RFPlasmid: Predicting plasmid sequences from short read assembly data using machine learning

Abstract

Antimicrobial resistance (AMR) genes in bacteria are often carried on plasmids and these plasmids can transfer AMR genes between bacteria. For molecular epidemiology purposes and risk assessment, it is important to know if the genes are located on highly transferable plasmids or in the more stable chromosomes. However, draft whole genome sequences are fragmented, making it difficult to discriminate plasmid and chromosomal contigs. Current methods that predict plasmid sequences from draft genome sequences rely on single features, like k-mer composition, circularity of the DNA molecule, copy number or sequence identity to plasmid replication genes, all of which have their drawbacks, especially when faced with large single copy plasmids, which often carry resistance genes. With our newly developed prediction tool RFPlasmid, we use a combination of multiple features, including k-mer composition and databases with plasmid and chromosomal marker proteins, to predict if the likely source of a contig is plasmid or chromosomal. The tool RFPlasmid supports models for 17 different bacterial species, including Campylobacter, E. coli, and Salmonella, and has a species agnostic model for metagenomic assemblies or unsupported organisms. RFPlasmid is available both as standalone tool and via web interface.

Recovering Escherichia coli Plasmids in the Absence of Long-Read Sequencing Data

“…Mlplasmids outperformed both cBAR and PlasFlow when classifying data derived from whole-genome sequencing experiments, and it can also accurately predict the plasmid localization of several antimicrobial resistance genes [ 29 ]. RFPlasmid [ 32 ], a recently released tool, uses a random forest classifier trained with a hybrid approach by identifying chromosomal and plasmids marker genes using two databases and also pentamer frequencies. This tool also works with metagenomic assemblies, albeit only for contigs from the 17 different species for which classifiers were trained.…”

Re-Emergence of Salmonellosis in Hog Farms: Outbreak and Bacteriological Characterization

“…All regions identified as non-core were compared with previously identified complete plasmids from an NCBI-based plasmid database [ 44 ] with BLAST [ 45 ]. In addition, RFPlasmid ( , accessed on 12 April 2021) was used to predict whether a contig in each assembly was likely part of a plasmid [ 46 ].…”

Identification and Characterization of a Novel Genomic Island Harboring Cadmium and Arsenic Resistance Genes in Listeria welshimeri

“…Circular maps were drawn with CGView Server ( (accessed on 18 March 2021)) [ 35 ]. The contigs did not include plasmid-associated contigs identified via RFPlasmid ( (accessed on 18 March 2021)) [ 33 ], were re-ordered with Mauve Contig Mover [ 34 ], and were manually inspected as described in Materials and Methods. The genomes were annotated with Prokka [ 32 ], yielding genetic features that are shown in the circular maps as rectangular boxes in the two outermost circles separated by the dark gray line.…”

“…The contigs were annotated with Prokka (version 1.14.5; included in the Docker image available at https://hub.docker.com/r/staphb/prokka (accessed on 18 March 2021)) [32]. Plasmid-borne contigs were identified using RFPlasmid (http://klif.uu.nl/rfplasmid/ (accessed on 18 March 2021)) [33] and eliminated from further analysis. The plasmid-free contigs were re-ordered with Mauve Contig Mover included in Mauve (version 20150226 build 10 (c)) [34] against the complete genome of L. welshimeri SLCC5334 (accession no.…”

“…To calculate the size and GC content of the plasmid-free re-ordered contigs, all contigs were concatenated into a single sequence, and bioawk (https://github.com/lh3/bioawk (accessed on 18 March 2021)) was used to calculate genome size and GC content. The contigs did not include plasmid-associated contigs identified via RFPlasmid (http://klif.uu.nl/rfplasmid/ (accessed on 18 March 2021)) [33], were re-ordered with Mauve Contig Mover [34], and were manually inspected as described in Materials and Methods. The genomes were annotated with Prokka [32], yielding genetic features that are shown in the circular maps as rectangular boxes in the two outermost circles separated by the dark gray line.…”