NEW YORK (GenomeWeb) – Researchers from the London School of Hygiene & Tropical Medicine and other institutions have developed a prototype of a tool called TB-Profiler that uses whole genome sequence from Mycobacterium tuberculosis strains to more accurately predict bacterial resistance to current tuberculosis treatments.
Full details of the online tool were published last week in Genome Medicine. According to the paper, TB-Profiler comprises a library of 1,325 curated drug resistance markers for 11 anti-tuberculosis drugs that were validated using genomic-phenotypic data from 792 isolates gleaned from six countries. The resource also includes an informatics pipeline that uses an algorithm called Snap to map raw sequences from clinical specimens such as sputum to a modified tuberculosis reference genome, and then uses samtool/vcf tools to call SNPs as well as insertions and deletions.
TB-Profiler provides a resource for researchers that are using next-generation sequencing technologies in TB testing. Whole-genome sequencing, they argue in Genome Medicine, can improve on existing methods for detecting sensitivity to TB treatments by providing a more fine-grained look at the resistance-conferring mutations in TB.
Increasing resistance to multiple first- and second-line TB treatments threatens efforts to curb the disease worldwide. According to numbers from the World Health Organization, 5 percent of the 11 million TB cases observed globally have multi drug-resistance disease (MDR-TB) — these are patients whose infections are resistant to first line drugs isoniazid and rifampicin. In 2013 alone, approximately 480,000 new MDR-TB cases were reported, according to WHO. Of those new cases, roughly 9 percent had extensively resistant tuberculosis (XDR-TB). These are cases where patients are infected with bacterial strains that are resistant to the aforementioned drugs as well as to two classes of second line drugs currently used to treat MDR-TB — fluoroquinolones and aminoglycosides.
In some cases, TB resistance is determined using traditional culture techniques which can take weeks or months to grow and test the effects of multiple treatments on bacterial colonies. Alternatives to cultures include molecular assays that are used to test sputum samples directly for resistance to key drugs, such as Cepheid's GeneXpert assay, which tests samples for resistance to rifampicin. Also effective are line probe assays which have been developed to detect resistance to both first- and second-line TB treatments. However tests in both of these categories are "limited in the number of loci they examine and they lack capacity to differentiate silent mutations from those that effect drug efficacy, leading to false positive results," the researchers wrote in their paper.
Whole-genome sequencing is a potential solution to the issues faced by all three testing mechanisms, offering a more in-depth look at more loci than previous methods, the paper states. Moreover, "sequencing already assists patient management for a number of conditions such as HIV, but now that it is possible to sequence M. tuberculosis from sputum from suspected multi-drug resistance patients, [the technology can now be used] in the management of tuberculosis," Taane Clark, a reader in genetic epidemiology and statistical genomics and one of the lead authors on the paper, said in a statement.
Recent technological advancements have made it possible to sequence bacterial samples in a matter of hours and dropping sequencing costs have put benchtop instruments within the reach of most labs and hospitals. However data analysis is still challenging for some as clinical labs often lack the requisite expertise needed to install and run analysis applications, according to the researchers.
And that's where TB-Profiler comes in. For input sequence, the software identifies mutations that confer resistance to particular drugs and also provides data on the type and lineage of the TB strain in question, making it possible to track things like transmission events, Clark told GenomeWeb. The solution returns results within minutes and users aren't required to have specialized bioinformatics expertise to run TB-Profiler, making it readily available for use in settings where those particular skills may be lacking, he said.
TB-Profiler is currently intended for research use and is currently being used by a number of research groups globally to try to identify TB strains from samples they collect, Clark told GenomeWeb. However, hospitals and clinical testing laboratories that have genome sequencing facilities in place could use the resource to gain some insights into potential treatment options that might work for patients, he said.