vaccinology studies have thus incorporated the use of web-based
tools or open-source programs for the identification of epitopes in
the process of selection of potential vaccine candidates [16–
23]. These tools have been integrated to subtractive and reductionist analyses to identify antigenic and immunogenic MHC class I,
MHC class II and/or B-cell epitopes present in preselected candidates, potentially eliciting T- and/or B-cell mediated immune
responses. Data-driven and structure-based methods have been
developed for the identification of T-cell antigenic regions or epitopes in pathogen proteins, in most cases relying on the prediction
of peptide-MHC binding as a surrogate of TCR recognition
[24, 25]. Data-driven methods for peptide-MHC binding prediction use experimental binding data to infer peptide-sequence motifs
or profiles or to train advanced machine-learning algorithms. There
are also in silico tools for prediction of humoral immunity [24];
however, the identification of B-cell epitopes is still complex since
they can be conformational in nature and discontinuous. The
immune epitope database (IEDB) is a freely available, manually
curated resource that contains epitope-specific experimental assays
and hosts methodologically diverse tools for the prediction and
analysis of both B- and T-cell epitopes [26, 27].
Combination of the genome-based and immunoinformatic
methodology with structural and systems-biology approaches has
consolidated further the standing of reverse vaccinology as a rational vaccine design process. The final objective of structural vaccinology, beyond its contribution to antigen and epitope
identification, is the rational, structure-based optimization of vaccine candidates using three-dimensional data on the antigens, epitopes, and their interactions with immune-system components, as
well as on potential scaffolds for antigen engineering [16, 28–
31]. On the other hand, systems vaccinology exploits datasets
retrieved from systems-based studies to facilitate rational design of
safe and efficacious vaccines [32, 33]. This approach combines
results from functional omics (transcriptomics, proteomics, metabolomics, etc.) with data from preclinical and clinical studies and
bioinformatic analyses of large-scale protein–protein interactions or
evolutionary dynamics of antigens.
What makes an antigen a good vaccine candidate is, however, a
matter of fine balances. For example, although conservation is a
desirable property to ensure population coverage, highly conserved
proteins tend to be poor immunogens [34, 35]. In contrast, highly
immunogenic, immunodominant epitopes tend to be highly variable [36]. This is not unexpected, since pathogens are naturally
under strong selection within the host and proteins under positive
selection are therefore more likely to be involved in immunogenicity [37]. Currently available sequencing data and bioinformatics
tools enable the assessment of genetic variation of candidate antigens and their epitopes among large amounts of closely related
Bacterial Pan-Proteome-Based Antigen Discovery
45
tools or open-source programs for the identification of epitopes in
the process of selection of potential vaccine candidates [16–
23]. These tools have been integrated to subtractive and reductionist analyses to identify antigenic and immunogenic MHC class I,
MHC class II and/or B-cell epitopes present in preselected candidates, potentially eliciting T- and/or B-cell mediated immune
responses. Data-driven and structure-based methods have been
developed for the identification of T-cell antigenic regions or epitopes in pathogen proteins, in most cases relying on the prediction
of peptide-MHC binding as a surrogate of TCR recognition
[24, 25]. Data-driven methods for peptide-MHC binding prediction use experimental binding data to infer peptide-sequence motifs
or profiles or to train advanced machine-learning algorithms. There
are also in silico tools for prediction of humoral immunity [24];
however, the identification of B-cell epitopes is still complex since
they can be conformational in nature and discontinuous. The
immune epitope database (IEDB) is a freely available, manually
curated resource that contains epitope-specific experimental assays
and hosts methodologically diverse tools for the prediction and
analysis of both B- and T-cell epitopes [26, 27].
Combination of the genome-based and immunoinformatic
methodology with structural and systems-biology approaches has
consolidated further the standing of reverse vaccinology as a rational vaccine design process. The final objective of structural vaccinology, beyond its contribution to antigen and epitope
identification, is the rational, structure-based optimization of vaccine candidates using three-dimensional data on the antigens, epitopes, and their interactions with immune-system components, as
well as on potential scaffolds for antigen engineering [16, 28–
31]. On the other hand, systems vaccinology exploits datasets
retrieved from systems-based studies to facilitate rational design of
safe and efficacious vaccines [32, 33]. This approach combines
results from functional omics (transcriptomics, proteomics, metabolomics, etc.) with data from preclinical and clinical studies and
bioinformatic analyses of large-scale protein–protein interactions or
evolutionary dynamics of antigens.
What makes an antigen a good vaccine candidate is, however, a
matter of fine balances. For example, although conservation is a
desirable property to ensure population coverage, highly conserved
proteins tend to be poor immunogens [34, 35]. In contrast, highly
immunogenic, immunodominant epitopes tend to be highly variable [36]. This is not unexpected, since pathogens are naturally
under strong selection within the host and proteins under positive
selection are therefore more likely to be involved in immunogenicity [37]. Currently available sequencing data and bioinformatics
tools enable the assessment of genetic variation of candidate antigens and their epitopes among large amounts of closely related
Bacterial Pan-Proteome-Based Antigen Discovery
45