Rory Stark, Principal Scientist Computational Biology, University of Cambridge, Cancer Research UK Cambridge Institute
This workshop will demonstrate the steps involved in performing a quantitative analysis of ChIP-seq data in Bioconductor (specifically differential binding analysis), with some discussion of related assays such as ATAC-seq. Particular attention will be paid to processing of aligned reads, including blacklisting, greylisting, filtering for quality and duplication, and the particular challenges presented when normalizing these data. While the workshop follows the DiffBind package vignette, the use of a number of other Bioconductor packages is discussed, including csaw, ChIPQC, edgeR, DESeq2, and GreyListChIP.
Attendees should be familiar with the following:
Helpful background reading:
There is a book chapter which offers a detailed outline of the design and analysis of ChIP-seq experiments:
Email the instructor if you would like a copy.
Students will participate by following along an R markdown document, and asking questions throughout the workshop.
The conference organizers have arranged for attendees to have access to a RStudio session running inside a Docker instance that contains everything you need to follow the workshop.
NOTE: The workshop Docker instance does not have much compute power. As a result, it is not possible to run all the code in the time allowed. Two operations are affected: building greylists, and counting reads. In these cases, the workshop includes pre-computed data objects that can be loaded to avoid the time-consuming code. This is noted inline.
If you do not have access to the RStudio instance provided by the conference organizers, or you would like to run the Docker image locally and you have a Docker client installed on your machine, you run this workshop locally via:
docker pull crukcibioinformatics/quantitative_chip_workshop:latest
docker images
docker run -e PASSWORD=DiffBind -p 8787:8787 <imageid> Then open a browser and go to the URL localhost:8787. Log into RStudio with username:rstudio and password:DiffBind.
The workshop uses a Docker container with Bioconductor Release version 3.12. If you would like to install Bioconductor on your computer, install R and the required Bioconductor packages, including this workshop package.
First load the workshop package:
library(QuantitativeChIPseqWorkshop)There are a number of ways to follow the workshop:
You can open the fully built workshop (with all code and results) in the “Help” pane in RStudio as follows:
vignette("Quantitative-ChIPseq-Workshop","QuantitativeChIPseqWorkshop")If you are using the Docker image, You can open the fully built workshop (with all code and results) in a browser tab as follows:
rstudioapi::viewer(paste(system.file("doc",package="QuantitativeChIPseqWorkshop"),
"Quantitative-ChIPseq-Workshop.html", sep="/"))It may be helpful to move the tab into its own window so you can see the vignette and the RStudio window at the same time.
You can open the workshop “source” .Rmd file in the RStudio “Source” pane, and follow along, executing each code chunk in sequence, to build it as you go.
file.edit("vignettes/Quantitative-ChIPseq-Workshop.Rmd")The workshop duration is 90 min. Approximate timing of activities:
| Activity | Time |
|---|---|
| Introduction | 5m |
| QC of DNA enrichment assays | 5m |
| Peaks vs. Windows | 5m |
| DiffBind Sample sheet | 5m |
| Blacklists and Greylists | 10m |
| Counting aligned reads | 10m |
| Examining and normalizing the binding matrix | 10m |
| Modeling and Testing | 5m |
| Examining analysis results | 10m |
| Multi-factor Design | 5m |
| Normalization discussion | 15m |
| Conclusions | 5m |
Participants will learn a complete set of steps for conducting a differential binding analysis of data from a ChIP-seq or related DNA enrichment experiment within Bioconductor.