Platon: Plasmid contig classification and characterization.

Author: Oliver Schwengers (oliver.schwengers@computational.bio.uni-giessen.de)

Contents

• Description
• Input/Output
• Installation
• Usage
• Examples
• Database
• Dependencies
• Citation

Description

Platon classifies contigs from bacterial WGS short read assemblies as plasmid or chromosome contigs, i.e. they either originate from a plasmid or a chromosome, respectively. Therefore, Platon takes advantage of pre-computed protein distribution statistics and computes mean protein scores (MPS) for each contig and finally tests them against certain thresholds. Contigs below a sensitivity threshold get classified as chromosome, contigs above a specifivity threshold get classified as plasmid. Contigs which protein score lies in between these thresholds get comprehensively characterized and finally classified following an heuristic approach.

In detail Platon conducts three analysis steps. First, it predicts open reading frames and searches the coding sequences against a database of marker genes. These are based on the NCBI RefSeq PCLA clusters to which we automatically pre-computed individual protein scores capturing the probability on which kind of replicon a certain protein is rather to be found on, i.e. on a plasmid or a chromosome. Platon then calculates the MPS for each contig and either classifies them as chromosome if the MPS is below a sensitivity cutoff (counting for 95 % sensitivity) or as plasmid if the MPS is above a specificity cutoff, counting for 99.99 % specificity. These threshold have been calculated by Monte Carlo simulations of artifical contigs created from closed RefSeq chromosome and plasmid sequences. In a second step contigs passing the sensitivity filter get comprehensivley characterized. Hereby, Platon tries to circularize the contig sequences, searches for rRNA, replication, mobilization and conjugation genes as well as incompatibility group DNA probes and finally performs a BLAST search against a plasmid database. In a third step, Platon finally classifies all remaining contigs based on an heuristic approach, i.e. a decision tree of simple rules exploiting all information at hand.

Input/Output

Input

Platon accepts draft genomes in fasta format. If contigs have been assembled with SPAdes, Platon is able to extract the coverage information stored in contigs names.

Output

Contigs classified as plasmid sequences are printed as tab separated values to STDOUT comprising the following columns:

• Contig ID
• Length
• Coverage
• # ORFs
• Protein Score
• Circularity
• Incompatibility Type(s)
• # Replication Genes
• # Mobilization Genes
• # Conjugation Genes
• # rRNA Genes
• # Plasmid Database Hits

Additionally, Platon writes the following files into the output directory:

• <prefix>.plasmid.fasta: contigs classified as plasmids or plasmodal origin
• <prefix>.chromosome.fasta: contigs classified as chromosomal origin
• <prefix>.tsv: dense information as printed to STDOUT (see above)
• <prefix>.json: comprehensive results and information on each single plasmid contig. All files are prefixed (<prefix>) as the input genome fasta file.

Installation

Platon can be installed/used in 3 different ways.

Additionally, a separate database must be downloaded which we provide for download as a zipped tarball: https://s3.computational.bio.uni-giessen.de/swift/v1/platon/db.tar.gz

GitHub

1. clone the latest version of the repository
2. download and extract the database

Example:

git clone git@github.com:oschwengers/platon.git
wget db `path`
tar -xzf db.tar.gz
rm db.tar.gz
platon/bin/platon --db ./db ...

If you move the extracted database directory into the platon directory, PLATON will automatically recognise it. In this case, the database path doesn't need to be specified:

git clone git@github.com:oschwengers/platon.git
wget https://s3.computational.bio.uni-giessen.de/swift/v1/platon/db.tar.gz
tar -xzf db.tar.gz
rm db.tar.gz
mv db $PLATON_HOME
platon/bin/platon ...

Pip

1. install PLATON per pip
2. download and extract the database

Example:

pip3 install cb-platon
wget https://s3.computational.bio.uni-giessen.de/swift/v1/platon/db.tar.gz
tar -xzf db.tar.gz
rm db.tar.gz
platon --db ./db ...

Docker

1. download our Docker shell wrapper script
2. download and extract the database

wget https://raw.githubusercontent.com/oschwengers/platon/master/platon-docker.sh
wget https://s3.computational.bio.uni-giessen.de/swift/v1/platon/db.tar.gz
tar -xzf db.tar.gz
rm db.tar.gz
platon-docker.sh ./db <genome>

Alternatively, just use the Docker image (oschwengers/platon) in order to ease the setup process.

Usage

Usage:

usage: platon [-h] [--threads THREADS] [--verbose] [--output OUTPUT]
              [--version]
              <genome>

Plasmid contig classification and characterization

positional arguments:
  <genome>              draft genome in fasta format

optional arguments:
  -h, --help            show this help message and exit
  --threads THREADS, -t THREADS
                        number of threads to use (default = number of
                        available CPUs)
  --verbose, -v         print verbose information
  --output OUTPUT, -o OUTPUT
                        output directory (default = current working directory)
  --version             show program's version number and exit

Examples

Simple:

platon ecoli.fasta

Expert: writing results to results directory with verbose output using 8 threads:

platon --output ./results --verbose --threads 8 ecoli.fasta

With Docker shell script:

platon-docker.sh <PLATON_DB> <genome>

Database

Platon depends on a custom database based on NCBI RefSeq nonredundant proteins (NRP), PCLA clusters, RefSeq Plasmid database, PlasmidFinder db as well as custom HMM models. These databases (RefSeq release 90) can be downloaded here: (zipped 1.8, unzipped 2.5 Gb) www.lorem.ipsum

Dependencies

Platon was developed and tested on Python 3.5. It depends on BioPython (1.71).

Additionally, it depends on the following 3rd party executables:

• Prodigal (2.6.3) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2848648 https://github.com/hyattpd/Prodigal
• Ghostz (1.0.1) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4393512 http://www.bi.cs.titech.ac.jp/ghostz
• Blast+ (2.7.1) https://www.ncbi.nlm.nih.gov/pubmed/2231712 https://blast.ncbi.nlm.nih.gov
• MUMmer (4.0.0-beta2) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC395750/ https://github.com/gmarcais/mummer
• INFERNAL (1.1.2) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3810854 http://eddylab.org/infernal

Platon has been tested against aforementioned software versions.

Citation

A manuscript is in preparation... stay tuned!