Methods: Whole genome sequences were obtained for 71 Mtb clinical isolates. We reconstructed phylogenetic trees using maximum likelihood methods, estimated neutrality statistics in DnaSP, and used a custom Perl script to identify SNPs with alleles unique to given subpopulations. Gene-wise dN/dS values were estimated using the branch-site random effects likelihood model in HyPhy v2.
Results: Mtb subpopulations in NYC and New Jersey can be distinguished via phylogenomic reconstruction and by determination of subgroup-defining nucleotide substitutions. These methods identified known historical outbreak clusters and previously unidentified subpopulations within relatively monomorphic endemic clone groups. Neutrality statistics based on the site frequency spectrum were less useful for identifying Mtb subpopulations. We observed that isolates from New York City endemic clone groups have acquired multiple non-synonymous SNPs in virulence-associated pathways, and relatively few mutations in drug resistance-associated genes, suggesting that overall pathoadaptive fitness, rather than the acquisition of drug resistance mutations, has played a central role in the evolutionary history and epidemiology of M. tuberculosis subpopulations in NYC.
Conclusion: Our results demonstrate that some but not all WGS-based methods are useful for detection of emerging Mtb clone groups, and support the use of phylogenomic reconstruction in routine tuberculosis laboratory surveillance, particularly in areas with relatively less diverse Mtb populations. We conclude that wider-reaching phylogenomic and population genomic methods can be informative for tuberculosis control activities by identifying genetic polymorphisms contributing to epidemiological success in local Mtb populations.
J. Walker, None
P. Planet, None
P. Bifani, None
S. O. Kolokotronis, None
B. N. Kreiswirth, None
B. Mathema, None