Comparison of Sample Preparation Methods Used for the Next-Generation Sequencing of Mycobacterium tuberculosis

Abstract

The advent and widespread application of next-generation sequencing (NGS) technologies to the study of microbial genomes has led to a substantial increase in the number of studies in which whole genome sequencing (WGS) is applied to the analysis of microbial genomic epidemiology. However, microorganisms such as Mycobacterium tuberculosis (MTB) present unique problems for sequencing and downstream analysis based on their unique physiology and the composition of their genomes. In this study, we compare the quality of sequence data generated using the Nextera and TruSeq isolate preparation kits for library construction prior to Illumina sequencing-by-synthesis. Our results confirm that MTB NGS data quality is highly dependent on the purity of the DNA sample submitted for sequencing and its guanine-cytosine content (or GC-content). Our data additionally demonstrate that the choice of library preparation method plays an important role in mitigating downstream sequencing quality issues. Importantly for MTB, the Illumina TruSeq library preparation kit produces more uniform data quality than the Nextera XT method, regardless of the quality of the input DNA. Furthermore, specific genomic sequence motifs are commonly missed by the Nextera XT method, as are regions of especially high GC-content relative to the rest of the MTB genome. As coverage bias is highly undesirable, this study illustrates the importance of appropriate protocol selection when performing NGS studies in order to ensure that sound inferences can be made regarding mycobacterial genomes.

Exact mapping of Illumina blind spots in theMycobacterium tuberculosisgenome reveals platform-wide and workflow-specific biases

“…Across all sequencing workflows, tandem repeats are the most problematic attribute, both in combination with other attributes ( Figure 5A), and in isolation ( Figure 5B). This result is consistent with the literature consensus that repeats introduce mapping ambiguity, creating assembly gaps and depleting coverage (6,12,49). It is unsurprising that this problem persists irrespective of library preparation or sequencing instrument, as it is inherent to short-read assembly.…”

“…Several genomic features cause biases that reduce coverage when sequenced on Illumina SBS technologies, particularly in segments of the genome where they are prevalent. The most wellcharacterized of these is GC content (4,6,7). GC bias originates primarily in library preparation, during amplification by polymerase chain reaction (PCR) (8).…”

“…This fracture in how the M. tuberculosis WGS community handles Illumina bias highlights a need for specific, empirically determined exclusion criteria. In an initial step towards this, Tyler and colleagues (6) studied the fidelity of Illumina-sequenced Mycobacterium tuberculosis genomes, and reported differential coverage bias between genomes prepared with Nextera and TruSeq library preps. They also reported that samples prepared with TruSeq resolved genomes into fewer contigs than Nextera and that certain regions, particularly GC-rich regions, could not be resolved with either library preparation method.…”

The variability and reproducibility of whole genome sequencing technology for detecting resistance to anti-tuberculous drugs

“…However, we did not investigate the impact of the different library preparation methods used (mechanical (MiSeq) and enzymatic (Ion PGM) processing). Sample quality and the mode or preparation have been shown to influence the depth of coverage in high GC regions [27], and further work is required to investigate this. The Ion PGM platform has previously been used to characterise mutations in XDR-TB strains [6], but the minimum read depth used to call alleles (fourfold) were less stringent than the tenfold coverage threshold adopted here.…”

Proteomics for the Investigation of Mycobacteria

“…It is worth noting that isolation of DNA and RNA from mycobacteria and further manipulations have been described in a large number of protocols [8-10] which are applied in various laboratories [11-16]. The situation with protein isolation, especially isolation of the total protein fraction required to obtain the proteome, is quite the opposite.…”