plasmidVerify classifies contigs (output of metaplasmidSPAdes or other assemblers) as plasmidic or non-plasmidic, based on gene content.
plasmidVerify predicts genes in the contig using Prodigal in the metagenomic mode, runs hmmsearch on the predicted proteins and classifies the contig as plasmidic or chromosomal by applying the Naive Bayes classifier (NBC). For the set of predicted HMMs, plasmidVerify uses trained NBC to classify this set to be plasmidic or chromosomal.
Since plasmidVerify is designed to separate chromosome and plasmid fragments, it does not always work correctly when there are viruses in the metagenome.
To obtain more accurate results for plasmid analysis, we recommend using our novel tool viralVerify with the -p option.
plasmidVerify is a Python script, thus, installation is not required. However, it has the following dependencies:
To work properly, plasmidVerify require Prodigal and hmmsearch in your PATH environment variable.
You can verify your output by BLAST to check if you found novel plasmid. In this case, you need to have blastn in your $PATH, Biopython installed, and provide a path to the local copy of nt database.
./plasmidverify.py
-f Input fasta file
-o output_directory
--hmm HMM Path to Pfam-A HMM database
Optional arguments:
-h, --help Show the help message and exit
--db DB Run BLAST on input contigs against provided dbOutput file: comma-separated table
Output format: contig name, prediction result, log-likelihood ratio, list of predicted HMMs
Please cite plasmidVeirfy as follows:
Antipov, D., Raiko, M., Lapidus, A., & Pevzner, P. A. (2019). Plasmid detection and assembly in genomic and metagenomic data sets. Genome Research, 29(6), 961-968.