Applying multi-omics analyses to investigate Mycobacterium tuberculosis infection: a dual perspective

Mycobacterium tuberculosis (Mtb) is the causative agent of tuberculosis (TB) and is the second highest cause of death from an infectious disease worldwide. This is despite years of research to understand and tackle TB, leading to a vaccine, diagnostic tests, and treatments. Despite these interventions, several aspects of the disease are poorly understood or neglected in this field of research. A greater understanding of the infection in these areas will enable the development of more effective prevention and treatment strategies. This thesis addressed these concerns through using a multi-targeted approach investigating several aspects of Mtb infection which are understudied. First, the characterisation of paediatric TB was performed through differential gene expression analysis using a cohort of paediatric TB patients from South Africa. This cohort was then used to validate previously existing host transcriptomic biomarker signatures for TB to identify if the current adult or paediatric signatures meet the guidelines set out by the World Health Organisation (WHO) for development of new non-sputum based diagnostic tests for TB. None of these signatures met these guidelines using their original models. Optimisation of the signature models increased the performance of some of these signatures to reach the WHO guidelines, suggesting that there is potential for certain signatures to be developed into diagnostic tests in the future. Second, the differences between Mtb lineages was investigated through the integration of multi-omics datasets. Differences in non-synonymous single nucleotide polymorphisms (nsSNPs) and metabolites from 14 Mtb strains corresponding to 3 lineages were identified. The differences in metabolites was then correlated with nsSNP blocks (groups of correlated nsSNPS). This identified a novel negative corelation between ergothioneine, a metabolite involved in redox pathways, and a nsSNP block only present in Mtb strains from lineage 1. Finally, host-pathogen interactions were investigated through dual RNA-seq of whole blood infected with BCG lux, a surrogate strain for Mtb. Dual RNA-seq is a relatively new technique which has never been formed from whole blood samples infected with Mtb. A workflow to combine the previously defined whole blood assay with dual RNA-seq was developed and optimised. A pilot dual RNA-seq experiment was performed to identify the number of reads, and the number of genes, obtained from both the host and bacteria through this workflow. Further characterisation of the whole blood assay was also performed through the evaluation of live immune cell proportions across the time-course of the assay and through comparison of the transcriptome of Mtb infected whole blood with the transcriptomes of adult and paediatric TB patients. Overall, this thesis provides new insights into the complex biology of Tuberculosis.