Skip to content

Installation

Note the following instructions are for pre-release installation for bohra version 3

bohra is a large pipeline with many dependencies, including databases required for speciation. Currently bohra is only available for installation with conda.

Create the environment

1. Install conda

If you do not already have conda installed, you can check out the documentation here. We recommend you install miniconda

Below are instructions to install miniconda on a linux machine, however there are other distributions for Windows and MacOS here

  1. Download and install miniconda 

mkdir -p ~/miniconda3
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
rm ~/miniconda3/miniconda.sh
2. Restart your terminal and then 

source ~/miniconda3/bin/activate
  1. Initialise conda
conda init --all

2. Setup the bohra environment and install the pipeline co-ordinator

  1. You can download the environment.yml file here. Please note that on the last line you may need to update the path to where your conda environments are stored. For example if you installed miniconda as above you should chnage this to line to:

prefix: ~/miniconda/envs/bohra-3pr
2. Create the environment 

conda env create -f environment.yml
  1. Install the bohra co-ordinator

pip3 install git+https://github.com/MDU-PHL/bohra.git@rethink_structure
4. Test this has worked as expected 
bohra --help
Should result in 

Usage: bohra [OPTIONS] COMMAND [ARGS]...

Options:
  --version  Show the version and exit.
  --help     Show this message and exit.

Commands:
  check           Check that dependencies are installed correctly.
  generate-input  Generare input files for the Bohra pipeline.
  install-deps    Install dependencies for Bohra - Highly recommended to...
  run             Run the Bohra pipeline.
  test            Check that bohra is installed correctly

3. Install dependencies and setup environment variables

It is highly recommended that you allow bohra to setup the required dependencies for the pipeline.

This step will setup conda environments in your ~/.conda or ~/.miniconda3 (depending on how you have configured your conda installation). These environments will be prefixed with the name of the environment that you have installed bohra into. For example if you used the environment.yml file in step 2 your prefix will be bohra-3pr. This will ensure consistency and prevent duplication of environments across a file system.

This is also useful for public health users, were the versions of software and databases needs to be strictly controlled.

Additionally, bohra depends upon either a correctly configure kraken2 compatible database OR a sylph compatible database. The bohra install-deps command will optionally download the databases for your and also set the appropriate environment variables for you.

  1. Activate your environment from step 2 above 

    conda activate bohra-3pr

  2. Run the bohra dependency installation

bohra install-deps

The initial creation of the conda environments may take some time. Once the environments are set up, bohra will try to set up your database environment variables. Although this is not essential it is HIGHLY recommended for ease of running and reproducibility.

Please be aware of the following:

  • If you elect to say no to setting up a BOHRA_KRAKEN2_DB or BOHRA_SYLPH_DB, you will need to provide EITHER the --kraken2_db OR a --sylph_db flag each time you run bohra if you wish to do speciation or use any features which depend on species (typing and AMR).

  • If you are happy to use the mlst database that comes bundled with the mlst tool, then you can decline setting the BOHRA_PUBMLST_DB and BOHRA_BLAST_DB.

  • If you are working in your own environments (~/.conda or ~/.miniconda) you skip setting the BOHRA_MOBSUITE_DB environment variable. However, please note if you are setting up bohra in a share environment, you may run into permissions issues if mobsuite database requires updating at the time of running (this can happen) - so please ensure that you have a properly set up mob_suite database and provide this path when requested during set up.

  • If you already have a kraken2 database and/or a sylph database you can press 'n' when asked if you want to download. If you select 'y' please make sure that you have enough storage space for the databases. Some are quite large, up to 600 GB (half-terabyte).