Enhanced analysis of the genomic diversity of Mycobacterium bovis in Great Britain to aid control of bovine tuberculosis

Abstract

Bovine tuberculosis (bTB) is an endemic disease in Great Britain (GB) that affects mainly cattle but also other livestock and wild mammal species, leading to significant economic and social impact. Traditional genotyping of Mycobacterium bovis (M. bovis) isolates, which cause bTB, had been used routinely since the late 1990s as the main resource of genetic information in GB to describe their population and to understand their epidemiology. Since 2017, whole-genome sequencing (WGS) has been implemented on M. bovis isolates collected during routine surveillance. In this study, we analysed genome sequences from 3,052 M. bovis isolates from across GB to characterise their diversity and population structure in more detail. Our findings show that the M. bovis population in GB, based on WGS, is more diverse than previously indicated by traditional genotyping and can be divided into seven major clades, with one of them subdivided further into 29 clades that differ from each other by at least 70 single-nucleotide polymorphisms (SNPs). Based on the observed phylogenetic structure, we present a SNP-based classification system that replaces the genotype scheme that had been used until recently in GB. The predicted function and associated processes of the genes harbouring these SNPs are discussed with potential implications for phenotypic/functional differences between the identified clades. At the local scale, we show that WGS provides greater discriminatory power and that it can reveal the origin of infection and associated risk pathways even in areas of high bTB prevalence. The difficulty in determining transmission pathways due to the limited discrimination of isolates by traditional typing methods has compromised bTB control, as without such information it is harder to determine the relative efficacy of potential intervention measures. This study demonstrates that the higher resolution provided by WGS data can improve determination of infection sources and transmission pathways, provide important insights that will inform and shape bTB control policies in GB, as well as improve farm specific advice on interventions that are likely to be effective.

Keywords: Mycobacterium bovis; bovine tuberculosis (bTB); clade; epidemiology; phylogenetics; whole genome sequencing.

Figure 1

Maximum-likelihood phylogeny of trimmed Mycobacterium bovis dataset (n = 938 sequences) from across Great Britain. The outgroup of the tree (M. caprae sequence SRR7617662) has been removed for better visualisation. The scale bar represents units of substitutions per site. Major clades (B1–B7) are labelled on the tree and minor clades (B6-11–B6-92) are shown next to the tips of the tree. WGS clade and in silico predicted spoligotype for each isolate are indicated in the bars next to the phylogeny (from left to right). Grey bars on the in silico predicted spoligotype annotations mean that a unique spoligotype pattern was identified that had not been observed for any other isolate (singletons). Bootstrap support values for major internal nodes including nodes that represent the most recent common ancestor of each of the WGS clades are over 95%.

Figure 2

Frequency distribution of pairwise SNP distances between isolates from select WGS clades (A) and between isolates within a WGS clade (B).

Figure 3

Frequency distributions of within-clade pairwise SNP distances (top) and geographical ranges (bottom) for clades B6-11 (A), B6-16 (B), B6-62 (C) and B6-41 (D). B6-11 and B6-62 (A, D) represent widespread clades with relatively high and low levels of genetic diversity, respectively, while B6-16 and B6-41 (B, C) represent clades with constrained geographical ranges and medium or low diversity, respectively. The y-axis in the frequency distribution plots shows the number of pairwise comparisons and the x-axis shows pairwise genetic distances in number of SNPs. Note that the scale of the x-axis differs in each of the clades.

Figure 4

Maximum parsimony tree (A) and map (B) showing the phylogenetic relationships and relative location of the three isolates collected from a recent outbreak (farm A) in the SW of England. The tree (A) is a snippet from the phylogeny of WGS clade B6-51. Branch length indicates number of SNPs. Relative locations are shown on the map (B) to comply with data protection law in the UK.