A simple guide/pipeline of RNA-seq data analysis with real data example using two different approach: (1) conventional genome alignment (reads mapping) with STAR/stringtie/Deseq2 and (2) pseudoalignment with kallisto/sleuth.

Sequence Read Archive (SRA) is a database that stores raw sequencing data and alignment information. SRA enhance reproducibility and facilitate new discoveries through data analysis. SRA is also a file format, the default file format SRA database. However, most RNA-seq software use .fasta-based file format and .sam file format. Therefore, we need a software that can manipulate .sra file. You can download SRA toolkit from NCBI SRA Toolkit . I use Ubuntu with sudo permission.

Documentation of all SRA toolkit software can be read at SRA Toolkit Documentation

Find SRA accession number from NCBI SRA. Type your query in the search bar. You have to find all the relevant SRA accession number from a certain experiment. For example, I search for (MCF-7) AND "Homo sapiens"[orgn:__txid9606]. This query means “search for MCF-7 from Homo sapiens origin. I look for complete experiment with only two group and two or three replicate (for simplicity of the project and “potato-ness” of my laptop). I found MCF-7 OVOL2 gene knockout experiment.

This experiment have two group: OVOL2 gene knockout and wild-type (control group). Click on one of those results. Click “All Experiment” to see all the RNA-seq data from OVOL2 gene knockout experiment.

Indeed, this experiment only have four RNA-seq data. After that, create a SRA file list using your favorite text editor.

Save as OVOL2_sra_list.txt. Then, download using prefetch command from SRA toolkit.

Enter your SRA directory, and then use fastq-dump command to convert.

It is easier to have all required files in one directory, thus we create symbolic link. In Linux, symbolic link is usually used instead of copying/moving file.

FastQC aims to provide a simple way to do some quality control checks on raw sequence data coming from high throughput sequencing pipelines. With FastQC, you can:

FastQC need Java. In Ubuntu, you can install with the following command.

Download FastQC from the official website. Unzip, place it into $PATH, and run it.

You can open GUI by typing fastqc or give it argument to run in cli (see below).

MultiQC aggregate results from bioinformatics analyses across many samples into a single report (a HTML file). It's a general use tool, perfect for summarising the output from numerous bioinformatics tools. MultiQC will use FastQC output file and many other RNA-seq software output file as an input.

MultiQC need python and pip. Install pip before installing MultiQC

To run multiple QC analysis with fastqc, just type multiple filename as argument.

You can directly open the html output.

We can see that there is 3 bad results and one questionable results. Let’s dive in... we will compare the bad one with the reference “good results” from FastQC website.

A little bit chaos in the beginning, but not much different with the reference.

It should be perfectly follow the normal distribution.

Use MultiQC to summarize all outpout from FastQC as beautiful as possible. Find fastqc output data (html & zip file) in current directory with MultiQC.

Screenshot

Kallisto is a program for quantifying abundances of transcripts from bulk and single-cell RNA-Seq data, or more generally of target sequences using high-throughput sequencing reads. It is based on the novel idea of pseudoalignment for rapidly determining the compatibility of reads with targets, without the need for complete alignment.

Download release for Linux (I use the latest v0.46.1), then move into $PATH.

To do quantification of transcripts abundance, kallisto need index file and RNA-seq experiment file. To build index file, we need transcriptome reference file specific to our organism. In this case, it’s Homo sapiens. We will use trimmed RNA-seq .fastq file.

Download transcriptome reference file from Ensembl:

After that, we can build kallisto’s index file. Note that index file is the output of this operation. You can name it whatever you like, just make sure that it has .idx file format.

For quantification of transcript abundance, we need an index file that we’ve built before and an .fastq file that we’ve trimmed. Here is the argument used:

After run, three output file will be created.

abundance.h5 is ..., abundance.tsv is ..., run_info.json is ...,

Sleuth is a program for analysis of RNA-Seq experiments for which transcript abundances have been quantified with kallisto. sleuth provides tools for exploratory data analysis utilizing Shiny by RStudio, and implements statistical algorithms for differential analysis that leverage the boostrap estimates of kallisto.

To use sleuth and many software explained afterward, you should install R (at least v.4.0). In conjunction with R, it is advised to use RStudio as your IDE. You can download RStudio from its official website.

Since R is a fast-moving project, the latest stable version isn’t always available from Ubuntu’s repositories, so we’ll install from CRAN repository. This guide assume you have Ubuntu 20.04 LTS.

Add GPG-key, then add repository into ppa:

Install R:

Open R by typing R in the terminal. We will install BioconductoR, a package manager that distributes software for bioinformatic and biology-related data analysis. In this guide, BioconductoR was used to install several R packages such as rhdf5, DESeq2, EdgeR, ggplot2, and many other!

sleuth is dependence to package devtools and rhdf5. However, to install devtools, we need to install some dependencies (especially true in Ubuntu, all of its version. See this StackOverFlow discussion). In the terminal,

Back to the R console, we will install rhdf5 and sleuth.

However, as 26/12/2021, there is currently an issue in rhdf5. In R v.4.1.2, after installing with ‘proper’ way, R console give us error warning:

This issue was addressed in StackOverFlow and GitHub. Thanks to mschilli87 for easy sleuth installation!