Whole-genome and targeted sequencing of drug-resistant Mycobacterium tuberculosis on the iSeq100 and MiSeq: A performance, ease-of-use, and cost evaluation

Abstract

Background

Accurate, comprehensive, and timely detection of drug-resistant tuberculosis (TB) is essential to inform patient treatment and enable public health surveillance. This is crucial for effective control of TB globally. Whole-genome sequencing (WGS) and targeted next-generation sequencing (NGS) approaches have potential as rapid in vitro diagnostics (IVDs), but the complexity of workflows, interpretation of results, high costs, and vulnerability of instrumentation have been barriers to broad uptake outside of reference laboratories, especially in low- and middle-income countries. A new, solid-state, tabletop sequencing instrument, Illumina iSeq100, has the potential to decentralize NGS for individual patient care.

Methods and findings

In this study, we evaluated WGS and targeted NGS for TB on both the new iSeq100 and the widely used MiSeq (both manufactured by Illumina) and compared sequencing performance, costs, and usability. We utilized DNA libraries produced from Mycobacterium tuberculosis clinical isolates for the evaluation. We conducted WGS on three strains and observed equivalent uniform genome coverage with both platforms and found the depth of coverage obtained was consistent with the expected data output. Utilizing the standardized, cloud-based ReSeqTB bioinformatics pipeline for variant analysis, we found the two platforms to have 94.0% (CI 93.1%–94.8%) agreement, in comparison to 97.6% (CI 97%–98.1%) agreement for the same libraries on two MiSeq instruments. For the targeted NGS approach, 46 M. tuberculosis–specific amplicon libraries had 99.6% (CI 98.0%–99.9%) agreement between the iSeq100 and MiSeq data sets in drug resistance–associated SNPs. The upfront capital costs are almost 5-fold lower for the iSeq100 ($19,900 USD) platform in comparison to the MiSeq ($99,000 USD); however, because of difference in the batching capabilities, the price per sample for WGS was higher on the iSeq100. For WGS of M. tuberculosis at the minimum depth of coverage of 30x, the cost per sample on the iSeq100 was $69.44 USD versus $28.21 USD on the MiSeq, assuming a 2 × 150 bp run on a v3 kit. In terms of ease of use, the sequencing workflow of iSeq100 has been optimized to only require 27 minutes total of hands-on time pre- and post-run, and the maintenance is simplified by a single-use cartridge–based fluidic system. As these are the first sequencing attempts on the iSeq100 for M. tuberculosis, the sequencing pool loading concentration still needs optimization, which will affect sequencing error and depth of coverage. Additionally, the costs are based on current equipment and reagent costs, which are subject to change.

Conclusions

The iSeq100 instrument is capable of running existing TB WGS and targeted NGS library preparations with comparable accuracy to the MiSeq. The iSeq100 has reduced sequencing workflow hands-on time and is able to deliver sequencing results in <24 hours. Reduced capital and maintenance costs and lower-throughput capabilities also give the iSeq100 an advantage over MiSeq in settings of individualized care but not in high-throughput settings such as reference laboratories, where sample batching can be optimized to minimize cost at the expense of workflow complexity and time.

Abstract

Rebecca E. Colman and colleagues assess performance, cost, and throughput of a new sequencer designed for detecting drug-resistant tuberculosis strains.

Author summary

Why was this study done?

• M. tuberculosis is the leading cause of death from a single infectious agent.

• Diagnosis of drug-resistant M. tuberculosis is imperative for reducing tuberculosis (TB) mortality.

• Next-generation sequencing (NGS) can provide comprehensive genetic information on drug resistance for supporting rapid clinical decision-making but currently can only be performed at the reference-laboratory level.

• Illumina’s solid-state benchtop sequencer iSeq100 adds diversity to the intended use settings in which NGS can be practically utilized for drug-resistant TB detection, but it has not yet been demonstrated that existing TB NGS workflows will perform on the new instrument.

What did the researchers do and find?

• The study was focused on the performance of existing TB whole-genome sequencing (WGS) and targeted NGS approaches that work on the Illumina MiSeq to determine if they could be also utilized on the iSeq100 platforms.

• WGS library preparations were performed on three strains, and the same libraries were run on two MiSeq instruments and one early-access iSeq100 instrument. The percentage of genome covered was similar across the instruments and the three isolates, ranging from 99.20% to 99.92% of the genome covered.

• Targeted NGS preparations were performed on 46 strains and run on two MiSeq instruments and one iSeq100 instrument with 99.6% (CI. 98.0%–99.9%) agreement on drug resistance–associated SNPs between iSeq100 and MiSeq data sets.

• Because of batching-based pricing constraints for TB sequencing, the sequencing cost per sample is higher on the iSeq100 platform than for the same coverage on the MiSeq when batched optimally. However, the iSeq100 allows for a more open-access workflow due to a lower number of samples being in one sequencing batch. This relationship is the same for WGS and targeted NGS, although the scale is different depending on the sequencing approach chosen.

What do these findings mean?

• In settings of individualized TB care, lower throughput will drive a different need for batching, and the iSeq100, with its lower-throughput capabilities, has an advantage over MiSeq. However, in high-throughput settings such as reference laboratories, where sample batching can be optimized to minimize cost at the expense of workflow complexity and time, the MiSeq will make the most sense.