alienzj / wfi

whofluIRMA - influenza assembly pipeline

wfi (WhoFlu IRMA)

Snakemake pipeline for running IRMA. Designed for Influenza and RSV Illumina Sequencing.

Note: This pipeline is not ready for general use, while it generally works, it is very brittle.

Requirements:

• Linux Distro (or *unix system like MacOS)
• Conda
• Snakemake
• Cutadapt
• IRMA
• R
• R Packages: ggplot2 dplyr stringr tidyr cowplot gridExtra furrr

---

Automatic Installation - Experimental

To install wfi, use the automatic installation script:

wget https://raw.githubusercontent.com/ammaraziz/wfi/master/tools/auto_install.sh
bash auto_install.sh

Currently in testing phase, please create an issue if you run into any problems.

Manual Installation

1. Install miniconda: https://docs.conda.io/en/latest/miniconda.html
2. Install snakemake: https://snakemake.readthedocs.io/en/stable/getting_started/installation.html

   conda install -n base -c conda-forge mamba
   mamba install -c bioconda -c conda-forge snakemake-minimal
   

3. Install cutadapt and biopython:

   mamba install -c bioconda cutadapt
   mamba install -c conda-forge biopython 
   

4. Install R (>3.6 should work) and R packages:

   mamba install -c conda-forge r-base 
   mamba install -c r r-ggplot2 r-dplyr r-tidyr r-cowplot r-gridExtra r-optparse r-furrr
   

Note: installing r packages through conda is troublesome for some, if so install manually in R.

5. Final step! Download the latest wfi release (see the side ->)
   • Put this in your /home/$USERNAME/bin/
   • Uncompress the .zip file
   • Navigate to /wfi/bin/
   • Uncompress flu-amd.zip
   • Copy the RSV module from /wfi/bin/custom_modules/ to /wfi/bin/flu-amd/IRMA_RES/modules/

In the future the installation process will be simplified, I promise! Any issues please email, I will be happy to assist.

---

Usage

To use the pipeline, follow these steps:

1. Navigate to wfi_config.yaml and modify as appropriate:

Params	Values	Information
input_dir	path	input directory - location of the raw fastq files for input
output_dir	path	output directory - location to output results - same dir where the config sits
second_assembly	True/False	if you suspect mixtures, set to True . It will increase run time substantially
subset	True/False	if you are only sequencing HA/NA/MP set this to True else leave as False
trim_prog	standard/tile	Trimming program to use, tile (bbduk) or standard (cutadapt)
trim_org	h1/h3	Influenza only, Flu subtype
technology	illumina/ont/pgm	seq technology used, will change the module by IRMA

2. Check snakemake is installed, if an error is produced it means snakemake was not found or it is not installed.

% snakemake --version
% 5.10.0 

3. Test the pipeline, this will output all the commands that will be run. Look for errors (red).

% snakemake -np

4. Run the pipeline, with option -j to specify number of cores to use.

% snakemake -j 8

---

Output structure:

1. Pipeline will output correctly formatted names located in:

{output_dir}/assemblies/rename/

2. Sorted by subtype - most likely the disired output:

{output_dir}/assemblies/rename/type/FLU{A|B}

3. IRMA assembly specific files, see: https://wonder.cdc.gov/amd/flu/irma/output.html

{output_dir}/assemblies/{sampleID}/

4. Files for depth and summary info located in:

{output_dir}/assemblies/{sampleID}/figures/ {output_dir}/assemblies/{sampleID}/tables/

Dependencies

• BLAT for the match step
• LABEL, which also packages certain resources used by IRMA:
  • Sequence Alignment and Modeling System (SAM) for both the rough align and sort steps
  • Shogun Toolbox, which is an essential part of LABEL, is used in the sort step
• SSW for the final assembly step, download our minor modifications to SSW
• samtools for BAM-SAM conversion as well as BAM sorting and indexing
• GNU Parallel for single node parallelization
• R and these R packages: optparse, ggplot2, dplyr, tidyr, stringr, cowplot, gridExtra

---

Troubleshooting problems:

1. Error regarding path directories		Check input and output directorys you've specified end with a '/'

2. Error: Nothing to be done			Check config file and ensure you've changed the input/output directories. 

3. A job crashed. What do I do?			Two options, delete the output directory so snakemake can run everything again. 
						Or find out where it crashed and delete the whole folder/sample. 
						Example, sometimes IRMA produces errors, find the sample which crashed, 
						go to assemblies and delete the corresponding {sampleID} folder. Rerun snakemake.
4. I'm very confusd 										
   or I need more help			
   or I've screwed something up badly!		Shoot me an email

Contact Ammar via email: [email protected] or go bug him in person for help at any time.