Link Search Menu Expand Document
The Knutt Pipeline

Installation

The pipelines only requires a working installation of Conda for Python 3, git and Snakemake. This guide has been tested on a clean, minimal installation of Ubuntu 19.10 and Debian Buster and should work on most Linux distributions.

MacOS and the Windows 10 Linux subsystem haven’t been tested.

Check your current setup with the following commands.

conda --version # => 5.10.0
snakemake --version # => 4.8.2

Table of contents

  1. System Requirements
  2. Miniconda
  3. Snakemake
  4. Knutt

System Requirements

They strongly depend on the executed steps of the pipeline. You should be fine with around 200 GB of memory, 32 vCPUs and 1.5 TB of disk space for a few medium sized metagenomes.

Certain steps also can be run on rented AWS EC2 instances for example.

Miniconda

To install the Conda distribution Miniconda follow the instructions on this site.

Commands:

wget -q https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
sh Miniconda3-latest-Linux-x86_64.sh

You will need to accept the licencse and select an installation path.

If you know you accepted the licencse, you can run a non-interactive installation like this:

wget -q https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
sh Miniconda3-latest-Linux-x86_64.sh -b -p $HOME/miniconda3
# Activate base env
eval "$($HOME/miniconda3/bin/conda shell.bash hook)"
# Init for Shell functions & automatic base activation
conda init

Conda environment activation can take a while, slowing the start of a new shell/login. You can disable automatic activation with:

# From the installer, just as a reminder
conda config --set auto_activate_base false
# Afterwards you can still just run
conda activate base

Snakemake

Snakemake can be installed with the required conda distribution. Install it from the Bioconda and conda-forge channel.

# From the installer, just as a reminder
conda install -c bioconda -c conda-forge snakemake

You could also create an environment like this

conda create -n snake -c bioconda -c conda-forge snakemake
# will be placed in `~/miniconda3/envs/snake`
conda activate snake

or like this

conda create -p ./snake -c bioconda -c conda-forge snakemake
# will be placed in `snake`
conda activate ./snake

Knutt

Clone the git repository for the pipeline you want to use:

git clone https://github.com/KnuttPipeline/KnuttReads2Bins.git
git clone https://github.com/KnuttPipeline/KnuttBinAnno.git
git clone https://github.com/KnuttPipeline/KnuttBinPhylo.git

Now you can edit the configuration files (config.yml).