Assessing the heterogeneity of in silico plasmid predictions based on whole-genome-sequenced clinical isolates

One day in Denmark: whole-genome sequence-based analysis of Escherichia coli isolates from clinical settings

BACKGROUND

WGS can potentially be routinely used in clinical microbiology settings, especially with the increase in sequencing accuracy and decrease in cost. Escherichia coli is the most common bacterial species analysed in those settings, thus fast and accurate diagnostics can lead to reductions in morbidity, mortality and healthcare costs.

OBJECTIVES

To evaluate WGS for diagnostics and surveillance in a collection of clinical E. coli; to examine the pool of antimicrobial resistance (AMR) determinants circulating in Denmark and the most frequent STs; and to evaluate core-genome MLST (cgMLST) and SNP-based clustering approaches for detecting genetically related isolates.

METHODS

We analysed the genomes of 699 E. coli isolates collected throughout all Danish Clinical Microbiology Laboratories. We used rMLST and KmerFinder for species identification, ResFinder for prediction of AMR, and PlasmidFinder for plasmid identification. We used Center for Genomic Epidemiology MLST, cgMLSTFinder and CSI Phylogeny to perform typing and clustering analysis.

RESULTS

Genetic AMR determinants were detected in 56.2% of isolates. We identified 182 MLSTs, most frequently ST-69, ST-73, ST-95 and ST-131. Using a maximum 15-allele difference as the threshold for genetic relatedness, we identified 23 clusters. SNP-based phylogenetic analysis within clusters revealed from 0 to 13 SNPs, except two cases with 111 and 461 SNPs.

CONCLUSIONS

WGS data are useful to characterize clinical E. coli isolates, including predicting AMR profiles and subtyping in concordance with surveillance data. We have shown that it is possible to adequately cluster isolates through a cgMLST approach, but it remains necessary to define proper interpretative criteria.

Detection of resistance and virulence plasmids in Campylobacter coli and Campylobacter jejuni isolated from North Carolina food animal production, 2018-2019

Campylobacter remains the leading cause of bacterial foodborne illness in the U.S. and worldwide. Campylobacter plasmids may play a significant role in antimicrobial resistance (AMR) and virulence factor distribution, and potentially drive rapid adaptation. C. coli (n = 345) and C. jejuni (n = 199) isolates collected from live cattle, swine, turkey, and chickens, poultry carcasses at production, and retail meat in N.C. were analyzed to determine plasmid prevalence, extrachromosomal virulence and AMR genes, and the phylogeny of assembled plasmids. Putative plasmids ranging from <2 kb to 237kb were identified with virulence factors present in 66.1% (228/345) C. coli and 88.4% (176/199) C. jejuni plasmids (promoting adherence, invasion, exotoxin production, immune modulation, chemotaxis, mobility, and the type IV secretion system). AMR genes were identified in 21.2% (73/345) C. coli and 28.1% C. jejuni plasmids (conferring resistance to tetracyclines, aminoglycosides, beta-lactams, nucleosides, and lincosamides). Megaplasmids (>100 kb) were present in 25.7% (140/544) of the isolates and carried genes previously recognized to be involved with interspecies recombination. Our study highlights the extensive distribution and diversity of Campylobacter plasmids in food animal production and their role in the dissemination of biomedically important genes. Characterizing Campylobacter plasmids within the food animal production niche is important to understanding the epidemiology of potential emerging strains.

Empirically derived sequence similarity thresholds to study the genomic epidemiology of plasmids shared among healthcare-associated bacterial pathogens

BACKGROUND

Healthcare-associated bacterial pathogens frequently carry plasmids that contribute to antibiotic resistance and virulence. The horizontal transfer of plasmids in healthcare settings has been previously documented, but genomic and epidemiologic methods to study this phenomenon remain underdeveloped. The objectives of this study were to apply whole-genome sequencing to systematically resolve and track plasmids carried by nosocomial pathogens in a single hospital, and to identify epidemiologic links that indicated likely horizontal plasmid transfer.

METHODS

We performed an observational study of plasmids circulating among bacterial isolates infecting patients at a large hospital. We first examined plasmids carried by isolates sampled from the same patient over time and isolates that caused clonal outbreaks in the same hospital to develop thresholds with which horizontal plasmid transfer within a tertiary hospital could be inferred. We then applied those sequence similarity thresholds to perform a systematic screen of 3074 genomes of nosocomial bacterial isolates from a single hospital for the presence of 89 plasmids. We also collected and reviewed data from electronic health records for evidence of geotemporal links between patients infected with bacteria encoding plasmids of interest.

FINDINGS

Our analyses determined that 95% of analyzed genomes maintained roughly 95% of their plasmid genetic content and accumulated fewer than 15 SNPs per 100 kb of plasmid sequence. Applying these similarity thresholds to identify horizontal plasmid transfer identified 45 plasmids that potentially circulated among clinical isolates. Ten highly preserved plasmids met criteria for geotemporal links associated with horizontal transfer. Several plasmids with shared backbones also encoded different additional mobile genetic element content, and these elements were variably present among the sampled clinical isolate genomes.

INTERPRETATION

Evidence suggests that the horizontal transfer of plasmids among nosocomial bacterial pathogens appears to be frequent within hospitals and can be monitored with whole genome sequencing and comparative genomics approaches. These approaches should incorporate both nucleotide identity and reference sequence coverage to study the dynamics of plasmid transfer in the hospital.

FUNDING

This research was supported by the US National Institute of Allergy and Infectious Disease (NIAID) and the University of Pittsburgh School of Medicine.

Plasmids of the urinary microbiota

Studies of the last decade have identified a phylogenetically diverse community of bacteria within the urinary tract of individuals with and without urinary symptoms. Mobile genetic elements (MGEs), including plasmids and phages, within this niche have only recently begun to be explored. These MGEs can expand metabolic capacity and increase virulence, as well as confer antibiotic resistance. As such, they have the potential to contribute to urinary symptoms. While plasmids for some of the bacterial taxa found within the urinary microbiota (urobiome) have been well characterized, many urinary species are under-studied with few genomes sequenced to date. Using a two-pronged bioinformatic approach, we have conducted a comprehensive investigation of the plasmid content of urinary isolates representative of 102 species. The bioinformatic tools plasmidSPAdes and Recycler were used in tandem to identify plasmid sequences from raw short-read sequence data followed by manual curation. In total, we identified 603 high-confidence plasmid sequences in 20 different genera of the urobiome. In total, 70 % of these high-confidence plasmids exhibit sequence similarity to plasmid sequences from the gut. This observation is primarily driven by plasmids from E. coli , which is found in both anatomical niches. To confirm our bioinformatic predictions, long-read sequencing was performed for 23 of the E. coli isolates in addition to two E. coli strains that were sequenced as part of a prior study. Overall, 66.95 % of these predictions were confirmed highlighting the strengths and weaknesses of current bioinformatic tools. Future studies of the urobiome, especially concerning under-studied species in the urobiome, should employ long-read sequencing to expand the catalogue of plasmids for this niche.

OUP accepted manuscript

Despite recent methodology and reference database improvements for taxonomic profiling tools, metagenomic assembly and genomic binning remain important pillars of metagenomic analysis workflows. In case reference information is lacking, genomic binning is considered to be a state-of-the-art method in mixed culture metagenomic data analysis. In this light, our previously published tool BusyBee Web implements a composition-based binning method efficient enough to function as a rapid online utility. Handling assembled contigs and long nanopore generated reads alike, the webserver provides a wide range of supplementary annotations and visualizations. Half a decade after the initial publication, we revisited existing functionality, added comprehensive visualizations, and increased the number of data analysis customization options for further experimentation. The webserver now allows for visualization-supported differential analysis of samples, which is computationally expensive and typically only performed in coverage-based binning methods. Further, users may now optionally check their uploaded samples for plasmid sequences using PLSDB as a reference database. Lastly, a new application programming interface with a supporting python package was implemented, to allow power users fully automated access to the resource and integration into existing workflows. The webserver is freely available under: https://www.ccb.uni-saarland.de/busybee.

Genetic evaluation of ESBL-producing Escherichia coli urinary isolates in Otago, New Zealand

OBJECTIVES

The incidence of infections with ESBL-producing Escherichia coli (ESBL-Ec) in New Zealand is increasing. ESBL-Ec most commonly cause urinary tract infections and are seen in both community and hospitalized patients. The reason for the increasing incidence of ESBL-Ec infections is unknown.

METHODS

In this study, 65 urinary ESBL-Ec isolates from the Otago region in 2015 were fully genetically characterized to understand the mechanisms of transmission. The ESBL gene, E. coli STs, plasmid types and genetic context (e.g. insertion sequences) of ESBL genes were determined by a combination of whole genome and plasmid sequencing. The phylogenetic relationships of the isolates were compared with ESBL-Ec isolates sequenced as part of the 2016 nationwide survey.

RESULTS

Significant diversity of E. coli strains, plasmids, and the genetic context of ESBL genes was seen. However, there was evidence of common mobile genetic elements in unrelated ESBL-Ec.

CONCLUSIONS

Multiple introductions of ESBL resistance genes or resistant bacterial strains with limited horizontal transmission of mobile genetic elements accounts for the increased incidence of ESBL-Ec in this low prevalence area. Future studies should investigate modes of transmission of ESBL-Ec in the Otago region.

A genomic epidemiological study shows that prevalence of antimicrobial resistance in Enterobacterales is associated with the livestock host, as well as antimicrobial usage

Enterobacterales from livestock are potentially important reservoirs for antimicrobial resistance (AMR) to pass through the food chain to humans, thereby increasing the AMR burden and affecting our ability to tackle infections. In this study 168 isolates from four genera of the order Enterobacterales , primarily Escherichia coli , were purified from livestock (cattle, pigs and sheep) faeces from 14 farms in the United Kingdom. Their genomes were resolved using long- and short-read sequencing to analyse AMR genes and their genetic context, as well as to explore the relationship between AMR burden and on-farm antimicrobial usage (AMU), in the three months prior to sampling. Although E. coli isolates were genomically diverse, phylogenetic analysis using a core-genome SNP tree indicated pig isolates to generally be distinct from sheep isolates, with cattle isolates being intermediates. Approximately 28 % of isolates harboured AMR genes, with the greatest proportion detected in pigs, followed by cattle then sheep; pig isolates also harboured the highest number of AMR genes per isolate. Although 90 % of sequenced isolates harboured diverse plasmids, only 11 % of plasmids (n =58 out of 522) identified contained AMR genes, with 91 % of AMR plasmids being from pig, 9 % from cattle and none from sheep isolates; these results indicated that pigs were a principle reservoir of AMR genes harboured by plasmids and likely to be involved in their horizontal transfer. Significant associations were observed between AMU (mg kg−1) and AMR. As both the total and the numbers of different antimicrobial classes used on-farm increased, the risk of multi-drug resistance (MDR) in isolates rose. However, even when AMU on pig farms was comparatively low, pig isolates had increased likelihood of being MDR; harbouring relatively more resistances than those from other livestock species. Therefore, our results indicate that AMR prevalence in livestock is not only influenced by recent AMU on-farm but also livestock-related factors, which can influence the AMR burden in these reservoirs and its plasmid mediated transmission.

Outcome of Different Sequencing and Assembly Approaches on the Detection of Plasmids and Localization of Antimicrobial Resistance Genes in Commensal Escherichia coli

Antimicrobial resistance (AMR) is a major threat to public health worldwide. Currently, AMR typing changes from phenotypic testing to whole-genome sequence (WGS)-based detection of resistance determinants for a better understanding of the isolate diversity and elements involved in gene transmission (e.g., plasmids, bacteriophages, transposons). However, the use of WGS data in monitoring purposes requires suitable techniques, standardized parameters and approved guidelines for reliable AMR gene detection and prediction of their association with mobile genetic elements (plasmids). In this study, different sequencing and assembly strategies were tested for their suitability in AMR monitoring in Escherichia coli in the routines of the German National Reference Laboratory for Antimicrobial Resistances. To assess the outcomes of the different approaches, results from in silico predictions were compared with conventional phenotypic- and genotypic-typing data. With the focus on (fluoro)quinolone-resistant E.coli, five qnrS-positive isolates with multiple extrachromosomal elements were subjected to WGS with NextSeq (Illumina), PacBio (Pacific BioSciences) and ONT (Oxford Nanopore) for in depth characterization of the qnrS1-carrying plasmids. Raw reads from short- and long-read sequencing were assembled individually by Unicycler or Flye or a combination of both (hybrid assembly). The generated contigs were subjected to bioinformatics analysis. Based on the generated data, assembly of long-read sequences are error prone and can yield in a loss of small plasmid genomes. In contrast, short-read sequencing was shown to be insufficient for the prediction of a linkage of AMR genes (e.g., qnrS1) to specific plasmid sequences. Furthermore, short-read sequencing failed to detect certain duplications and was unsuitable for genome finishing. Overall, the hybrid assembly led to the most comprehensive typing results, especially in predicting associations of AMR genes and mobile genetic elements. Thus, the use of different sequencing technologies and hybrid assemblies currently represents the best approach for reliable AMR typing and risk assessment.

Genomic Analysis of Antimicrobial Resistance and Resistance Plasmids in Salmonella Serovars from Poultry in Nigeria

Antimicrobial resistance is a global public health concern, and resistance genes in Salmonella, especially those located on mobile genetic elements, are part of the problem. This study used phenotypic and genomic methods to identify antimicrobial resistance and resistance genes, as well as the plasmids that bear them, in Salmonella isolates obtained from poultry in Nigeria. Seventy-four isolates were tested for susceptibility to eleven commonly used antimicrobials. Plasmid reconstruction and identification of resistance and virulence genes were performed with a draft genome using in silico approaches in parallel with plasmid extraction. Phenotypic resistance to ciprofloxacin (50.0%), gentamicin (48.6%), nalidixic acid (79.7%), sulphonamides (71.6%) and tetracycline (59.5%) was the most observed. Antibiotic resistance genes (ARGs) detected in genomes corresponded well with these observations. Commonly observed ARGs included sul1, sul2, sul3, tet (A), tet (M), qnrS1, qnrB19 and a variety of aminoglycoside-modifying genes, in addition to point mutations in the gyrA and parC genes. Multiple ARGs were predicted to be located on IncN and IncQ1 plasmids of S. Schwarzengrund and S. Muenster, and most qnrB19 genes were carried by Col (pHAD28) plasmids. Seventy-two percent (19/24) of S. Kentucky strains carried multidrug ARGs located in two distinct variants of Salmonella genomic island I. The majority of strains carried full SPI-1 and SPI-2 islands, suggesting full virulence potential.

Detection of extended-spectrum beta-lactamase (ESBL) genes and plasmid replicons in Enterobacteriaceae using PlasmidSPAdes assembly of short-read sequence data

Knowledge of the epidemiology of plasmids is essential for understanding the evolution and spread of antimicrobial resistance. PlasmidSPAdes attempts to reconstruct plasmids using short-read sequence data. Accurate detection of extended-spectrum beta-lactamase (ESBL) genes and plasmid replicon genes is a prerequisite for the use of plasmid assembly tools to investigate the role of plasmids in the spread and evolution of ESBL production in Enterobacteriaceae. This study evaluated the performance of PlasmidSPAdes plasmid assembly for Enterobacteriaceae in terms of detection of ESBL-encoding genes, plasmid replicons and chromosomal wgMLST genes, and assessed the effect of k-mer size. Short-read sequence data for 59 ESBL-producing Enterobacteriaceae were assembled with PlasmidSPAdes using different k-mer sizes (21, 33, 55, 77, 99 and 127). For every k-mer size, the presence of ESBL genes, plasmid replicons and a selection of chromosomal wgMLST genes in the plasmid assembly was determined. Out of 241 plasmid replicons and 66 ESBL genes detected by whole-genome assembly, 213 plasmid replicons [88 %; 95 % confidence interval (CI): 83.9-91.9] and 43 ESBL genes (65 %; 95 % CI: 53.1-75.6) were detected in the plasmid assemblies obtained by PlasmidSPAdes. For most ESBL genes (83.3 %) and plasmid replicons (72.0 %), detection results did not differ between the k-mer sizes used in the plasmid assembly. No optimal k-mer size could be defined for the number of ESBL genes and plasmid replicons detected. For most isolates, the number of chromosomal wgMLST genes detected in the plasmid assemblies decreased with increasing k-mer size. Based on our findings, PlasmidSPAdes is not a suitable plasmid assembly tool for short-read sequence data for ESBL-encoding plasmids of Enterobacteriaceae.

Reconstruction of plasmids by shotgun sequencing from environmental DNA: which bioinformatic workflow?

Plasmids play important roles in microbial evolution and also in the spread of antibiotic resistance. Plasmid sequences are extensively studied from clinical isolates but rarely from the environment with a metagenomic approach focused on the plasmid fraction referred to as the plasmidome. A clear challenge in this context is to define a workflow for discriminating plasmids from chromosomal contaminants existing in the plasmidome. For this purpose, we benchmarked existing tools from assembly to detection of the plasmids by reference-free methods (cBar and PlasFlow) and database-guided approaches. Our simulations took into account short-reads alone or combined with moderate long-reads like those actually generated in environmental genomics experiments. This benchmark allowed us to select the best tools for limiting false-positives associated to plasmid prediction tools and a combination of reference-guided methods based on plasmid and bacterial databases.

Plasmid ATLAS: plasmid visual analytics and identification in high-throughput sequencing data

Plasmid ATLAS (pATLAS, http://www.patlas.site) provides an easy-to-use web accessible database with visual analytics tools to explore the relationships of plasmids available in NCBI's RefSeq database. pATLAS has two main goals: (i) to provide an easy way to search for plasmids deposited in NCBI RefSeq and their associated metadata; (ii) to visualize the relationships of plasmids in a graph, allowing the exploration of plasmid evolution. pATLAS allows searching by plasmid name, bacterial host taxa, antibiotic resistance and virulence genes, plasmid families, and by sequence length and similarity. pATLAS is also able to represent in the plasmid network, plasmid sets identified by external pipelines using mapping, mash screen or assembly from high-throughput sequencing data. By representing the identified hits within the network of relationships between plasmids, allowing the possibility of removing redundant results, and by taking advantage of the browsing capabilities of pATLAS, users can more easily interpret the pipelines' results. All these analyses can be saved to a JSON file for sharing and future re-evaluation. Furthermore, by offering a REST-API, the pATLAS database and network display are easily accessible by other interfaces or pipelines.

PlaScope: a targeted approach to assess the plasmidome from genome assemblies at the species level

Plasmid prediction may be of great interest when studying bacteria of medical importance such as Enterobacteriaceae as well as Staphylococcus aureus or Enterococcus. Indeed, many resistance and virulence genes are located on such replicons with major impact in terms of pathogenicity and spreading capacities. Beyond strain outbreak, plasmid outbreaks have been reported in particular for some extended-spectrum beta-lactamase- or carbapenemase-producing Enterobacteriaceae. Several tools are now available to explore the ‘plasmidome’ from whole-genome sequences with various approaches, but none of them are able to combine high sensitivity and specificity. With this in mind, we developed PlaScope, a targeted approach to recover plasmidic sequences in genome assemblies at the species or genus level. Based on Centrifuge, a metagenomic classifier, and a custom database containing complete sequences of chromosomes and plasmids from various curated databases, PlaScope classifies contigs from an assembly according to their predicted location. Compared to other plasmid classifiers, PlasFlow and cBar, it achieves better recall (0.87), specificity (0.99), precision (0.96) and accuracy (0.98) on a dataset of 70 genomes of Escherichia coli containing plasmids. In a second part, we identified 20 of the 21 chromosomal integrations of the extended-spectrum beta-lactamase coding gene in a clinical dataset of E. coli strains. In addition, we predicted virulence gene and operon locations in agreement with the literature. We also built a database for Klebsiella and correctly assigned the location for the majority of resistance genes from a collection of 12 Klebsiella pneumoniae strains. Similar approaches could also be developed for other well-characterized bacteria.

Reconsidering plasmid maintenance factors for computational plasmid design

Plasmids are genetic parasites of microorganisms. The genomes of naturally occurring plasmids are expected to be polished via natural selection to achieve long-term persistence in the microbial cell population. However, plasmid genomes are extremely diverse, and the rules governing plasmid genomes are not fully understood. Therefore, computationally designing plasmid genomes optimized for model and nonmodel organisms remains challenging. Here, we summarize current knowledge of the plasmid genome organization and the factors that can affect plasmid persistence, with the aim of constructing synthetic plasmids for use in gram-negative bacteria. Then, we introduce publicly available resources, plasmid data, and bioinformatics tools that are useful for computational plasmid design.

Various Sequence Types of Enterobacteriaceae Isolated from Commercial Chicken Farms in China and Carrying the blaNDM-5 Gene

A total of 108 meropenem-resistant Enterobacteriaceae isolates were obtained from 1,658 rectal swabs collected from 15 unrelated commercial chicken farms in China between 2014 and 2016. These samples yielded 16 Escherichia coli and 2 Klebsiella pneumoniae isolates of diverse sequence types carrying a bla_NDM-5-bearing IncX3 plasmid. K. pneumoniae strain sequence type 709 (ST709) has two bla_NDM-5-carrying plasmids that were transferred together to E.coli.

MOB-suite: software tools for clustering, reconstruction and typing of plasmids from draft assemblies

Large-scale bacterial population genetics studies are now routine due to cost-effective Illumina short-read sequencing. However, analysing plasmid content remains difficult due to incomplete assembly of plasmids. Bacterial isolates can contain any number of plasmids and assembly remains complicated due to the presence of repetitive elements. Numerous tools have been developed to analyse plasmids but the performance and functionality of the tools are variable. The MOB-suite was developed as a set of modular tools for reconstruction and typing of plasmids from draft assembly data to facilitate characterization of plasmids. Using a set of closed genomes with publicly available Illumina data, the MOB-suite identified contigs of plasmid origin with both high sensitivity and specificity (95 and 88 %, respectively). In comparison, plasmidfinder demonstrated high specificity (99 %) but limited sensitivity (50 %). Using the same dataset of 377 known plasmids, MOB-recon accurately reconstructed 207 plasmids so that they were assigned to a single grouping without other plasmid or chromosomal sequences, whereas plasmidSPAdes was only able to accurately reconstruct 102 plasmids. In general, plasmidSPAdes has a tendency to merge different plasmids together, with 208 plasmids undergoing merge events. The MOB-suite reduces the number of errors but produces more hybrid plasmids, with 84 plasmids undergoing both splits and merges. The MOB-suite also provides replicon typing similar to plasmidfinder but with the inclusion of relaxase typing and prediction of conjugation potential. The MOB-suite is written in Python 3 and is available from https://github.com/phac-nml/mob-suite.

PlasmidTron: assembling the cause of phenotypes and genotypes from NGS data

Increasingly rich metadata are now being linked to samples that have been whole-genome sequenced. However, much of this information is ignored. This is because linking this metadata to genes, or regions of the genome, usually relies on knowing the gene sequence(s) responsible for the particular trait being measured and looking for its presence or absence in that genome. Examples of this would be the spread of antimicrobial resistance genes carried on mobile genetic elements (MGEs). However, although it is possible to routinely identify the resistance gene, identifying the unknown MGE upon which it is carried can be much more difficult if the starting point is short-read whole-genome sequence data. The reason for this is that MGEs are often full of repeats and so assemble poorly, leading to fragmented consensus sequences. Since mobile DNA, which can carry many clinically and ecologically important genes, has a different evolutionary history from the host, its distribution across the host population will, by definition, be independent of the host phylogeny. It is possible to use this phenomenon in a genome-wide association study to identify both the genes associated with the specific trait and also the DNA linked to that gene, for example the flanking sequence of the plasmid vector on which it is encoded, which follows the same patterns of distribution as the marker gene/sequence itself. We present PlasmidTron, which utilizes the phenotypic data normally available in bacterial population studies, such as antibiograms, virulence factors, or geographical information, to identify traits that are likely to be present on DNA that can randomly reassort across defined bacterial populations. It is also possible to use this methodology to associate unknown genes/sequences (e.g. plasmid backbones) with a specific molecular signature or marker (e.g. resistance gene presence or absence) using PlasmidTron. PlasmidTron uses a k-mer-based approach to identify reads associated with a phylogenetically unlinked phenotype. These reads are then assembled de novo to produce contigs in a fast and scalable-to-large manner. PlasmidTron is written in Python 3 and is available under the open source licence GNU GPL3 from https://github.com/sanger-pathogens/plasmidtron.