No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Computational methods for prediction of T-cell epitopes - A framework for modelling, testing, and applications
Abstract
Computational models complement laboratory experimentation for efficient identification of MHC-binding peptides and T-cell epitopes. Methods for prediction of MHC-binding peptides include binding motifs, quantitative matrices, artificial neural networks, hidden Markov models, and molecular modelling. Models derived by these methods have been successfully used for prediction of T-cell epitopes in cancer, autoimmunity, infectious disease, and allergy. For maximum benefit, the use of computer models must be treated as experiments analogous to standard laboratory procedures and performed according to strict standards. This requires careful selection of data for model building, and adequate testing and validation. A range of web-based databases and MHC-binding prediction programs are available. Although some available prediction programs for particular MHC alleles have reasonable accuracy, there is no guarantee that all models produce good quality predictions. In this article, we present and discuss a framework for modelling, testing, and applications of computational methods used in predictions of T-cell epitopes.
... Previously, cholera, smallpox, bubonic plague and influenza are some of the most brutal killers in human history and killed millions of people all over the world. Nowadays, the world is scared of the coronavirus disease 2019 (COVID- 19), and its outbreak continues to spread from China to all over the world, and we do not yet know when it will stop. It is a contagious disease caused by a SARS family virus named 'severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)', which has not been previously identified in humans. ...
... At present, unfortunately, no vaccine or specific treatment is available. However, the WHO listed (as per 19 Moreover, there are no chemotherapeutic agents available to curb this menace; however, few agents are being used, including natural compounds [4][5][6], western medicines [7,8] and traditional Chinese medicines (TCN) [9,10], which may have potential efficacy against the SARS-CoV-2. Moreover, other drugs like interferon-α (IFN-α), lopinavir/ritonavir, chloroquine phosphate, ribavirin, favipiravir, disulfiram, arbidol and hydroxychloroquine are recommended as the tentative treatments for COVID-19 [11,12]. ...
... These methods had earlier been used in the development of vaccines against several diseases, including dengue [13], malaria [14], influenza [15], multiple sclerosis [16] and tumor [17]. However, this approach generally works through the identification of major histocompatibility complex (MHC)-1 and II molecules and thymus cells (T-cell) epitopes (CD8 + and CD4 + ) [18], which particularize the selection of the potential vaccine agents related to the transporter of antigen presentation (TAP) molecules [19,20]. ...
The recurrent and recent global outbreak of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has turned into a global concern which has infected more than 42 million people all over the globe, and this number is increasing in hours. Unfortunately, no vaccine or specific treatment is available, which makes it more deadly. A vaccine-informatics approach has shown significant breakthrough in peptide-based epitope mapping and opens the new horizon in vaccine development. In this study, we have identified a total of 15 antigenic peptides [including thymus cells (T-cells) and bone marrow or bursa-derived cells] in the surface glycoprotein (SG) of SARS-CoV-2 which is nontoxic and nonallergenic in nature, nonallergenic, highly antigenic and non-mutated in other SARS-CoV-2 virus strains. The population coverage analysis has found that cluster of differentiation 4 (CD4+) T-cell peptides showed higher cumulative population coverage over cluster of differentiation 8 (CD8+) peptides in the 16 different geographical regions of the world. We identified 12 peptides ((LTDEMIAQY, WTAGAAAYY, WMESEFRVY, IRASANLAA, FGAISSVLN, VKQLSSNFG, FAMQMAYRF, FGAGAALQI, YGFQPTNGVGYQ, LPDPSKPSKR, QTQTNSPRRARS and VITPGTNTSN) that are 80--90% identical with experimentally determined epitopes of SARS-CoV, and this will likely be beneficial for a quick progression of the vaccine design. Moreover, docking analysis suggested that the identified peptides are tightly bound in the groove of human leukocyte antigen molecules which can induce the T-cell response. Overall, this study allows us to determine potent peptide antigen targets in the SG on intuitive grounds, which opens up a new horizon in the coronavirus disease (COVID-19) research. However, this study needs experimental validation by in vitro and in vivo.
... Although many experimentally confirmed HLA ligands of ZEBOV have been reported, only a limited number of human T-cell epitopes are known [16]. Bioinformatics tools for predictions of HLA-binding peptides have been proven to minimize the cost and time for experimental T-cell epitope mapping [17]. These tools utilize a plethora of advanced algorithms for the prediction of HLA binding peptides [17][18][19][20], and allow the prediction for a wide range of HLA alleles. ...
... Bioinformatics tools for predictions of HLA-binding peptides have been proven to minimize the cost and time for experimental T-cell epitope mapping [17]. These tools utilize a plethora of advanced algorithms for the prediction of HLA binding peptides [17][18][19][20], and allow the prediction for a wide range of HLA alleles. Prediction in the context of HLA supertypes is offered by a number of the tools, such as Hotspot Hunter [21], MAPPP [22], MULTIPRED2 [23], PEPVAC [24], and NetMHC [25], among others. ...
... Colored in orange were the epitopes, while white were non-epitope sequences. Completely conserved epitope positions are not shown valid, relevant, and properly assessed for accuracy are useful for planning of complementary laboratory experiments [17,58]. The prediction system NetCTLpan, which was used herein to predict HLA-A2, -A3, and -B7 supertype-restricted epitopes has been trained and rigorously tested using experimentally known peptides [45]. ...
Background
Ebolavirus (EBOV) is responsible for one of the most fatal diseases encountered by mankind. Cellular T-cell responses have been implicated to be important in providing protection against the virus. Antigenic variation can result in viral escape from immune recognition. Mapping targets of immune responses among the sequence of viral proteins is, thus, an important first step towards understanding the immune responses to viral variants and can aid in the identification of vaccine targets. Herein, we performed a large-scale, proteome-wide mapping and diversity analyses of putative HLA supertype-restricted T-cell epitopes of Zaire ebolavirus (ZEBOV), the most pathogenic species among the EBOV family. Methods
All publicly available ZEBOV sequences (14,098) for each of the nine viral proteins were retrieved, removed of irrelevant and duplicate sequences, and aligned. The overall proteome diversity of the non-redundant sequences was studied by use of Shannon’s entropy. The sequences were predicted, by use of the NetCTLpan server, for HLA-A2, -A3, and -B7 supertype-restricted epitopes, which are relevant to African and other ethnicities and provide for large (~86%) population coverage. The predicted epitopes were mapped to the alignment of each protein for analyses of antigenic sequence diversity and relevance to structure and function. The putative epitopes were validated by comparison with experimentally confirmed epitopes. Results & discussionZEBOV proteome was generally conserved, with an average entropy of 0.16. The 185 HLA supertype-restricted T-cell epitopes predicted (82 (A2), 37 (A3) and 66 (B7)) mapped to 125 alignment positions and covered ~24% of the proteome length. Many of the epitopes showed a propensity to co-localize at select positions of the alignment. Thirty (30) of the mapped positions were completely conserved and may be attractive for vaccine design. The remaining (95) positions had one or more epitopes, with or without non-epitope variants. A significant number (24) of the putative epitopes matched reported experimentally validated HLA ligands/T-cell epitopes of A2, A3 and/or B7 supertype representative allele restrictions. The epitopes generally corresponded to functional motifs/domains and there was no correlation to localization on the protein 3D structure. These data and the epitope map provide important insights into the interaction between EBOV and the host immune system.
... The antigen processing channels (TAP) are required to present the peptide-MHC complex on the surface of the cell for immune response. Therefore, considering the c-terminal cleavage activity and TAP efficiency greatly help in the selection of effective vaccine candidates [24][25][26][27]. ...
... Detail analysis of the predicted epitopes such as percentile rank, MHC binding affinity, TAP, C-terminal cleavage activity, antigenicity and allergenic profiling was carried out to select the most promising epitopes as this criteria is experimentally validated for immune response potential [64][65][66]. The epitopes predicted in this study could become a clinical candidate sooner or later for the treatment of HPVs infection and cervical cancer, as epitopes such as KLPQLCTEL [18][19][20][21][22][23][24][25][26] and FAFRDLCIV 52-60 of E6 proteins, have been tested in a transgenic mice for IC 50 value which resulted in immune response in experimental conditions [67]. Previous studies already verified that peptide FAFRDLCIVYR 52-62 possess antitumor effect and is reported to be processed by T-cell endogenously [68,69]. ...
High-risk human papillomaviruses (hrHPVs) are the most prevalent viruses in human diseases including cervical cancers. Expression of E6 protein has already been reported in cervical cancer cases, excluding normal tissues. Continuous expression of E6 protein is making it ideal to develop therapeutic vaccines against hrHPVs infection and cervical cancer. Therefore, we carried out a meta-analysis of multiple hrHPVs to predict the most potential prophylactic peptide vaccines. In this study, immunoinformatics approach was employed to predict antigenic epitopes of hrHPVs E6 proteins restricted to 12 Human HLAs to aid the development of peptide vaccines against hrHPVs. Conformational B-cell and CTL epitopes were predicted for hrHPVs E6 proteins using ElliPro and NetCTL. The potential of the predicted peptides were tested and validated by using systems biology approach considering experimental concentration. We also investigated the binding interactions of the antigenic CTL epitopes by using docking. The stability of the resulting peptide-MHC I complexes was further studied by molecular dynamics simulations. The simulation results highlighted the regions from 46-62 and 65-76 that could be the first choice for the development of prophy-lactic peptide vaccines against hrHPVs. To overcome the worldwide distribution, the predicted epitopes restricted to different HLAs could cover most of the vaccination and would help to explore the possibility of these epitopes for adaptive immunotherapy against HPVs infections.
... The antigen processing channels (TAP) are required to present the peptide-MHC complex on the surface of the cell for immune response. Therefore, considering the c-terminal cleavage activity and TAP efficiency greatly help in the selection of effective vaccine candidates [24][25][26][27]. ...
... Detail analysis of the predicted epitopes such as percentile rank, MHC binding affinity, TAP, C-terminal cleavage activity, antigenicity and allergenic profiling was carried out to select the most promising epitopes as this criteria is experimentally validated for immune response potential [64][65][66]. The epitopes predicted in this study could become a clinical candidate sooner or later for the treatment of HPVs infection and cervical cancer, as epitopes such as KLPQLCTEL [18][19][20][21][22][23][24][25][26] and FAFRDLCIV 52-60 of E6 proteins, have been tested in a transgenic mice for IC 50 value which resulted in immune response in experimental conditions [67]. Previous studies already verified that peptide FAFRDLCIVYR 52-62 possess antitumor effect and is reported to be processed by T-cell endogenously [68,69]. ...
High-risk human papillomaviruses (hrHPVs) are the most prevalent viruses in human diseases including cervical cancers. Expression of E6 protein has already been reported in cervical cancer cases, excluding normal tissues. Continuous expression of E6 protein is making it ideal to develop therapeutic vaccines against hrHPVs infection and cervical cancer. Therefore, we carried out a meta-analysis of multiple hrHPVs to predict the most potential prophylactic peptide vaccines. In this study, immunoinformatics approach was employed to predict antigenic epitopes of hrHPVs E6 proteins restricted to 12 Human HLAs to aid the development of peptide vaccines against hrHPVs. Conformational B-cell and CTL epitopes were predicted for hrHPVs E6 proteins using ElliPro and NetCTL. The potential of the predicted peptides were tested and validated by using systems biology approach considering experimental concentration. We also investigated the binding interactions of the antigenic CTL epitopes by using docking. The stability of the resulting peptide-MHC I complexes was further studied by molecular dynamics simulations. The simulation results highlighted the regions from 46-62 and 65-76 that could be the first choice for the development of prophylactic peptide vaccines against hrHPVs. To overcome the worldwide distribution, the predicted epitopes restricted to different HLAs could cover most of the vaccination and would help to explore the possibility of these epitopes for adaptive immunotherapy against HPVs infections.
... However computational procedures and in silico analysis facilitate the prediction of epitopes due to reducing the number of candidate epitopes to be experimentally tested, still cannot replace experimental approaches and only can be a starting point [38] . Finally, the laboratory experimentation should be performed to validate the accuracy of the in silico prediction [30,34,39] . ...
... Finally, the laboratory experimentation should be performed to validate the accuracy of the in silico prediction [30,34,39] . The combination of computational and experimental procedures is a desirable method for efficient selection of epitopes [38,40] . ...
... However computational procedures and in silico analysis facilitate the prediction of epitopes due to reducing the number of candidate epitopes to be experimentally tested, still cannot replace experimental approaches and only can be a starting point [38] . Finally, the laboratory experimentation should be performed to validate the accuracy of the in silico prediction [30,34,39] . ...
... Finally, the laboratory experimentation should be performed to validate the accuracy of the in silico prediction [30,34,39] . The combination of computational and experimental procedures is a desirable method for efficient selection of epitopes [38,40] . ...
... Conventional approaches for T-cell epitope identification have depended entirely upon experimental technologies and experiences and are obviously time-consuming and costly. As a result, alternative computational approaches to implement antigen epitope identification have become powerful methods in immunology and vaccinology research and have significantly decreased the experimental load associated with epitope identification (Brusic et al., 2004;Zhang et al., 2012). To date, most T-cell epitope prediction tools have been developed using machine learning algorithms to train various experimental data, which are generally available in specialized epitope databases, such as the Immune Epitope Database (IEDB) (Vita et al., 2019). ...
Computational prediction to screen potential vaccine candidates has been proven to be a reliable way to provide guarantees for vaccine discovery in infectious diseases. As an important class of organisms causing infectious diseases, pathogenic eukaryotes (such as parasitic protozoans) have evolved the ability to colonize a wide range of hosts, including humans and animals; meanwhile, protective vaccines are urgently needed. Inspired by the immunological idea that pathogen-derived epitopes are able to mediate the CD8⁺ T-cell-related host adaptive immune response and with the available positive and negative CD8⁺ T-cell epitopes (TCEs), we proposed a novel predictor called CD8TCEI-EukPath to detect CD8⁺ TCEs of eukaryotic pathogens. Our method integrated multiple amino acid sequence-based hybrid features, employed a well-established feature selection technique, and eventually built an efficient machine learning classifier to differentiate CD8⁺ TCEs from non-CD8⁺ TCEs. Based on the feature selection results, 520 optimal hybrid features were used for modeling by utilizing the LightGBM algorithm. CD8TCEI-EukPath achieved impressive performance, with an accuracy of 79.255% in ten-fold cross-validation and an accuracy of 78.169% in the independent test. Collectively, CD8TCEI-EukPath will contribute to rapidly screening epitope-based vaccine candidates, particularly from large peptide-coding datasets. To conduct the prediction of CD8⁺ TCEs conveniently, an online web server is freely accessible (http://lab.malab.cn/∼hrs/CD8TCEI-EukPath/).
... These methods are time-consuming and costly. Therefore, various computational methods for discovering the antigen epitope have become a powerful tool in immunological and vaccination research, significantly reducing the research burden related with epitope prediction (Brusic et al. 2004;Zhang et al. 2012). Popular computational methods based on Machine learning (ML) and deep learning (DL) techniques have been used for the development of T-cell epitope prediction tools, which typically collect their experimental data from particular epitope repository such as Immune Epitope Database (IEDB) (Vita et al. 2019). ...
The healthcare sector is advancing with emerging technologies that help to detect new diseases or viruses worldwide. Generally, virus detection is based on various symptomatic tests and analysis of data samples that contain raw data (blood sample, protein sequence, etc.). Raw data cannot be directly processed by any machine learning (ML) technique. Therefore, feature extraction (FE) techniques are required to generate feature vectors. FE methods yield high dimensional data. Also, ML alone cannot provide accurate results due to large feature vectors. Thus, the feature selection (FS) technique is apt to solve this problem. The paper proposes a feature engineering module in which the FE module is developed and a new FS technique is proposed. The feature engineering module can help in the vaccine development process as accurate disease diagnosis is critical as any vaccine's success depends on precise disease detection. For the first time, this paper uses a metaheuristic method for FS in T-cell prediction. A new variant of the original fruit fly optimization algorithm (FFOA) is proposed in which the problem of local optima and slow convergence of FFOA is solved. The proposed variant of FFOA is named modified binary fruit fly optimization algorithm (MBFFOA) in which the mutation operator is updated. MBFFOA is first tested on benchmark functions and then tested on dengue and Zika virus datasets to predict CD4+ and CD8+ T-cells epitopes. The performance of MBFFOA is compared with that of competing metaheuristic algorithms. The performance of MBFFOA is statistically analyzed by Bonferroni–Dunn post-hoc test.
... B-cell epitopes are the antibody binding sites on antigens. T-cell epitopes bind well with MHC class-I molecules, and they have the potential to elicit a primary immunological response in the hosts [43]. B-cell epitopes and dengue-specific CD8+ and CD4+ lymphocytes are not only involved in pathogenesis and immunological research but also are main targets for vaccine and diagnostic reagent development against dengue virus [44,45]. ...
Dengue virus belonging to the family Flaviviridae and its four serotypes are responsible for dengue infections, which extend over 60 countries in tropical and subtropical areas of the world including Pakistan. During the ongoing dengue outbreak in Pakistan (2022), over 30,000 cases have been reported, and over 70 lives have been lost. The only commercialized vaccine against DENV, Dengvaxia, cannot be administered as a prophylactic measure to cure this infection due to various complications. Using machine learning and reverse vaccinology approaches, this study was designed to develop a tetravalent modified nucleotide mRNA vaccine using NS1, prM, and EIII sequences of dengue virus from Pakistani isolates. Based on high antigenicity, non-allergenicity, and toxicity profiling, B-cell epitope, cytotoxic T lymphocyte (CTL), and helper T lymphocyte (HTL) putative vaccine targets were predicted. Molecular docking confirmed favorable interactions between T-cell epitopes and their respective HLA alleles, while normal mode analysis validated high-affinity interactions of vaccine proteins with immune receptors. In silico immune simulations confirmed adequate immune responses to eliminate the antigen and generate memory. Codon optimization , physicochemical features, nucleotide modifications, and suitable vector availability further ensured better antigen expression and adaptive immune responses. We predict that this vaccine construct may prove to be a good vaccinal candidate against dengue virus in vitro as well.
... The development of AI algorithms for determining whether a peptide binds numerous HLA molecules is extremely important in making the design of vaccines more time-efficient. The proposed systems include systems based on hidden Markov models (HMMs), artificial neural networks (ANNs) (Brusic et al., [97]), and SVMs (Bozic et al., [98]). SVM has been used to predict antigens in an RV problem (Heinson et al., [79]). ...
The global COVID-19 (coronavirus disease 2019) pandemic, which was caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has resulted in a significant loss of human life around the world. The SARS-CoV-2 has caused significant problems to medical systems and healthcare facilities due to its unexpected global expansion. Despite all of the efforts, developing effective treatments, diagnostic techniques, and vaccinations for this unique virus is a top priority and takes a long time. However, the foremost step in vaccine development is to identify possible antigens for a vaccine. The traditional method was time taking, but after the breakthrough technology of reverse vaccinology (RV) was introduced in 2000, it drastically lowers the time needed to detect antigens ranging from 5-15 years to 1-2 years. The different RV tools work based on machine learning (ML) and artificial intelligence (AI). Models based on AI and ML have shown promising solutions in accelerating the discovery and optimization of new antivirals or effective vaccine candidates. In the present scenario, AI has been extensively used for drug and vaccine research against SARS-COV-2 therapy discovery. This is more useful for the identification of potential existing drugs with inhibitory human coronavirus by using different datasets. The AI tools and computational approaches have led to speedy research and the development of a vaccine to fight against the coronavirus. Therefore, this paper suggests the role of artificial intelligence in the field of clinical trials of vaccines and clinical practices using different tools.
... It is important for T-cell epitopes to be found in order for adaptive immunity to initiate. The NetCTL and Tepitool servers predicted 9-mer CTL and 15-mer HTL epitope that triggered adaptive immunity functions appropriately with MHC molecules [71]. So, finding epitopes that can be cooperated by MHC is very important in predicting which T-cell epitopes are potent [27]. ...
Kaposi sarcoma-associated herpesvirus (KSHV) is the etiologic agent of Kaposi Sarcoma (KS) and other two B-cell originated malignancies. Despite its familiarity as a direct carcinogen, still there is no permanent treatment or approved vaccine. This work intended to develop a multi-epitope vaccine aiming KSHV's key glycoproteins involved in viral entry. After applying rigorous immunoinformatics algorithms and numerous immunological filters, a unique vaccine containing multiple CTL, HTL, and BCL epitopes was created. Further, the putative vaccine's overall stability was demonstrated by three molecular dynamics simulations, along with a series of computational evaluations. Docking experiments revealed that the vaccination can make stable interactions with Toll-Like Receptor. Codon optimization and insertion into the cloning vector revealed that the newly designed vaccine candidate could be proficiently expressed in the E. coli system. Finally, an immune simulation was carried out to calibrate the vaccine's potency to trigger an immune response of the host.
... As stated in the previous section, Immunoinformatics tools aid in discovering the critical T-cell epitopes on the selected antigens that interact with the host T cells. These Immunoinformatics tools for epitope mapping utilize algorithms that exploit threading, non-linear functions, and neural networks [23][24][25][26][27][28]. These tools allow to scan of the sequences from protein targets and predict potential T-cell epitopes. ...
The host immune system recognizes and responds to the selective antigens or epitopes (immunome) of the intruding pathogen over an entire organism. The immune response so generated is ample to confer the desired immunity and protection to the host. This led to the conception of immunome-derived vaccines that exploit selective genome-derived antigens or epitopes from the pathogen's immunome and not its entire genome or proteome. These are designed to elicit the required immune response and confer protection against future invasions by the same pathogen. Immunoinformatics through its epitope mapping tools allows direct selection of antigens from a pathogen's genome or proteome, which is critical for the generation of an effective vaccine. This paved way for novel vaccine design strategies based on the mapped epitopes for translational applications that includes prophylactic, therapeutic, and personalized vaccines. In this chapter, various Immunoinformatics tools for epitope mapping are presented along with their applications. The methodology for immunoinformatics-assisted vaccine design is also outlined.
... Peptide vaccines are based on a chemical approach to synthesize identified epitope fragments that are highly immunogenic and can be tailored to produce the desired immunogenic responses [31,32]. However, identification of an immunogenic peptide can be highly complex by conventional means, to negate this problem bioinformatic prediction methodologies evolved, and were used in this work for the NP epitope design [33,32]. ...
... It is important for T-cell epitopes to be found in order for adaptive immunity to initiate. The NetCTL and Tepitool servers predicted 9mer CTL and 15-mer HTL epitope that triggered adaptive immunity functions appropriately with MHC molecules [67]. So, nding epitopes that can be cooperated by MHC is very important in predicting which T-cell epitopes are potent [26]. ...
The Kaposi sarcoma-associated herpesvirus (KSHV) is a virus that is identified as a direct carcinogen, causes Kaposi sarcoma, primary effusion lymphoma and multicentric Castleman disease. Despite this, there is no permanent treatment. This work intended to develop a multi-epitope vaccine aiming KSHV's key glycoproteins involved in viral entry. After applying rigorous immunoinformatics study and numerous immunological filters, the multi epitope vaccine was created, comprising of potent CTL, HTL and BCL epitope. A series of computational evaluations established the general dependability of the putative vaccine, and a molecular dynamics simulation established the vaccine's overall stability. Docking experiments revealed that the vaccination can make stable interactions with Toll-Like Receptors. Codon optimization and insertion into the cloning vector revealed that the vaccine could be expressed proficiently in the E. coli expression system. Finally, an immune simulation was done to assess the vaccine's potency to trigger an immune response.
... Fish pathogens Edwardsiella tarda and Flavobacterium columnare was successfully targeted by computer-aided epitopebased vaccine candidates (Mahendran et al. 2016). T-cell epitopes can induce primary immune response in the hosts as they have good interactional activity with MHC Class-I molecules (Brusic et al. 2004). In present study we Ramachandran plot for Epitope based vaccine (EBV) obtained from Molprobity server successfully found six epitopes by deploying RNA dependent RNA polymerase and Coat proteins of NNV (neural necrosis virus) that mostly affect Groupers (Epinephelus spp.), Gray mullet (Mugil cephalus), Rainbow trout (Onchorynchus mykiss) and Sea bass (Lates calcarifer) fish species. ...
Neural necrosis virus (NNV) of family Nodaviridae affect wide range of fish species with viral encephalopathy and retin-opathy causing mass mortality up to 100%. Currently there is no effective treatment and depopulation is only suggested recommendation. New avenues and approach are required to control this harmful malady. In this study we developed an epitope-based vaccine (EBV), against NNV using computation approach. We have selected two conserved proteins RNA-dependent RNA polymerase (RdRP) and capsid proteins. Based on more than ~ 1000 epitopes we selected six antigenic epitopes. These were conjugated to adjuvant and linker peptides to generate a full-length vaccine candidate. Biochemical structural properties were analyzed by Phyre2 server. ProtParam, Molprobity. Ramachandran plot results indicate that 98.7% residues are in a favorable region and 93.4% residues in the favored region. The engineered EBV binds to toll like receptor-5 (TLR5) an important elicitor of immune response. Further molecular docking by PatchDock server reveals the atomic contact energy (i.e. − 267.08) for the best docked model of EBV and TLR5 receptor. The molecular simulation results suggest a stable interaction; the RMSD and RMSF values are 1-4 Ǻ and 1-12Ǻ, respectively. Further we have suggested the best possible codon optimized sequence for its cloning and subsequent purification of the protein. Overall, this is a first report to suggest an in-silico method for generation of an EBV candidate against NNV. We surmise that the method and approach suggested could be used as a promising cure for NNVs. Keywords Nervous necrosis virus (NNV) · Epitope based vaccine · Peptide vaccine · Toll like receptor · Molecular simulation Abbreviations AI Artificial intelligence ANN Artificial neural networks HMM Hidden Markov model RMSD Root mean square deviation RMSF Root mean square fluctuation
... The development of DNA vaccines is a complicated process that needs molecular biology and immunology (34). In recent years, computational methods along with in silico studies compared to wet-lab methods have shown strong potential for the development of effective vaccines (35,36). As a crucial step in the preparation of vaccines and antibodies, we require the necessary data for the identification of B and T cell epitopes. ...
: Toxoplasmosis is a worldwide infection that can lead to serious problems in immune-compromised individuals and fetuses. A DNA vaccine strategy would be an ideal tool against Toxoplasma gondii. One of the necessary measures to provide an effective vaccine is the selection of proteins with high antigenicity. The SAG1-related sequence 3 (SRS3) protein is a major surface antigen in T. gondii that can be used as a vaccine candidate. In the present study, bioinformatics and computational methods were utilized to predict protein characteristics, as well as secondary and tertiary structures. The in silico approach is highly suited to analyze, design, and evaluate DNA vaccine strategies. Hence, in silico prediction was used to identify B and T cell epitopes and compare the antigenicity of SRS3 and other candidate genes of Toxoplasma previously applied in the production of vaccines. The results of the analysis theoretically showed that SRS3 has multiple epitopes with high antigenicity, proposing that SRS3 is a promising immunogenic candidate for the development of DNA vaccines against toxoplasmosis.
... B) Cytotoxic T cell antigenic epitopes including predicted epitopes that require antigen processing by nucleated somatic cells and presentation on class I MHC. The basis for predicting T-cell epitopes is that they can be presented on MHC molecules [12,48]. ...
Toxoplasmosis caused by an obligatory intracellular protozoan parasite of Toxoplasma gondii threats a wide spectrum of human and animal hosts. It has been shown that the intensity of the disease in humans depends on the host's immune responses. Immunological investigations on whole protein molecules of T. gondii have shown that these antigens are not fully responsible for the immune response, which leads to a decrease in specificity and affinity of the antigen (epitope)-antibody (paratope) binding. Currently, epitopes have shown promising entities to stimulate B, T, cytotoxic T lymphocyte, and NK cells resulting in enhancement of protective immunity against toxoplasmosis patients. Thus, the accurate designing, prediction, and conducting of antigenic epitopes of T. gondii (with linear and/or spatial structures (can augment our understanding about development of new serological diagnostic kits and vaccines. The current review provides an update on the latest advances of current epitopes described against toxoplasmosis including B cell/T cell epitopes, antigen types, parasite strains, epitope sequences, assay settings (in vitro and/or in vivo), and target strategy. Present results disclosed that the designing of effective multiepitopes of T. gondii by in silico modeling and immunoinformatics tools can strengthen our knowledge about triggering of epitope-based vaccine/diagnosis strategies in future perspectives.
... Molecular modelling utilizes detailed knowledge of the crystal structure of MHC molecules and of proteinpeptide interactions. 26,27 Molecular modelling provides a detailed insight into specific 3-D structures and interactions. Docking work can be extended to the prediction of peptide binding affinities using free energy scoring functions. ...
Background:
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative viral strain for the contagious pandemic respiratory illness in humans which is a public health emergency of international concern. There is a desperate need for vaccines and antiviral strategies to combat the rapid spread of SARS-CoV-2 infection.
Methods: The present study based on computational methods has identified novel conserved cytotoxic T-lymphocyte epitopes as well as linear and discontinuous B-cell epitopes on the SARS-CoV-2 spike (S) protein. The predicted MHC class I and class II binding peptides were further checked for their antigenic scores, allergenicity, toxicity, digesting enzymes and mutation.
Results: A total of fourteen linear B-cell epitopes where GQSKRVDFC displayed the highest antigenicity-score and sixteen highly antigenic 100% conserved T-cell epitopes including the most potential vaccine candidates MHC class-I peptide KIADYNYKL and MHC class-II peptide VVFLHVTYV were identified. Furthermore, the potential peptide QGFSALEPL with high antigenicity score attached to larger number of human leukocyte antigen alleles. Docking analyses of the allele HLA-B*5201 predicted to be immunogenic to several of the selected epitopes revealed that the peptides engaged in strong binding with the HLA-B*5201 allele.
Conclusions: Collectively, this research provides novel candidates for epitope-based peptide vaccine design against SARS-CoV-2 infection.
... Numerous computer algorithms have been developed for finding the location of linear epitopes restricted to MHC molecules and T-cell epitopes (Brusic et al. 2004;De Groot and Moise 2007;Zhang et al. 2008). For instance SYFPEI-THI (Rammensee et al. 1999) contains more than 7000 MHC-binding peptides. ...
Interferon beta (IFNβ) is naturally occurring cytokine made and secreted by immune cells in response to stimuli. Non-glycosylated interferon beta Ser 17 mutein (IFNβ-Ser17) is widely used for treatment of active relapsing multiple sclerosis. Despite all efforts to humanize this protein, it is still immunogenic by increasing of anti-IFNβ antibodies in patients. In order to decrease its antigenicity, identification and modification of epitopes is an effective step. We used a peptide microarray to detect linear epitopes by screening a synthetic peptide library. The interaction between synthetic peptides and anti-IFNβ antibodies presented in enriched plasma of transgenic and non-transgenic mice, were detected using chemiluminescence. With this technique, three 10-mer peptides were identified as linear epitopes. Computational algorithms were used to select residues suitable for point mutations in each epitope. These three 10-mer epitopes were mutated with the aim to reduce antigenicity of IFNβ. Three variants of IFNβ each with one mutated epitope were produced. The results showed no binding affinity of the mutated epitopes to anti-IFNβ antibodies compared with native epitopes. Biological activities of these epitope variants were measured; equal antiviral activity of the C-terminus mutated and the near N-terminus mutated were evaluated compare to the standard. Our results showed that the antigenicity of IFNβ epitope variants were reduced in vitro.
... Concerning MHC-peptide binding, several tools such as IEDB analysis resource (TepiTool) (http://tools.iedb.o rg/tepitool/) and ANN method have been developed to select T-cell epitopes [15][16][17]. Predicting antigenicity and allergenicity of T-cell epitope of the hypervariable regions was based on MHC-1 and MHC-II of the HAdV-D8 hexon by Vaxijen and Alegpred. The obtained results are presented in Tables 4 and 5. ...
Aim: In silico analysis of the hexon protein of human adenovirus serotype D-8 isolated from a patients with keratoconjunctivitis in Iran. Materials & methods: The hexon gene of HAdV-D8 was amplified by PCR. HAdV-D8 recovered from EKC outbreak was isolated by growing in A549 cells. Results: The hexon gene isolated from a patient with EKC comprised 2829 nt and 942 aa. The analyses of selected B-cell epitopes prediction (KTFQPEPQIGENNWQD) and T-cell epitopes prediction (TENFDIDLAFFDIPQ), showed high score immunogenicity, which may prove this to be a promising candidate for epitope vaccine development. Conclusion: In silico analysis of selected B-cell epitopes prediction (KTFQPEPQIGENNWQD) and T-cell epitopes prediction (TENFDIDLAFFDIPQ) are immunogenic and provoke B- and T-cell responses.
... MHC class I alleles exhibit broader binding specificity for peptides between 1,000 and 10,000 amino acids with high promiscuity. 86 Neural networks based on computational approaches have played a vital role in the prediction of the binding affinity of a peptide for specific MHC molecules. 87 In the present study, the FTFPHAFPF, CVSYWGVYY, LTAEVMSYI, and FSDPSIIEV antigenic epitopes were found to interact with more number of HLA alleles. ...
Background and aim: Human papillomavirus (HPV) is an oncogenic agent that causes over
90% of cases of cervical cancer in the world. Currently available prophylactic vaccines are
type specific and have less therapeutic efficiency. Therefore, we aimed to predict the potential
species-specific and therapeutic epitopes from the protein sequences of HPV45 by using different immunoinformatics tools.
Methods: Initially, we determined the antigenic potential of late (L1 and L2) and early (E1, E2,
E4, E5, E6, and E7) proteins. Then, major histocompatibility complex class I-restricted CD8 +
T-cell epitopes were selected based on their immunogenicity. In addition, epitope conservancy,
population coverage (PC), and target receptor-binding affinity of the immunogenic epitopes
were determined. Moreover, we predicted the possible CD8 + , nested interferon gamma (IFN-γ)-
producing CD4 + , and linear B-cell epitopes. Further, antigenicity, allergenicity, immunogenicity,
and system biology-based virtual pathway associated with cervical cancer were predicted to
confirm the therapeutic efficiency of overlapped epitopes.
Results: Twenty-seven immunogenic epitopes were found to exhibit cross-protection (≥55%)
against the 15 high-risk HPV strains (16, 18, 31, 33, 35, 39, 51, 52, 56, 58, 59, 68, 69, 73, and
82). The highest PC was observed in Europe (96.30%), North America (93.98%), West Indies
(90.34%), North Africa (90.14%), and East Asia (89.47%). Binding affinities of 79 docked
complexes observed as global energy ranged from -10.80 to -86.71 kcal/mol. In addition, CD8 +
epitope-overlapped segments in CD4 + and B-cell epitopes demonstrated that immunogenicity
and IFN-γ-producing efficiency ranged from 0.0483 to 0.5941 and 0.046 to 18, respectively.
Further, time core simulation revealed the overlapped epitopes involved in pRb, p53, COX-2,
NF-X1, and HPV45 infection signaling pathways.
Conclusion: Even though the results of this study need to be confirmed by further experimental peptide sensitization studies, the findings on immunogenic and IFN-γ-producing CD8 + and
overlapped epitopes provide new insights into HPV vaccine development.
... Targets of immune response pre-selected by computational analysis minimize the number of experiments required for validation [58]. Web servers/tools based on a number of algorithms are available for reliable prediction of promiscuous or HLA-supertype restricted peptides. ...
Background:
Viral vaccine target discovery requires understanding the diversity of both the virus and the human immune system. The readily available and rapidly growing pool of viral sequence data in the public domain enable the identification and characterization of immune targets relevant to adaptive immunity. A systematic bioinformatics approach is necessary to facilitate the analysis of such large datasets for selection of potential candidate vaccine targets.
Results:
This work describes a computational methodology to achieve this analysis, with data of dengue, West Nile, hepatitis A, HIV-1, and influenza A viruses as examples. Our methodology has been implemented as an analytical pipeline that brings significant advancement to the field of reverse vaccinology, enabling systematic screening of known sequence data in nature for identification of vaccine targets. This includes key steps (i) comprehensive and extensive collection of sequence data of viral proteomes (the virome), (ii) data cleaning, (iii) large-scale sequence alignments, (iv) peptide entropy analysis, (v) intra- and inter-species variation analysis of conserved sequences, including human homology analysis, and (vi) functional and immunological relevance analysis.
Conclusion:
These steps are combined into the pipeline ensuring that a more refined process, as compared to a simple evolutionary conservation analysis, will facilitate a better selection of vaccine targets and their prioritization for subsequent experimental validation.
... Epitopes or antigenic determinants, which represent the immune-active regions of antigen molecules, are the regions of an antigen, which are recognized by the immune system, specifically by antibodies and lymphocyte (B-cell or T-cell) surface antigen receptors. The properties of the antigen epitope, their number and their spatial configuration determine antigen specificity (42,43). Epitopes usually contain 6-8 amino acids residues and in general contain <20 amino acid residues. ...
House dust mite allergens can cause allergic diseases, including asthma, atopic dermatitis and rhinitis. Der f 20 is a novel allergen of Dermatophagoides farina (Der f), which is an arginine kinase. In the present study, the B‑cell and T‑cell epitopes of Der f 20 were predicted. The protein attribution, patterns, physicochemical properties and secondary structure of Der f 20 were also predicted. Der f 20 is a member of the ATP:guanido phosphotransferase family and contains a phosphagen kinase pattern. Using homology modeling, the present study constructed a reasonable tertiary structure of Der f 20. Using BcePred, ABCpred, BCPred and BPAP systems, B‑cell epitopes at 20‑25, 41‑49, 111‑118, 131‑141, 170‑174 and 312‑321 were predicted. Using NetMHCIIpan‑3.0 and NetMHCII‑2.2, T‑cell epitopes were predicted at 194‑202, 239‑247 and 274‑282. These results provide a theoretical basis for the design off Der f 20 epitope‑based vaccines.
... Many computational algorithms have been developed for predicting the binding of peptides to MHC molecules [20,21] including quantitative matrices [22,23], arti ficial neural netw orks [24], hidden-markov models [25] and molecular modelling [26,27]. These approaches could be used for prediction of antigenic epitopes. ...
Objective: To predict the immunogenic epitopes from human papillomavirus (HPV) virus using matrix based computational tools.Methods: In the present study, three matrix based algorithms, SYFPETHI, BIMAS and RANKPEP were used to predict the cytotoxic T lymphocyte (CTL) epitopes of HPV 16 and 18. The ability of the peptides to bind HLA A_0201, a most common allele, was evaluated using these algorithms. High scoring peptides were considered as potential binders.Results: Evaluation of HPV 16 proteome resulted in the prediction of 249 peptides as potential binders. Out of these only 25 peptides were predicted as binders by all three algorithms. Analysis of HPV 18 predicted 215 peptides, as potential binders. Among the 215 peptides only 20 peptides were predicted as binders by all three algorithms.Conclusion: The efficacy of these peptides in inducing a stronger immune response needs to be tested using in vitro and in vivo assays. The identified epitopes could be used in designing a novel epitope vaccine for HPV.
... Therefore, new approaches to optimize the screening process are needed. In this study we established an in silico framework by using CBS and IEDB prediction servers, which are usually employed to find MHC epitopes inside protein sequences [32]. Hereby, we propose an innovative use of this tools to screen mutational libraries of defined MHC-I epitopes. ...
... Therefore, new approaches to optimize the screening process are needed. In this study we established an in silico framework by using CBS and IEDB prediction servers, which are usually employed to find MHC epitopes inside protein sequences [32]. Hereby, we propose an innovative use of this tools to screen mutational libraries of defined MHC-I epitopes. ...
Tolerance towards tumor antigens, which are shared by normal tissues, have often limited the efficacy of cancer vaccines. However, wild type epitopes can be tweaked to activate cross-reactive T-cell clones, resulting in anti-tumor activity. The design of these analogues (i.e. heteroclitic peptides) can be difficult and time-consuming since no automated in silico tools are available. Hereby we describe the development of an in silico framework to improve the selection of heteroclitic peptides. The Epitope Discovery and Improvement System (EDIS) was first validated by studying the model antigen SIINFEKL. Based on ANN predictions, we selected two mutant analogues that are characterized by an increased MHC-I binding affinity (SIINFAKL) or increased TCR stimulation (SIIWFEKL). Therapeutic vaccination using optimized peptides resulted in enhanced anti-tumor activity and against B16.OVA melanomas in vivo. The translational potential of the EDIS platform was further demonstrated by studying the melanoma-associated antigen tyrosinase related protein 2 (TRP2). Following therapeutic immunization with the EDIS-derived epitope SVYDFFAWL, a significant reduction in the growth of established B16.F10 tumors was observed, suggesting a break in the tolerance towards the wild type epitope. Finally, we tested a multi vaccine approach, demonstrating that combination of wild type and mutant epitopes targeting both TRP2 and OVA antigens increases the anti-tumor response.
In conclusion, by taking advantage of available prediction servers and molecular dynamics simulations, we generated an innovative platform for studying the initial sequences and selecting lead candidates with improved immunological features. Taken together, EDIS is the first automated algorithm-driven platform to speed up the design of heteroclitic peptides that can be publicly queried online.
Routine vaccination has always been a center of controversy due to its lack of acceptability among the masses. Adverse event (AE) study relating to vaccine making is termed ‘adversomics,’ focusing on studying the adverse events associated with vaccination. As the technology improves, with the advancement of recombinant DNA technology, we get to see biology from the molecular perspective and manipulate it according to our wishes for betterment. The complete knowledge of the human genomic sequence now presents new unexploited fields. The advent of high-dimension assay technology and system biology presents a plethora of potential in developing vaccines with a newer approach that incorporates the recent discoveries about iRNA, transcriptome, immunogenetics, and many more. The major problem associated with complex diseases is low immunogenicity and difficulty to isolate antigens. For example, in the case of meningococcus B, an endless four decades of efforts using conventional methods of microbiology and biochemistry could not provide us with an effective and universal vaccine. It was later decided to obtain genomic sequencing of the serogroup B Neisseria meningitidis (MenB) and use this information for vaccine manufacturing. This was a novel approach denoted as “reverse vaccinology.”
Background
Epstein-Barr virus is commonly known as human herpes virus 4; an oncovirus belonging to the herpes virus family. The pathogen is extremely ubiquitous and infects more than 90% of population once in a lifetime.
Methods
The current study has employed a computational pipeline to develop a multiepitope vaccine design by targeting the most antigenic glycoproteins of the virus. The proteins were separately processed to retrieve B-cell and T-cell epitopes. The most suitable epitopes were scrutinized to design the peptide vaccine using appropriate linkers and adjuvants. The designed chimeric vaccines were further analyzed for their molecular interactions with TLR-4 and CD21 receptor. Consequently, the structural motion of the docked complexes was analyzed by molecular dynamics simulation approach followed by immune simulation.
Results
Our results showed promising outcomes in terms of vaccine antigenicity, population coverage and significantly lower free binding energies with potential receptors tested on 4 different docking platforms.
Conclusion
The conducted in silico study concludes that peptide vaccines could be a suitable alternative to traditional vaccinology approaches. Hence, our study will aid in the better formulation of vaccines in future by targeting the suitable drug or vaccine candidates.
Due to the growing importance of immunotherapy, especially in the treatment of cancer or designing personalized vaccines, there is a need to understand the mechanisms of the adaptive immune response. Based on the available data on the immunogenic response of T cells (CD8 +) to viral peptides presented on the molecules of the major histocompatibility complex (MHC) class I, models of predicting immunogenicity were developed using methods such as: Decision Tree, Support Vector Machine and Extreme Gradient Boosting. Models were compared and validated to choose the best method of predicting immunogenicity.KeywordsPrediction modelMachine learningImmunogenicityXGBoostSVMDecision tree
The article presents a description of the results of interdisciplinary cooperation. It is related to the implementation of research carried out in the form of Project Based Learning as a part of the project ‘Silesian University of Technology as the Centre for Modern Education based on research and innovation’. Employees and students of the Faculty of Biomedical Engineering and the Faculty of Architecture of the Silesian University of Technology have attempted to design some furniture for senior integrated with control systems located in the LeonardoLab—2014a room for testing technological solutions for the elderly.KeywordsElderly dependentsLeonardoLab
The ongoing pandemic of coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has globally devastated public health, the economies of many countries and quality of life universally. The recent emergence of immune-escaped variants and scenario of vaccinated individuals being infected has raised the global concerns about the effectiveness of the current available vaccines in transmission control and disease prevention. Given the high rate mutation of SARS-CoV-2, an efficacious vaccine targeting against multiple variants that contains virus-specific epitopes is desperately needed. An immunoinformatics approach is gaining traction in vaccine design and development due to the significant reduction in time and cost of immunogenicity studies and increasing reliability of the generated results. It can underpin the development of novel therapeutic methods and accelerate the design and production of peptide vaccines for infectious diseases. Structural proteins, particularly spike protein (S), along with other proteins have been studied intensively as promising coronavirus vaccine targets. Numbers of promising online immunological databases, tools and web servers have widely been employed for the design and development of next generation COVID-19 vaccines. This review highlights the role of immunoinformatics in identifying immunogenic peptides as potential vaccine targets, involving databases, and prediction and characterization of epitopes which can be harnessed for designing future coronavirus vaccines.
Marine birnavirus (MABV) has been the most pathogenic virus for marine and shellfish species across the world in terms of production and financial benefits. An outbreak of the disease may lead to entire mortality within a short period under standard culture conditions. The sequencing and characterization of several MABV strains have begun to reveal information regarding pathogen biology and pathogenicity. The major capsid protein VP2 and RNA-dependent RNA polymerase (RdRp), as well as polyprotein of Lates calcarifer birnavirus, was determined from several marine birnavirus (MABV) strains from different host or geographic origins. Despite the devastating complications, there is very limited prevention or control for the virus. In this regard, an immunoinformatics method was used to generate an epitope-based vaccine against this pathogen. The immunodominant T-cell epitopes were identified using the most antigenic and pathogenic proteins of MABV. The final constructed vaccine sequence was developed to be immunogenic, non-allergenic as well as have better solubility. Molecular dynamics simulation revealed significant binding stability and structural compactness. Finally, using Escherichia coli K12 as a model, codon optimization yielded ideal GC content and a higher CAI value, which was then included in the cloning vector pET2+ (a). Overall, our findings suggest that the suggested peptide vaccine might be a viable approach for MABV prevention.
In the context of the COVID-19 disease outbreak, organizations such as the universities are at risk of being essentially shut around the world if the overall condition does not improve. The other name for COVID-19 is a serious acute respiratory syndrome, a virus that causes serious respiratory problems. Corona virus-2 is a contagious agent spread through droplets in the air from an affected patient. This spreads easily by direct contact with affected patients or touching the objects which all already touched by the affected patients. Even if there are many vaccines available to defend against COVID-19 across the globe, still there is a high necessity to consider the precautions for avoiding infection. The major aspect for preventing the infection using a facemask that protects a person from entering the virus into the body through the nose and mouth of a person. The other major aspect for preventing the infection by washing hands using and washes or sanitizers. In the present article, the major and popular advanced technique used for image-based detection and classification is the Deep Learning-based VGG-16 technique. The deep learning technology is used in the analysis to identify face mask recognition and determine whether or not the individual is carrying a facemask. VGG-16 is the CNN (Convolutional Neural Network) framework is utilized for the present study. The Kaggle dataset considered consists of 25,000 images with each of the images having 225 × 225 pixels as the resolution, and the proposed model performed with a 96% accuracy. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
Machine learning and AI plays a crucial role in the recent advancements of multiple science disciplines. The medical field has seen new developments and achievements with the boost of technology. Breast cancer is one of the most dominant types of cancer in women. Recent studies indicate that there is an increase in the numbers globally. Many researchers started using the power of deep learning and created a model which helps doctors to diagnose and treat this cancer effectively. In this paper, we have used Kaggle dataset of histopathology images which contains 2,77,524 images. A deep learning CNN model is created and used with 80% training and 20% testing split. Without tuning hyperparameters, 61.01% accuracy is achieved. However, with parameters tuning 81% accuracy is achieved.
Many biological processes are governed by protein-ligand interactions. One such example is the recognition of self and non-self cells by the immune system. This immune response process is regulated by the major histocompatibility complex (MHC) protein which is encoded by the human leukocyte antigen (HLA) complex. Understanding the binding potential between MHC and peptides can lead to the design of more potent, peptide-based vaccines and immunotherapies for infectious autoimmune diseases.
We apply machine learning techniques from the natural language processing (NLP) domain to address the task of MHC-peptide binding prediction. More specifically, we introduce a new distributed representation of amino acids, name HLA-Vec, that can be used for a variety of downstream proteomic machine learning tasks. We then propose a deep convolutional neural network architecture, name HLA-CNN, for the task of HLA class I-peptide binding prediction. Experimental results show combining the new distributed representation with our HLA-CNN architecture acheives state-of-the-art results in the majority of the latest two Immune Epitope Database (IEDB) weekly automated benchmark datasets. We further apply our model to predict binding on the human genome and identify 15 genes with potential for self binding. Codes are available at https://github.com/uci-cbcl/HLA-bind .
Introduction and objectives
Allergen-specific immunotherapy (ASIT) is the only allergic disease-modifying therapy available for children and adults, and recombinant allergens are an interesting approach to improve allergy diagnosis and ASIT. Tyrophagus putrescentiae is a common storage mite that produces potent allergens. The aim of this study was to express and characterize recombinant group 4 allergen protein of T. putrescentiae (Tyr p 4), and to further investigate allergenicity and potential epitopes of Tyr p 4.
Materials and methods
The cDNA encoding Tyr p 4 was generated by RT-PCR and subcloned into pET-28a(+) plasmid. The plasmid was then transformed into E. coli cells for expression. After purification by nickel affinity chromatography and identification by SDS-PAGE, recombinant Tyr p 4 protein was used for a skin prick test and an ELISA to determine the allergic response.
Results
Study participants’ allergic response rate to Tyr p 4 protein was 13.3% (16/120). Eight B-cell epitopes and three T-cell epitopes of Tyr p 4 were predicted.
Conclusions
Similar to group 4 allergens of other species of mite, allergenicity of Tyr p 4 is weak. The expression, characterization and epitope prediction of recombinant Tyr p 4 protein provide a foundation for further study of this allergen in the diagnosis and ASIT of storage mite allergy.
Epstein–Barr virus (EBV), also known as human herpesvirus 4 (HHV-4), is a member of the Herpesviridae family and causes infectious mononucleosis, Burkitt’s lymphoma, and nasopharyngeal carcinoma. Even in the United States of America, the situation is alarming, as EBV affects 95% of the young population between 35 and 40 years of age. In this study, both linear and conformational B-cell epitopes as well as cytotoxic T-lymphocyte (CTL) epitopes were predicted by using the ElliPro and NetCTL.1.2 webservers for EBV proteins (GH, GL, GB, GN, GM, GP42 and GP350). Molecular modeling tools were used to predict the 3D coordinates of peptides, and these peptides were then docked against the MHC molecules to obtain peptide-MHC complexes. Studies of their postdocking interactions helped to select potential candidates for the development of peptide vaccines. Our results predicted a total of 58 T-cell epitopes of EBV. Of which the most potential were selected based on their TAP, MHC binding and C-terminal Cleavage score. The topmost peptides were subjected to MD simulation and stability analysis. Validation of our predicted epitopes using a 0.45 µM concentration was carried out by using a systems biology approach. Our results suggest a panel of epitopes that could be used to immunize populations to provide protection against the multiple diseases caused by EBV.
Development of an effective machine-learning model for T-cell Mycobacterium tuberculosis (M. tuberculosis) epitopes is beneficial for saving biologist’s time and effort for identifying epitope in a targeted antigen. Existing NetMHC 2.2, NetMHC 2.3, NetMHC 3.0 and NetMHC 4.0 estimate binding capacity of peptide. This is still a challenge for those servers to predict whether a given peptide is M. tuberculosis epitope or non-epitope. One of the servers, CTLpred, works in this category but it is limited to peptide length of 9-mers. Therefore, in this work direct method of predicting M. tuberculosis epitope or non-epitope has been proposed which also overcomes the limitations of above servers. The proposed method is able to work with variable length epitopes having size even greater than 9-mers. Identification of T-cell or B-cell epitopes in the targeted antigen is the main goal in designing epitope-based vaccine, immune-diagnostic tests and antibody production. Therefore, it is important to introduce a reliable system which may help in the diagnosis of M. tuberculosis. In the present study, computational intelligence methods are used to classify T-cell M. tuberculosis epitopes. The caret feature selection approach is used to find out the set of relevant features. The ensemble model is designed by combining three models and is used to predict M. tuberculosis epitopes of variable length (7–40-mers). The proposed ensemble model achieves 82.0% accuracy, 0.89 specificity, 0.77 sensitivity with repeated k-fold cross-validation having average accuracy of 80.61%. The proposed ensemble model has been validated and compared with NetMHC 2.3, NetMHC 4.0 servers and CTLpred T-cell prediction server.
T-cell epitopes are specific peptide sequences derived from foreign or own proteins that can initiate an immune response and which are recognized by specific T-cells when displayed on the surface of other cells. The prediction of T-cell epitopes is of particular interest in vaccine design, disease prevention and the development of immunotherapeutics. There are two principal categories of predictive methods: peptide-sequence based and peptide-structure-based. Sequence-based methods make use of various approaches to identify likely immunogenic amino acid sequences, such as sequence motifs, decision trees, partial least squares (PLS), quantitative matrices (QM), artificial neural networks (ANN), hidden Markov models (HMM), and support vector machines (SVM). Structure-based methods are more diverse in nature and involve approaches such as quantitative structure-activity relationships (QSAR), molecular modelling, molecular docking and molecular dynamics simulations (MD). This review highlights the key features of all of these approaches, provides some key examples of their application, and compares and contrasts the most important methods currently in use.
Lassa virus (LASV) is responsible for an acute viral hemorrhagic fever known as Lassa fever. Sequence analyses of LASV proteome identified the most immunogenic protein that led to predict both T-cell and B-cell epitopes and further target and binding site depiction could allow novel drug findings for drug discovery field against this virus. To induce both humoral and cell-mediated immunity peptide sequence SSNLYKGVY, conserved region 41–49 amino acids were found as the most potential B-cell and T-cell epitopes, respectively. The peptide sequence might intermingle with 17 HLA-I and 16 HLA-II molecules, also cover 49.15–96.82% population coverage within the common people of different countries where Lassa virus is endemic. To ensure the binding affinity to both HLA-I and HLA-II molecules were employed in docking simulation with suggested epitope sequence. Further the predicted 3D structure of the most immunogenic protein was analyzed to reveal out the binding site for the drug design against Lassa Virus. Herein, sequence analyses of proteome identified the most immunogenic protein that led to predict both T-cell and B-cell epitopes and further target and binding site depiction could allow novel drug findings for drug discovery field against this virus.
Hemophilia A is a rare bleeding disorder characterized by defective blood clotting due to diminished levels or absence of coagulation Factor VIII (FVIII). The preferred treatment option is FVIII replacement therapy. However, in 20–30% of the patients neutralizing (inhibitory) anti-FVIII antibodies develop rendering patients dependent on other treatment modalities such as the bypassing agent recombinant factor VIIa (rFVIIa). rFVIIa has a 20-year safety track record with no reports of immunogenicity in congenital hemophilia patients with inhibitors. To improve treatment efficacy of rFVIIa, the recombinant analog vatreptacog alpha was developed by Novo Nordisk A/S and taken into clinical development in 2006. Despite differing from rFVIIa by only three amino acid substitutions, results from the phase III trial demonstrated that some patients developed anti-drug antibodies. In this chapter, we give an introduction to hemophilia with focus on rFVIIa and the development of vatreptacog alfa. In addition, we summarize the findings from the clinical trials and characterization of the identified anti-drug antibodies. Finally, we show how various immunogenicity prediction tools have been used to investigate the immunogenicity risk of vatreptacog alfa leading to the identification of a potential new T-cell epitope that could contribute to the observed immunogenicity of the compound in humans.
Summary In this study we used TEPITOPE, a new epitope prediction software, to identify sequence seg- ments on the MAGE-3 protein with promiscuous binding to histocompatibility leukocyte an- tigen (HLA)-DR molecules. Synthetic peptides corresponding to the identified sequences were synthesized and used to propagate CD4 1 T cells from the blood of a healthy donor. CD4 1 T cells strongly recognized MAGE-3 281-295 and, to a lesser extent, MAGE-3 141-155 and MAGE- 3 146-160 . Moreover, CD4 1 T cells proliferated in the presence of recombinant MAGE-3 after processing and presentation by autologous antigen presenting cells, demonstrating that the MAGE-3 epitopes recognized are naturally processed. CD4 1 T cells, mostly of the T helper 1 type, showed specific lytic activity against HLA-DR11/MAGE-3-positive melanoma cells. Cold target inhibition experiments demonstrated indeed that the CD4 1 T cells recognized MAGE-3 281-295 in association with HLA-DR11 on melanoma cells. This is the first evidence that a tumor-specific shared antigen forms CD4 1 T cell epitopes. Furthermore, we validated the use of algorithms for the prediction of promiscuous CD4 1 T cell epitopes, thus opening the possibility of wide application to other tumor-associated antigens. These results have direct implications for cancer immunotherapy in the design of peptide-based vaccines with tumor- specific CD4 1 T cell epitopes.
The human Melan-A/MART-1 gene encodes an HLA-A2-restricted peptide epitope recognized by melanoma-reactive CD8(+) cytotoxic T lymphocytes. Here we report that this gene also encodes at least one HLA-DR4-presented peptide recognized by CD4(+) T cells. The Melan-A/MART-1(51-73) peptide was able to induce the in vitro expansion of specific CD4(+) T cells derived from normal DR4(+) donors or from DR4(+) patients with melanoma when pulsed onto autologous dendritic cells. CD4(+) responder T cells specifically produced IFN-gamma in response to, and also lysed, T2.DR4 cells pulsed with the Melan-A/MART-1(51-73) peptide and DR4(+) melanoma target cells naturally expressing the Melan-A/MART-1 gene product. Interestingly, CD4(+) T cell immunoreactivity against the Melan-A/MART-1(51-73) peptide typically coexisted with a high frequency of anti-Melan-A/MART-1(27-35) reactive CD8(+) T cells in freshly isolated blood harvested from HLA-A2(+)/DR4(+) patients with melanoma. Taken together, these data support the use of this Melan-A/MART-1 DR4-restricted melanoma epitope in future immunotherapeutic trials designed to generate, augment, and quantitate specific CD4(+) T cell responses against melanoma in vivo.
Prediction methods for identifying binding peptides could minimize the number of peptides required to be synthesized and assayed, and thereby facilitate the identification of potential T-cell epitopes. We developed a bioinformatic method for the prediction of peptide binding to MHC class II molecules.
Experimental binding data and expert knowledge of anchor positions and binding motifs were combined with an evolutionary algorithm (EA) and an artificial neural network (ANN): binding data extraction --> peptide alignment --> ANN training and classification . This method, termed PERUN, was implemented for the prediction of peptides that bind to HLA-DR4(B1*0401). The respective positive predictive values of PERUN predictions of high-, moderate-, low- and zero-affinity binders were assessed as 0.8, 0.7, 0.5 and 0.8 by cross-validation, and 1.0, 0.8, 0.3 and 0.7 by experimental binding. This illustrates the synergy between experimentation and computer modeling, and its application to the identification of potential immunotherapeutic peptides.
Software and data are available from the authors upon request.
vladimir@wehi.edu. au
A set of overlapping peptides corresponding to the L1, E6, and E7 proteins of human papilloma virus 16 was tested for their ability to bind to major histocompatibility complex class I molecules and to stimulate cytotoxic T-lymphocyte (CTL) responses in vitro. A class I binding assay using intact RMA-S cells showed that 20 of the 99 human papilloma virus peptides bound to H-2Kb and/or Db molecules. Fifteen of the 20 class I-binding peptides stimulated primary CTL responses, whereas peptides that were negative in the binding assay failed to do so. Peptide-induced CTLs recognized the immunizing peptide very efficiently, requiring no more than 1-10 nM peptide for target cell lysis. However, two observations were made that have important implications for the design of peptide-based vaccines for inducing CTLs. (i) Not all major histocompatibility complex-binding peptides that contained known motifs characteristic of naturally processed peptides induced CTLs. (ii) The efficiency of CTL lysis was strongly decreased when the size of the target peptide differed by only one amino acid residue from that of the immunizing peptide. We conclude that peptides chosen for vaccination must correspond in length to naturally processed peptides.
A new approach to rapid sequence comparison, basic local alignment search tool (BLAST), directly approximates alignments that optimize a measure of local similarity, the maximal segment pair (MSP) score. Recent mathematical results on the stochastic properties of MSP scores allow an analysis of the performance of this method as well as the statistical significance of alignments it generates. The basic algorithm is simple and robust; it can be implemented in a number of ways and applied in a variety of contexts including straightforward DNA and protein sequence database searches, motif searches, gene identification searches, and in the analysis of multiple regions of similarity in long DNA sequences. In addition to its flexibility and tractability to mathematical analysis, BLAST is an order of magnitude faster than existing sequence comparison tools of comparable sensitivity.
We have developed three computer programs for comparisons of protein and DNA sequences. They can be used to search sequence data bases, evaluate similarity scores, and identify periodic structures based on local sequence similarity. The FASTA program is a more sensitive derivative of the FASTP program, which can be used to search protein or DNA sequence data bases and can compare a protein sequence to a DNA sequence data base by translating the DNA data base as it is searched. FASTA includes an additional step in the calculation of the initial pairwise similarity score that allows multiple regions of similarity to be joined to increase the score of related sequences. The RDF2 program can be used to evaluate the significance of similarity scores using a shuffling method that preserves local sequence composition. The LFASTA program can display all the regions of local similarity between two sequences with scores greater than a threshold, using the same scoring parameters and a similar alignment algorithm; these local similarities can be displayed as a "graphic matrix" plot or as individual alignments. In addition, these programs have been generalized to allow comparison of DNA or protein sequences based on a variety of alternative scoring matrices.
We describe here a new method for predicting class II major histocompatibility complex-binding peptides, based on the preferences observed in a systematic series of peptide binding experiments where each position in a "minimal" peptide was replaced individually by every amino acid. The DRB1*0401 peptide binding preferences were determined and incorporated into a computer program that looks through sequences for potential epitopes and assigns each a score. These scores correlate well with previously determined T cell epitopes of foreign antigens and endogenous peptides from self proteins. Our findings hold implications for the design of subunit vaccines and in the identification of autoantigenic peptide regions within self proteins.
A method to predict the relative binding strengths of all possible nonapeptides to the MHC class I molecule HLA-A2 has been developed based on experimental peptide binding data. These data indicate that, for most peptides, each side-chain of the peptide contributes a certain amount to the stability of the HLA-A2 complex that is independent of the sequence of the peptide. To quantify these contributions, the binding data from a set of 154 peptides were combined together to generate a table containing 180 coefficients (20 amino acids x 9 positions), each of which represents the contribution of one particular amino acid residue at a specified position within the peptide to binding to HLA-A2. Eighty peptides formed stable HLA-A2 complexes, as assessed by measuring the rate of dissociation of beta 2m. The remaining 74 peptides formed complexes that had a half-life of beta 2m dissociation of less than 5 min at 37 degrees C, or did not bind to HLA-A2, and were included because they could be used to constrain the values of some of the coefficients. The "theoretical" binding stability (calculated by multiplying together the corresponding coefficients) matched the experimental binding stability to within a factor of 5. The coefficients were then used to calculate the theoretical binding stability for all the previously identified self or antigenic nonamer peptides known to bind to HLA-A2. The binding stability for all other nonamer peptides that could be generated from the proteins from which these peptides were derived was also predicted. In every case, the previously described HLA-A2 binding peptides were ranked in the top 2% of all possible nonamers for each source protein. Therefore, most biologically relevant nonamer peptides should be identifiable using the table of coefficients. We conclude that the side-chains of most nonamer peptides to the first approximation bind independently of one another to the HLA-A2 molecule.
MHCPEP (http://wehih.wehi.edu.au/mhcpep/) is a curated database comprising over 13 000 peptide sequences known to bind MHC molecules. Entries are compiled from published
reports as well as from direct submissions of experimental data. Each entry contains the peptide sequence, its MHC specificity
and where available, experimental method, observed activity, binding affinity, source protein and anchor positions, as well
as publication references. The present format of the database allows text string matching searches but can easily be converted
for use in conjunction with sequence analysis packages. The database can be accessed via Internet using WWW or FTP.
Computational methods were used to predict the sequences of peptides that bind to the MHC class I molecule, K(b). The rules for predicting binding sequences, which are limited, are based on preferences for certain amino acids in certain positions of the peptide. It is apparent though, that binding can be influenced by the amino acids in all of the positions of the peptide. An artificial neural network (ANN) has the ability to simultaneously analyze the influence of all of the amino acids of the peptide and thus may improve binding predictions. ANNs were compared to statistically analyzed peptides for their abilities to predict the sequences of K(b) binding peptides. ANN systems were trained on a library of binding and nonbinding peptide sequences from a phage display library. Statistical and ANN methods identified strong binding peptides with preferred amino acids. ANNs detected more subtle binding preferences, enabling them to predict medium binding peptides. The ability to predict class I MHC molecule binding peptides is useful for immunolological therapies involving cytotoxic-T cells.
Most pockets in the human leukocyte antigen-group DR (HLA-DR) groove are shaped by clusters of polymorphic residues and, thus, have distinct chemical and size characteristics in different HLA-DR alleles. Each HLA-DR pocket can be characterized by "pocket profiles," a quantitative representation of the interaction of all natural amino acid residues with a given pocket. In this report we demonstrate that pocket profiles are nearly independent of the remaining HLA-DR cleft. A small database of profiles was sufficient to generate a large number of HLA-DR matrices, representing the majority of human HLA-DR peptide-binding specificity. These virtual matrices were incorporated in software (TEPITOPE) capable of predicting promiscuous HLA class II ligands. This software, in combination with DNA microarray technology, has provided a new tool for the generation of comprehensive databases of candidate promiscuous T-cell epitopes in human disease tissues. First, DNA microarrays are used to reveal genes that are specifically expressed or upregulated in disease tissues. Second, the prediction software enables the scanning of these genes for promiscuous HLA-DR binding sites. In an example, we demonstrate that starting from nearly 20,000 genes, a database of candidate colon cancer-specific and promiscuous T-cell epitopes could be fully populated within a matter of days. Our approach has implications for the development of epitope-based vaccines.
Although atopic allergy affects </=20% of the total population, the relationship between the protein structure and immunogenic activity of the allergens is still largely unknown. We observed that group 5 grass allergens are characterized by repeated structural motifs. Using a new algorithm, TEPITOPE, we predicted promiscuous HLA-DR ligands within the repeated motifs of the Lol p5a allergen from rye grass. In vitro binding studies confirmed the promiscuous binding characteristics of these peptides. Moreover, most of the predicted ligands were novel T cell epitopes that were able to stimulate T cells from atopic patients. We generated a panel of Lol p5a-specific T cell clones, the majority of which recognized the peptides in a cross-reactive fashion. The computational prediction of DR ligands might thus allow the design of T cell epitopes with potential useful application in novel immunotherapy strategies.
Evidence has accumulated that the immune system can play a significant role in the defense against tumors in humans. Especially melanoma and renal cell carcinoma (RCC) are considered immunogenic tumors. In contrast to melanoma, hardly any RCC-associated antigens have been identified as targets for RCC-reactive T cells. Here, we report the identification of a human leukocyte antigen (HLA)-A2.1-restricted T-cell epitope within the G250 antigen. This antigen is expressed in 85% of RCCs but not by neighboring normal kidney tissue and has recently been molecularly defined and shown to be identical to MN/CA IX. Computer-aided motif prediction revealed the presence of 60 potential HLA-A2.1-binding peptides within the G250 antigen. Subsequent binding analysis showed that 13 of these peptides bound to HLA-A2.1 with high-to-intermediate affinity. Analysis of their immunogenicity in HLA-A2.1Kb transgenic mice indicated that 4 of the 13 peptides gave rise to cytotoxic T lymphocytes (CTLs) capable of lysing peptide-loaded target cells. However, only the G250 peptide 254-262 induced CTLs that recognized target cells that endogenously expressed the G250 antigen. Similarly, we were also able to raise human CTLs against the G250 peptide 254-262, which lysed target cells that endogenously expressed the G250 antigen. These findings and the high prevalence of this antigen in RCC patients makes G250 a potential target for anti-RCC immunotherapy.
The binding of a major histocompatibility complex (MHC) molecule to a peptide originating in an antigen is essential to recognizing antigens in immune systems, and it has proved to be important to use computers to predict the peptides that will bind to an MHC molecule. The purpose of this paper is twofold: First, we propose to apply supervised learning of hidden Markov models (HMMs) to this problem, which can surpass existing methods for the problem of predicting MHC-binding peptides. Second, we generate peptides that have high probabilities to bind to a certain MHC molecule, based on our proposed method using peptides binding to MHC molecules as a set of training data. From our experiments, in a type of cross-validation test, the discrimination accuracy of our supervised learning method is usually approximately 2–15% better than those of other methods, including backpropagation neural networks, which have been regarded as the most effective approach to this problem. Furthermore, using an HMM trained for HLA-A2, we present new peptide sequences that are provided with high binding probabilities by the HMM and that are thus expected to bind to HLA-A2 proteins. Peptide sequences not shown in this paper but with rather high binding probabilities can be obtained from the author (E-mail: mami@ccm.cl.nec.co.jp). Proteins 33:460–474, 1998. © 1998 Wiley-Liss, Inc.
We have developed three computer programs for comparisons of protein and DNA sequences. They can be used to search sequence data bases, evaluate similarity scores, and identify periodic structures based on local sequence similarity. The FASTA program is a more sensitive derivative of the FASTP program, which can be used to search protein or DNA sequence data bases and can compare a protein sequence to a DNA sequence data base by translating the DNA data base as it is searched. FASTA includes an additional step in the calculation of the initial pairwise similarity score that allows multiple regions of similarity to be joined to increase the score of related sequences. The RDF2 program can be used to evaluate the significance of similarity scores using a shuffling method that preserves local sequence composition. The LFASTA program can display all the regions of local similarity between two sequences with scores greater than a threshold, using the same scoring parameters and a similar alignment algorithm; these local similarities can be displayed as a "graphic matrix" plot or as individual alignments. In addition, these programs have been generalized to allow comparison of DNA or protein sequences based on a variety of alternative scoring matrices.
This article reviews the newly released JenPep database and two new powerful techniques for T-cell epitope prediction: (i) the additive method; and (ii) a 3D-Quantitative Structure Activity Relationships (3D-QSAR) method, based on Comparative Molecular Similarity Indices Analysis (CoMSIA). The JenPep database is a family of relational databases supporting the growing need of immunoinformaticians for quantitative data on peptide binding to major histocompatibility complexes and to the Transporters associated with Antigen Processing (TAP). It also contains an annotated list of T-cell epitopes. The database is available free via the Internet (http://www.jenner.ac.uk/JenPep). The additive prediction method is based on the assumption that the binding affinity of a peptide depends on the contributions from each amino acid as well as on the interactions between the adjacent and every second side-chain. In the 3D-QSAR approach, the influence of five physicochemical properties (steric bulk, electrostatic potential, local hydrophobicity, hydrogen-bond donor and hydrogen-bond acceptor abilities) on the affinity of peptides binding to MHC molecules were considered. Both methods were exemplified through their application to the well-studied problem of peptides binding to the human class I MHC molecule HLA-A*0201.
We describe here a new method for predicting class II major histocompatibility complex-binding peptides, based on the preferences observed in a systematic series of peptide binding experiments where each position in a "minimal" peptide was replaced individually by every amino acid. The DRB1*0401 peptide binding preferences were determined and incorporated into a computer program that looks through sequences for potential epitopes and assigns each a score. These scores correlate well with previously determined T cell epitopes of foreign antigens and endogenous peptides from self proteins. Our findings hold implications for the design of subunit vaccines and in the identification of autoantigenic peptide regions within self proteins.
Specific binding of antigenic peptides to major
histocompatibility complex (MHC) class I molecules is a
prerequisite for their recognition by cytotoxic T-cells.
Prediction of MHC-binding peptides must therefore be incorporated
in any predictive algorithm attempting to identify immunodominant
T-cell epitopes, based on the amino acid sequence of the
protein antigen. Development of predictive algorithms based
on experimental binding data requires experimental testing
of a very large number of peptides. A complementary approach
relies on the structural conservation observed in crystallographically
solved peptide-MHC complexes. By this approach, the peptide
structure in the MHC groove is used as a template upon
which peptide candidates are threaded, and their compatibility
to bind is evaluated by statistical pairwise potentials.
Our original algorithm based on this approach used the
pairwise potential table of Miyazawa and Jernigan (Miyazawa
S, Jernigan RL, 1996, J
Mol
Biol
256:623–644)
and succeeded to correctly identify good binders only for
MHC molecules with hydrophobic binding pockets, probably
because of the high emphasis of hydrophobic interactions
in this table. A recently developed pairwise potential
table by Betancourt and Thirumalai (Betancourt MR, Thirumalai
D, 1999, Protein
Sci
8:361–369) that is
based on the Miyazawa and Jernigan table describes the
hydrophilic interactions more appropriately. In this paper,
we demonstrate how the use of this table, together with
a new definition of MHC contact residues by which only
residues that contribute exclusively to sequence specific
binding are included, allows the development of an improved
algorithm that can be applied to a wide range of MHC class
I alleles.
Using panels of peptides well characterized for their ability to bind to HLA DR1, DRB1∗1101, or DRB1∗0401 molecules, algorithms were deduced to predict binding to these molecules. These algorithms consist of blocks of 8 amino acids containing an amino acid anchor (Tyr, Phe, Trp, Leu, Ile, or Val) at position i and different amino acid combinations at positions i+2 to i+7 depending on the class II molecule. The sensitivity (% of correctly predicted binder peptides) and specificity (% of correctly predicted non-binder peptides) of these algorithms, were tested against different independent panels of peptides and compared to other algorithms reported in the literature. Similarly, using a panel of 232 peptides able to bind to one or more HLA molecules as well as 43 non-binder peptides, we deduced a general motif for the prediction of binding to HLA-DR molecules. The sensitivity and specificity of this general motif was dependent on the threshold score used for the predictions. For a score of 0.1, the sensitivity and specificity were 84.7% and 69.8%, respectively. This motif was validated against several panels of binder and non-binder peptides reported in the literature, as well as against 35, 15-mer peptides from hepatitis C virus core protein, that were synthesized and tested in a binding assay against a panel of 19 HLA-DR molecules. The sensitivities and specificities against these panels of peptides were similar to those attained against the panels used to deduce the algorithm. These results show that comparison of binder and non-binder peptides, as well as correcting for the relative abundance of amino acids in proteins, is a useful approach to deduce performing algorithms to predict binding to HLA molecules.
Computer models can be combined with laboratory experiments for the efficient determination of (i) peptides that bind MHC molecules and (ii) T-cell epitopes. For maximum benefit, the use of computer models must be treated as experiments analogous to standard laboratory procedures. This requires the definition of standards and experimental protocols for model application. We describe the requirements for validation and assessment of computer models. The utility of combining accurate predictions with a limited number of laboratory experiments is illustrated by practical examples. These include the identification of T-cell epitopes from IDDM-, melanoma- and malaria-related antigens by combining computational and conventional laboratory assays. The success rate in determining antigenic peptides, each in the context of a specific HLA molecule, ranged from 27 to 71%, while the natural prevalence of MHC-binding peptides is 0.1–5%.
Peptides that induce and recall T-cell responses are called T-cell epitopes. T-cell epitopes may be useful in a subunit vaccine against malaria. Computer models that simulate peptide binding to MHC are useful for selecting candidate T-cell epitopes since they minimize the number of experiments required for their identification. We applied a combination of computational and immunological strategies to select candidate T-cell epitopes. A total of 86 experimental binding assays were performed in three rounds of identification of HLA-A11 binding peptides from the six pre-erythrocytic malaria antigens. Thirty-six peptides were experimentally confirmed as binders. We show that the cyclical refinement of the ANN models results in a significant improvement of the efficiency of identifying potential T-cell epitopes.
The peptide binding specificities of HLA-DRB1*0401, DRB1*0101, and DRB1*0701 have been analyzed by the use of large collections of synthetic peptides corresponding to naturally occurring sequences. The results demonstrated that nearly all peptides binding to these DR molecules bear a motif characterized by a large aromatic or hydrophobic residue in position 1 (Y, F, W, L, I, V, M) and a small, noncharged residue in position 6 (S, T, C, A, P, V, I, L, M). In addition, allele-specific secondary effects and secondary anchors were defined, and these parameters were utilized to derive allele-specific motifs and algorithms. By the combined use of such algorithms, peptides capable of degenerate DRB1*0101, DRB1*0401, and DRB1*0701 binding were identified. Additional experiments utilizing a panel of quantitative assays specific for nine additional common DR molecules identified a large set of DR molecules, which includes at least the DRB1*0101, DRB1*0401, DRB1*0701, DRB5*0101, DRB1*1501, DRB1*0901, and DRB1*1302 allelic products, characterized by overlapping peptide-binding repertoires. These results have implications for understanding the molecular interactions involved in peptide-DR binding, as well as the genetic and structural basis of MHC polymorphism. These results also have potential practical implications for the development of epitope-based prophylactic and therapeutic vaccines.
Epstein-Barr virus (EBV) nuclear antigen 1 (EBNA1) and latent membrane proteins (LMP) are the only antigens consistently expressed in malignancies such as nasopharyngeal carcinoma (NPC) and Hodgkin's disease (HD). Since EBNA1 is not recognized by EBV-specific cytotoxic T lymphocytes (CTL), there is increasing interest in the identification of the potential target epitopes within LMP1. Although LMP1-specific CTL have been isolated from seropositive individuals, earlier attempts to identify the peptide epitopes recognized by these T cells have been unsuccessful. In the present report we used a novel protocol to identify CTL epitopes within LMP1 which can be recognized by both polyclonal and clonal CTL. Firstly, a computer-based program was employed to identify the potential HLA-binding peptides within LMP1. Polyclonal CD8+ CTL were then isolated from seropositive donors that recognized the peptide epitopes YLLEMLWRL and YLQQNWWTL from LMP1 in association with HLA A2. Limiting dilution analysis of the memory CTL response revealed that the LMP1-specific CTL response constitutes a minor component of the CTL response in healthy virus carriers. Interestingly, analysis of YLLEMLWRL-specific CTL revealed that these CTL were able to lyse EBV-infected B cells expressing different HLA A2 supertype alleles including A*0201, A*0202, A*0203, A*0204, A*0206, A*6802 and A*6901. These data strongly support the notion that HLA class I supertype-restricted CTL may be of significant use in the development of peptide-based immunotherapeutics against EBV-associated malignancies in different ethnic populations.
The antigenic determinant recognized by a HLA-DPw4-restricted human T cell clone specific for rabies virus was identified by using a vaccinia-rabies nonstructural phosphoprotein recombinant virus and synthetic peptides of the sequence of rabies nonstructural Ag. These peptides were selected on the basis of three models that predict T cell epitopes. The antigenic determinant recognized by the rabies virus-specific T cell clone contained a five-amino acid segment highly homologous to a sequence found in a hepatitis B surface Ag epitope that stimulates human T cells in the context of the HLA-DPw4. A preliminary model of DPw4-restricted T cell determinants is elaborated based on a hypothesis of how the 2 alpha-helical peptides may bind to this MHC molecule. Results are further discussed in the context of the usefulness in identifying DPw4-restricted T cell epitopes for the production of synthetic vaccines because this MHC class II molecule is found with high frequency in the population.
We have previously experimentally analyzed the structural requirements for interaction between peptide antigens and mouse major histocompatibility complex (MHC) molecules of the d haplotype. We describe here two procedures devised to predict specifically the capacity of peptide molecules to interact with these MHC class II molecules (IAd and IEd). The accuracy of these procedures has been tested on a large panel of synthetic peptides of eukaryotic, prokaryotic, and viral origin, and also on a set of overlapping peptides encompassing the entire staphylococcal nuclease molecule. For both sets of peptides, IAd and IEd binding was successfully predicted in approximately 75% of the cases. This suggests that definition of such sequence "motifs" could be of general use in predicting potentially immunogenic peptide regions within proteins.
We have designed two computer-based algorithms for T cell epitope prediction, OptiMer and EpiMer, which incorporate current knowledge of MHC-binding motifs. OptiMer locates amphipathic segments of protein antigens with a high density of MHC-binding motifs. EpiMer identifies peptides with a high density of MHC-binding motifs alone. These algorithms exploit the striking tendency for MHC-binding motifs to cluster within short segments of each protein. Putative epitopes predicted by these algorithms contain motifs corresponding to many different MHC alleles, and may contain both class I and class II motifs, features thought to be ideal for the peptide components of synthetic subunit vaccines. In this study, we describe the use of OptiMer and EpiMer for the prediction of putative T cell epitopes from Mycobacterium tuberculosis and human immunodeficiency virus protein antigens, and demonstrate that these two algorithms may provide sensitive and efficient means for the prediction of promiscuous T cell epitopes that may be critical to the development of vaccines against these and other pathogens.
The development of T-cell therapy for the treatment of human malignancy has been hindered, in large part, by a lack of identifiable tumor antigens. Studies to identify potential T-cell targets in humans have been difficult because of practical problems limiting the use of in vivo immunization and a lack of reproducible in vitro priming methods. Oncogenic proteins are involved in malignant transformation and maintenance of the transformed phenotype and theoretically are potential targets to T-cell therapy. HER-2/neu protein is a protooncogene product overexpressed in a variety of human malignancies and is associated with malignant transformation and aggressive disease in human breast cancer. Previous studies have shown that some patients with breast cancer have existent helper/inducer T-cell immunity to p185HER-2/neu protein and peptides. The current study represents initial attempts to identify candidate cytotoxic T-lymphocyte (CTL) epitopes. Synthetic peptides were constructed identical to HER-2/neu protein segments with amino acid motifs similar to the published motif for HLA-2.1-binding peptides. Four peptides were synthesized and two were shown to be avid binders to HLA-A2.1. Two of the four peptides could be shown to elicit peptide-specific CTL by primary in vitro immunization in a culture system using peripheral blood lymphocytes from a normal individual homozygous for HLA-A2. p185HER-2/neu protooncogene protein contains immunogenic epitopes capable of generating human CD8+ CTL. The identification of candidate CTL epitopes will allow studies to determine whether some cancer patients have existent CTL immunity to HER-2/neu protein. The demonstrated ability to generate human peptide-specific CTL in vitro allows screening of other oncogenic proteins to identify candidate T-cell epitopes potentially useful for future immunotherapy studies.
We have measured the binding affinity for five HLA-A alleles: HLA-A1 (A*0101), A2.1 (A*0201), A3 (A*0301), A11 (A*1101), and A24 (A*2401); of a set of all possible nonamer peptides (n = 240) of human papillomavirus type 16 E6 and E7 proteins. High affinity binding peptides were identified for each of the alleles, thus allowing us to select several candidates for CTL-based vaccines. Moreover, this unbiased set of peptides allowed an evaluation of the predictive value of HLA motifs derived either from the analysis of sequencing of pools of naturally processed peptides or from the binding analysis of polyalanine nonameric peptides that differed in the amino acids (aa) present at the anchor positions. Whereas pool sequencing-derived motifs were present in only 27% of high affinity binders, the more expanded motif, based on analysis of different aa substitutions at the anchor positions, was present in 73% of high affinity binders. Furthermore, it was found that the presence of anchor residues in a peptide was in itself not sufficient to determine binding to MHC class I molecules, because the majority of motif-containing peptides failed to bind to the relevant MHC. Finally, specific HLA motifs were used to predict peptide binders of 8, 10, and 11 aa in length. Several high affinity binding peptides were identified for each of the various peptide lengths, indicating a significant size heterogeneity in peptides capable of high affinity binding to HLA-A molecules.
A quantitative peptide binding assay using purified DQA1*0301/DQB1*0301 (DQ3.1) molecules was developed and validated by examining the correlation between the data obtained in the binding assay with those obtained in inhibition of Ag presentation assays. By the combined use of large libraries of synthetic peptides and of substitution and truncation analogues, a putative DQ3.1 motif was defined. Its most prominent feature is the requirement for two small and/or hydrophobic residues spaced at positions i + 2 and i + 4. This motif is quite different from the motif recognized by DR molecules, but similar to the motif previously defined for certain IA alleles (the putative mouse homologue of DQ). These data suggest that various class II isotypes have evolved to present different peptide structures to each other, thus maximizing the repertoire of different epitopes available to T cell scrutiny.
The class I major histocompatibility complex-encoded HLA-B*2705 protein was simulated in complex with six different peptides exhibiting unexpected structure-activity relationships. Various structural and dynamical properties of the solvated protein-peptide complexes (atomic fluctuations, solvent-accessible surface areas, hydrogen bonding pattern) were found to be in qualitative agreement with the available binding data. Peptides that have been experimentally shown to bind to the protein remained tightly anchored to the MHC molecule, whereas nonbinders were significantly more weakly complexes to the protein and progressively dissociate from it at their N- and C-terminal ends. The molecular dynamics simulations emphasize the unexpectedly important role of secondary anchors (positions 1 and 3) in influencing the MHC-bound conformation of antigenic nonapeptides. Furthermore, it confirms that dominant anchor residues cannot solely account for peptide binding to a class I MHC molecule. The molecular dynamics method could be used as a complementary tool to T cell epitope predictions from the primary sequences of proteins of immunological interest. It is better suited to MHC proteins for which a crystal structure already exists. Furthermore, it may facilitate the engineering of T cell epitopes as well as the rational design of new MHC inhibitors designed to fit optimally the peptide binding cleft.
The binding of antigenic peptide sequences to major histocompatibility complex (MHC) molecules is a prerequisite for stimulation of cytotoxic T cell responses. Neural networks are here used to predict the binding capacity of polypeptides to MHC class I molecules encoded by the gene HLA-A*0201. Given a large database of 552 nonamers and 486 decamers and their known binding capacities, the neural networks achieve a predictive hit rate of 0.78 for classifying peptides which might induce an immune response (good or intermediate binders) vs. those which cannot (weak or non-binders). The neural nets also depict specific motifs for different binding capacities. This approach is in principle applicable to all MHC class I and II molecules, given a suitable set of known binding capacities. The trained networks can then be used to perform a systematic search through all pathogen or tumor antigen protein sequences for potential cytotoxic T lymphocyte epitopes.
The binding capacity of large sets of peptides corresponding to naturally occurring sequences and carrying previously defined A24-specific motifs was analyzed. It was found that only a minority (9-25%) of the motif-carrying peptides bound the relevant HLA-A molecule with good affinity (IC 50% < or = 50 nM), while the majority of them bound only weakly or not at all (IC 50% > or = 500 nM). By correlating the presence of specific residue types at each position along the peptide sequence with average binding affinity, the prominent influence of specific secondary interactions (secondary anchor residues) was revealed. Moreover, secondary interactions appeared to be size-dependent in that the specific effects detected differed in 9-mer and 10-mer peptide sets. Based on these observations, A24-specific refined motifs were also established for both 9-mer and 10-mer ligands, and their merit was verified by testing the binding capacity of independent sets of synthetic peptides. Such refined motifs should facilitate accurate prediction of potential A24-restricted peptide epitopes. It was also noted that certain crucial secondary interactions appear to be remarkably similar in the case of A24 and other HLA-A molecules previously analyzed (A*0201, A3, A11, and others). This may reflect contributions to binding affinity of relatively invariant residues located within the polymorphic pockets of the HLA binding groove.
We used the human processing defective cell line 174CEM.T2 (T2) to identify potential cytotoxic T lymphocyte (CTL) epitopes of human proteins. Exogenously added peptides can increase the number of properly folded HLA-A2.1 molecules on the cell surface of T2 cells, as shown by immunofluorescence measurements using the mouse monoclonal antibody BB7.2 (anti-HLA-A2.1) and fluorescein isothiocyanate-labeled goat anti-mouse F(ab')2 antibody. The peptides were selected on the basis of a computer score derived from the recently described HLA-A2.1 specific motif. Analysis of the influenza matrix protein showed that 15 out of 35 high-scoring peptides up-regulate the expression of HLA-A2.1 molecules on the T2 cell surface. The combination of the computer scoring program and an immunofluorescence-based peptide binding assay allows rapid detection of potential CTL target peptides.
Complexes of five peptides (from HIV-1, influenza A virus, HTLV-1, and hepatitis B virus proteins) bound to the human class I MHC molecule HLA-A2 have been studied by X-ray crystallography. While the peptide termini and their second and C-terminal anchor side chains are bound similarly in all five cases, the main chain and side chain conformations of each peptide are strikingly different in the center of the binding site, and these differences are accessible to direct TCR recognition. Each of the central peptide residues is seen to point up for some bound peptides, but down or sideways for others. Thus, although fixed at its ends, the structure of an MHC-bound peptide appears to be a highly complex function of its entire sequence, potentially sensitive to even small sequence differences. In contrast, MHC structural variation is relatively limited. These results offer a structural framework for understanding the role of nonanchor peptide side chains in both peptide-MHC binding affinity and TCR recognition.
Proteins can interact with short peptide sequences in a variety of ways that can be sequence dependent or independent. The bound peptides are frequently in an extended conformation but may also adopt beta-turns or alpha-helices as motifs for recognition. The peptides can be completely buried in cavities, bound in grooves or pockets, or form beta-strand type interactions at the protein surface. These various recognition motifs are illustrated by peptide interactions with antibodies, calmodulin, OppA periplasmic binding protein, PapD chaperone, MHC class I and class II molecules, and Src homology (SH) domains 2 and 3.
The structural requirements for the interaction between antigens and class I molecules was investigated through the use of a quantitative assay to measure peptide binding to different MHC class I alleles. We determined the permissiveness of the main anchors reported by Rammensee and his group for peptide binding and defined an extended motif for peptides binding to the HLA-A2.1 allele, including the role of non-anchor positions. It was found that the main anchors were necessary, but not sufficient, for good binding. Certain non-anchor positions contributed significantly to overall binding and were referred to a secondary anchors. This finding allowed a better prediction of high affinity binding peptides selected from libraries of different viral and tumor proteins. Furthermore, our data allowed correlation of the structural requirements for binding of peptides with crystallographic data of the MHC molecule. In order to characterize allele-specific motifs for a larger number of alleles, the HLA-A alleles A1, A3, A11, and A24, which represent some of the most common alleles found in different ethnic populations, were chosen. Here, most motifs were found to be highly exclusive; however, HLA-A3 and A11 shared a common motif. The defined motifs were validated further by using naturally processed peptides. Those peptides were also synthesized and tested for binding to the appropriate HLA alleles, giving a binding affinity from 0.3 to 200 nM for sequences of naturally processed peptides. Finally, a set of all possible 9-mer peptides from HPV 16 proteins were synthesized and tested for binding to the five class I alleles. For each allele, high affinity binders were identified, thus allowing for selection of possible peptide candidates for a CTL based vaccine.
MHC class II molecules assemble in the endoplasmic reticulum in a chaperone-mediated fashion to form a nine-chain structure consisting of three alpha beta dimers associated with an invariant chain trimer. This complex is transported through the Golgi apparatus and into the endosomal system. The signal for endosomal targeting resides in the cytoplasmic tail of the invariant chain. Current evidence argues that the segregation of the class II-invariant chain complex from the constitutive pathway of membrane protein transport occurs in the trans-Golgi network. However, class II-invariant chain complexes that reach the cell surface are also rapidly internalized into endosomes. Within the endosomal system, probably in a late endosome/pre-lysosome, the invariant chain is degraded, releasing alpha beta dimers that bind peptides predominantly derived from endocytosed proteins. Evidence suggests that many of these peptides are actually generated in lysosomes. The precise mechanisms involved in forming class II-peptide complexes are unclear, although the existence of antigen-processing mutants argues that additional gene products, at least one of which is encoded in the MHC, are involved. After binding peptides, class II molecules are transported by an unknown route to the cell surface, where their primary function of presenting antigenic peptides to CD4+ T cells is carried out.
Hydrogen bonding between conserved amino acids in the HLA DR and the peptide backbone of the ligand both provide the majority of free energy of binding and force the peptide ligands to adopt a similar extended conformation. Consequently the corresponding side chains of all peptides interact with similar pockets in the binding site. For peptides of a common length the contribution of the peptide backbone can be treated as a constant and the differential affinity can be viewed as a simple sum of the side chain interactions. These can be quantified by measuring the effects of each of the naturally occurring amino acids in the context of a simplified polyalanine backbone containing an aromatic amino acid to orient the peptide unequivocally in the binding site. The dataset of the relative contributions can be used to predict quantitatively the affinity of any peptide sequence.
The recent determination of the structure of a class II MHC molecule complexed to a specific peptide reveals both similarities and differences with peptide binding by class I MHC.
An influenza virus peptide binds to HLA-DR1 in an extended conformation with a pronounced twist. Thirty-five per cent of the peptide surface is accessible to solvent and potentially available for interaction with the antigen receptor on T cells. Pockets in the peptide-binding site accommodate five of the thirteen side chains of the bound peptide, and explain the peptide specificity of HLA-DR1. Twelve hydrogen bonds between conserved HLA-DR1 residues and the main chain of the peptide provide a universal mode of peptide binding, distinct from the strategy used by class I histocompatibility proteins.
MHC class I molecules are peptide receptors of stringent specificity which however still allow millions of different ligands. This is achieved by the following specificity characteristics summarized as allele specific peptide motifs: Peptides are of defined length, depending on the class I allele (either 8 or 9 residues; exceptions have been observed). Typically, 2 of the 8 or 9 positions are anchors that can only be occupied by a single amino acid residue, or by residues with closely related side chains. Location and characteristics of anchors vary with class I alleles. The C terminus of the peptide ligands is frequently an aliphatic or charged residue. Such allele-specific class I peptide ligand motifs, known so far for H-2Kd, Kb, Kk, Kkm1, Db, HLA-A*0201, A*0205, and B*2705, are useful to predict natural T cell epitopes. The latter can be determined by extraction from cells recognized by the T cell of interest. It is not known how the class I ligands are produced in the cell, although speculative models exist. The peptide specificity of class I molecules and experimental evidence indicate that T cells are tolerant to only a small fraction of the expressed genomic sequences and are not tolerant to the remainder. The function of class I molecules is to present a collection of self-peptide samples at the cell surface for surveillance by T cells.
Peptides that bind to major histocompatibility complex products (MHC) are known to exhibit certain sequence motifs which, though common, are neither necessary nor sufficient for binding: MHCs bind certain peptides that do not have the characteristic motifs and only about 30% of the peptides having the required motif, bind. In order to develop and test more accurate methods we measured the binding affinity of 463 nonamer peptides to HLA-A2.1. We describe two methods for predicting whether a given peptide will bind to an MHC and apply them to these peptides. One method is based on simulating a neural network and another, called the polynomial method, is based on statistical parameter estimation assuming independent binding of the side-chains of residues. We compare these methods with each other and with standard motif-based methods. The two methods are complementary, and both are superior to sequence motifs. The neural net is superior to simple motif searches in eliminating false positives. Its behavior can be coarsely tuned to the strength of binding desired and it is extendable in a straightforward fashion to other alleles. The polynomial method, on the other hand, has high sensitivity and is a superior method for eliminating false negatives. We discuss the validity of the independent binding assumption in such predictions.
Analysis of peptides derived from HLA class I molecules indicates that thousands of unique peptides are bound by a single molecular type, and sequence examination of the pooled constituents yields a motif which collectively defines the peptides bound by a given class I molecule. Motifs resulting from pooled sequencing are then used to infer whether particular viral and tumor protein fragments might serve as class I-presented peptide therapeutics. Still undetermined from a pooled motif is the breadth or range of peptides in the population which are brought together to form the pooled motif, and it is therefore not yet known how representative of the population a pooled motif is. By employing hollow fiber bioreactors for large-scale production of HLA class I molecules, sufficient peptides are produced to investigate individual subsets of peptides comprising a motif. Edman sequencing and mass spectrometric analysis of peptides eluted from HLA-B*1501 reveal that many peptide sequences fail to align with either the N- or C-terminal anchors predicted for the B*1501 peptide motif through whole pool sequencing. These analyses further reveal auxiliary anchors not previously detected and peptides significantly larger and smaller than the predicted nonamer, ranging from 6 to 12 amino acids in length. These results demonstrate that constituents of the B*1501 peptide pool vary markedly in comparison with one another and therefore in comparison with previously established B*1501 motifs, and such complexity indicates that many of the peptide ligands presented to CTL cannot be predicted using class I consensus motifs as search criteria.
Activation of T cells requires recognition by T-cell receptors of specific peptides bound to major histocompatibility complex (MHC) molecules on the surface of either antigen-presenting or target cells. These peptides, T-cell epitopes, have potential therapeutic applications, such as for use as vaccines. Their identification, however, usually requires that multiple overlapping synthetic peptides encompassing a protein antigen be assayed, which in humans, is limited by volume of donor blood. T-cell epitopes are a subset of peptides that bind to MHC molecules. We use an artificial neural network (ANN) model trained to predict peptides that bind to the MHC class II molecule HLA-DR4(*0401). Binding prediction facilitates identification of T-cell epitopes in tyrosine phosphatase IA-2, an autoantigen in DR4-associated type1 diabetes. Synthetic peptides encompassing IA-2 were tested experimentally for DR4 binding and T-cell proliferation in humans at risk for diabetes. ANN-based binding prediction was sensitive and specific, and reduced the number of peptides required for T-cell assay by more than half, with only a minor loss of epitopes. This strategy could expedite identification of candidate T-cell epitopes in diverse diseases.
This preliminary study was undertaken to identify new human leucocyte antigens (HLA) ligands from human immunodeficiency virus type 1 (HIV-1) which are highly conserved across HIV-1 clades and which may serve to induce cross-reactive cytotoxic T lymphocytes (CTLs). EpiMatrix was used to predict putative ligands from HIV-1 for HLA-A2 and HLA-B27. Twenty-six peptides that were both likely to bind and also highly conserved across HIV-1 strains in the Los Alamos HIV sequence database were selected for binding assays using the T2 stabilization assay. Two peptides that were also highly likely to bind (for A2 and B27, as determined by EpiMatrix) and well conserved across HIV-1 strains, and had previously been described to bind in the published literature, were also selected to serve as positive controls for the assays. Ten new major histocompatibility complex (MHC) ligands were identified among the 26 study peptides. The control peptides bound, as expected. These data confirm that EpiMatrix can be used to screen HIV-1 protein sequences for highly conserved regions that are likely to bind to MHC and may prove to be highly conserved HIV-1 CTL epitopes.
The binding of a major histocompatibility complex (MHC) molecule to a peptide originating in an antigen is essential to recognizing antigens in immune systems, and it has proved to be important to use computers to predict the peptides that will bind to an MHC molecule. The purpose of this paper is twofold: First, we propose to apply supervised learning of hidden Markov models (HMMs) to this problem, which can surpass existing methods for the problem of predicting MHC-binding peptides. Second, we generate peptides that have high probabilities to bind to a certain MHC molecule, based on our proposed method using peptides binding to MHC molecules as a set of training data. From our experiments, in a type of cross-validation test, the discrimination accuracy of our supervised learning method is usually approximately 2-15% better than those of other methods, including backpropagation neural networks, which have been regarded as the most effective approach to this problem. Furthermore, using an HMM trained for HLA-A2, we present new peptide sequences that are provided with high binding probabilities by the HMM and that are thus expected to bind to HLA-A2 proteins. Peptide sequences not shown in this paper but with rather high binding probabilities can be obtained from the author.
Background:
The binding of T-cell antigenic peptides to MHC molecules is a prerequisite for their immunogenicity. The ability to identify binding peptides based on the protein sequence is of great importance to the rational design of peptide vaccines. As the requirements for peptide binding cannot be fully explained by the peptide sequence per se, structural considerations should be taken into account and are expected to improve predictive algorithms. The first step in such an algorithm requires accurate and fast modeling of the peptide structure in the MHC-binding groove.
Results:
We have used 23 solved peptide-MHC class I complexes as a source of structural information in the development of a modeling algorithm. The peptide backbones and MHC structures were used as the templates for prediction. Sidechain conformations were built based on a rotamer library, using the 'dead end elimination' approach. A simple energy function selects the favorable combination of rotamers for a given sequence. It further selects the correct backbone structure from a limited library. The influence of different parameters on the prediction quality was assessed. With a specific rotamer library that incorporates information from the peptide sidechains in the solved complexes, the algorithm correctly identifies 85% (92%) of all (buried) sidechains and selects the correct backbones. Under cross-validation, 70% (78%) of all (buried) residues are correctly predicted and most of all backbones. The interaction between peptide sidechains has a negligible effect on the prediction quality.
Conclusions:
The structure of the peptide sidechains follows from the interactions with the MHC and the peptide backbone, as the prediction is hardly influenced by sidechain interactions. The proposed methodology was able to select the correct backbone from a limited set. The impairment in performance under cross-validation suggests that, currently, the specific rotamer library is not satisfactorily representative. The predictions might improve with an increase in the data.