Nanopore adaptive sampling effectively enriches bacterial plasmids

Abstract

Bacterial plasmids play a major role in the spread of antibiotic resistance genes. However, their characterization via DNA sequencing suffers from the low abundance of plasmid DNA in those samples. Although sample preparation methods can enrich the proportion of plasmid DNA before sequencing, these methods are expensive and laborious, and they might introduce a bias by enriching only for specific plasmid DNA sequences. Nanopore adaptive sampling could overcome these issues by rejecting uninteresting DNA molecules during the sequencing process. In this study, we assess the application of adaptive sampling for the enrichment of low-abundant plasmids in known bacterial isolates using two different adaptive sampling tools. We show that a significant enrichment can be achieved even on expired flow cells. By applying adaptive sampling, we also improve the quality of de novo plasmid assemblies and reduce the sequencing time. However, our experiments also highlight issues with adaptive sampling if target and non-target sequences span similar regions.

Importance: Antimicrobial resistance causes millions of deaths every year. Mobile genetic elements like bacterial plasmids are key drivers for the dissemination of antimicrobial resistance genes. This makes the characterization of plasmids via DNA sequencing an important tool for clinical microbiologists. Since plasmids are often underrepresented in bacterial samples, plasmid sequencing can be challenging and laborious. To accelerate the sequencing process, we evaluate nanopore adaptive sampling as an in silico method for the enrichment of low-abundant plasmids. Our results show the potential of this cost-efficient method for future plasmid research but also indicate issues that arise from using reference sequences.

Keywords: adaptive sampling; bacteria; enrichment; nanopore sequencing; plasmid; readuntil.

Fig 1

Comparison of flow cell yield in terms of sequenced base pairs and reads after 24 hours. (a) Yield in megabases for each flow cell separated by control and adaptive sampling region (depletion). (b) Yield in megabases for each flow cell region separated by plasmid and chromosome. (c) Number of sequenced reads for each flow cell separated by control and adaptive sampling region (depletion). (d) Number of sequenced reads for each flow cell region separated by plasmid and chromosome.

Fig 2

Comparison of plasmid yield in megabases for each flow cell regarding sequenced base pairs after 24 hours. There is a small increase in plasmid yield for the two Campylobacter samples from (a) ReadBouncer1 and (b) MinKNOW1. (c) Plasmid yield is increased for all three bacterial samples from ReadBouncer2. (d) Plasmid yield is increased for two of the three samples from flow cell MinKNOW2. There is a decreased plasmid yield for K. pneumoniae with adaptive sampling using MinKNOW.

Fig 3

Comparison of relative plasmid abundances in five bacterial samples. Adaptive sampling with MinKNOW was used on flow cells MinKNOW1 and MinKNOW2, and ReadBouncer was used as an adaptive sampling tool on flow cells ReadBouncer1 and ReadBouncer2. For all experiments, plasmid abundances for each sample were measured after 24 hours of sequencing for control regions and adaptive sampling regions (depletion). Plasmid abundances are highest when using MinKNOW for the depletion of chromosomal nanopore reads.

Fig 4

Scatterplots for relative plasmid enrichment by composition and yield. (a) Observed enrichment factor by composition against relative abundance. Each point represents a bacterial sample, with the position on the x-axis indicating the original relative abundance of plasmids in the sample and the position on the y-axis indicating the enrichment factor obtained. Points above the dashed line indicate enrichment, and points below the line indicate plasmid sequence depletion. (b) Correlation between observed enrichment values by composition and predicted enrichment values by the mathematical model (Pearson’s r of 0.55). (c) Enrichment factor by yield against relative abundance. Relative enrichment of plasmids is independent of the plasmid abundance in the sample. (d) Correlation between observed enrichment values by yield and predicted enrichment values by the mathematical model (Pearson’s r of −0.07). The model fails to predict relative plasmid enrichment by yield.

Fig 5

Comparison of enrichment in five bacterial samples. (a) Enrichment by the number of plasmid reads for the five bacterial strains across all four sequencing runs. (b) Enrichment by the number of sequenced plasmid bases for all five bacterial strains across the four sequencing runs. (c) Enrichment by mean depth of coverage of plasmid references for the five bacterial strains across the four sequencing runs. The dashed line indicates the enrichment factor threshold, which values above 1.0, implying effective enrichment. Values below 1.0 imply depletion of plasmid sequences. All strains but the Klebsiella pneumoniae sample show a slight enrichment, where MinKNOW was used for adaptive sampling.