We performed 12 loci MIRU–VNTR on 327 Mycobacterium tuberculosis (Mtb) isolates belonging to three major spoligotypes MANU1, CAS1_Delhi and Beijing from Mumbai, western India and two proximal rural locations. Complete allele and drug susceptibility data was available for 232 isolates. These included 143 MANU1 (ST100), 65 CAS1_Delhi (ST26) and 24 Beijing (ST1) isolates. Of the 232 isolates, 26 were rural consisting 6 CAS1_Delhi and 20 MANU1 isolates.

Using eBURST multi-locus sequence typing (MLST), cluster analyses was performed for each of the spoligotypes and drug susceptibility profiles. MANU1 MLST consisted of 90 related isolates (clustered and grouped) and 53 singletons; CAS1_Delhi MLST consisted of 44 related isolates and 21 singletons; Beijing MLST consisted of 10 related isolates and 14 singletons. Although the number of related isolates were different in MANU1 (63%), CAS1_Delhi (68%) and Beijing (42%) clusters, it was not statistically significant. Furthermore, it was observed that while MANU1 and CAS1_Delhi singletons (n = 74) had only 12 (16%) MDR isolates, the Beijing MLST had 8/14 (57%) MDR singleton isolates.

Phylogenetic ananlysis using minimum spanning tree (MST) and a UPGMA radial tree revealed MANU1 had the largest number of nodes as compared to the CAS1_Delhi and Beijing spoligotypes. Additionally the CAS isolates were more homogeneous than the MANU1 isolates.

The 12 loci MIRU–VNTR was used to provide greater discrimination than spoligotyping, but 6 of the 12 loci provided less than 50% discriminatory power. The highest discrimination was achieved using locus 26 (80%). Our results concur with recent reports that the most discriminatory MIRU–VNTR combination varied across different lineages. The results also highlight the need for more robust genetic markers for studying the transmission of Mtb in endemic regions like India.