Search

Extrachromosomal elements of bacterial cells such as plasmids are notorious for their importance in evolution and adaptation to changing ecology. However, high-resolution population-wide analysis of plasmids has only become accessible recently with the advent of scalable long-read sequencing technology. Current typing methods for the classification of plasmids remain limited in their scope which motivated us to develop a computationally efficient approach to simultaneously recognize novel types and classify plasmids into previously identified groups. Our method can easily handle thousands of input sequences which are compressed using a unitig representation in a de Bruijn graph. We provide an intuitive visualization, classification and clustering scheme that users can explore interactively. This provides a framework that can be easily distributed and replicated, enabling a consistent labelling of plasmids across past, present, and future sequence collections. We illustrate the attractive features of our approach by the analysis of population-wide plasmid data from the opportunistic pathogenEscherichia coliand the distribution of the colistin resistance genemcr-1.1in the plasmid population.

IntroductionThe human gut microbiota is an important reservoir of ESBL-producing Escherichia coli (ESBL-Ec). Community surveillance studies of ESBL-Ec to monitor circulating clones and ESBL genes are logistically challenging and costly.ObjectivesTo evaluate if isolates obtained in routine clinical practice can be used as an alternative to monitor the distribution of clones and ESBL genes circulating in the community.MethodsWGS was performed on 451 Dutch ESBL-Ec isolates (2014–17), including 162 community faeces and 289 urine and blood isolates. We compared proportions of 10 most frequently identified STs, PopPUNK-based sequence clusters (SCs) and ESBL gene subtypes and the degree of similarity using Czekanowski’s proportional similarity index (PSI).ResultsNine out of 10 most prevalent STs and SCs and 8/10 most prevalent ESBL genes in clinical ESBL-Ec were also the most common types in community faeces. The proportions of ST131 (39% versus 23%) and SC131 (40% versus 25%) were higher in clinical isolates than in community faeces (P ConclusionsDistributions of the 10 most prevalent clones and ESBL genes from ESBL-Ec community gut colonization and extra-intestinal infection overlapped in majority, indicating that isolates from routine clinical practice could be used to monitor ESBL-Ec clones and ESBL genes in the community.

Horizontal gene transfer (HGT) plays a critical role in the evolution and diversification of many microbial species. The resulting dynamics of gene gain and loss can have important implications for the development of antibiotic resistance and the design of vaccine and drug interventions. Methods for the analysis of gene presence/absence patterns typically do not account for errors introduced in the automated annotation and clustering of gene sequences. In particular, methods adapted from ecological studies, including the pangenome gene accumulation curve, can be misleading as they may reflect the underlying diversity in the temporal sampling of genomes rather than a difference in the dynamics of HGT. Here, we introduce Panstripe, a method based on Generalised Linear Regression that is robust to population structure, sampling bias and errors in the predicted presence/absence of genes. We demonstrate using simulations that Panstripe can effectively identify differences in the rate and number of genes involved in HGT events, and illustrate its capability by analysing several diverse bacterial genome datasets representing major human pathogens. Panstripe is freely available as an R package at https://github.com/gtonkinhill/panstripe.

Background Bacterial whole-genome sequencing based on short-read technologies often results in a draft assembly formed by contiguous sequences. The introduction of long-read sequencing technologies permits those contiguous sequences to be unambiguously bridged into complete genomes. However, the elevated costs associated with long-read sequencing frequently limit the number of bacterial isolates that can be long-read sequenced. Here we evaluated the recently released 96 barcoding kit from Oxford Nanopore Technologies (ONT) to generate complete genomes on a high-throughput basis. In addition, we propose an isolate selection strategy that optimizes a representative selection of isolates for long-read sequencing considering as input large-scale bacterial collections. Results Despite an uneven distribution of long reads per barcode, near-complete chromosomal sequences (assembly contiguity = 0.89) were generated for 96 Escherichia coli isolates with associated short-read sequencing data. The assembly contiguity of the plasmid replicons was even higher (0.98), which indicated the suitability of the multiplexing strategy for studies focused on resolving plasmid sequences. We benchmarked hybrid and ONT-only assemblies and showed that the combination of ONT sequencing data with short-read sequencing data is still highly desirable (i) to perform an unbiased selection of isolates for long-read sequencing, (ii) to achieve an optimal genome accuracy and completeness, and (iii) to include small plasmids underrepresented in the ONT library. Conclusions The proposed long-read isolate selection ensures the completion of bacterial genomes that span the genome diversity inherent in large collections of bacterial isolates. We show the potential of using this multiplexing approach to close bacterial genomes on a high-throughput basis.

Enterococcus faecium is a nosocomial, multidrug-resistant pathogen. Whole genome sequence studies revealed that hospital-associated E. faecium isolates are clustered in a separate clade A1. Here, we investigated the distribution, integration site and function of a putative iol gene cluster that encodes for myo-inositol (MI) catabolism. This iol gene cluster was found as part of an ~20 kbp genetic element (iol element), integrated in ICEEfm1 close to its integrase gene in E. faecium isolate E1679. Among 1644 E. faecium isolates, ICEEfm1 was found in 789/1227 (64.3 %) clade A1 and 3/417 (0.7 %) non-clade A1 isolates. The iol element was present at a similar integration site in 180/792 (22.7 %) ICEEfm1-containing isolates. Examination of the phylogenetic tree revealed genetically closely related isolates that differed in presence/absence of ICEEfm1 and/or iol element, suggesting either independent acquisition or loss of both elements. E. faecium iol gene cluster containing isolates E1679 and E1504 were able to grow in minimal medium with only myo-inositol as carbon source, while the iolD-deficient mutant in E1504 (E1504∆iolD) lost this ability and an iol gene cluster negative recipient strain gained this ability after acquisition of ICEEfm1 by conjugation from donor strain E1679. Gene expression profiling revealed that the iol gene cluster is only expressed in the absence of other carbon sources. In an intestinal colonization mouse model the colonization ability of E1504∆iolD mutant was not affected relative to the wild-type E1504 strain. In conclusion, we describe and functionally characterise a gene cluster involved in MI catabolism that is associated with the ICEEfm1 island in hospital-associated E. faecium isolates. We were unable to show that this gene cluster provides a competitive advantage during gut colonisation in a mouse model. Therefore, to what extent this gene cluster contributes to the spread and ecological specialisation of ICEEfm1-carrying hospital-associated isolates remains to be investigated.

Summary Plasmids can horizontally transmit genetic traits, enabling rapid bacterial adaptation to new environments and hosts. Short-read whole-genome sequencing data are often applied to large-scale bacterial comparative genomics projects but the reconstruction of plasmids from these data is facing severe limitations, such as the inability to distinguish plasmids from each other in a bacterial genome. We developed gplas, a new approach to reliably separate plasmid contigs into discrete components using sequence composition, coverage, assembly graph information and network partitioning based on a pruned network of plasmid unitigs. Gplas facilitates the analysis of large numbers of bacterial isolates and allows a detailed analysis of plasmid epidemiology based solely on short-read sequence data. Availability and implementation Gplas is written in R, Bash and uses a Snakemake pipeline as a workflow management system. Gplas is available under the GNU General Public License v3.0 at https://gitlab.com/sirarredondo/gplas.git. Supplementary information Supplementary data are available at Bioinformatics online.

The incidence of infections caused by multidrug-resistant E. coli strains has risen in the past years. Antibiotic resistance in E. coli is often mediated by acquisition and maintenance of plasmids. The study of E. coli plasmid epidemiology and genomics often requires long-read sequencing information, but recently a number of tools that allow plasmid prediction from short-read data have been developed. Here, we reviewed 25 available plasmid prediction tools and categorized them into binary plasmid/chromosome classification tools and plasmid reconstruction tools. We benchmarked six tools (MOB-suite, plasmidSPAdes, gplas, FishingForPlasmids, HyAsP and SCAPP) that aim to reliably reconstruct distinct plasmids, with a special focus on plasmids carrying antibiotic resistance genes (ARGs) such as extended-spectrum beta-lactamase genes. We found that two thirds (n = 425, 66.3%) of all plasmids were correctly reconstructed by at least one of the six tools, with a range of 92 (14.58%) to 317 (50.23%) correctly predicted plasmids. However, the majority of plasmids that carried antibiotic resistance genes (n = 85, 57.8%) could not be completely recovered as distinct plasmids by any of the tools. MOB-suite was the only tool that was able to correctly reconstruct the majority of plasmids (n = 317, 50.23%), and performed best at reconstructing large plasmids (n = 166, 46.37%) and ARG-plasmids (n = 41, 27.9%), but predictions frequently contained chromosome contamination (40%). In contrast, plasmidSPAdes reconstructed the highest fraction of plasmids smaller than 18 kbp (n = 168, 61.54%). Large ARG-plasmids, however, were frequently merged with sequences derived from distinct replicons. Available bioinformatic tools can provide valuable insight into E. coli plasmids, but also have important limitations. This work will serve as a guideline for selecting the most appropriate plasmid reconstruction tool for studies focusing on E. coli plasmids in the absence of long-read sequencing data.

Background Enterococcus faecium is a commensal of the gastrointestinal tract of animals and humans but also a causative agent of hospital-acquired infections. Resistance against glycopeptides and to vancomycin has motivated the inclusion of E. faecium in the WHO global priority list. Vancomycin resistance can be conferred by the vanA gene cluster on the transposon Tn1546, which is frequently present in plasmids. The vanA gene cluster can be disseminated clonally but also horizontally either by plasmid dissemination or by Tn1546 transposition between different genomic locations. Methods We performed a retrospective study of the genomic epidemiology of 309 vancomycin-resistant E. faecium (VRE) isolates across 32 Dutch hospitals (2012–2015). Genomic information regarding clonality and Tn1546 characterization was extracted using hierBAPS sequence clusters (SC) and TETyper, respectively. Plasmids were predicted using gplas in combination with a network approach based on shared k-mer content. Next, we conducted a pairwise comparison between isolates sharing a potential epidemiological link to elucidate whether clonal, plasmid, or Tn1546 spread accounted for vanA-type resistance dissemination. Results On average, we estimated that 59% of VRE cases with a potential epidemiological link were unrelated which was defined as VRE pairs with a distinct Tn1546 variant. Clonal dissemination accounted for 32% cases in which the same SC and Tn1546 variants were identified. Horizontal plasmid dissemination accounted for 7% of VRE cases, in which we observed VRE pairs belonging to a distinct SC but carrying an identical plasmid and Tn1546 variant. In 2% of cases, we observed the same Tn1546 variant in distinct SC and plasmid types which could be explained by mixed and consecutive events of clonal and plasmid dissemination. Conclusions In related VRE cases, the dissemination of the vanA gene cluster in Dutch hospitals between 2012 and 2015 was dominated by clonal spread. However, we also identified outbreak settings with high frequencies of plasmid dissemination in which the spread of resistance was mainly driven by horizontal gene transfer (HGT). This study demonstrates the feasibility of distinguishing between modes of dissemination with short-read data and provides a novel assessment to estimate the relative contribution of nested genomic elements in the dissemination of vanA-type resistance.

Enterococcus faecium is a gut commensal of the gastro-digestive tract, but also known as nosocomial pathogen among hospitalized patients. Population genetics based on whole-genome sequencing has revealed that E. faecium strains from hospitalized patients form a distinct clade, designated clade A1, and that plasmids are major contributors to the emergence of nosocomial E. faecium . Here we further explored the adaptive evolution of E. faecium using a genome-wide co-evolution study (GWES) to identify co-evolving single-nucleotide polymorphisms (SNPs). We identified three genomic regions harbouring large numbers of SNPs in tight linkage that are not proximal to each other based on the completely assembled chromosome of the clade A1 reference hospital isolate AUS0004. Close examination of these regions revealed that they are located at the borders of four different types of large-scale genomic rearrangements, insertion sites of two different genomic islands and an IS30-like transposon. In non-clade A1 isolates, these regions are adjacent to each other and they lack the insertions of the genomic islands and IS30-like transposon. Additionally, among the clade A1 isolates there is one group of pet isolates lacking the genomic rearrangement and insertion of the genomic islands, suggesting a distinct evolutionary trajectory. In silico analysis of the biological functions of the genes encoded in three regions revealed a common link to a stress response. This suggests that these rearrangements may reflect adaptation to the stringent conditions in the hospital environment, such as antibiotics and detergents, to which bacteria are exposed. In conclusion, to our knowledge, this is the first study using GWES to identify genomic rearrangements, suggesting that there is considerable untapped potential to unravel hidden evolutionary signals from population genomic data.

Enterococcus faecium is a gut commensal of the gastro-digestive tract, but also known as nosocomial pathogen among hospitalized patients. Population genetics based on whole-genome sequencing has revealed that E. faecium strains from hospitalized patients form a distinct clade, designated as clade A1 and that plasmids are major contributors to the emergence of nosocomial E. faecium. Here we further explored the adaptive evolution of E. faecium using a genome-wide co-evolution study (GWES) to identify co-evolving SNPs. We identified three genomic regions harboring large numbers of SNPs in tight linkage which are not proximal to each other based on the completely assembled chromosome of clade A1 reference hospital isolate AUS0004. Close examination of these regions revealed that they are located at the borders of four different types of large-scale genomic rearrangements, insertion sites of two different genomic islands and an IS30-like transposon. In non-clade A1 isolates, these regions are adjacent to each other and they lack the insertions of the genomic islands and IS30-like transposon. Additionally, among the clade A1 isolates there is one group of pet isolates lacking the genomic rearrangement and insertion of the genomic islands, suggesting a distinct evolutionary trajectory. In silico analysis of the biological functions of the genes encoded in three regions revealed a common link to a stress response. This suggests that these rearrangements may reflect adaptation to the stringent conditions in the hospital environment, such as antibiotics and detergents, to which bacteria are exposed. In conclusion, to our knowledge, this is the first study using GWES to identify genomic rearrangements, suggesting that there is considerable untapped potential to unravel hidden evolutionary signals from population genomic data.Impact statementEnterococcus faecium has emerged as an important nosocomial pathogen around the world. Population genetics revealed that clinical E. faecium strains form a distinct clade, designated as clade A1 and that plasmids are major contributors to the emergence of nosocomial E. faecium. Here, the adaptive evolution of E. faecium was further explored using an unsupervised machine learning method (SuperDCA) to identify genome-wide co-evolving SNPs. We identified three genomic regions harboring large numbers of SNPs in tight linkage which are separated by a large chromosomal distance in a clinical clade A1 reference isolate, but appeared adjacent to each other in non-clade A1 isolates. We identified four different types of large-scale genomic rearrangements and in all cases, we found insertion of two different genomic islands and an insertion element at the border. In contrast, no genomic rearrangement and insertions were identified among a group of clade A1 pet isolates, suggesting a distinct evolutionary trajectory. Based on the in silico predicted biological functions, we found a common link to a stress response for the genes encoded in three regions. This suggests that these rearrangements may reflect adaptation to the stringent conditions in the hospital environment, such as antibiotics and detergents, to which bacteria are exposed.Data summaryRaw core-genome alignment (1.1 MB, Harvest suite v1.1.2), including the 1,644 Clade A isolates and the complete E. faecium AUS0004 (accession number CP003351) as a reference is available under the following gitlab repository https://gitlab.com/sirarredondo/efm_gwes.

Plasmids are extrachromosomal elements that can disseminate between strains or even between different bacterial species. Plasmids can also acquire novel genetic traits by acquisition of transposons or cointegration with other plasmids thereby providing host bacterial strains novel adaptive traits. These inherent characteristics make plasmids optimal vectors for disseminating antimicrobial resistance (AMR). Plasmid-mediated AMR dissemination often follows a Russian-Doll model in which nested genomic elements intervene in resistance propagation by vertical and horizontal transmission. However, current epidemiological studies on AMR dissemination often solely focus on clonal outbreaks. Traditional typing methods such as MLST, but also current whole genome sequence (WGS)-based epidemiological studies fail to resolve the dynamics of plasmids and thus challenge the monitoring of plasmid-mediated AMR. A main reason for this focus on clonal dissemination is the challenge of reconstructing plasmid sequences from short-read WGS as explained in chapter 2. The scope of this thesis was to develop a new set of tools in order to overcome this limitation and accurately detect and trace plasmid sequences, from short-read WGS data, in the nosocomial pathogen Enterococcus faecium.This was especially relevant to investigate the dissemination of antimicrobial resistance (AMR), in particular vancomycin resistance, as well as other disseminated genes.

BackgroundEnterococcus faecium is a commensal of the gastrointestinal tract of animals and humans but also a causative agent of hospital-acquired infections. Resistance against glycopeptides and especially to vancomycin, a first-line antibiotic to treat infections with multidrug-resistant Gram-positive pathogens, has motivated the inclusion of E. faecium in the WHO global priority list. Vancomycin resistance can be conferred by the vanA gene cluster on the transposon Tn1546, which is frequently present in plasmids. The vanA gene cluster can be disseminated clonally but also horizontally either by plasmid dissemination or Tn1546 transposition between different genomic locations. Here, we reconstructed all nested genetic elements (clone, plasmid, transposon) to study how the dissemination of vanA-type vancomycin resistance occurred in Dutch hospitals (2012-2015).MethodsWe performed a retrospective study of the genomic epidemiology of 309 vancomycin-resistant E. faecium (VRE) isolates across 32 Dutch hospitals (2012-2015). Genomic information regarding clonality and Tn1546 characterisation was extracted using hierBAPS sequence clusters (SC) and TETyper, respectively. Plasmids were predicted using gplas in combination with a network approach based on shared k-mer content. This allowed determining all nested genomic elements (clone, plasmid and transposon) involved in the dissemination of the vanA gene cluster. Next, we conducted an “all vs. all” pairwise comparison between isolates sharinga potential epidemiological link to elucidate whether clonal, plasmid or Tn1546 spread accounted for the dissemination of vanA resistance.ResultsThe 309 VRE isolates belonged to 18 different SCs of which SC13 (n = 102, 33%), SC17 (n = 52, 16.8%) and SC18 (n = 42, 13.6%) were predominant. We identified seven different plasmid types bearing the vanA gene cluster, four of which were highly similar (identity ∼99%, coverage∼84%) to previously described complete plasmid sequences. We estimated that clonal dissemination contributed most (∼45%) to the spread of vancomycin-resistance in Dutch hospitals, followed by Tn1546 mobilisation (∼12%) and plasmid dissemination (∼6%).ConclusionsThe dissemination of the vanA gene cluster in Dutch hospitals between 2012 and 2015 was dominated by clonal spread. However, we also identified outbreak settings with high frequencies of Tn1546 transposition and/or plasmid dissemination in which the spread of resistance was mainly driven by horizontal gene transfer (HGT). This study demonstrates the feasibility of distinguishing between modes of dissemination with short-read data and provides one of the first quantitative assessments to estimate the relative contribution of nested genomic elements in the dissemination of vanA-type vancomycin resistance cluster.

Background The nosocomial pathogen Enterococcus faecium can survive for prolonged periods of time on surfaces in the absence of nutrients. This trait is thought to contribute to the ability of E. faecium to spread among patients in hospitals. There is currently a lack of data on the mechanisms that are responsible for the ability of E. faecium to survive in the absence of nutrients. Results We performed a high-throughput transposon mutant library screening (Tn-seq) to identify genes that have a role in long-term survival during incubation in phosphate-buffered saline (PBS) at 20 °C. A total of 24 genes were identified by Tn-seq to contribute to survival in PBS, with functions associated with the general stress response, DNA repair, metabolism, and membrane homeostasis. The gene which was quantitatively most important for survival in PBS was usp (locus tag: EfmE745_02439), which is predicted to encode a 17.4 kDa universal stress protein. After generating a targeted deletion mutant in usp, we were able to confirm that usp significantly contributes to survival in PBS and this defect was restored by in trans complementation. The usp gene is present in 99% of a set of 1644 E. faecium genomes that collectively span the diversity of the species. Conclusions We postulate that usp is a key determinant for the remarkable environmental robustness of E. faecium. Further mechanistic studies into usp and other genes identified in this study may shed further light on the mechanisms by which E. faecium can survive in the absence of nutrients for prolonged periods of time.

Background Whole genome sequence (WGS)-based strain typing finds increasing use in the epidemiologic analysis of bacterial pathogens in both public health as well as more localized infection control settings. Aims This minireview describes methodologic approaches that have been explored for WGS-based epidemiologic analysis and considers the challenges and pitfalls of data interpretation. Sources Personal collection of relevant publications. Content When applying WGS to study the molecular epidemiology of bacterial pathogens, genomic variability between strains is translated into measures of distance by determining single nucleotide polymorphisms in core genome alignments or by indexing allelic variation in hundreds to thousands of core genes, assigning types to unique allelic profiles. Interpreting isolate relatedness from these distances is highly organism specific, and attempts to establish species-specific cutoffs are unlikely to be generally applicable. In cases where single nucleotide polymorphism or core gene typing do not provide the resolution necessary for accurate assessment of the epidemiology of bacterial pathogens, inclusion of accessory gene or plasmid sequences may provide the additional required discrimination. Implications As with all epidemiologic analysis, realizing the full potential of the revolutionary advances in WGS-based approaches requires understanding and dealing with issues related to the fundamental steps of data generation and interpretation.

Enterococcus faecium is one of the most frequent nosocomial pathogens of hospital-acquired infections. E. faecium has gained resistance against most commonly available antibiotics, most notably, against ampicillin, gentamicin, and vancomycin, which renders infections difficult to treat. Many antibiotic resistance traits, in particular, vancomycin resistance, can be encoded in autonomous and extrachromosomal elements called plasmids. These sequences can be disseminated to other isolates by horizontal gene transfer and confer novel mechanisms to source specificity. In our study, we elucidated the total plasmid content, referred to as the plasmidome, of 1,644 E. faecium isolates by using short- and long-read whole-genome technologies with the combination of a machine-learning classifier. This was fundamental to investigate the full collection of plasmid sequences present in our collection (pan-plasmidome) and to observe the potential transfer of plasmid sequences between E. faecium hosts. We observed that E. faecium isolates from hospitalized patients carried a larger number of plasmid sequences compared to that from other sources, and they elucidated different configurations of plasmidome populations in the hospital environment. We assessed the contribution of different genomic components and observed that plasmid sequences have the highest contribution to source specificity. Our study suggests that E. faecium plasmids are regulated by complex ecological constraints rather than physical interaction between hosts., Enterococcus faecium is a gut commensal of humans and animals but is also listed on the WHO global priority list of multidrug-resistant pathogens. Many of its antibiotic resistance traits reside on plasmids and have the potential to be disseminated by horizontal gene transfer. Here, we present the first comprehensive population-wide analysis of the pan-plasmidome of a clinically important bacterium, by whole-genome sequence analysis of 1,644 isolates from hospital, commensal, and animal sources of E. faecium. Long-read sequencing on a selection of isolates resulted in the completion of 305 plasmids that exhibited high levels of sequence modularity. We further investigated the entirety of all plasmids of each isolate (plasmidome) using a combination of short-read sequencing and machine-learning classifiers. Clustering of the plasmid sequences unraveled different E. faecium populations with a clear association with hospitalized patient isolates, suggesting different optimal configurations of plasmids in the hospital environment. The characterization of these populations allowed us to identify common mechanisms of plasmid stabilization such as toxin-antitoxin systems and genes exclusively present in particular plasmidome populations exemplified by copper resistance, phosphotransferase systems, or bacteriocin genes potentially involved in niche adaptation. Based on the distribution of k-mer distances between isolates, we concluded that plasmidomes rather than chromosomes are most informative for source specificity of E. faecium.

Enterococcus faecium is a gut commensal of many mammals but is also recognized as a major nosocomial human pathogen, as it is listed on the WHO global priority list of multi-drug resistant organisms. Previous research has suggested that nosocomial strains have multiple zoonotic origins and are only distantly related to those involved in human commensal colonization. Here we present the first comprehensive population-wide joint genomic analysis of hospital, commensal and animal isolates using both short- and long-read sequencing techniques. This enabled us to investigate the population plasmidome, core genome variation and genome architecture in detail, using a combination of machine learning, population genomics and genome-wide co-evolution analysis. We observed a high level of genome plasticity with large-scale inversions and heterogeneous chromosome sizes, collectively painting a high-resolution picture of the adaptive landscape of E. faecium, and identified plasmids as the main indicator for host-specificity. Given the increasing availability of long-read sequencing technologies, our approach could be widely applied to other human and animal pathogen populations to unravel fine-scale mechanisms of their evolution.

Assembly of bacterial short-read whole genome sequencing (WGS) data frequently results in hundreds of contigs for which the origin, plasmid or chromosome, is unclear. Long-read sequencing has emerged as a solution to resolve plasmid structures and to obtain complete genomes for most bacterial species. This information can be used to generate and label datasets from short-read based contigs as plasmid- or chromosome-derived. We investigated the use of several popular machine learning methods to classify short-read contigs with known plasmid- or chromosome-origin from Enterococcus faecium, Klebsiella pneumoniae and Escherichia coli using pentamer frequencies. Based on resulting F1-scores we selected support-vector machine (SVM) models as best classifier for all three bacterial species (F1-score E. faecium = 0.94, F1-score K. pneumoniae = 0.90, F1-score E. coli = 0.76), which outperformed other existing plasmid tools using an independent set of isolates (precision E. faecium = 0.92, precision K. pneumoniae = 0.86, precision E. coli = 0.82). We demonstrated the scalability of our model by accurately predicting the plasmidome of a large collection of 1,644 E. faecium isolates with only short-read WGS available using a standard laptop with a single core. A low number of false positive predicted sequences suggests that the assignment of a particular gene of interest as plasmid- or chromosome-encoded by the models is plausible. The SVM classifiers are publicly available as a new R package called ‘mlplasmids’ at https://gitlab.com/sirarredondo/mlplasmids under the GNU General Public License v3.0. We additionally developed a graphical-user interface using the Shiny package which can be accessed at https://sarredondo.shinyapps.io/mlplasmids/. Single genomes can easily be predicted by uploading genome assemblies. We anticipate that this tool may significantly facilitate research on the dissemination of plasmids encoding antibiotic resistance and/or contributing to host adaptation.

To benchmark algorithms for automated plasmid sequence reconstruction from short-read sequencing data, we selected 42 publicly available complete bacterial genome sequences spanning 12 genera, containing 148 plasmids. We predicted plasmids from short-read data with four programs (PlasmidSPAdes, Recycler, cBar and PlasmidFinder) and compared the outcome to the reference sequences. PlasmidSPAdes reconstructs plasmids based on coverage differences in the assembly graph. It reconstructed most of the reference plasmids (recall=0.82), but approximately a quarter of the predicted plasmid contigs were false positives (precision=0.75). PlasmidSPAdes merged 84 % of the predictions from genomes with multiple plasmids into a single bin. Recycler searches the assembly graph for sub-graphs corresponding to circular sequences and correctly predicted small plasmids, but failed with long plasmids (recall=0.12, precision=0.30). cBar, which applies pentamer frequency analysis to detect plasmid-derived contigs, showed a recall and precision of 0.76 and 0.62, respectively. However, cBar categorizes contigs as plasmid-derived and does not bin the different plasmids. PlasmidFinder, which searches for replicons, had the highest precision (1.0), but was restricted by the contents of its database and the contig length obtained from de novo assembly (recall=0.36). PlasmidSPAdes and Recycler detected putative small plasmids (50 kbp) containing repeated sequences remains challenging and limits the high-throughput analysis of plasmids from short-read whole-genome sequencing data.

Plasmids are autonomous extra-chromosomal elements in bacterial cells that can carry genes that are important for bacterial survival. To benchmark algorithms for automated plasmid sequence reconstruction from short read sequencing data, we selected 42 publicly available complete bacterial genome sequences which were assembled by a combination of long- and short-read data. The selected bacterial genome sequence projects span 12 genera, containing 148 plasmids. We predicted plasmids from short-read data with four different programs (PlasmidSPAdes, Recycler, cBar and PlasmidFinder) and compared the outcome to the reference sequences.PlasmidSPAdes reconstructs plasmids based on coverage differences in the assembly graph. It reconstructed most of the reference plasmids (recall = 0.82) but approximately a quarter of the predicted plasmid contigs were false positives (precision = 0.76). PlasmidSPAdes merged 83 % of the predictions from genomes with multiple plasmids in a single bin. Recycler searches the assembly graph for sub-graphs corresponding to circular sequences and correctly predicted small plasmids but failed with long plasmids (recall = 0.12, precision = 0.30). cBar, which applies pentamer frequency composition analysis to detect plasmid-derived contigs, showed an overall recall and precision of 0.78 and 0.64. However, cBar only categorizes contigs as plasmid-derived and does not bin the different plasmids correctly within a bacterial isolate. PlasmidFinder, which searches for matches in a replicon database, had the highest precision (1.0) but was restricted by the contents of its database and the contig length obtained from de novo assembly (recall = 0.36).Surprisingly, PlasmidSPAdes and Recycler detected single isolated components corresponding to putative novel small plasmids (This study shows that it is possible to automatically predict plasmid sequences, but only for small plasmids. The reconstruction of large plasmids (>50 kbp) containing repeated sequences remains challenging and limits the high-throughput analysis of WGS data.Author SummaryShort read sequencing of the DNA of bacteria is often used to understand characteristics such as antibiotic resistance. However the assembly of short read sequencing data with the goal of reconstructing a complete genome is often fragmented and leaves gaps. Therefore independently replicating DNA fragments called plasmids cannot easily be identified from an assembly. Lately a number of programs have been developed to enable the automated prediction of the sequences of plasmids. Here we tested these programs by comparing their outcomes with complete genome sequences. None of the tested programs were able to fully and unambiguously predict distinct plasmid sequences. All programs performed best with the prediction of plasmids smaller than 50 kbp. Larger plasmids were only correctly predicted if they were present as a single contig in the assembly. While predictions by PlasmidSPAdes and cBar contained most of the plasmids, they were merged with or indistinguishable from other plasmids and sometimes chromosome sequences. PlasmidFinder missed most plasmids but all its predictions were correct. Without manual steps or long-read sequencing information, plasmid reconstruction from short read sequencing data remains challenging.