WHO-recommended protocols and clinical data [180, 181]. In
order to assert quality and validity of the analysis, ReSeqTB provides a MTBC-validated Unified Variant Pipeline (UVP) to identify
variants and assign lineages. UVP has been tested in large studies
that established associations between genotype and phenotype for
multiple TB drugs and developed a confidence-graded list of mutations to interpret global sequencing-based drug resistance surveillance data [158, 168, 182]. While the clinical utility if ReSeqTB is
still being defined, this tool provides valuable information to assay
developers and researchers by allowing easy analysis of aggregated
data tracking 179 genetic loci associated with antimicrobial resistance. Users can search these genomic mutations and correlate their
presence/absence with multiple drug resistance phenotypes. This
knowledge base aggregates the information submitted by its users
and grows about 500 genomes per month. Users are encouraged to
submit their WGS and the associated metadata, not only to allow
analysis of their own isolates but also to enrich the depth and
breadth—and thus the future utility—of the knowledge base. One
of the long-term goals of ReSeqTB is to allow clinicians to establish
WGS-based drug-resistant MTBC surveillance programs, obviating
the need for culturing clinical mycobacterial isolates and improving
feasibility of WGS technologies to support timely high-quality
patient management decisions.
5.4 Genomic
Epidemiology
The tracking of pathogens using genetic information, termed
genomic or molecular epidemiology, can encompass various scenarios such as local transmission in a population up to global spread
and timescales. The genetic information underlying these analyses
also ranges from specific genes or markers (e.g., microsatellites) up
to whole-genome sequences. As with most mycobacterial research,
genomic epidemiology has been most widely implemented for
M. tuberculosis [183]. RFLP, spoligotyping, and MIRU-VNTR
pattern matching have all been used in the past to try to determine
if two isolates are potentially linked in a transmission chain
[184, 185]. However, when searching for recent transmission
(e.g., in the past decade), these techniques are known to vastly
overestimate transmission cluster sizes [186–188].
It is now recommended that WGS data be used for transmission clustering as it gives the best resolution [186, 188–191]. Since
M. tuberculosis mutates very slowly, strains tend to be nearly identical if they are linked in a transmission chain. Current best practice
states that if two strains differ by fewer than 5 SNPs, they are likely
in a transmission cluster; if they differ by more than 12 SNPs, they
are highly unlikely to be in a transmission cluster [192]. These
cutoffs have been applied to examination of both withinpopulation transmission dynamics and intercountry transmissions,
demonstrating their usefulness for surveillance and research purposes [192–194]. However, this approach was devised using data
Mycobacterial ID and Characterization Using WGS
443
order to assert quality and validity of the analysis, ReSeqTB provides a MTBC-validated Unified Variant Pipeline (UVP) to identify
variants and assign lineages. UVP has been tested in large studies
that established associations between genotype and phenotype for
multiple TB drugs and developed a confidence-graded list of mutations to interpret global sequencing-based drug resistance surveillance data [158, 168, 182]. While the clinical utility if ReSeqTB is
still being defined, this tool provides valuable information to assay
developers and researchers by allowing easy analysis of aggregated
data tracking 179 genetic loci associated with antimicrobial resistance. Users can search these genomic mutations and correlate their
presence/absence with multiple drug resistance phenotypes. This
knowledge base aggregates the information submitted by its users
and grows about 500 genomes per month. Users are encouraged to
submit their WGS and the associated metadata, not only to allow
analysis of their own isolates but also to enrich the depth and
breadth—and thus the future utility—of the knowledge base. One
of the long-term goals of ReSeqTB is to allow clinicians to establish
WGS-based drug-resistant MTBC surveillance programs, obviating
the need for culturing clinical mycobacterial isolates and improving
feasibility of WGS technologies to support timely high-quality
patient management decisions.
5.4 Genomic
Epidemiology
The tracking of pathogens using genetic information, termed
genomic or molecular epidemiology, can encompass various scenarios such as local transmission in a population up to global spread
and timescales. The genetic information underlying these analyses
also ranges from specific genes or markers (e.g., microsatellites) up
to whole-genome sequences. As with most mycobacterial research,
genomic epidemiology has been most widely implemented for
M. tuberculosis [183]. RFLP, spoligotyping, and MIRU-VNTR
pattern matching have all been used in the past to try to determine
if two isolates are potentially linked in a transmission chain
[184, 185]. However, when searching for recent transmission
(e.g., in the past decade), these techniques are known to vastly
overestimate transmission cluster sizes [186–188].
It is now recommended that WGS data be used for transmission clustering as it gives the best resolution [186, 188–191]. Since
M. tuberculosis mutates very slowly, strains tend to be nearly identical if they are linked in a transmission chain. Current best practice
states that if two strains differ by fewer than 5 SNPs, they are likely
in a transmission cluster; if they differ by more than 12 SNPs, they
are highly unlikely to be in a transmission cluster [192]. These
cutoffs have been applied to examination of both withinpopulation transmission dynamics and intercountry transmissions,
demonstrating their usefulness for surveillance and research purposes [192–194]. However, this approach was devised using data
Mycobacterial ID and Characterization Using WGS
443