mlplasmids: a user-friendly tool to predict plasmid- and chromosome-derived sequences for single species

Abstract

Assembly of bacterial short-read whole-genome sequencing data frequently results in hundreds of contigs for which the origin, plasmid or chromosome, is unclear. Complete genomes resolved by long-read sequencing can be used to generate and label short-read contigs. These were used to train several popular machine learning methods to classify the origin of contigs from Enterococcus faecium, Klebsiella pneumoniae and Escherichia coli using pentamer frequencies. We selected support-vector machine (SVM) models as the best classifier for all three bacterial species (F1-score E. faecium=0.92, F1-score K. pneumoniae=0.90, F1-score E. coli=0.76), which outperformed other existing plasmid prediction tools using a benchmarking set of isolates. We demonstrated the scalability of our models by accurately predicting the plasmidome of a large collection of 1644 E. faecium isolates and illustrate its applicability by predicting the location of antibiotic-resistance genes in all three species. The SVM classifiers are publicly available as an R package and graphical-user interface called ‘mlplasmids’. We anticipate that this tool may significantly facilitate research on the dissemination of plasmids encoding antibiotic resistance and/or contributing to host adaptation.

Optimised use of Oxford Nanopore Flowcells for Hybrid Assemblies

“…There was however a major discrepancy in one sample where a ~ 868kb region was called as chromosomal in library 3 and a circularised super-plasmid-like component in library 4 ( figure 4). As expected, ML plasmids [15] predicted with high confidence (97% probability) this contig was of chromosomal origin. Interestingly this error was fixed after filtering with Filtlong, suggesting it may have arisen from low quality reads.…”

“…Minimap2 was used to map contigs from long-read to hybrid assemblies. ML plasmids [15] was used as a further arbitrator of the chromosomal/plasmid origin of sequences. Human reads were detected using Centrifuge [16] as part of the Crumpit [17] pipeline.…”

Complete assembly ofEscherichia coliST131 genomes using long reads demonstrates antibiotic resistance gene variation within diverse plasmid and chromosomal contexts

“…Contigs were classified as chromosomal or plasmid-derived using mlplasmids given a probability threshold of 60% [22], with further screening for plasmid-related gene content using MARA, CARD and PlasmidFinder (Supplementary Table 2). The largest plasmid was a 156.3 Kb IncFIA one in VREC1073, its sole plasmid.…”

“…Assembly graphs were visualized with Bandage [33]. The resulting contigs in each assembly were classified as chromosomal or plasmid using machine learning algorithms implemented in mlplasmids [22].…”

“…Complex transposable units (TUs) consisting of multiple TEs or ISs can mobilise AMR genes by transposition, and this can sometimes be followed by recombination within the TU between one of the inverted repeats (IRs) flanking the TE and the IR of another local TE or an adjacent homologous sequence, resulting in different TU structures, locations and copy numbers. At present, the exact resolution of complex structural rearrangements of repetitive TUs containing AMR genes may be impossible with short reads [22]. Consequently, plasmid assembly is a challenge requiring accurate long reads and sufficient coverage to distinguish between independent plasmids with regions of sequence identity [21,23].…”