An in silico approach to identify vaccine candidates against Neospora caninum inspired by reverse vaccinology
- Publication Type:
- Issue Date:
NO FULL TEXT AVAILABLE. This thesis contains 3rd party copyright material. ----- Neospora caninum is an apicomplexan protozoan parasite that can cause significant economic and reproductive loss to cattle industries worldwide. Wasteful culling of seropositive cattle is currently the only effective infection control, but potential vaccination is considered more cost-effective. The thesis research objective was to exclusively use in silico techniques, inspired by reverse vaccinology, to identify protein-based vaccine candidates against N. caninum. This objective, in effect, was to develop a bioinformatics pipeline using freely available programs and biological data to predict candidates. Five main challenges confronted the research project. First, no consensus is in the literature on protein antigen types assured to mediate protective immunity (i.e. unknown prediction targets). Second, most data for N. caninum are predicted and uncharacterised (i.e. unknown data reliability). Third, no consensus exists on the type of programs and input data required for in silico vaccine discovery against eukaryotic pathogens (i.e. unknown computational requirements). Fourth, prediction programs are typically inaccurate (i.e. unknown program reliability). Fifth, no known program exists to validate candidates in a host-vaccine interaction. Eight publications and one unsubmitted manuscript are presented in nine thesis chapters, which summarise the body of work that endeavours to address the five challenges. Each chapter is an independent sub-study that contributes knowledge and methodology towards collectively fulfilling the research objective. The main research outcome provides, for the first time using exclusively a high-throughput pipeline, a prioritized list of every known N. caninum protein in accordance with its predicted vaccine candidacy potential. A desired percentage for validation can be selected from the list that is consistent with laboratory capability and budget. Furthermore, this freely available pipeline can be adapted for other eukaryotic pathogens. Prediction targets were protein characteristics suggesting antigenicity, based on majority opinion and current trends from the literature. Importantly, an ensemble of machine learning algorithms and ranking strategies alleviated data and program reliability concerns. The best deemed interim option to validate candidates was to validate the in silico process using proteins with known immunogenicity. A pipeline proof-of-concept distinguished known immunogenic proteins with a sensitivity and specificity of 0.97 and 0.98, respectively. Seven of 14 known N. caninum candidates were ranked in the top 1% for laboratory investigation, although three were poorly ranked. Top ranked proteins are judged the optimum candidates within the constraints of available data, current knowledge, and existing programs. Ultimately, it is trusted that this thesis provides vital contributions towards identifying N. caninum candidates and improving in silico vaccine discovery methodology.
Please use this identifier to cite or link to this item: