ARIBA

Antimicrobial Resistance Identification By Assembly

For how to use ARIBA, please see the ARIBA wiki page.

PLEASE NOTE: we currently do not have the resources to provide support for Ariba - see the Feedback/Issues section.

Introduction
Quick Start
Installation
Temporary files
Usage
License
Feedback/Issues
Citation

Introduction

ARIBA is a tool that identifies antibiotic resistance genes by running local assemblies. It can also be used for MLST calling.

The input is a FASTA file of reference sequences (can be a mix of genes and noncoding sequences) and paired sequencing reads. ARIBA reports which of the reference sequences were found, plus detailed information on the quality of the assemblies and any variants between the sequencing reads and the reference sequences.

Quick Start

Get reference data, for instance from CARD. See getref for a full list.

ariba getref ncbi out.ncbi

Prepare reference data for ARIBA:

ariba prepareref -f out.ncbi.fa -m out.ncbi.tsv out.ncbi.prepareref

Run local assemblies and call variants:

ariba run out.ncbi.prepareref reads1.fastq reads2.fastq out.run

Summarise data from several runs:

ariba summary out.summary out.run1/report1.tsv out.run2/report2.tsv out.run3/report3.tsv

Please read the ARIBA wiki page for full usage instructions.

Tutorials

The Jupyter notebook tutorial

Installation

If you encounter an issue when installing ARIBA please contact your local system administrator. If you encounter a bug you can log it here.

Required dependencies

Python3 version >= 3.6.0
Bowtie2 version >= 2.1.0
CD-HIT version >= 4.6
MUMmer version >= 3.23

ARIBA also depends on several Python packages, all of which are available via pip. Installing ARIBA with pip3 will get these automatically if they are not already installed:

dendropy >= 4.2.0
matplotlib>=3.1.0
pyfastaq >= 3.12.0
pysam >= 0.9.1
pymummer >= 0.10.1
biopython

Using pip3

Install ARIBA using pip:

pip3 install ariba

From Source

Download the latest release from this github repository or clone it. Run the tests:

python3 setup.py test

Note for OS X: The tests require gawk which will need to be installed separately, e.g. via Homebrew.

If the tests all pass, install:

python3 setup.py install

Alternatively, install directly from github using:

pip3 install git+https://github.com/sanger-pathogens/ariba.git #--user

Docker

ARIBA can be run in a Docker container. First install Docker, then install the latest version of ARIBA:

docker pull gchr.io/sanger-pathogens/ariba:latest

All Docker images are listed in the packages page.

To use ARIBA use a command like this (substituting in your directories), where your files are assumed to be stored in /home/ubuntu/data:

docker run --rm -it -v /home/ubuntu/data:/data sangerpathogens/ariba ariba -h

When calling Ariba via Docker (as above) you’ll also need to add /data/ in front of all the passed in file or directory names (e.g. /data/my_output_folder).

Singularity

ARIBA can be run in a Singularity container. First install Singularity. Releases include a Singularity image to download.

Alternatively, build your own Singularity image:

singularity build ariba.simg Singularity.def

Debian (Ariba version may not be the latest)

ARIBA is available in the latest version of Debian, and over time will progressively filter through to Ubuntu and other distributions which use Debian. To install it as root:

sudo apt-get install ariba

Ubuntu

You can use apt-get (see above), or to ensure you get the latest version of ARIBA, the following commands can be used to install ARIBA and its dependencies. This was tested on a new instance of Ubuntu 16.04.

sudo  apt-get update
sudo apt-get install -y python3-dev python3-pip python3-tk zlib1g-dev bowtie2 mummer cd-hit
export ARIBA_CDHIT=cdhit-est
sudo pip3 install ariba

Dependencies and environment variables

By default, ARIBA will look for the dependencies in your $PATH, using the names in the table below. This behaviour can be overridden and point ARIBA to a specific program using environment variables. The environment variable is checked first and is used if it is set. Otherwise ARIBA looks in your $PATH for the default name. This applies to the following dependencies.

Dependency	Default executable	Environment variable name
Bowtie2	`bowtie2`	`$ARIBA_BOWTIE2`
CD-HIT (est)	`cd-hit-est`	`$ARIBA_CDHIT`

For example, you could specify an exact version of a bowtie2 executable that you compiled and downloaded in your home directory (assuming BASH):

export ARIBA_BOWTIE2=$HOME/bowtie2-2.1.0/bowtie2

Note that ARIBA also runs bowtie2-build, for which it uses the bowtie2 executable with -build appended. So in this case it would try to use

$HOME/bowtie2-2.1.0/bowtie2-build

Temporary files

ARIBA can temporarily make a large number of files whilst running, which are put in a temporary directory made by ARIBA. The total size of these files is small, but there can be a many of them. This can be a problem when running large numbers (100s or 1000s) of jobs simultaneously on the same file system. The parent directory of the temporary directory is determined in the following order of precedence:

The value of the option --tmp_dir (if that option was used)
The environment variable $ARIBA_TMPDIR (if it is set)
The environment variable $TMPDIR (if it is set)
If none of the above is found, then use the run’s output directory.

Each temporary directory is unique to one run of ARIBA, and is automatically deleted at the end of the run (even if ARIBA was killed by the user or crashed). For example,

export $ARIBA_TMPDIR=/tmp

will result in the creation of a new directory inside /tmp, which will have a name of the form

/tmp/ariba.tmp.abcdef

where the suffix abcdef is a random string of characters, chosen such that /tmp/ariba.tmp.abcdef does not already exist.

The exception to the above is if the option --noclean is used. This forces the temporary directory to be placed in the output directory, and temporary files are kept. It is intended for debugging.

Usage

usage: ariba <command> <options>

optional arguments:
  -h, --help      show this help message and exit

Available commands:

aln2meta      Converts multi-aln fasta and SNPs to metadata
expandflag    Expands flag column of report file
flag          Translate the meaning of a flag
getref        Download reference data
micplot       Make violin/dot plots using MIC data
prepareref    Prepare reference data for input to "run"
pubmlstget    Download species from PubMLST and make db
pubmlstspecies
          Get list of available species from PubMLST
refquery      Get cluster or sequence info from prepareref output
run           Run the local assembly pipeline
summary       Summarise multiple reports made by "run"
test          Run small built-in test dataset
version       Get versions and exit