Encode Chip-Seq Data To Sequence Logos In R

Question

[Pascal]

For example, http://www.factorbook.org/mediawiki/index.php/TCF7L2

I have downloaded all the files of this data set (i.e. .bam, .bam.bai, .narrowPeak, .fastq and .bigWig) and I would now like to produce the first of the shown sequence logos. How exactly can I do that in R?

Answer 1

[Steve Lianoglou]

Here are a few steps you can follow to begin to figure it out:

1. Go to the list of analysis packages provided by bioconductor.
2. Use the search functionality in your browser to look for "motif"
3. Read through the vignettes of those packages to identify the one (of several) that do these sorts of things.
4. Optionally: please update your question with a comment letting us know which one works best for you.

Answer 2

[Alex Reynolds]

I have used bits and piece of the seqLogo package in my own scripts. To use this R package to make sequence logos with your data, I suspect you would need to:

1. Get the sequence data of the motif associated with your ChIP-seq hits from one of your datasets, or use coordinate data with other databases (UCSC Genome Browser, for instance) to get sequence data
2. Compute the PFM from your motif sequence data
3. Compute the PWM from your PFM (more reading)
4. Feed the PWM into seqLogo and make the sequence logo per documentation

Another option is to contact the Factorbook people and ask for the exact recipe used to recapitulate the sequence logos on their page.

A third option, once you have the PWM, is to use the Python-based WebLogo kit to make your own sequence logos on the command line. The logos you see on Factorbook look like they were made with WebLogo, for instance. (Though, on closer examination, it looks like they were built with the sequence logo code in MEME.)