getting acquainted with MotifDb

PriceLab / chip-seq-motif-study

to determine the how TF motifs do and do not match ChIP-seq assays

0 stars 0 forks source link

getting acquainted with MotifDb #2

Open paul-shannon opened 5 years ago

paul-shannon commented 5 years ago

Hi @mariam16548,

I propose sequence matching as your next topic.

There are several things to learn:

MotifDb: a bioconductor package, see http://bioconductor.org/packages/release/bioc/html/MotifDb.html
obtaining sequence for a specified region of the human genome (more on this later)
querying for match statistics, using FIMO (and maybe other tools). see http://meme-suite.org/doc/fimo-tutorial.html

I will give you some simple working code later, by noon I think. Please read into the above topics this morning.

paul-shannon commented 5 years ago

@mariam16548 Motif-to-sequence matching is demonstrated in the new file explore/firstLook/fimoClientDemo.R.

Installation of the FimoClient package, required by the demo, is best approached via

  git clone https://github.com/PriceLab/fimoService.git

paul-shannon commented 5 years ago

@mariam16548 I realize now that, after you clone the repo, you need to cd into the FimoClient subdirectory, install a websocket package (pbdZMQ) then build the FimoClient package itself. Complicated! Let's work on this when I arrive about noon.

paul-shannon commented 5 years ago

@mariam16548

 query(MotifDb, c("CTCF", "hsapiens"))

returns 10 motifs for CTCF. How similar are they? A visual inspection is a good place to start.

If you have notes (as I hope you do) you may already have example code with which you can render these motifs as logos. If you do not yet have a cumulative notes file, then this is a good time to begin!

paul-shannon commented 5 years ago

@mariam16548 There are algorithms and standard software tools to do mathematical (including information theoretic) ways to compare motif logos. I will send you some links later.

mariam16548 commented 5 years ago

Okay, I’ll try my best!

On Jul 23, 2019, at 9:33 AM, Paul Shannon notifications@github.com wrote:

@mariam16548 https://github.com/mariam16548 There are algorithms and standard software tools to do mathematical (including information theoretic) ways to compare motif logos. I will send you some links later.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/PriceLab/chip-seq-motif-study/issues/2?email_source=notifications&email_token=AMNCZQS33KC5XQV7TF4R2ETQA4XGVA5CNFSM4H4TWVQ2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2TWLVY#issuecomment-514287063, or mute the thread https://github.com/notifications/unsubscribe-auth/AMNCZQWMJXUCAANYEBYEH63QA4XGVANCNFSM4H4TWVQQ.