Researchers use ChIP binding data to identify potential transcription factor binding sites. They use gene expression data from sequencing or microarrays to quantify the effect of the factor over-expression or knockdown on its targets. The integration of the binding and expression data therefore can be used to improve the understanding of a transcription factor function. In this workshop, I present a complete workflow for integrating the gene expression (RNA-seq) and DNA-binding data (ChIP-seq) to predict the combined function of two transcription factors using R/Bioconductor. The example we will be using in the workshop is from real datasets of two functionally and evolutionary related transcription factors YY1 and YY2 in HeLa cells. We will try to identify the factor-specific and the shared targets of the factors in this particular cell line. Then we will use a technique find out the aggregate functions of the factors on their individual (inducer or repressor) and common targets (cooperative or competitive). The first half of the workshop would be dedicated to introduce the target package followed by a walk through the workflow interactively in a live demo in the second half.
Participants are expected to walk through the code (rmarkdown document). A brief introduction will be given at the beginning to introduce the package and discussion will be at the end.
Live instances of the workshops can be launched in the cloud freely at http://app.orchestra.cancerdatasci.org/.
To run the materials locally, use the docker image mahshaaban/targetshop and knit the
Rmd files in
vignettes/ from within Rstudio.
An Rstudio session will be accessable at https://localhost:8787/ in the browser. The login username is always
rstudio and the password is
|Walk through the code||20m|
The workshop aims to teach participants how to use R/Bioconductor packages to read in differential expression and binding peaks data, run a predictive analysis and explore its output. I hope that by providing a complete realistic example, participants would develop an understanding of the issues and the importance of integrating those two types of data. Ideally, participants would be able to adapt this code and the workflow to apply this kind of analysis to their own datasets.
targetoutput through the package visualization and testing tools
data.frames and Bioconductor
The workshop is divided into two parts
This workshop is based on a workflow article: a draft