species-level identification of these three genes showed correct
identification of 86.79% and 81.55% of strains with hsp65 or rpoB,
respectively, compared to only 71.3% for 16S [73]. Between certain
species, however, any of these genes individually show little to no
interspecies divergence [73], limiting their usefulness for definitive
species-level identification. Concatenations of these gene sequences
can provide additional resolution, particularly if combined with
additional genes such as sodA and recA [74].
With increasing combinations of genes, the use of sequence
from many different loci becomes known as multilocus sequence
analysis (MLSA). MLSA is a technique that applies the comparison
of numerous genes that encode proteins to determining interspecies taxonomic relationships and delineating species. This is related
to, but distinct from, multilocus sequence typing, which is used for
intraspecies molecular epidemiology to place strains within taxonomically extant species [75, 76]. Numerous MLSA schemes for
various mycobacteria have been developed, and the technique has
been applied to M. avium [77], M. tuberculosis [78], and the
classification of M. abscessus subspecies [79]. MLST has been
applied to epidemiological elucidation of M. xenopi [80],
M. avium [81], and M. abscessus [82–84].
4.3 WGS-Based
Identification
Techniques
At this time, the “gold standard” technique for genomic circumscription of bacterial species remains DNA-DNA hybridization
(DDH) [85]. In brief, DDH is a wet-lab technique by which the
genomic DNA from two strains are denatured and allowed to
anneal to each other as hybridized strains, and the melting temperature (T m ) is calculated. Because the amount of energy required to
separate the DNA strands is proportional to the number of hydrogen bonds between them, which in turn is proportional to the
sequence homology between the two strains, a higher similarity
between the genomes of the two strains results in a higher T m ,
while more distantly related strains will have a lower T m . The
recommended DDH threshold for determining that the genomes
from two strains belong to the same species is 70% [86].
However, the DDH technique is labor-intensive, error-prone,
poorly reproducible, and generally only available from relatively few
specialized laboratories [87]. Instead, recent focus has been on
leveraging WGS to accomplish comparisons at the whole-genome
level and ultimately provide elucidation of a genome-based taxonomy [16, 86, 88–92]. Collectively, such methods comprise a technique known as overall genome relatedness index (OGRI)
[89]. The two primary methods for direct sequence comparison
of two genomes are average nucleotide identity (ANI) and digital
DNA-DNA hybridization (dDDH).
ANI was first proposed in 2005 by Konstantinidis and Tiedje
[93]. In order to more accurately correspond to the physical fragmentation experienced by genomic DNA during DDH, a later
422
Marco A. Riojas et al.
identification of 86.79% and 81.55% of strains with hsp65 or rpoB,
respectively, compared to only 71.3% for 16S [73]. Between certain
species, however, any of these genes individually show little to no
interspecies divergence [73], limiting their usefulness for definitive
species-level identification. Concatenations of these gene sequences
can provide additional resolution, particularly if combined with
additional genes such as sodA and recA [74].
With increasing combinations of genes, the use of sequence
from many different loci becomes known as multilocus sequence
analysis (MLSA). MLSA is a technique that applies the comparison
of numerous genes that encode proteins to determining interspecies taxonomic relationships and delineating species. This is related
to, but distinct from, multilocus sequence typing, which is used for
intraspecies molecular epidemiology to place strains within taxonomically extant species [75, 76]. Numerous MLSA schemes for
various mycobacteria have been developed, and the technique has
been applied to M. avium [77], M. tuberculosis [78], and the
classification of M. abscessus subspecies [79]. MLST has been
applied to epidemiological elucidation of M. xenopi [80],
M. avium [81], and M. abscessus [82–84].
4.3 WGS-Based
Identification
Techniques
At this time, the “gold standard” technique for genomic circumscription of bacterial species remains DNA-DNA hybridization
(DDH) [85]. In brief, DDH is a wet-lab technique by which the
genomic DNA from two strains are denatured and allowed to
anneal to each other as hybridized strains, and the melting temperature (T m ) is calculated. Because the amount of energy required to
separate the DNA strands is proportional to the number of hydrogen bonds between them, which in turn is proportional to the
sequence homology between the two strains, a higher similarity
between the genomes of the two strains results in a higher T m ,
while more distantly related strains will have a lower T m . The
recommended DDH threshold for determining that the genomes
from two strains belong to the same species is 70% [86].
However, the DDH technique is labor-intensive, error-prone,
poorly reproducible, and generally only available from relatively few
specialized laboratories [87]. Instead, recent focus has been on
leveraging WGS to accomplish comparisons at the whole-genome
level and ultimately provide elucidation of a genome-based taxonomy [16, 86, 88–92]. Collectively, such methods comprise a technique known as overall genome relatedness index (OGRI)
[89]. The two primary methods for direct sequence comparison
of two genomes are average nucleotide identity (ANI) and digital
DNA-DNA hybridization (dDDH).
ANI was first proposed in 2005 by Konstantinidis and Tiedje
[93]. In order to more accurately correspond to the physical fragmentation experienced by genomic DNA during DDH, a later
422
Marco A. Riojas et al.