Shallow seq pipeline for optimal shotgun data usage
These installation instructions are streamlined for Linux and macOS systems. The tool SHOGUN is installable on windows with a few minor tweaks to this tutorial. This package requires anaconda, which is a system agnostic package and virtual environment manager. Follow the installation instructions for your system at http://conda.pydata.org/miniconda.html.
Once anaconda is installed, get the environment file:
wget https://raw.githubusercontent.com/knights-lab/SHOGUN/master/environment.yml
Then install the requirements into the environment 'shogun':
conda env create -f environment.yml
Once anaconda is installed, create an environment:
conda create -n shogun python=3
Now activate the environment.
# OSX, Linux
source activate shogun
With the shogun environment activated, install the developmental SHOGUN toolchain.
# If you want to use bowtie2
conda install -c bioconda bowtie2
# NINJA-utils
pip install git+https://github.com/knights-lab/NINJA-utils.git --no-cache-dir --upgrade
# DOJO
pip install git+https://github.com/knights-lab/DOJO.git --no-cache-dir --upgrade
# SHOGUN
pip install git+https://github.com/knights-lab/SHOGUN.git --no-cache-dir --upgrade
With the flags provided to pip, copying and pasting any of these commands will redo the installation if a failure happened.
If you are installing SHOGUN for BugBase, you are done. The database is provided for you.
Next, to test the installation, download the test data.
wget https://www.dropbox.com/s/b5w4xe08x7snm93/shogun_test_files.zip?dl=1
Extract the folder using your favorite extraction utility.
7z x <downloaded file>
Next you create the database.
shogun_bt2_db -i ./test.hmp_species.fna -x '>, '
This will take some time, the DOJO software is lazy loading the NCBI Taxonomy.
shogun_bt2_lca -i ./mock_communities -b ./annotated/bt2/test.hmp_species
The results of the taxonomy counts will be in the taxon_counts.csv 🐱👤
To run it with UTree
shogun_utree_db -i ./test.hmp_species.fna -x '>, '
The run LCA:
shogun_utree_lca -i ./mock_communities -u ./annotated/utree/test.hmp_species.ctr
As of 1/10/17 the only supported functional profiling is through bowtie2 and the IMG database.
# Align reads to IMG
# Input directory has one FASTA (.fna extension) file per sample
shogun_functional -i <input directory> -o <output directory> -l False -b /project/flatiron/tonya/img_bowtie_builds/img.gene.bacteria.bowtie
# Input is a folder filled with SAM files, one SAM file per sample
kegg_parse_img_ids -i <output folder from shogun_functional> -o <location of the kegg.csv file>
# Input file is the KEGG csv file from the kegg_parse_img_ids
# -m mapping file for IMG gene to ko-map was generated by you in the spring
kegg_predictions -i <kegg.csv file> -o <final output> --algorithm intersection -m /project/flatiron/data/img/img-gene-ko-map.txt