User Tools

Site Tools


Differential expression analysis with DESeq2


It is easy to follow the DESeq2 instructions even if you are not very familiar with R. Most of the commands in the .R script we'll use today are verbatim taken from the Vignette:

The project directory

Start a project on your laptop.:

  • Give the project directory a name like PROJ08_yeastDESeq

Initiate some directories:

  • 01_input
  • 02_scripts
  • 03_output

Download these two directories and put them in 01_input


Explore the files you have acquired.


Open the script file in Rstudio

We will switch over to the DESeq2 script file now.

Learning objectives:

  • Learn about DESeq2 and what it does.
  • Know the five steps of what DESeq2 does.
  • Understand that DESeq2 does not use a Poisson distribution to model the spread of the data. Instead it uses a negative binomial distribution because RNA-seq data naturally is overdispersed.
  • Understand why multiple testing correction is required and that the Benjamini-Hochberg Correction is the type that is used by DESeq2.
  • Know how to read and interpret three types of plots:
    • MA-plot
    • Volcano plot
    • Correlation matrix

Extra stuff:

Clustering lesson from 2017

Reproducible Research Lesson from 2017

Downstream experimentation Revisit the original powerpoint from the first day called RNA-seq Intro & Experimental Design. There are a few slides at the end that illustrate some examples of what you can do with the differentially expressed genes to verify, study, and understand them. The point of RNA-seq should be to stimulate some hypotheses that can be experimentally tested.

wiki/deseq2.txt · Last modified: 2018/12/06 09:29 by erin