Tips for setting up an RNA-seq project

From the Microbial Genomics and Transcriptomics hands-on workshop , Sep 24-25, 2015 – Presented by: Ben Johnson (Michigan State University) & C. Titus Brown (UC Davis).

So you’ve decided to do an RNA-seq experiment. Some things you will need to think about are:

  • What is the end-goal? (differential gene expression, variant calling, de novo assembly)
  • How many replicates do I need per condition? (related to first question; 3 is good, 6 is better)
  • How am I going to analyze this data?

You will save yourself a lot of headache if you take the time to set up your project in a coherent manner.

Some tips:

  1. Put your project in a single directory/folder (named something informative without spaces)
  2. Start a virtual notebook for your analysis, commands, parameters, etc. – plain text or markdown – I like markdown because you can readily export to HTML to share with collaborators
  3. Keep your raw data raw (e.g. make a working copy)
  4. Name files something concise (NO SPACES!) but informative (tab auto-complete is your friend)
  5. Make several separate folders within the main project folder to house different steps of the analysis
  6. In each subfolder create a README to describe the contents of that folder, how they got there, etc. (metadata)
  7. If possible, automate (script) analysis to allow you to re-run analyses on (many) files quickly and reproducibly

To put it another way, treat it as you would setting up a bench experiment and documenting progress in a lab notebook. You will write down dates, times, reagents (lot numbers, purity, etc.), buffer/media recipes, purpose of the experiment, experimental procedure, results (and if you believe them), and what you plan to do next to continue to investigate the question or if there are new questions to be answered.

You want to approach bioinformatics analyses in a similar manner. Document everything as best as you can. It will be useful for when you share data with others (publish, collaborate, etc.) and when you have to go back in 6 months to reassess what it was you did to generate graphs x, y, and z.

Also, be aware that tools and technology evolve quickly in this field. It’s worth getting a Twitter account to interact with people who develop/perform these analyses and probing the literature for new/improved tools (they come out frequently). Further, Google is your friend and there are all kinds of resources (forums, etc.) to ask questions and get answers (usually pretty quickly). In particular, you should check out and biostars.

Note, on Twitter, for microbiology, we suggest following Nick Loman, Torsten Seemann, and Jonathan Eisen.

Start a virtual notebook

Let’s have a look at the Markdown syntax and how you can use it to document your work.

Markdown is a plain text format that can be rendered to HTML and is quite nice if working collaboratively, like on GitHub.

I will use an example from Vince Buffalo’s book “Bioinformatics Data Skills” (highly recommended):

# *Zea Mays* SNP Calling

We sequenced three lines of *zea mays*, using paired-end
sequencing. This sequencing was done by our sequencing core and we
received the data on 2013-05-10. Each variety should have **two**
sequences files, with suffixes `_R1.fastq` and `_R2.fastq`, indicating
which member of the pair it is.

## Sequencing Files

All raw FASTQ sequences are in `data/seqs/`:

$ find data/seqs -name "*.fastq"

## Quality Control Steps

After the sequencing data was received, our first stage of analysis
was to ensure the sequences were high quality. We ran each of the
three lines' two paired-end FASTQ files through a quality diagnostic
and control pipeline. Our planned pipeline is:

1. Create base quality diagnostic graphs.
2. Check reads for adapter sequences.
3. Trim adapter sequences.
4. Trim poor quality bases.

Recommended trimming programs:

- Trimmomatic
- Scythe

When this is rendered, it looks like this.

Here are some Markdown editors for

If you use GitHub to work collaboratively on things, you can copy and paste things right into a Gist and it will render for you (if you add the .md file extension when you name your file) and you can share privately or publicly!

Markdown syntax guide

Next: Quality control and trimming of reads

Source –

Leave a Reply

Your email address will not be published. Required fields are marked *


Time limit is exhausted. Please reload CAPTCHA.