PAN-GWES (with Conda installation)

Introduction

The original PAN-GWES repo is located here. This duplicate repo provides a quick and easy interface for installation and testing.

Installation

Bioconda (Recommended)

1. If you haven’t already, first, you will need to install conda or recommended: mamba.

conda create -n pangwes
conda activate pangwes
mamba install -c bioconda -c conda-forge pangwes

If you installed conda, run the last command using conda inplace of mamba.

If you are an ARM-64 user (Macs with M series chips), please follow these steps to change your channel to osx-64:

conda create -n pangwes
conda activate pangwes
conda config --env --set subdir osx-64
conda install -c bioconda -c conda-forge pangwes

It is recommended to use the conda command (instead of mamba) for installation after channel change.

Local Conda Installation

1. If you haven’t already, first, you will need to install conda.
2. Next, install conda build tools:

conda activate base
conda install conda-build

3. Clone the repo

git clone https://github.com/Sudaraka88/PAN-GWES

Alternatively, you can click the <>Code button at the top right and click Download ZIP. Afterwards, unzip the repo.

Linux-64 and osx-64 (Intel Macs)

4. Build the repo

cd PAN-GWES
conda build -c bioconda -c conda-forge sw

5. Create a new environment and install the package

conda create -n pangwes -c bioconda -c conda-forge pangwes --use-local

When prompted, enter y to confirm the installation of pangwes and dependencies

ARM-64 (Macs with M series chips)

SpydrPick and Cuttlefish are available via bioconda as osx-64 dependencies. You can create a conda environment that uses the osx-64 channel. Follow steps 1-3 first, then:

4. Create a new conda environment and change channels

conda create -n pangwes
conda activate pangwes
conda config --env --set subdir osx-64

5. Build the package

cd PAN-GWES
conda build -c bioconda -c conda-forge sw

6. Install the package

conda install -c bioconda -c conda-forge pangwes --use-local

Make and Install

Please refer to the original repo

Example

There is a sample dataset available here. This compressed file contains a single folder called efcls_assemblies, which contains 337 E. faecalis assemblies. Uncompress the downloaded file and move the efcls_assemblies folder into your working directory.

1. Open a terminal in your working directory and populate a list of these assemblies using the following command.

ls -d efcls_assemblies/* > efcls_assemblies.txt

2. Activate the pangwes conda environment

conda activate pangwes

3. Build the pangenome graph using cuttlefish

cuttlefish build --list efcls_assemblies.txt --kmer-len 61 --output efcls --threads 16 -f 1

You can provide an optional work directory to store temporary files using –work-dir and adjust the number of threads depending on your system resources. If you run into the error “Cannot open temporary file ./kmc_01021.bin”, run ulimit -n 2048 in the terminal. See here and here.

4. Prase the built gfa1 file

gfa1_parser efcls.gfa1 efcls

Warning! The next couple of steps will take a bit of time. Remember to adjust the number of threads depending on your system resources.

5. Run SpydrPick on the unitig fasta alignment

SpydrPick --alignmentfile efcls.fasta --maf-threshold 0.05 --mi-values 50000000 --threads 16 --verbose

6. Calculate unitig distances

unitig_distance --unitigs-file efcls.unitigs --edges-file efcls.edges --k-mer-length 61 --sgg-paths-file efcls.paths --queries-file efcls.*spydrpick_couplings*edges --threads 16 --queries-one-based --run-sggs-only --output-stem efcls --verbose

7. Generate the pangenome Manhattan Plot

./gwes_plot.r -i efcls.ud_sgg_0_based -n 337

Warning! This script will try to install the required dependencies. Use --no-deps to generate the plot without installing any packages (slower!).

If all went to plan, this example should generate the following figure: