Accessibility of B-Cells Peptides was predicted via the Emini Surface Accessibility tool [21] on the IEDB [19] at the default threshold holding value.

To identify the antigenic sites within the Nucleocapsid phosphoprotein and Spike Glycoprotein, the Kolaskar and Tongaonker method's on the IEDB [19] at the default threshold value.

To predict the interaction with different MHC I alleles, the Major Histocompatibility Complex class I (MHC I) binding prediction tool on the IEDB [19] was used. All peptide length was set as 9amino acid. To predict the binding affinity, the Artificial Neural Network (ANN) prediction method was selected with a half-maximal inhibitory concentration (IC50) value of less than 100.

In contrast, to predict the interaction with different MHC II alleles, The Major Histocompatibility Complex class II (MHC II) binding prediction was used. To predict the binding affinity, the NN align algorithm was selected with an IC50 value of less than 500.

The Human allele reference sets (HLA DR, DP, and DQ) were included in the prediction.

To predict the percentage of peptides binding with various MHC I and MHC II alleles that cover the world population, the population coverage tool on the IEDB [19] was used.

To predict the peptides' allergicity, the AllergenFP v.1.0 [22] and AllerCatPro v. 1.7 [23] servers were used. In contrast, to predict the peptides' toxicity, the ToxinPred server [24] was used.

To model the 3D structure of the Nucleocapsid, Spike, and MHC molecules, the SWISS-MODEL server [25] , and the Phyre2 web portal [26] were used. To visualize the modeled structures, the USCF Chimera 1.8 software [27] was used.

The predicted peptides were docked with MHC I alleles HLA-A*02:01, HLA-B15:03, HLA-C*12:03, and MHC II alleles HLA-DPA1*01:03/DPB1*02:01, HLA-DQA1*01:02/DQB1*06:02, HLA-DRB1. The modeled MHC structures were prepared for the docking via Cresset Flare software [28] at the normal type calculation method. To dock the predicted peptides, the MDockPeP [29] and HPEPDOCK [30] [31] [32] [33] [34] servers were used. The predicted peptides were submitted as amino acid sequences. The 2D and 3D interactions were visualized using the PoseView [35] at the ProteinPlus web portal [36] and Cresset Flare viewer [28] , respectively.

According to the IEDB [19] prediction, the average binding score for the Nucleocapsid phosphoprotein and Spike Glycoprotein were 0.558, 0.470, respectively. All values equal to or greater than the default threshold were predicted as potential B-cell binders. Regarding the cytotoxic T-lymphocyte peptides, the MHC I binding prediction tool predicts 46 peptides from the Nucleocapsid and 192 peptides from the Spike Glycoprotein could interact with the different MHC I alleles. The most promising peptides were listed in Table 3 .

In contrast, the MHC II binding prediction tool predicts 774 peptides from the Nucleocapsid and 1111 peptides from the Spike Glycoprotein could interact with the different MHC II alleles. The most promising peptides were listed in Table 4 According to the MDockPeP [29] and HPEPDOCK [30] servers prediction, the spike peptide (FTISVTTEI) has the highest binding affinity to the MHC I HLA-B1503 allele and the spike peptide (MIAQYTSAL) has the highest binding affinity to the MHC I HLA-C1203 allele. The Nucleocapsid peptide (KTFPPTEPK) has the highest binding affinity to the MHC I HLA-A0202 allele ( Table 6) According to the AllergenFP v.1.0 [12] , AllerCatPro v. 1.7 [13] , and ToxinPred servers [14] , all the predicted peptides except the spike peptide (EVFNATRFASVYAWN) were Nonallergen and Non-Toxin (Tables 8and 9) . The Hydrophobicity scores of the Spike Glycoprotein peptides were not available. 

Due to the current COVID-19 pandemic, the rapid discovery of a safe and effective vaccine is an essential issue [37] .

Since the successful vaccine relies on the selection of the most antigenic parts and the best approaches [38] , COVID-19 Nucleocapsid phosphoprotein (N) and Spike Glycoprotein (S) were selected to design a peptide vaccine. The antigenicity of the Nucleocapsid and Spike is well predicted [8] , and the advantages of the peptide vaccines are well established [9, 10] .

The peptide design via the Immunoinformatics approach is achieved through multiple steps including the prediction of; B-Cells and T-cell Peptides, the surface accessibility, antigenic sites, and the Population coverage. After the selection of the candidate peptides, their interaction with the MHC molecules is simulated and their safety is predicted [39] .

Regarding the B-Cells peptides prediction, the successful candidates must pass the threshold scores in the Bepipred, Parker hydrophilicity, Kolaskar and Tongaonkar antigenicity, as well as Emini surface accessibility tests [40] . The IEDB Bepipred test [19] on the Nucleocapsid showed that eleven peptides were predicted, however, the peptide DAYKTFPPTEPKKDK-KKKADETQALPQRQKKQQTVTLLPAADLDD was the only one that passed all the tests. In contrast, the IEDB Bepipred test [18] on the Spike showed that forty-two peptides were predicted, but the peptides SQCVNLTTRTQLPPAYTNSFTRGVY and LGKY the only two that passed all the tests. As the length of effective B-cell peptides varies from 5-30 amino acids [41] , the peptide LGKY is too short and the peptide DAYKTFPPTEPKKDK-KKKADETQALPQRQKKQQTVTLLPAADLDD is too long. Consequently, the Spike peptide peptides SQCVNLTTRTQLPPAYTNSFTRGVY is predicted to have the highest binding affinity to the B-Cells (Table 2) .

Concerning the T-Cells peptides prediction, the test measures the peptides' binding affinity to the MHC molecules [19] . The available MHC I alleles HLA A, HLA B, HLA C, HLA E, and MHC II alleles HLA-DR, HLA-DQ, and HLA-DP were used. The MHC I IEDB tests [19] The results of collective The IEDB tests [19] revealed that the Spike glycoprotein peptides FTISVTTEI, MIAQYTSAL, and the Nucleocapsid peptide KTFPPTEPK are the most promised MHC1 peptides. On the other hand, the Spike peptides EVFNATRFASVYAWN, VFRSSVLHSTQDLFL, and the Nucleocapsid peptides AALALLLLDRLNQLE, ALALL-LLDRLNQLES, PRWYFYYLGTGPEAG, RWYFYYLGTGPEAGL are the most promised MHC II peptides.

To stimulate better immunological responses by the predicted peptides, they must interact and bind effectively with the MHC1 and MHC II molecules [42] , therefore we must study their interaction with the MHC molecules.

The simulation and prediction of the interaction between the predicted peptides and the MHC molecules are conducted using molecular docking studies that rely on the calculation of the binding free energy. The lowest binding energy scores of the MHC-Peptide complex will indicate the best interaction and the highest stability [43] .

To validate the results of molecular docking, MDockPeP [29] and HPEPDOCK [30] [31] [32] [33] [34] servers were used. The MDockPeP server predicts the MHC-Peptides interaction by "docking the peptides onto the whole surface of protein independently and flexibly using a novel the conformation restriction in its novel iterative approach. It ranks the docked Peptides via the ITScorePeP scoring function that uses the known protein-peptide complex structures in the calculations" [29] . In contrast, HPEPDOCK uses "a hierarchical flexible peptide docking approach" to predict the MHC-Peptides interaction [30] . The MHC I, HLA-A0202, HLA-B1503, HLA-C1203 was predicted to present the highly conserved SARS-CoV-2 peptides more effectively [44] , hence, they were used in molecular docking study.

The molecular docking results showed that the Spike peptide FTISVTTEI has the lowest docking energy score with the MHC I HLA-B1503 allele, hence it is predicted to have the highest binding affinity. The Spike peptide MIAQYTSAL showed the lowest docking energy score with the MHC I HLA-C1203, consequently, it is predicted to have the highest binding affinity to the MHC I HLA-C1203 allele. In contrast, regarding the MHC I HLA-A0202; the results of the MDockPeP [29] server showed that the Nucleocapsid peptide KTFPPTEPK has the lowest docking energy score, but the results of HPEPDOCK [30] server showed the Spike peptide MIAQYTSAL has the lowest docking energy score ( Table  6 ).

To illustrate the MHC-Peptide interaction, the PoseView [35] at the ProteinPlus web portal [36] that illustrate the 2D interactions and Cresset Flare viewer [8] that illustrate the 3D interaction were used.

The Spike peptide FTISVTTEI interacts with the MHC I HLA-B1503 allele by forming hydrogen bonds with the amino acids Thr2, Ser4, Ser101A, Asn104A, Tyr123A, Thr167A, Lys170A, Trp171A, Glu176A and hydrophobic bonds with the amino acids Thr2, Thr6, Trp98A, Ser101A, Trp171A, Ala174A, Leu180A.

The Spike peptide MIAQYTSAL interacts with the MHC I HLA-C1203 allele by forming hydrogen bonds with the amino acids Tyr33A, Arg86A, Lys90A, Gln179A, Thr187A and hydrophobic bonds with the amino acids Ile2, Gln4, Gln94A, Ala8, Arg86A, Tyr183A, Trp191A. In comparison, it interacts with the MHC I HLA-A0202 allele by forming hydrogen bonds with the amino acids Thr6, Glu87A, Arg121A, Trp180A, and hydrophobic bonds with the amino acids Ile2, Trp171A, Tyr183A, Trp191A. It forms six hydrogen bonds and seven hydrophobic bonds with the MHC I HLA-C1203 allele that is more than its bonds with the MHC I HLA-A0202 allele. This finding indicates the higher binding affinity of Spike peptide MIAQYTSAL to the MHC I HLA-C1203 allele and supports the MDockPeP [3] server score (Table 6 and Figures 5, 6 ).

The Nucleocapsid peptide KTFPPTEPK interacts with the MHC I HLA-A0202 allele by forming hydrogen bonds with the amino acids Pro4, Glu82A, Lys90A, Asp101A, Trp171A and hydrophobic bonds with the amino acids Pro4, Pro8, Leu105A, Thr167A, Trp171A, Ala174A, Val176A (Figures 5 and 6 ).

Among the reported MHC I alleles, the HLA-B1503 allele was predicted to have "the greatest ability to present the highly conserved SARS-CoV-2 peptides" [44] , therefore, the Spike peptide FTISVTTEI is predicted to make the highest response, since the binding with the MHC I stimulates the natural killer and the cytotoxic T cells [45] .

Regarding the interaction with the MHC II molecule, the Spike peptide EVFNATRFASVYAWN showed the lowest docking energy score with the three MHC II alleles HLA-DPA1*01:03/DPB1*02:01, HLA-DQA1*01:02/DQB1*06:02, and HLA-DRB1, hence it is predicted to have the highest binding affinity to the three alleles. Hence, it predicted to stimulate the CD4+ (helper) T cells more effectively, since the MHC II molecule presents the antigenic peptides to the CD4+ (helper) T cells [46] .

On the contrary, the Nucleocapsid peptide (RWYFYYLGTGPEAGL) showed lower docking energy scores with the MHC II allele HLA-DQA1*01:02/DQB1*06:02. The results of the MDockPeP [29] server showed that the peptide (PRWYFYYLGTGPEAG) has the lowest docking energy score, however, the results of HPEPDOCK [30] server differed from it in the MHC II alleles HLA-DPA1*01:03/DPB1*02:01 and HLA-DRB1 (Table 7) .

Spike peptide (EVFNATRFASVYAWN) interacts with HLA-DPA1*01:03/DPB1*02:01 allele by forming hydrogen bonds with the amino acids Val2, Ser10, Tyr12, Tyr40A, Glu59A, hydrophobic bonds with the amino acids Val2, Phe8, Tyr40A, Ala41A, Ala42A, Asn93A, Leu101A, Leu104A, Trp14, Tyr174A, and pi-pi bonds with the aromatic amino acid Phe8, Ser10, Tyr12, Trp14, Tyr174A. It interacts with the HLA-DRB1 allele by forming hydrogen bonds with the amino acids Val2, Thr6, Ser10, Phe46B, Tyr107B, Tyr152B, Gly154B, hydrophobic bonds with the amino acids Val2, Asn4, Thr6, Phe8, Tyr12, Trp14, Glu43B, Cys44B, Phe46B, Asn111B, Tyr112B, Pro153B, Trp182B, and pi-pi bonds with the aromatic amino acid Phe8, Tyr12, Trp14. Its interaction with the HLA-DQA1*01:02/DQB1*06:02 allele was not obtained from the server.

The Nucleocapsid peptide (RWYFYYLGTGPEAGL) interacts with HLA-DPA1*01:03/DPB1*02:01 allele by forming hydrogen bonds with the amino acids Glu12, Leu101A, hydrophobic bonds with the amino acids Trp2, Phe4, Tyr6, Gly10, Glu12, Gln45A, Leu97A, Leu101A, Leu104A, Phe144A, Pro146A, Pro170A, Tyr174A, and pi-pi bonds with the aromatic amino acid Trp2, Phe4, Tyr6, Phe144A, Tyr174A. It interacts with the HLA-DQA1*01:02/DQB1*06:02 allele by forming hydrogen bonds with the amino acids Glu12, Asn8A, Asn95A, Leu96A, Thr109A, Asn110A, hydrophobic bonds with the amino acids Trp2, Phe4, Tyr6, Gly10, Glu12, Gly14, Gln40A, Phe41A, Ala92A, Tyr103A, Thr106A, Ala107A, Ala108A, and pi-pi bonds with the aromatic amino acid Trp2, Phe4, Tyr6, Tyr103A. It interacts with the HLA-DRB1 allele by forming hydrogen bonds with the amino acids Glu12, Val40B, Cys44B, Tyr59B, Tyr107B, Asn111B, Gln178B, hydrophobic bonds with the amino acids Trp2, Phe4, Tyr6, Gly10, Glu12, Lys41B, His42B, Tyr107B, Asn111B, Tyr112B, Trp182B, and pi-pi bonds with the aromatic amino acid Trp2, Phe4, Tyr6. The Nucleocapsid peptide (RWYFYYLGTGPEAGL) interacts most effectively with the HLA-DQA1*01:02/DQB1*06:02 allele, hence it showed the highest binding affinity to it (Figures 7 and 8) .

Concerning the two Nucleocapsid peptides (RWYFYYLGTGPEAGL) and (PRWYFYYLGTGPEAG) in the interaction with the MHC II alleles HLA-DPA1*01:03/DPB1*02:01 and HLA-DRB1, the 2D and 3D interaction results showed that the peptide (PRWYFYYLGTGPEAG) more effectively with the HLA-DPA1*01:03/DPB1*02:01 allele and the peptide (RWYFYYLGTGPEAGL) more effectively with the HLA-DRB1 allele, however, they differ only in the first and last amino acid (Figures 9 and 10 ).

In comparison between the Spike and Nucleocapsid peptides, the Spike peptide (FTISVTTEI) showed a higher binding affinity to the MHC I HLA-B1503 allele. The Nucleocapsid peptides (KTFPPTEPK) and (RWYFYYLGTGPEAGL) showed a higher binding affinity to the MHC I HLA-A0202 allele and the three MHC II alleles HLA-DPA1*01:03/DPB1*02:01, HLA-DQA1*01:02/DQB1*06:02, HLA-DRB1, respectively, however, the total population coverage of the peptides FTISVTTEI and KTFPPTEPK is not high (Table 10) . Joshi A, et al. in their predictive COVID-19 Peptide-based vaccine study found that the ORF-7A protein's peptide (ITLCFTLKR) binds most effectively with the MHC I HLA alleles HLA-A*11:01, HLA-A*68:01 [13] . Enayatkhani M, et al. in their predictive study included Nucleocapsid N, but they studied its interaction with the MHC I HLA-A*11:01 allele [11] . Kalita P, et al. included the Nucleocapsid protein and Spike glycoprotein. They used the predicted Peptides from Nucleocapsid, Spike, and Membrane glycoprotein to design a subunit vaccine [12] . Singh A, et al. used the Nucleocapsid protein to design multi-Peptides vaccine [14] . Since we didn't include the two alleles HLA-A*11:01, HLA-A*68:01 in our study and didn't design subunit or multi-Peptides vaccine, the logical comparison will not be applied.

Besides the binding with the MHC molecules, the predicted peptide must be non-toxic and non-allergen, hence, their safety was predicted using the AllergenFP v.1.0 [22] , AllerCatPro [23] v. 1.7, ToxinPred [24] servers. The result showed that all the peptides were non-toxic. The AllerCatPro v. 1.7 [23] server results showed there is no evidence about the allergicity of all peptides, however, the AllergenFP v.1.0 [22] server predicts Spike peptide (EVFNATRFASVYAWN) as an allergen (Tables 8 and 9 ).

A potential COVID-19 Peptide-based vaccine was predicted from the Nucleocapsid phosphoprotein (N) and Spike Glycoprotein (S) via the Immunoinformatics approach. The Spike peptide peptides SQCVNLTTRTQLPPAYTNSFTRGVY is predicted to have the highest binding affinity to the B-Cells. The Spike peptide FTISVTTEI has the highest binding affinity to the MHC I HLA-B1503 allele. The Nucleocapsid peptides KTFPPTEPK and RWYFYYLGTGPEAGL have the highest binding affinity to the MHC I HLA-A0202 allele and the three MHC II alleles HLA-DPA1*01:03/DPB1*02:01, HLA-DQA1*01:02/DQB1-*06:02, HLA-DRB1, respectively. Furthermore, those peptides were predicted as non-toxic and non-allergen. Therefore, the combination of those peptides is predicted to stimulate better immunological responses.

Since the study is an in silico predictive work, further experimental studies are recommended to validate the obtained results.

A comparative sequence analysis to revise the current taxonomy of the family Coronaviridae

Novel Coronavirus ( 2019-nCoV) : situation report, 3

Coronavirus disease 2019 (COVID-19): Epidemiology, virology, clinical features, diagnosis, and prevention

SARS-CoV-2: an Emerging Coronavirus that Causes a Global Threat

COVID-19 situation update worldwide

WHO Solidarity Trial -Accelerating a safe and effective COVID-19 vaccine

The COVID-19 vaccine development landscape

Genomic characterization of the 2019 novel human-pathogenic coronavirus isolated from a patient with atypical pneumonia after visiting Wuhan

Peptide Vaccine: Progress and Challenges

Peptide-based vaccine successfully induces protective immunity against canine visceral leishmaniasis

Reverse vaccinology approach to design a novel multi-epitope vaccine candidate against COVID-19: an in silico study

Design of a Peptide-Based Subunit Vaccine against Novel Coronavirus SARS-CoV-2

Epitope based vaccine prediction for SARS-COV-2 by deploying immuno-informatics approach

Designing a multi-epitope peptide-based vaccine against SARS-CoV-2

Database resources of the national center for biotechnology information

IPD-IMGT/HLA Database

CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties, and weight matrix choice

BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT

The immune epitope database (IEDB) 3.0

Improved method for predicting linear Bcell epitopes

Induction of hepatitis A virusneutralizing antibody by a virus-specific synthetic peptide

AllergenFP: allergenicity prediction y descriptor fingerprints

AllerCatPro-prediction of protein allergenicity potential from the protein sequence

In silico approach for predicting toxicity of peptides and proteins

SWISS-MODEL: an automated protein homology-modeling server

The Phyre2 web portal for protein modeling, prediction, and analysis

UCSF Chimera--a visualization system for exploratory research and analysis

A. Molecular Field Extrema as Descriptors of Biological Activity: Definition and Validation

MDockPeP: An ab-initio protein-peptide docking server

HPEPDOCK: a web server for blind peptide-protein docking based on a hierarchical algorithm

HHblits: lightning-fast iterative protein sequence searching by HMM-HMM alignment

Improved tools for biological sequence comparison

Comparative Protein Structure Modeling of Genes and Genomes

The Protein Data Bank

PoseView--molecular interaction patterns at galance

ProteinsPlus: a web portal for structure analysis of macromolecules

Developing Covid-19 Vaccines at Pandemic Speed

Vaccines: From Empirical Development to Rational Design

Immunoinformatics Approach for Epitope-Based Peptide Vaccine Design and Active Site Prediction against Polyprotein of Emerging Oropouche Virus

Antibody Epitope Prediction -Tutorial

Present Yourself! By MHC Class I and MHC Class II Molecules

Relative binding free energy calculations in drug discovery: recent advances and practical considerations

Human leukocyte antigen susceptibility map for SARS-CoV-2," medRxiv

Major Histocompatibility Complex (MHC) Class I and MHC Class II Proteins: Conformational Plasticity in Antigen Presentation

Function and regulation of MHC class II molecules in T-lymphocytes: of mice and men

The authors declare that have no conflict of interest.