candidate proteins with their properties for the user to evaluate.
The fact that all programs are independent and a large number of
proteins may be examined prompts for straightforward parallelization of proteins/tasks so that if a computer cluster is available the
entire process can be completed in a very acceptable time.
4 Conclusions and Key Challenges
The classical in silico reverse vaccinology strategy has been widely
used for the identification of surface-exposed and secreted extracellular proteins as potential candidates for vaccines seeking to promote a dominantly humoral immune response. This simple analysis
may be further complicated by the incorporation of computational
methods to identify B-cell epitopes, often conformational and discontinuous in nature, in candidate proteins. Most methods for the
prediction of B-cell epitopes require the three-dimensional structure of the antigen [25]. The recognition of specific protein fragments as T-cell epitopes (both HLA class I and HLA class II
restricted) has been also successfully integrated into antigen discovery pipelines. Although HLA (particularly HLA class I) epitope
prediction is more advanced and reliable than that of B-cell epitope
prediction, HLA presentation does not guarantee TCR recognition. This results in a systematic overprediction of T-cell epitopes
(false positive epitopes), in addition to the usual limitations of
predictors based on far from perfect experimental datasets
[25]. Therefore, we recommend combining different HLA epitope
prediction algorithms and selecting those epitopes showing a relative consensus. As mentioned, for HLA class I preliminary
Fig. 2 Proposed stages for genome-wide identification of positive selection and recombination in bacterial
pangenomes (a). Sequence conservation analysis for two Burkholderia pseudomallei candidate proteins
(BPSL1626 and BPSS1727) in 344 strains (b). Asterisks indicate positively selected sites. Same color in the
graphic background indicates same recombination fragments
56
Daniel Yero et al.
The fact that all programs are independent and a large number of
proteins may be examined prompts for straightforward parallelization of proteins/tasks so that if a computer cluster is available the
entire process can be completed in a very acceptable time.
4 Conclusions and Key Challenges
The classical in silico reverse vaccinology strategy has been widely
used for the identification of surface-exposed and secreted extracellular proteins as potential candidates for vaccines seeking to promote a dominantly humoral immune response. This simple analysis
may be further complicated by the incorporation of computational
methods to identify B-cell epitopes, often conformational and discontinuous in nature, in candidate proteins. Most methods for the
prediction of B-cell epitopes require the three-dimensional structure of the antigen [25]. The recognition of specific protein fragments as T-cell epitopes (both HLA class I and HLA class II
restricted) has been also successfully integrated into antigen discovery pipelines. Although HLA (particularly HLA class I) epitope
prediction is more advanced and reliable than that of B-cell epitope
prediction, HLA presentation does not guarantee TCR recognition. This results in a systematic overprediction of T-cell epitopes
(false positive epitopes), in addition to the usual limitations of
predictors based on far from perfect experimental datasets
[25]. Therefore, we recommend combining different HLA epitope
prediction algorithms and selecting those epitopes showing a relative consensus. As mentioned, for HLA class I preliminary
Fig. 2 Proposed stages for genome-wide identification of positive selection and recombination in bacterial
pangenomes (a). Sequence conservation analysis for two Burkholderia pseudomallei candidate proteins
(BPSL1626 and BPSS1727) in 344 strains (b). Asterisks indicate positively selected sites. Same color in the
graphic background indicates same recombination fragments
56
Daniel Yero et al.