TP_experimental_biologists.md

title: "TP for experimental biologists"
author: Laurent Modolo [laurent.modolo@ens-lyon.fr](mailto:laurent.modolo@ens-lyon.fr)
date: 6 Jun 2018
output:
  pdf_document:
    toc: true
    toc_depth: 3
    number_sections: true
    highlight: tango
    latex_engine: xelatex

The Goal of this practical is to learn how to build your own pipeline with nextflow and using the tools already wrapped. For this we are going to build a small RNASeq analysis pipeline that should run the following steps:

• remove Illumina adaptors
• trim reads by quality
• build the index of a reference genome
• estimate the number of RNA fragments mapping to the transcript of this genome

Initialize your own project

You are going to build a pipeline for you or your team. So the first step is to create your own project.

Forking

Instead of reinventing the wheel, you can use the pipelines/nextflow as a template. To easily do so, go to the pipelines/nextflow repository and click on the fork button.

In git, the action of forking means that you are going to make your own private copy of a repository. You can then write modifications in your project, and if they are of interest for the source repository (here pipelines/nextflow) create a merge request. Merge request are send to the source repository to ask the maintainers to integrate modifications.

Project organisation

This project (and yours) follow the guide of good practices for the LBMC

You are now on the main page of your fork of the pipelines/nextflow. You can explore this project, all the code in it is under the CeCILL lience (in the LICENCE file).

The README.md file contains instructions to run your pipeline and test it's installation.

The CONTRIBUTING.md file contains guidelines to follow if you want to contribute to the pipelines/nextflow (making a merge request for example).

The data folder will be the place were you store the raw data for your analysis. The results folder will be the place were you store the results of your analysis. Note that the content of these two folders should never be saved on git.

The doc folder contains the documentation of this practical course.

And most interestingly for you, the src contains code to wrapp tools. This folder contains two subdirectory. A docker_modules, an nf_modules and an sge_modules folder.

docker_modules

The src/docker_modules contains the code to wrapp tools in Docker. Docker is a framework that allow you to execute software withing containers. The docker_modules contains directory corresponding to tools and subdirectories corresponding to their version.

ls -l src/docker_modules/
rwxr-xr-x  3 laurent  _lpoperator   96 May 25 15:42 BEDtools/
drwxr-xr-x  4 laurent  _lpoperator  128 Jun  5 16:14 Bowtie2/
drwxr-xr-x  3 laurent  _lpoperator   96 May 25 15:42 FastQC/
drwxr-xr-x  4 laurent  _lpoperator  128 Jun  5 16:14 HTSeq/

To each tools/version corresponds two files:

ls -l src/docker_modules/Bowtie2/2.3.4.1/
-rw-r--r--  1 laurent  _lpoperator  283 Jun  5 15:07 Dockerfile
-rwxr-xr-x  1 laurent  _lpoperator   79 Jun  5 16:18 docker_init.sh*

The Dockerfile is the Docker recipe to create a container containing Bowtie2 in it's 2.3.4.1 version. And the docker_init.sh file is a small script to create the container from this recipe.

By running this script you will be able to easily install tools in different version on your personal computer and use it in your pipeline. Some of the advantages are:

• Whatever the computer, the installation and the results will be the same
• You can keep container for old version of tools and run it on new systems (science = reproducibility)
• You don't have to bother with tedious installation procedure, somebody else already did the job and wrote a Dockerfile.
• You can easily keep container for different version of the same tools.

sge_modules

The src/sge_modules folder is not really there. It's a submodule of the project PSMN/modules. To populate it locally you can use the following command:

git submodule init

Like for the src/docker_modules the PSMN/modules project describe recipes to install tools and use them. The main difference is that you cannot use Docker on the PSMN. Instead you have to use another framework Environment Module which allows you to load modules for specific tools and version. The README.md file of the PSMN/modules respository contains all the instruction to be able to load the modules maintained by the LBMC en present in the PSMN/modules respository.

nf_modules

The src/nf_modules folder contains templates of nextflow wrapper for the tools available in Docker and SGE. The details of the nextflow wrapper will be presented in the next section. Alongside the .nf and .config there is a tests folder that contains a tests.sh script to run test on the tool.

Nextflow pipeline

A pipeline is a succession of process. Each process has data input(s) and optional data output(s). Data flow are modeled as channels.

Processes

Here are an example of process:

process sample_fasta {
  input:
    file fasta from fasta_file

  output:
    file "sample.fasta" into fasta_sample

  script:
"""
head ${fasta} > sample.fasta
"""
}

We have the process sample_fasta that take as fasta_file channel as imput and output a fasta_sample channel. The process itself is deffined in the script: block and within """.

  input:
    file fasta from fasta_file

When we zoom on the input: block we see that we define a variable fasta of type file from the fasta_file channel. This mean that groovy is going to write a file named as the content of the variable fasta in the root of the folder where script: is executed.

  output:
    file "sample.fasta" into fasta_sample

At the end of the script, a file named sample.fasta is found in the root the folder where script: is executed and send into the pipeline fasta_sample

Using the WebIDE of Gitlab create a file src/fasta_sampler.nf with this process and commit to your repository.

Channels

Why bother with channels ? In the above example, the advantages of channels are not really clear. We could have just given the fasta file to the process. But what if we have many fasta file to process ? What if we have sub processes to run on each of the sampled fasta files ? Nextflow can easily deal with these problems with the help of channels.

Channels are streams of items that are emitted by a source and consumed by a process. A process with a channel as input will be run on every items send through the channel.

Channel
  .fromPath( "data/tiny_dataset/fasta/*.fasta" )
  .set { fasta_file }

Here we defined a channel fasta_file that is going to send every fasta file from the folder data/fasta/ into the process that take it as input.

Add the definition of the channel to the src/fasta_sampler.nf file and commit to your repository.

Run your pipeline locally

After writing this first pipeline, you may want to test it. To do that first clone your repository. To easily do that set visibility level to public in the settings of your project.

You can then run the following commands to download your project on your computer :

git clone -c http.sslVerify=false https://gitlab.biologie.ens-lyon.fr/<usr_name>/nextflow.git
cd nextflow
src/install_nextflow.sh

We also need data to run our pipeline :

cd data
git clone -c http.sslVerify=false https://gitlab.biologie.ens-lyon.fr/LBMC/tiny_dataset.git
cd ..

We can run our pipeline with the following command:

./nextflow src/fasta_sampler.nf

Getting your results

Our pipeline seems to work but we don't know where is the sample.fasta. To get results out of a process, we need to tell nextflow to write it somewhere (we may don't need to get every intermediate files in our results).

To do that we need to add the following line before the input: section:

  publishDir "results/sampling/", mode: 'copy'

Every file described in the output: section will be copied from nextflow to the folder results/sampling/.

Add this to you src/fasta_sampler.nf file with the WebIDE and commit to your repository. Pull your modifications locally with the command:

git pull origin master

You can run you pipeline again and check the content of the folder results/sampling.

Fasta everywhere

We ran our pipeline on one fasta file. How nextflow would handle 100 of them ? To test that we need to duplicate the tiny_v2.fasta file:

for i in {1..100}
do
 cp data/tiny_dataset/fasta/tiny_v2.fasta data/tiny_dataset/fasta/tiny_v2_${i}.fasta
done

You can run you pipeline again and check the content of the folder results/sampling.

Every fasta_sampler process write a sample.fasta file. We need to make the name of the output file dependent of the name of the input file.

  output:
    file "*_sample.fasta" into fasta_sample

  script:
"""
head ${fasta} > ${fasta.baseName}_sample.fasta
"""

Add this to you src/fasta_sampler.nf file with the WebIDE and commit to your repository before pulling your modifications locally. You can run you pipeline again and check the content of the folder results/sampling.

Build your own RNASeq pipeline

In this section you are going to build your own pipeline for RNASeq analysis from the code available in the src/nf_modules folder.

Create your Docker containers

For this practical, we are going to need the following tools :

• For Illumina adaptor removal : cutadapt
• For reads trimming by quality : UrQt
• For mapping and quantifying reads : BEDtools and Kallisto

To initialize these tools, follow the Installing section of the README.md file.

If you are using a CBP computer don't forget to cleanup your docker containers at the end of the practical with the following command:

docker rm $(docker stop $(docker ps -aq))
docker rmi $(docker images -qf "dangling=true")

Cutadapt

The first step of the pipeline is to remove any Illumina adaptor left in your reads files.

Open the WebIDE and create a src/RNASeq.nf file. Browse for src/nf_modules/cutadapt/cutadapt.nf, this file contains example for cutadapt. We are interested in the Illumina adaptor removal,for paired-end data section of the code. Copy this code in your pipeline and commit.

Compared to before, we have few new lines:

params.fastq = "$baseDir/data/fastq/*_{1,2}.fastq"

We declare a variable that contain the path of the fastq file to look for. The advantage of using params.fastq is that now the option --fastq in our call to the pipeline allow us to define this variable:

./nextflow src/RNASeq.nf --fastq "data/tiny_dataset/fastq/*_R{1,2}.fastq"

log.info "fastq files : ${params.fastq}"

This line simply display the value of the variable

Channel
  .fromFilePairs( params.fastq )

As we are working with paired-end RNASeq data we tell nextflow to send pairs of fastq in the fastq_file channel.

cutadapt.config

For the fastq_sampler.nf pipeline we used the command head present in most base UNIX systems. Here we want to use cutadapt which is not. Therefore, we have three main options:

• install cutadapt locally so nextflow can use it
• launch the process in a Docker container that have cutadapt installed
• launch the process with SGE while loading the correct module to have cutadapt available

We are not going to use the first option which requiere no configuration for nextflow but tedious tools installation. Instead, we are going to use existing wrappers and tell nextflow about it. This is what the src/nf_modules/cutadapt/cutadapt.config is used for.

Copy the content of this config file to an src/RNASeq.config file. This file is structured in process blocks. Here we are only interested in configuring adaptor_removal process not trimming process. So you can remove the trimming block and commit.

You can test your pipeline with the following command:

./nextflow src/RNASeq.nf -c src/RNASeq.config -profile docker --fastq "data/tiny_dataset/fastq/*_R{1,2}.fastq"

UrQt

The second step of the pipeline is to trim reads by quality.

Browse for src/nf_modules/UrQt/urqt.nf, this file contains example for UrQt. We are interested in the for paired-end data section of the code. Copy the process section code in your pipeline and commit.

This code won't work if you try to run it: the fastq_file channel is already consumed by the adaptor_removal process. In nextflow once a channel is used by a process, it cease to exist. Moreover, we don't want to trim the input fastq, we want to trim the fastq that come from the adaptor_removal process.

Therefore, you need to change the line :

  set pair_id, file(reads) from fastq_files

In the trimming process to:

  set pair_id, file(reads) from fastq_files_cut

The two processes are now connected by the channel fastq_files_cut.

Add the content of the src/nf_modules/UrQt/urqt.config file to your src/RNASeq.config file and commit.

You can test your pipeline.

BEDtools

Kallisto need the sequences of the transcript that need to be quantified. We are going to extract these sequences from the reference data/tiny_dataset/fasta/tiny_v2.fasta with the bed annotation data/tiny_dataset/annot/tiny.bed.

You can copy to your src/RNASeq.nf file the content of src/nf_modules/BEDtools/bedtools.nf and to your src/RNASeq.config file the content of src/nf_modules/BEDtools/bedtools.config.

Commit your work and test your pipeline with the following command :

./nextflow src/RNASeq.nf -c src/RNASeq.config -profile docker --fastq "data/tiny_dataset/fastq/*_R{1,2}.fastq" --fasta "data/tiny_dataset/fasta/tiny_v2.fasta" --bed "data/tiny_dataset/annot/tiny.bed"

Kallisto

Kallisto run in two step: the indexation of the reference and the quantification on this index.

You can copy to your src/RNASeq.nf file the relevant content of src/nf_modules/Kallisto/kallisto.nf and to your src/RNASeq.config file the content of src/nf_modules/Kallisto/kallisto.config.

We are going to work with paired-end so only copy the relevant processes. The index_fasta process need to take as input the output of your fasta_from_bed process. The fastq input of your mapping_fastq process need to take as input the output of your index_fasta process and the trimming process.

Commit your work and test your pipeline.