Skip to content

2.0 Introduction

Part 2 builds on fundamental concepts learned in Part 1 and provides you with hands-on experience in Nextflow workflow development. Throughout the session we will be working with a bulk RNAseq dataset to build our workflow.

We will construct channels that control how our data flows through processes that we will progressively construct to build our workflow. Each lesson in Part 2 will build on the previous lessons, so you can gain a deeper understanding of the techniques and the impact they have on your resulting workflow.

In Part 2 of this workshop, we will explore a scenario of creating a multi-sample Nextflow workflow for preparing RNAseq data. We will build the workflow, step-by-step, by converting a series of provided bash scripts into small workflow components.

Along the way, you will encounter Nextflow concepts (from Part 1, and some new) and our best practice recommendations for developing your own pipeline.

Part 2 is based off the Simple RNA-Seq workflow training material developed by Seqera.

2.0.1 Log back into your instance

Re-connect to your Virtual Machine by following the "Connect to the VM" section from the setup page.

Once connected, in your VSCode terminal, change directories into the part2/ directory: 

cd ~/part2/

All Part 2 activities will be conducted in this folder.

2.0.2 Our scenario: from bash scripts to scalable workflows

Imagine you are a bioinformatician in a busy research lab. Your team will be receiving a large batch of samples that need to be processed through a series of analysis steps.

You have inherited a set of bash scripts from a former colleague, which were used to process a handful of samples manually. These scripts are robust and well-tested, but they were not designed with scalability in mind.

As more samples come in, running these scripts one by one is becoming increasingly tedious and error-prone.

You need a way to automate this process, ensuring consistency and efficiency across many samples.

You decide to use Nextflow.

Exercise

View the bash scripts your colleague provided:

  1. Use either the VSCode File Explorer, or the integrated terminal to navgiate to the ~/part2/bash_scripts/ directory.
  2. Inspect the scripts (open in a VSCode tab, or text editor in the terminal).

Each script runs a single data processing step and are run in order of the prefixed number.

Poll

What are some limitations of these scripts in terms of running them in a pipeline and monitoring it?

2.0.3 Our workflow: RNAseq data processing

Don't worry if you don't have prior knowledge of RNAseq!

The focus of this workshop is on learning Nextflow, the RNAseq data we are using in this part are just a practical example to help you understand how the workflow system works.

RNAseq is used to study gene expression and has many applications across biomedicine, agriculture and evolutionary studies. In our scenario we are going to run through some basic core steps that allow us to explore different aspects of Nextflow.

Data

The data we will use includes:

  • *.fastq: Paired-end RNAseq reads from three different samples (gut, liver, lung).
  • transcriptome.fa: A transcriptome file.
  • samplesheet*.csv: CSV files that help us track which files belong to which samples.

Tools

We will be implementing and integrating three commonly used bioinformatics tools:

  1. Salmon is a tool for quantifying molecules known as transcripts through RNA-seq data.
  2. FastQC is a tool for quality analysis of high throughput sequence data. You can think of it as a way to assess the quality of your data.
  3. MultiQC searches a given directory for analysis logs and compiles an HTML report for easy viewing. It's a general use tool, perfect for summarising the output from numerous bioinformatics tools.

These tools will be run using Docker containers. We will not explore how the data and tools work further, and focus on how they should be implemented in a Nextflow workflow.

2.0.4 Pipeline structure and design

Having reviewed the bash scripts, we've decided to keep its modular structure and will build the following four processes (discrete steps):

  1. INDEX - Transcriptome indexing (tool: Salmon): create an index of the reference transcriptome for faster and efficient data processing.
  2. FASTQC - Raw data quality control (tool: FastQC): assess the quality of fastq files to ensure our data is usable.
  3. QUANTIFICATION - Gene quantification (tool: Salmon): counting how many reads map to each gene in the transcriptome.
  4. MULTIQC - Summarise results in a report (tool: MultiQC): generate a report that summarises quality control and gene quantification results.

2.0.5 Nextflowing the workflow

Each lesson in part 2 of our workshop focuses on implementing one process of the workflow at a time. We will iteratively build the workflow and processes in a single main.nf file and lightly use a nextflow.config file for configuration.

main.nf

The main.nf file is the core script that defines the steps of your Nextflow workflow. It outlines each process (the individual commands, or, data processing steps) and how they are connected to each other. This main.nf script focuses on what the workflow does.

Most of the code you will write in this Part will go in main.nf.

We will follow an ordered approach for each step of the workflow building off the process structure from Part 1.3. You will be using this process template for each step of the workflow, adding them to the main.nf script: 

process < name > {
  [ directives ]

  input:
    < process inputs >

  output:
    < process outputs >

  script:
  """
  < script to be executed >
  """
}

nextflow.config

The nextflow.config file is a key part of any Nextflow workflow. While main.nf outlines the steps and processes of the workflow, nextflow.config allows you to define important settings and configurations that control how your workflow should run.

This script will be intermittently used in the following lessons to control the use of Docker containers, how much resources (CPUs) should be used, and reporting of the workflow after it has finished running.