Background: The genus Mycobacterium contains nearly 200 species and among them, some species of clinical interest due to their severe human pathogenicity. Mycobacterium taxonomy has recently being reviewed and the genus is now divided into 5 distinct groups, Mycolicibacterium, Mycobacteroides, Mycolicibacter, Mycolicibacterium and Mycobacterium, the last including all the human pathogens. Taken into account this new taxonomy, a molecular analysis was performed within each clade to set up an easy molecular method to identify the best relevant species. Public genomes were analyzed by similarity and target sequences on partial genes of 16S, rpoB and hsp65 genes.
Methods: Whole genome analysis was done with an ANI approach (average nucleotide identity) on 176 NCBI genomes from type strains and collection strains available (89 Mycobacterium, 53 Mycolicibacterium and 34 Mycobacteroides), the two other clades being not clinically relevant. Target sequencing was done by phylogenic classification using UPGMA/Neighbor-joining with public sequences of 16S, partial rpoB (711pb) and partial hsp65 (603pb) genes.
Results: The ANI study showed more than 99.9% homology between subspecies of Mycobacterium abscessus and those of Mycobacterium avium. The analysis reveals also a high similarity between some species which have recently been reclassified (M. intracellulare and M. paraintracellulare, M. bouchedurhonense and M. avium, M. intracellulare and M. chimaera).
The target sequencing analysis showed that 16S was not sufficiently discriminant, especially for the Mycobacteroides group. Analysis of rpoB and hsp65 sequences separately allows a correct identification at the species level but is not sufficient at the subspecies level. The concatenation of rpoB and hsp65 allows the discrimination of subspecies such as those of Mycobacteroides abscessus and Mycobacterium intracellulare.
Conclusions: ANI programs on whole genomes are not discriminative enough to identify all Mycobacterium species of interest, especially at the subspecies level. Target sequencing using partial rpoB and hsp65 concatened genes allows to identify all relevant pathogenic Mycobacterium species, at the subspecies level taken into account the new taxonomy. This in-silico analysis has to be completed with internal strain sequencing before being used in our laboratory.