Whole genome sequencing reveals genomic heterogeneity and antibiotic purification in Mycobacterium tuberculosis isolates

• DOI: 10.1186/s12864-015-2067-2
• Publication Date: 2015-10-24T19:46:55Z
• Authors: PA Black; M. de Vos; GE Louw; RG van der Merwe; A. Dippenaar; EM Streicher; AM Abdallah; SL Sampson; TC Victor; T. Dolby; JA Simpson; PD van Helden; RM Warren; A. Pain
• Keywords: 0301 basic medicine; 570; Antitubercular Agents; 610; Genomic heterogeneity; Evolution, Molecular; Genetic Heterogeneity; 03 medical and health sciences; Relaxed variant filtering; Genomes -- Effect of drugs on; Drug resistance in microorganisms; Next generation sequencing; Drug Resistance, Bacterial; Genetics; Humans; Tuberculosis, Pulmonary; Clinical isolates; Mycobacterium tuberculosis -- Genetic aspects; Genetic Variation; High-Throughput Nucleotide Sequencing; Genomics; Mycobacterium tuberculosis; Sequence Analysis, DNA; 3. Good health; ROC Curve; Whole genome sequencing; Genetic complexity; Heterogeneity; Genome, Bacterial; Biotechnology; Research Article

Whole genome sequencing has revolutionised the interrogation of mycobacterial genomes. Recent studies have reported conflicting findings on the genomic stability of Mycobacterium tuberculosis during the evolution of drug resistance. In an age where whole genome sequencing is increasingly relied upon for defining the structure of bacterial genomes, it is important to investigate the reliability of next generation sequencing to identify clonal variants present in a minor percentage of the population. This study aimed to define a reliable cut-off for identification of low frequency sequence variants and to subsequently investigate genetic heterogeneity and the evolution of drug resistance in M. tuberculosis.Genomic DNA was isolated from single colonies from 14 rifampicin mono-resistant M. tuberculosis isolates, as well as the primary cultures and follow up MDR cultures from two of these patients. The whole genomes of the M. tuberculosis isolates were sequenced using either the Illumina MiSeq or Illumina HiSeq platforms. Sequences were analysed with an in-house pipeline.Using next-generation sequencing in combination with Sanger sequencing and statistical analysis we defined a read frequency cut-off of 30% to identify low frequency M. tuberculosis variants with high confidence. Using this cut-off we demonstrated a high rate of genetic diversity between single colonies isolated from one population, showing that by using the current sequencing technology, single colonies are not a true reflection of the genetic diversity within a whole population and vice versa. We further showed that numerous heterogeneous variants emerge and then disappear during the evolution of isoniazid resistance within individual patients. Our findings allowed us to formulate a model for the selective bottleneck which occurs during the course of infection, acting as a genomic purification event.Our study demonstrated true levels of genetic diversity within an M. tuberculosis population and showed that genetic diversity may be re-defined when a selective pressure, such as drug exposure, is imposed on M. tuberculosis populations during the course of infection. This suggests that the genome of M. tuberculosis is more dynamic than previously thought, suggesting preparedness to respond to a changing environment.