Typing and classification of non-tuberculous mycobacteria isolates

Background: There are a large and growing number of non-tuberculous mycobacteria (NTM) species that have been isolated, identified, and described in the literature, yet there are many clinical isolates which are not assignable to known species even when the genome has been sequenced. Additionally, a recent manuscript has proposed the reclassification  of the Mycobacterium genus into five distinct genera. Methods: We describe using a fast average nucleotide identity (ANI) approximation method, MASH, for classifying NTM genomes by comparison to a resource of type strain genomes and proxy genomes. We evaluate the genus reclassification proposal in light of our ANI, MLST, and pan-genome work. Results: We describe here a sequencing study of hundreds of clinical NTM isolates. To aid in characterizing these isolates we defined a multi-locus sequence typing (MLST) schema for NTMs which can differentiate strains at the species and subspecies level using eight ribosomal protein genes. We determined and deposited the allele profiles for 2,802 NTM and Mycobacterium tuberculosis complex strains in PubMLST. Conclusions: The MLST schema and our pan-genome analysis of Mycobacteria can help inform the design of marker-gene diagnostics. The ANI comparisons likewise can assist in the classification of unknown genomes, even from previously unknown species.