Inference of events in Mycobacterium tuberculosis genome evolution from high-throughput sequencing data

Mycobacterium tuberculosis (MTB) is a slow-growing, gram-positive bacterium from the phylum Actinobacteria and is the causative agent of human tuberculosis disease. MTB is a strict host-associated pathogen thought to have infected one third of the human population. In addition, MTB strains exhibit diverse antibiotic resistance profiles and high transmissibility, with notably an ongoing spread of multi-drug resistance. MTB evolution is observed through the emergence of de novo variations such as substitutions, and no horizontal evolution has been described so far. In the absence of horizontal evolution, bacterial genomes are expected to undergo degradation and to exhibit high levels of epistasis. Notably, short-term MTB evolution is mostly studied by substitutions where other variants, such as insertions and deletions (indels), are generally overlooked. In this thesis, we systematically investigated the presence of all variant types in MTB to assess their prevalence and impact in MTB evolution. In the first part, we analyzed publicly available complete MTB genomes for signals of horizontal gene transfer and genetic recombination. We found that noise overcomes the recombination signal due to low quality of the available reference genomes. In particular, our analysis suggests that reference-guided assemblies have many uncertainties introduced by the reference, thereby biasing the phylogenetic signal. In the second part, we analyzed a multi-drug resistant MTB outbreak of 353 strains, longitudinally sampled over 20 years, for the presence of high-quality indels. Our main results show that indels are significantly enriched in genes conferring antibiotic resistance. Furthermore, indels are inherited in the reconstructed phylogeny and follow a molecular clock that is 23 times slower than that of substitutions. Finally, we highlight co-occurrences of substitutions and indels along related biological pathways, suggesting widespread epistatic interactions. This thesis first confirms that horizontal mechanisms are not reliably identifiable in current complete MTB genomes. If MTB evolution is impacted by horizontal mechanisms, the signal is expected to be low. Hence, high-quality genomes are needed to conclude on the presence or absence of horizontal evolution in MTB. Furthermore, we discovered that indels play an important role in short-term MTB evolution, notably in the emergence of antibiotic resistance. We hypothesize that epistasis governs the evolution of MTB, in particular regarding indels, which might be involved in so far undiscovered compensatory mechanisms. MTB evolution is intriguing on many levels, and studying more in details the interplay of all variants acting together is expected to shed light on the complex evolutionary mechanisms ruling MTB evolution.


Use and reproduction:

No license. The provisions of the German Copyright Act (UrhG) apply.

Please note that individual components of the publication may be subject to other licensing or copyright conditions.


Citation style:
Could not load citation form.