Tracking microbial pathogen transmission with high-accuracy nanopore sequencing

At the Nanopore Community Meeting 2024 (Boston, USA), Dr Samuel Shelburne (The University of Texas MD Anderson Cancer Center, USA) gave a talk about how a Master’s student changed his opinion about nanopore sequencing accuracy and its use for the investigation of potential healthcare-associated infection (HAI) transmission.

The impact of healthcare-associated infections

HAIs are defined as infections caused by pathogens acquired during a hospital stay. Samuel explained that these pathogens are important to track in order to inform infection prevention and control because they cause illness across millions of patients in the USA every year, resulting in lengthened hospital stays and increased mortality rates. This in turn creates a significant financial burden — ‘tens of millions of dollars can be at stake, in addition, of course, to the impact on patients’ lives’.

However, HAIs are becoming increasingly difficult to track due to complex modern healthcare systems and treatment plans that include more outpatients and individuals with multiple admissions. Samuel also highlighted that patients with HAIs will rarely display symptoms at the same time as other patients, with disease onset often occurring weeks or months after infection, further increasing the difficulty of tracking pathogens.

How pathogen transmissions go undetected

Currently, legacy short-read sequencing technology is used to investigate HAIs when there are unusually high outbreak patterns, taking approximately 70 days from sample to answer — a lengthy and expensive process that requires large sample batches. Despite not regularly sequencing HAIs, Samuel highlighted that the MD Anderson Cancer Center has previously sequenced large cohorts of pathogen isolates collected over many years to investigate molecular epidemiology. This showed clusters of highly genetically related strains and ‘identical isolates that are popping up in our patients, causing life-threatening infections — but some of them are as long as four, five years apart’. This illustrated that cryptic transmission of HAI pathogens had occurred undetected and over much longer periods than previously assumed.

Due to the long gaps between the infections, Samuel explained that the infection prevention and control (IPC) unit would struggle to investigate these outbreaks: ‘these patients were in the hospital several years apart. How do we even begin to investigate that?’. Furthermore, with the current legacy short-read sequencing process in place, results cannot be generated for every hospital patient fast enough for effective infection control, and the method is too costly to allow for routine sequencing.

Challenging preconceived ideas about nanopore sequencing

Chin-Ting Wu, a Master’s student with a background in infection control, approached Samuel because she wanted to use Oxford Nanopore sequencing to detect HAI pathogens and determine their potential transmission. Samuel was not convinced that this was a feasible idea because he thought the accuracy from nanopore-only data was too low and could only supplement legacy short-read sequencing data.

To challenge Samuel’s preconceptions, Chin-Ting showed him a paper from Sereika et al.¹, who had tested the newly released R10.4.1 Flow Cells. This publication highlighted the increased accuracy of nanopore sequencing data generated with this latest chemistry, enabling near-complete, closed microbial genomes. Since then, Oxford Nanopore data accuracy has further improved and complete, closed microbial genomes have been generated from nanopore-only sequencing reads².

Chin-Ting’s evidence convinced Samuel that it was worth investigating nanopore sequencing to find out if it could detect HAI pathogens; however, the lack of funding and laboratory staff were the next hurdles for the Master’s student. Nevertheless, Chin-Ting explained that she could do the sequencing herself — the workflows are simple and can be performed outside of a dedicated sequencing facility, and the MinION device and sequencing kits are cost efficient to use.

A further boost to nanopore sequencing data accuracy

The goals of the project were to develop an accurate workflow that took less than 72 hours from sample to answer using minimal resources and to generate data with sufficiently high accuracy to determine pathogen transmission. In a microbiology laboratory, Chin-Ting was able to sequence approximately ten isolates per week on a MinION sequencer in an average time of 24 hours. Furthermore, she was able to use the Oxford Nanopore consumables as efficiently as possible by reusing flow cells up to three times whilst still achieving 20 Gb of output per isolate, demonstrating the cost effectiveness of nanopore sequencing.

Samuel said that the sequencing worked ‘much better than I expected’. With an average of 80x coverage per isolate genome and 98.82–99.56% coverage of the alleles of interest, genetic relatedness was easily called between the isolates using nanopore data and the Dorado basecaller. Samuel explained that they started the project with the Guppy basecaller, but they switched to Dorado when it was released midway through the research project. The Dorado basecaller further boosted their sequencing accuracy, increasing their average Q score from 50.5 to 60, which was ‘really remarkable and statistically significant’. This accuracy boost from Dorado enabled them to confidently call potential transmission, which they could not achieve with Guppy.

The potential impact on pathogen transmission detection

Overall, Samuel ‘really thought this was remarkable data that Chin-Ting generated; that she can use Oxford Nanopore sequencing to efficiently and accurately assess genetic relatedness among major antimicrobial-resistant pathogens’. Using her nanopore workflow, she was able to achieve the research project goals, including developing a 24-hour workflow using minimal resources that generated data with very low error rates, enabling identification of likely pathogen transmission. From the nanopore data, she identified five genetically related clusters that ranged between 2–12 research samples and detected 21 pathogen strains that were potentially transmitted between patients — 15 of which had epidemiological links, suggesting undetected pathogen transmission occurred at the time of infection.

Samuel concluded that this was ‘remarkable data’ and suggested that nanopore sequencing could potentially be used to accurately and efficiently assess genetic relatedness among major antimicrobial-resistant pathogens. Samuel is now planning to utilise nanopore sequencing for several future research endeavours including developing an on-demand workflow that the MD Anderson Cancer Center could implement to detect HAI pathogen transmissions as well as respiratory viruses.

To learn more about pathogen surveillance, see the infectious disease application page.