The non-structural protein sequences for NS3, NS4a, NS4b, NS5a and NS5b of HCV genotype-1 were retrieved from NCBI database https://www.ncbi.nlm.nih.gov/. To capture a comprehensive view of the protein variations, we downloaded approximately 250 sequences from diverse regions, ensuring a broad representation of genetic variability. The inclusion Criteria for selecting the target sequences were: Genotype specificity: Sequences must belong to HCV genotype 1, as verified through BLAST analysis https://blast.ncbi.nlm.nih.gov/Blast.cgi; Protein specificity: Only sequences corresponding to the non-structural proteins NS3, NS4a, NS4b, NS5a, and NS5b were included; Sequence completeness: Full-length protein sequences were selected to ensure comprehensive analysis. Geographic diversity: Sequences were chosen to represent a broad range of geographic regions to capture genetic variability; and lastly Quality and annotation: Sequences must be well-annotated with minimal ambiguous bases, ensuring high-quality data. The exclusion criteria included Incomplete Sequences: Partial or truncated sequences were excluded to avoid incomplete data representation; Non-genotype 1 sequences: Any sequence not verified as belonging to HCV genotype 1 through BLAST analysis was excluded.; Low-quality sequences: Sequences with poor annotation, high levels of ambiguity, or those lacking sufficient metadata were excluded; Redundant sequences: Duplicates or highly similar sequences were excluded to avoid redundancy and ensure a diverse dataset; and Recombinant strains: Sequences identified as potential recombinants were excluded to maintain focus on standard genotype 1 variations. Initially, the sequences underwent BLAST analysis to confirm their identity and similarity. Subsequently, the aligned sequences were subjected to multiple sequence alignment using the COBALT tool (https://www.ncbi.nlm.nih.gov/tools/cobalt/re_cobalt.cgi) (Papadopoulos and Agarwala, 2007). This alignment was essential for identifying conserved and variable regions within the proteins. The physicochemical properties of the proteins like theoretical pI, mol. wt, extinction coefficient, estimated half-life, instability index, and grand average of hydropathicity (GRAVY) were then assessed using the PROTPARAM tool (https://web.expasy.org/protparam/) (Gasteiger et al., 2005).

Initially, the amino acid sequences of each non-structural protein were analyzed using the TEPITOOL server (https://tools.iedb.org/tepitool/) by IEDB, one of the most advanced and up-to-date tools for epitope prediction (Paul et al., 2016). This tool enabled the identification of top-scoring epitopes based on their potential to elicit a CTL response. Following this, the top-scoring epitopes with percentile rank ≤ 1 were further evaluated for their efficient TAP transport and C-terminal proteasomal cleavage. This analysis was performed using the combined TAP transport and MHC class-I predictor tool provided by IEDB, which integrates these processes to predict the likelihood of an epitope being effectively presented on the surface of infected cells (Tenzer et al., 2005). TAP transport and proteasomal cleavage analysis are essential for evaluating the potential of epitopes as CTL targets. These processes are critical because proteasomal cleavage ensures that the epitope is processed and presented by the MHC, while TAP transport facilitates the transfer of these processed peptides into the endoplasmic reticulum for MHC loading. Both assessments are key to determining whether an epitope can be effectively recognized by CTLs.

The initially identified epitopes were further refined based on several criteria to ensure their suitability for vaccine development.

Epitopes were examined for conservancy using the IEDB conservancy analysis tool (Bui et al., 2007). Epitopes demonstrating greater than 90% conservancy were selected for further analysis. Targeting conserved regions within the non-structural proteins of HCV genotype-1 is crucial because these regions are less likely to undergo mutations, which minimizes immune evasion and ensures the effectiveness of the vaccine against various viral strains, including drug-resistant ones.

The VaxiJen v2.0 server was utilized to evaluate the immunogenicity of the selected epitopes, effectively distinguishing between non-antigenic and antigenic epitopes (Doytchinova and Flower, 2007). Non-immunogenic epitopes were discarded. This was carried out in view that most of the peptides are less immunogenic, thus it becomes important to screen the epitopes for their immunogenicity.

The remaining epitopes were evaluated for allergenicity using the AllerTOP v2.0 server (https://www.ddg-pharmfac.net/AllerTOP/method.html) (Dimitrov et al., 2014). Epitopes identified as allergenic were excluded.

Epitopes were also checked for the presence of glycosylation sites using the NetNGlyc 1.0 server (https://services.healthtech.dtu.dk/services/NetNGlyc-1.0/) (Gupta and Brunak, 2001). Epitopes located in glycosylated regions were omitted to avoid potential issues with epitope processing and presentation.

The final list comprised epitopes that met all the criteria: high predictive scores, efficient TAP transport, proteasomal cleavage, conservancy, immunogenicity, non-allergenicity, and location in non-glycosylated regions. These promising epitopes were selected for further analysis.

The physicochemical properties of the epitopes as discussed in the methodology section were determined using the ProtParam server. The secondary structure of the epitopes was also analyzed. This analysis provided insights into the alpha helices, beta strands, and coils within the epitope structures, which are crucial for understanding their stability and potential interactions.

The epitopes were further assessed for their affinity to the HLA-A*02:01 allele. To achieve this, the epitopes were modelled using the PEPFOLD 2.0 server (https://bioserv.rpbs.univ-paris-diderot.fr/services/PEP-FOLD/), which is a de novo tool for predicting the 3D structure of peptides between 9 and 36 amino acids in length (Shen et al., 2014). The secondary structural properties were also noted for all the epitopes as provided by the PEPFOLD server. PEPFOLD 2.0 uses Hidden Markov Models and performs simulations to generate the most representative conformations based on energy and population metrics. The predicted 3D structures of the epitopes were docked with the HLA-A*02:01 allele using the ClusPro server (Kozakov et al., 2017). ClusPro evaluates balanced docking energies, which include the combination of van der Waals interactions, electrostatics, and desolvation contributions, to predict the binding affinity of each epitope. The 3D model of HLA Class-I allele: HLA-A*02:01 was obtained from Protein Data Bank (PDB ID: 1QEW).

To evaluate the immunogenic potential of the epitopes, their binding affinity to Toll-like receptor-3 (TLR-3) was further analyzed. The structure of the TLR-3 receptor was obtained from the Protein Data Bank (PDB ID: 7C76). The selected epitopes were docked with the TLR-3 receptor to evaluate their binding interactions. This step was crucial to ensure that the epitopes not only bind effectively to the HLA-A*02:01 allele but also engage with the TLR-3 receptor, which is vital for the innate immune response.

The immune response elicited by the lead epitope candidates, particularly the cytotoxic T-cell response, was assessed through in-silico immune simulations utilizing an ImmSim server (Rapin et al., 2010). The immune simulations were conducted using advanced in-silico tools with over 100 simulation steps. This comprehensive simulation aimed to model the immune response dynamics over time. The primary focus was on evaluating the cytotoxic T-cell response, which is crucial for targeting and eliminating HCV-infected cells. These immune simulations helped predict the potential effectiveness of the selected epitopes in generating a robust and targeted immune response, thereby validating their suitability as candidates for an epitope-based therapeutic strategy against HCV genotype-1.

Approximately 250 protein sequences for each of the HCV non-structural proteins (NS3, NS4a, NS4b, NS5a, and NS5b) were retrieved in FASTA format from the NCBI database. These sequences represented a wide range of variability and geographical regions. The sequences were initially aligned using the protein BLAST tool from NCBI. The aligned sequences were then processed using the COBALT multiple sequence alignment tool. The COBALT alignment allowed us to visualize the sequences based on a frequency-based difference method. In this method, residues in a column are scored based on their representation in the column’s frequency profile. Rarely occurring residues are highlighted darkly, and columns with any degree of mismatch are also highlighted ( Supplementary Figure S1 ). The physicochemical properties of these proteins were analyzed using the ProtParam tool. These analyses provided insights into various parameters such as molecular weight, theoretical pI etc, which are summarized in Supplementary Table S1 . This comprehensive sequence alignment and physicochemical analysis laid the foundation for the subsequent identification and evaluation of potential epitopes.

Epitopes targeting the HLA-A*02:01 allele were predicted using the TEPITOOL server, specifically focusing on 9-mer epitopes. The top-scoring epitopes were selected based on a consensus percentile rank of ≤1, ensuring high predictive accuracy. We employed the IEDB-recommended method for epitope prediction, which utilizes the default selection of predictors. This approach leverages the best available methods for a given MHC molecule, including artificial neural networks (ANN), stabilizing membrane models (SMM), and combinatorial libraries (CombLib) whenever applicable. Following the initial prediction, the selected epitopes were further evaluated for their likelihood of undergoing proteasomal cleavage and TAP transport. For this analysis, the sequences were evaluated using the Combined Predictor tool from IEDB, which integrates Proteasomal Cleavage, TAP Transport, and MHC Class I assessments. This comprehensive tool combines predictors for proteasomal processing, TAP transport, and MHC binding to deliver an overall score that indicates each peptide’s potential as a T cell epitope. This dual approach effectively led to the identification of potential epitopes characterized by high percentile ranks and favorable conditions for proteasomal cleavage, thereby enhancing their suitability as candidates for further development in an epitope-based therapeutic strategy ( Table 1 ).

The identified epitopes using the above-mentioned epitope prediction tools underwent a rigorous screening process based on several immunological parameters. Initially, the epitopes were assessed for conservancy, with only those exhibiting greater than 90% conservancy selected for further evaluation. The conserved epitopes were then subjected to immunogenicity testing using the VaxiJen v2.0. Non-immunogenic epitopes (below threshold 0.4) identified in this analysis were eliminated from consideration. The immunogenic epitopes were further subjected to allergenicity testing. The allergenic epitopes identified were also excluded from the final selection. Lastly, the epitopes were screened for their presence in glycosylation sites. Epitopes located in glycosylated regions were also eliminated from the list. As a result of this comprehensive screening process, we identified a set of epitopes that met the following criteria: high percentile scores, likelihood of undergoing proteasomal cleavage, high conservancy, immunogenicity, non-allergenicity, and non-glycosylation. The final predictions included two epitopes from NS3, two from NS4a, two from NS4b, thirteen from NS5a, and eight from NS5b ( Table 1 ).

The top-scoring epitopes were subjected to molecular docking analysis with the HLA-A02:01 allele using the ClusPro server ( Figure 1 ). The 3D model of the HLA-A02:01 allele was extracted from the PDB, which includes a 9-mer melanoma-associated antigen-3 (MAA-3) complexed with the HLA-A*02:01 allele. Before docking, the retrieved 3D model underwent energy minimization and was prepared by removing the MAA-3 peptide, along with water molecules and other heteroatoms. All the epitopes were successfully docked in the active site of the HLA-A02:01 allele showing varying degrees of docking energy scores ( Figures 1 , 2 ). In addition, the epitopes were also docked with TLR-3 in order to examine the affinities of epitopes towards immune receptors. The docked energy scores and docking patterns were examined and noted ( Figures 1 , 3 ). Previously several studies have also utilized the ClusPro tool for studying molecular interactions between epitope and HLA molecules for vaccine development (Chauhan et al., 2021; Dubey et al., 2024). This analysis highlighted the potential of the epitopes to engage effectively with key immune components, supporting their candidacy for further development in an epitope-based therapeutic strategy. In addition, the secondary structural properties of the epitopes as provided by the Pepfold server were also examined ( Figure 4 ).

The top five epitopes with the highest affinities to the HLA-A*02:01 allele and TLR-3 i.e. epitopes E7, E9, E13, E14 and E27 underwent immune simulation analysis. Among these, epitopes E9, E13 and E27 demonstrated the strongest potential for activating cytotoxic T cells ( Figure 5 ). This analysis facilitated the identification of lead epitopes that possess high affinities to both HLA and TLR, as well as key immunological characteristics: they are conserved, non-allergenic, immunogenic, non-glycosylated, and capable of inducing a robust cytotoxic T cell response.

While this study presents promising findings, the in-silico predictions must be validated through experimental and clinical studies. The identified epitopes should undergo further validation in preclinical settings, such as in vitro and in vivo immune response assays. Additionally, continuous monitoring of HCV mutations remains essential to ensure the ongoing relevance of the selected epitopes against emerging strains. Future research should focus on integrating these findings with population-based studies to determine the efficacy of these epitopes in diverse HCV-infected populations.