top of page


Routine whole genome sequencing of Mycobacterium tuberculosis when the consensus in not enough

R M Anthony(1) R de Zwaan(1) M Kamst-van Agterveld(1) A Mulder(1) J van den Dool(1) D van Soolingen(1)

1:RIVM National Tuberculosis Reference Laboratory, Centre for Infectious Disease Control, National Institute for Public Health and the Environment, Bilthoven, the Netherlands

The Netherlands whole genome sequencing (WGS) database currently contains almost 6000 M. tuberculosis local clinical isolates. The consensus sequence has proven utility to rule out recent transmission, for (sub)species identification and for the prediction of susceptibility. We routinely generate a consensus sequence (SNPs present in >80% of reads called) against the reference genotype to detect any mutations; a short list of SNPs(single nucleotide polymorphisms), including high confidence resistance mutations and clade specific SNPs, are additionally routinely screened for any mutant reads to rule out mixed infections and detect emerging resistance.

We have monitored the noise profile of the SNPs routinely screened for low frequency mutations in detail over time in our database. Notably, the well-known rpoB 761152 T>A noise  resolved when we switched (Illumina™) sequence provider.  In addition to our routine analysis sequence variability detectable in isolates with epidemiological links and serial isolates is analyzed. This more detailed analysis may provide additional insight into epidemiological and disease processes. For example; 1. previous treatment, particularly when treatment has been interrupted or suboptimal fragile regimens have been used, resulting in multiple sub populations of escape mutants; 2. chronic infections or reactivation of a temporally distant infection may yield genetically diversified sub populations within a single individual; 3. emerging mutations can sometimes be detected in isolates within recent clusters, allowing the inference of likely transmission links and direction from sequence data. We present examples of these effects and discuss the possibilities to expand these types of analysis with current and emerging sequencing approaches.

bottom of page