Metagenomic and genomic surveillance of antimicrobial resistance in hospital wastewater (ATTACK-AMR) Analysis Pipeline

The cooperative research project ATTACK-AMR aims to deliver alternative, non-antibiotic therapies to combat antimicrobial resistant (AMR) pathogens. AMR is one of the biggest challenges facing healthcare industries and is on a rapid rise as a result of the overuse of antibiotics. Replacing antibiotics with alternative products will delay the resistance and restore the activity of antibiotics that are no longer effective due to resistance.

This serves as a guide to run the analysis pipeline written in two versions, [1] Snakemake and [2] Workflow Description Language (WDL).

Installation for Snakemake

This Snakemake pipeline requires the package manager Conda and the workflow management system Snakemake. Additional dependencies not handled by Snakemake are described in Section 1.3.

1.1. Install Miniconda

$ curl -sL \
  "https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh" > \
  "Miniconda3.sh"
$ bash Miniconda3.sh
$ conda update conda
$ rm Miniconda3.sh
$ conda install wget

1.2. Install Mamba

$ conda config --add channels conda-forge
$ conda update -n base --all
$ conda install -n base mamba

1.3. Install Snakemake

$ mamba create -c conda-forge -c bioconda -n snakemake snakemake

This creates an isolated enviroment containing the latest Snakemake. To activate it:

$ conda activate snakemake

To test snakemake:

$ snakemake --help

1.4. Install Additional Dependencies

Install git and gawk. We require gawk to process the filtering stage of our databases.

$ mamba install git
$ mamba install gawk

1.5. Download the pipeline

Download ATTACK-AMR from the online repository, or using the command line:

$ git clone https://github.com/bioinfodlsu/attack_amr_pipeline

2. Quickstart Usage Guide

2.1. Input

The pipeline requires, at the very least: (1) Metagenomic sequences (sample sequences can be downloaded at ENA, and (2) reference protein databases for (Resfinder, MGE). These and other input parameters are specified via a YAML-format config file -- config.yaml is provided in the config folder.

2.2. Running the pipeline

After constructing a config.yaml file and with the snakemake conda environment activated, you can call the pipeline from the top-level directory of ATTACK-AMR:

$ cd attack_amr_pipeline
$ snakemake --use-conda --cores all

2.3. Ouput

Outputs are stored the top-level directory of ATTACK-AMR.

Installation for Workflow Description Language

If working on a non-Linux machine, the following installation guidelines will work using WSL software in your local Windows machine.

Follow installation guidelines at https://miniwdl.readthedocs.io/en/latest/getting_started.html.

Install miniWDL

via PyPI

  pip3 install miniwdl

via Conda

  conda install -c conda-forge miniwdl

Then open a command prompt and try,

  miniwdl run_self_test

…to test the installation with WDL's built in workflow. This should end with miniwdl run_self_test OK.

If there is an error prompt due to docker engine, follow these series of steps

1. Update the apt package index.

   sudo apt-get update

2. Install docker engine.

   sudo apt-get install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin

3. Verify that the Docker Engine installation is successful by running the hello-world image.

   sudo docker run hello-world

4. Activate docker daemon.

   sudo dockerd

5. Run the miniWDL test script on a separate command prompt.

   miniwdl run_self_test

Run WDL Script via Command Prompt

You can choose to run the miniWDL through your command prompt.

1. Check your WDL Script for any errors.

   miniwdl check main.wdl --no-shellcheck

2. If no errors are detected, run your pipeline. Note: Place all your input files in one json file.

   miniwdl run main.wdl --input inputs.json

Alternative: Running it via Jupyter

These are the steps to run it via Jupyter Notebook.

1. Install Jupyter Notebook.

   pip install jupyterlab

2. Run Jupyter Notebook.

   jupyter-lab

3. Run the WDL Script via Jupyter's terminal.

   miniwdl check main.wdl --no-shellcheck
   miniwdl run main.wdl --input inputs.json