Nanopore-based enrichment of antimicrobial resistance genes – a case-based study

Abstract

Rapid screening of hospital admissions to detect asymptomatic carriers of resistant bacteria can prevent pathogen outbreaks. However, the resulting isolates rarely have their genome sequenced due to cost constraints and long turn-around times to get and process the data, limiting their usefulness to the practitioner. Here we used real-time, on-device target enrichment (“adaptive”) sequencing as a highly multiplexed assay covering 1,147 antimicrobial resistance genes. We compared its utility against standard and metagenomic sequencing, focusing on an isolate of Raoultella ornithinolytica harbouring three carbapenemases (NDM, KPC, VIM). Based on this experimental data, we then modelled the influence of several variables on the enrichment results and predicted the large effect of nucleotide identity (higher is better) and read length (shorter is better). Lastly, we showed how all relevant resistance genes are detected using adaptive sequencing on a miniature (“Flongle”) flow cell, motivating its use in a clinical setting to monitor similar cases and their surroundings.

Graph-based Nanopore Adaptive Sampling with GNASTy enables sensitive pneumococcal serotyping in complex samples

“…However, accepted and rejected reads in adaptive channels for CBL enrichment were notably shorter than those in whole genome enrichment, whilst control channel read lengths were similar (Supplementary Table 2). This effect may be due to accepting of shorter reads when aligning to shorter targets, as observed in Viehweger et al ( 33 ). Size selection also had a positive significant effect on enrichment when targeting CBL, increasing the average enrichment from 3.4 to 4.8 (Supplementary Figure 6).…”

“…Supplementary Table 2 We additionally tested the effectiveness of size selection on enrichment of CBL. Previously, Viehweger et al (33) showed that enrichment of loci, in this case AMR genes, was possible using NAS, although performance was dependent on the library fragment length. Paradoxically, the authors suggested that when enriching for genes, libraries with a greater proportion of short fragments provide better enrichment, in contrast to our observation that size selection improves enrichment.…”

“…Paradoxically, the authors suggested that when enriching for genes, libraries with a greater proportion of short fragments provide better enrichment, in contrast to our observation that size selection improves enrichment. Viehweger et al ( 33 ) postulated that this was due to the higher probability of a short target locus being found in the middle of a longer read, rather than at the start, which is the region aligned during NAS. In longer DNA fragments, it is therefore more likely that the start region of a read will align outside of the target locus, and will therefore be incorrectly rejected, despite the target sequence being present later in the read.…”

“…However, there has been limited quantification of its accuracy, particularly in metagenomics. It has been previously shown that NAS sensitivity is highest when a target present in a metagenome is closely related to a sequence in the reference database ( 31 , 33 ). However, high genetic relatedness between non-target and target taxa in the same sample has the potential to negatively impact NAS specificity, as target and non-target reads will be more difficult to distinguish between during the rejection process.…”

Nanopore adaptive sampling effectively enriches bacterial plasmids

“…Recently, deep-learning-based tools like SquiggleNet and DeepSelectNet have also been developed, addressing host depletion in human microbiome samples (19,20). The potential enrichment reached by using adaptive sampling was already shown, and even mathematical models that predict the enrichment factor were recently described by some groups (17,21,22). In one study, Marquet et In the present study, we investigate the efficiency of adaptive sampling to enrich plasmid sequences in five different bacterial isolates.…”