Bayesian population structure analysis reveals presence of phylogeographically specific sublineages within previously ill-defined T group of Mycobacterium tuberculosis

Abstract

Mycobacterium tuberculosis genetic structure, and evolutionary history have been studied for years by several genotyping approaches, but delineation of a few sublineages remains controversial and needs better characterization. This is particularly the case of T group within lineage 4 (L4) which was first described using spoligotyping to pool together a number of strains with ill-defined signatures. Although T strains were not traditionally considered as a real phylogenetic group, they did contain a few phylogenetically meaningful sublineages as shown using SNPs. We therefore decided to investigate if this observation could be corroborated using other robust genetic markers. We consequently made a first assessment of genetic structure using 24-loci MIRU-VNTRs data extracted from the SITVIT2 database (n = 607 clinical isolates collected in Russia, Albania, Turkey, Iraq, Brazil and China). Combining Minimum Spanning Trees and Bayesian population structure analyses (using STRUCTURE and TESS softwares), we distinctly identified eight tentative phylogenetic groups (T1-T8) with a remarkable correlation with geographical origin. We further compared the present structure observed with other L4 sublineages (n = 416 clinical isolates belonging to LAM, Haarlem, X, S sublineages), and showed that 5 out of 8 T groups seemed phylogeographically well-defined as opposed to the remaining 3 groups that partially mixed with other L4 isolates. These results provide with novel evidence about phylogeographically specificity of a proportion of ill-defined T group of M. tuberculosis. The genetic structure observed will now be further validated on an enlarged worldwide dataset using Whole Genome Sequencing (WGS).

Local adaptive evolution of two distinct clades of Beijing and T families of Mycobacterium tuberculosis in Chongqing: a Bayesian population structure and phylogenetic study

“…In this study, our phylogenetic analysis suggested the clades BSP2 and TSL6 might evolve from BSP4 and TSL5, respectively, and this was consistent with the recent studies by Reynaud and Zheng [12,13]. However, according to the geographical distribution (Supplementary Table 4) and the results of the recent study by Reynaud and Zheng, the main epidemic region of TSL5 was in Jiangsu province [12], while BSP4 in the mainland of China was epidemic in Tibet [13]. Therefore, BSP2 and TSL6 might be caused by the genetic polymorphisms generated after the entry of extraneous isolates into Chongqing from Tibet and Jiangsu.…”

“…Several recent studies suggested that the Bayesian population structure analysis method based on the 24loci mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) could divide each family of MTB into more detailed clades [11][12][13]. The global T family was divided into eight T sub-lineage (TSL1-TSL8) by Bayesian population structure analysis [12], while the Beijing family in China into five Beijing sub-pedigree (BSP1-BSP5) [13].…”

“…Several recent studies suggested that the Bayesian population structure analysis method based on the 24loci mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) could divide each family of MTB into more detailed clades [11][12][13]. The global T family was divided into eight T sub-lineage (TSL1-TSL8) by Bayesian population structure analysis [12], while the Beijing family in China into five Beijing sub-pedigree (BSP1-BSP5) [13]. Interestingly, the clade distribution varies substantially with regions, especially for BSP2 and TSL6, since 99.49% of BSP2 and 89.79% of TSL6 isolates are distributed in Chongqing, compared with that only 0.51% of BSP2 isolates were in Xinjiang, and 4.08, 2.04 and 2.04% of TSL6 in Guizhou, Sichuan, and Jiangsu province, respectively [12,13].…”

“…The global T family was divided into eight T sub-lineage (TSL1-TSL8) by Bayesian population structure analysis [12], while the Beijing family in China into five Beijing sub-pedigree (BSP1-BSP5) [13]. Interestingly, the clade distribution varies substantially with regions, especially for BSP2 and TSL6, since 99.49% of BSP2 and 89.79% of TSL6 isolates are distributed in Chongqing, compared with that only 0.51% of BSP2 isolates were in Xinjiang, and 4.08, 2.04 and 2.04% of TSL6 in Guizhou, Sichuan, and Jiangsu province, respectively [12,13].…”

“…In order to clarify whether BSP2 and TSL6 originated by their local adaptive evolution from the specific clades of Beijing and T families in Chongqing, the genotyping data of 16 090 MTB isolates from laboratory collection in China, published literatures and an open database [19] were collected (Supplementary Table 1). The Bayesian population structure analysis was implemented to mold the consistent clades [12,13]. The Spacer Oligonucleotide Forest (Spoligoforest) analysis and the 24-loci MIRU-VNTR-based minimum spanning tree (MST) analysis were used to explore the phylogenetic pathways of different families and their clades [20,21] (Fig.…”