Concordance of variable-number tandem repeat (VNTR) and large sequence polymorphism (LSP) analyses of Mycobacterium tuberculosis strains.

Variable-number tandem repeat (VNTR) and large sequence polymorphism (LSP) analyses were compared to determine whether VNTR analysis was effective for population genetic analysis of Mycobacterium tuberculosis strains. A total of 682 strains, 510 Beijing genotype and 172 non-Beijing genotype strains, were studied. The number of repeats was investigated for 24 VNTR loci: the 15 loci of "optimized miru", the 8 loci of "Beijing option", and 1 locus for "JATA12". Six loci (miru31, Mtub4, QUB4156c, QUB3232, VNTR3820, and VNTR4120) showed significantly different median numbers of repeats in strains belonging to different lineages defined by LSP (P<0.01, Mann-Whitney U test). When a minimum-spanning tree (MST) was reconstructed using these 6 loci, most strains clustered in the expected branches in the MST branches. However, topology of the MST was not congruent with the evolutional hypothesis of M. tuberculosis, indicating that MST analysis using VNTR data should not use for phylogeny of the organism. When the standardized index of association (sI(A)) was calculated using data for the 6 VNTR loci, the value of sI(A) was significantly different from zero (Monte Carlo simulation with 10,000 resamplings) in every lineage, indicating the linkage disequilibrium in different lineage strains of M. tuberculosis. These results were consistent with the hypothesis that clonal evolution of lineages of the organism has occurred. Therefore, the 6 loci identified in this study would be effective for M. tuberculosis population genetic analysis due to their significantly different median numbers of repeat and linkage disequilibrium though VNTR data was not effective for phylogeny of the organism.