Extended workflow

Warning

This tutorial shows a full run of the workflow with all options activated. For testing, we ran this example locally on a large cloud instance. The data is likely too large for running locally on an average laptop.

We show a run of the full workflow, including deepribo predictions and differential expression analysis, on data available from NCBI. For this purpose, we use a salmonella enterica dataset available under the accession number PRJNA421559 [PGAR19].

Warning

Ensure that you have miniconda3 and singularity installed and a snakemake environment set-up. Please refer to the overview for details on the installation.

Setup

First of all, we start by creating the project directory and changing to it. (you can choose any directory name)

mkdir project
cd project

We then download the latest version of HRIBO into the newly created project folder and unpack it.

wget https://github.com/RickGelhausen/HRIBO/archive/1.7.0.tar.gz
tar -xzf 1.7.0.tar.gz; mv HRIBO-1.7.0 HRIBO; rm 1.7.0.tar.gz;

Retrieve and prepare input files

Before starting the workflow, we have to acquire and prepare several input files. These files are the annotation file, the genome file, the fastq files, the configuration file and the sample sheet.

Annotation and genome files

First, we want to retrieve the annotation file and the genome file. In this case, we can find both on NCBI using the accession number NC_016856.1.

../_images/HRIBO-2.png

Note

Ensure that you download the annotation for the correct strain str. 14028S.

On this page, we can directly retrieve both files by clicking on the according download links next to the file descriptions. Alternatively, you can directly download them using the following commands:

wget ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/022/165/GCF_000022165.1_ASM2216v1/GCF_000022165.1_ASM2216v1_genomic.gff.gz
wget ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/022/165/GCF_000022165.1_ASM2216v1/GCF_000022165.1_ASM2216v1_genomic.fna.gz

Then, we unpack and rename both files.

gunzip GCF_000022165.1_ASM2216v1_genomic.gff.gz && mv GCF_000022165.1_ASM2216v1_genomic.gff annotation.gff
gunzip GCF_000022165.1_ASM2216v1_genomic.fna.gz && mv GCF_000022165.1_ASM2216v1_genomic.fna genome.fa

.fastq files

Next, we want to acquire the fastq files. The fastq files are available under the accession number PRJNA421559 on NCBI. The files have to be downloaded using the Sequence Read Archive (SRA). There are multiple ways of downloading files from SRA as explained here.

As we already have conda installed, the easiest way is to install the sra-tools:

conda create -n sra-tools -c bioconda -c conda-forge sra-tools pigz

This will create a conda environment containing the sra-tools and pigz. Using these, we can simply pass the SRA identifiers and download the data:

conda activate sra-tools;
fasterq-dump SRR6359966; pigz -p 2 SRR6359966.fastq; mv SRR6359966.fastq.gz RIBO-WT-1.fastq.gz
fasterq-dump SRR6359967; pigz -p 2 SRR6359967.fastq; mv SRR6359967.fastq.gz RIBO-WT-2.fastq.gz
fasterq-dump SRR6359974; pigz -p 2 SRR6359974.fastq; mv SRR6359974.fastq.gz RNA-WT-1.fastq.gz
fasterq-dump SRR6359975; pigz -p 2 SRR6359975.fastq; mv SRR6359975.fastq.gz RNA-WT-2.fastq.gz
fasterq-dump SRR6359970; pigz -p 2 SRR6359970.fastq; mv SRR6359970.fastq.gz RIBO-csrA-1.fastq.gz
fasterq-dump SRR6359971; pigz -p 2 SRR6359971.fastq; mv SRR6359971.fastq.gz RIBO-csrA-2.fastq.gz
fasterq-dump SRR6359978; pigz -p 2 SRR6359978.fastq; mv SRR6359978.fastq.gz RNA-csrA-1.fastq.gz
fasterq-dump SRR6359979; pigz -p 2 SRR6359979.fastq; mv SRR6359979.fastq.gz RNA-csrA-2.fastq.gz
conda deactivate;

Note

we will use two conditions and two replicates for each condition. There are 4 replicates available for each condition, we run it with two as this is just an example. If you run an analysis always try to use as many replicates as possible.

Warning

If you have a bad internet connection, this step might take some time. It is advised to run this workflow on a cluster or cloud instance.

This will download compressed files for each of the required .fastq files. We will move them into a folder called fastq.

mkdir fastq;
mv *.fastq.gz fastq;

Sample sheet and configuration file

Finally, we will prepare the configuration file (config.yaml) and the sample sheet (samples.tsv). We start by copying templates for both files from the HRIBO/templates/ into the HRIBO/ folder.

cp HRIBO/templates/samples.tsv HRIBO/

The sample file looks as follows:

method

condition

replicate

fastqFile

fastqFile2

RIBO

A

1

fastq/RIBO-A-1.fastq.gz

RIBO

A

2

fastq/RIBO-A-2.fastq.gz

RIBO

B

1

fastq/RIBO-B-1.fastq.gz

RIBO

B

2

fastq/RIBO-B-2.fastq.gz

RNA

A

1

fastq/RNA-A-1.fastq.gz

RNA

A

2

fastq/RNA-A-2.fastq.gz

RNA

B

1

fastq/RNA-B-1.fastq.gz

RNA

B

2

fastq/RNA-B-2.fastq.gz

Note

When using your own data, use any editor (vi(m), gedit, nano, atom, …) to customize the sample sheet.

Warning

Please ensure not to replace any tabulator symbols with spaces while changing this file.

We will rewrite this file to fit the previously downloaded .fastq.gz files.

method

condition

replicate

fastqFile

fastqFile2

RIBO

WT

1

fastq/RIBO-WT-1.fastq.gz

RIBO

WT

2

fastq/RIBO-WT-2.fastq.gz

RIBO

csrA

1

fastq/RIBO-csrA-1.fastq.gz

RIBO

csrA

2

fastq/RIBO-csrA-2.fastq.gz

RNA

WT

1

fastq/RNA-WT-1.fastq.gz

RNA

WT

2

fastq/RNA-WT-2.fastq.gz

RNA

csrA

1

fastq/RNA-csrA-1.fastq.gz

RNA

csrA

2

fastq/RNA-csrA-1.fastq.gz

Next, we are going to set up the config.yaml.

cp HRIBO/templates/config.yaml HRIBO/

The config file can be used to easily change the parameters of HRIBO.

Note

For a detailed overview of the available options please refer to our workflow-configuration

In our small example, we will adjust the adapter sequence which will lead to the following changes in the config.yaml file:

biologySettings:
    # Adapter sequence used
    adapterS3: "AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC"

Running the workflow

Now that all the required files are prepared, we can start running the workflow, either locally or in a cluster environment.

Warning

if you have problems running deepribo, please refer to Activating DeepRibo.

Warning

before you start using snakemake remember to activate the environment first.

conda activate snakemake

Run the workflow locally

Use the following steps when you plan to execute the workflow on a single server, cloud instance or workstation.

Warning

Please be aware that some steps of the workflow require a lot of memory or time, depending on the size of your input data. To get a better idea about the memory consumption, you can have a look at the provided sge.yaml or torque.yaml files.

Navigate to the project folder containing your annotation and genome files, as well as the HRIBO folder. Start the workflow locally from this folder by running:

snakemake --use-conda --use-singularity --singularity-args " -c " --greediness 0 -s HRIBO/Snakefile --directory ${PWD} -j 10 --latency-wait 60

This will start the workflow locally.

  • --use-conda: instruct snakemake to download tool dependencies from conda.

  • -s: specifies the Snakefile to be used.

  • --directory: specifies your current path.

  • -j: specifies the maximum number of cores snakemake is allowed to use.

  • --latency-wait: specifies how long (in seconds) snakemake will wait for filesystem latencies until declaring a file to be missing.

Run Snakemake in a cluster environment

Use the following steps if you are executing the workflow via a queuing system. Edit the configuration file <cluster>.yaml according to your queuing system setup and cluster hardware.

Navigate to the project folder on your cluster system. Start the workflow from this folder by running (The following system call shows the usage with Grid Engine):

snakemake --use-conda --use-singularity --singularity-args " -c " -s HRIBO/Snakefile --directory ${PWD} -j 20 --cluster-config HRIBO/sge.yaml

Note

Ensure that you use an appropriate <cluster>.yaml for your cluster system. We provide one for SGE and TORQUE based systems.

Example: Run Snakemake in a cluster environment

Warning

Be advised that this is a specific example, the required options may change depending on your system.

We ran the tutorial workflow in a cluster environment, specifically a TORQUE cluster environment. Therefore, we created a bash script torque.sh in our project folder.

vi torque.sh

Note

Please note that all arguments enclosed in <> have to be customized. This script will only work if your cluster uses the TORQUE queuing system.

We proceeded by writing the queuing script:

#!/bin/bash
#PBS -N <ProjectName>
#PBS -S /bin/bash
#PBS -q "long"
#PBS -d <PATH/ProjectFolder>
#PBS -l nodes=1:ppn=1
#PBS -o <PATH/ProjectFolder>
#PBS -j oe
cd <PATH/ProjectFolder>
source activate HRIBO
snakemake --latency-wait 600 --use-conda --use-singularity --singularity-args " -c " -s HRIBO/Snakefile --directory ${PWD} -j 20 --cluster-config HRIBO/torque.yaml --cluster "qsub -N {cluster.jobname} -S /bin/bash -q {cluster.qname} -d <PATH/ProjectFolder> -l {cluster.resources} -o {cluster.logoutputdir} -j oe"

We then simply submitted this job to the cluster:

qsub torque.sh

Using any of the presented methods, this will run the workflow on the tutorial dataset and create the desired output files.

Results

The last step will be to aggregate all the results once the workflow has finished running. In order to do this, we provided a script in the scripts folder of HRIBO called makereport.sh.

bash HRIBO/scripts/makereport.sh <reportname>

Running this will create a folder where all the results are collected from the workflows final output, it will additionally create compressed file in .zip format. The <reportname> will be extended by report_HRIBOX.X.X_dd-mm-yy.

Note

A detailed explanation of the result files can be found in the result section.

Note

The final result of this example workflow, can be found here .

Warning

As many browsers stopped the support for viewing ftp files, you might have to use a ftp viewer instead.

Runtime

The total runtime of the extended workflow, using 12 cores of an AMD EPYC Processor (with IBPB), 1996 MHz CPUs and 64 GB RAM, was 5h51m14s.

The runtime contains the automatic download and installation time of all dependencies by conda and singularity. This step is mainly dependent on the available network bandwidth. In this case it took about 12 minutes.

The runtime difference compared to the example workflow is explained by the additional libraries and tools used.

References

GruningDSjodin+17

Björn Grüning, Ryan Dale, Andreas Sjödin, Jillian Rowe, Brad A. Chapman, Christopher H. Tomkins-Tinch, Renan Valieris, and Johannes Köster. Bioconda: a sustainable and comprehensive software distribution for the life sciences. bioRxiv, 2017. URL: https://www.biorxiv.org/content/early/2017/10/27/207092, arXiv:https://www.biorxiv.org/content/early/2017/10/27/207092.full.pdf, doi:10.1101/207092.

KosterR18

Johannes Köster and Sven Rahmann. Snakemake—a scalable bioinformatics workflow engine. Bioinformatics, ():bty350, 2018. URL: http://dx.doi.org/10.1093/bioinformatics/bty350, arXiv:/oup/backfile/content_public/journal/bioinformatics/pap/10.1093_bioinformatics_bty350/2/bty350.pdf, doi:10.1093/bioinformatics/bty350.

PGAR19

Anastasia H. Potts, Yinping Guo, Brian M. M. Ahmer, and Tony Romeo. Role of csra in stress responses and metabolism important for salmonella virulence revealed by integrated transcriptomics. PLOS ONE, 14(1):1–30, 01 2019. URL: https://doi.org/10.1371/journal.pone.0211430, doi:10.1371/journal.pone.0211430.

ZAA+18

Daniel R Zerbino, Premanand Achuthan, Wasiu Akanni, M Ridwan Amode, Daniel Barrell, Jyothish Bhai, Konstantinos Billis, Carla Cummins, Astrid Gall, Carlos García Girón, Laurent Gil, Leo Gordon, Leanne Haggerty, Erin Haskell, Thibaut Hourlier, Osagie G Izuogu, Sophie H Janacek, Thomas Juettemann, Jimmy Kiang To, Matthew R Laird, Ilias Lavidas, Zhicheng Liu, Jane E Loveland, Thomas Maurel, William McLaren, Benjamin Moore, Jonathan Mudge, Daniel N Murphy, Victoria Newman, Michael Nuhn, Denye Ogeh, Chuang Kee Ong, Anne Parker, Mateus Patricio, Harpreet Singh Riat, Helen Schuilenburg, Dan Sheppard, Helen Sparrow, Kieron Taylor, Anja Thormann, Alessandro Vullo, Brandon Walts, Amonida Zadissa, Adam Frankish, Sarah E Hunt, Myrto Kostadima, Nicholas Langridge, Fergal J Martin, Matthieu Muffato, Emily Perry, Magali Ruffier, Dan M Staines, Stephen J Trevanion, Bronwen L Aken, Fiona Cunningham, Andrew Yates, and Paul Flicek. Ensembl 2018. Nucleic Acids Research, 46(D1):D754–D761, 2018. URL: http://dx.doi.org/10.1093/nar/gkx1098, arXiv:/oup/backfile/content_public/journal/nar/46/d1/10.1093_nar_gkx1098/2/gkx1098.pdf, doi:10.1093/nar/gkx1098.