MACS2 peaks and MEME-suite analysis

[TRypdal]

Hi

Has anybody here used the results of MACS2 as an input to the MEME suite of tools for motif finding? 

I'm starting to do a simple analysis of this kind on my chipseq data and I'm interested in knowing what is the best way to pre-select a subset of peaks from the MACS bed/narrowpeak output to pass onto MEME. MEME seems to be very efficient when dealing with about 200/300 peak intervals If I understand correctly. I have about 2500. Would it be a good idea to sort the peaks by p-value, or maybe by signal (tracks 7/8 in the narrowPeak file). Alternatively what would experts recommend? Thanks

[pbczyd .]

Hi,

you may consider using IDR to get "top"or "optimal" picks from your data. 

See: https://sites.google.com/site/anshulkundaje/projects/idr

Regards,

Sheng

--
You received this message because you are subscribed to the Google Groups "MACS announcement" group.
To unsubscribe from this group and stop receiving emails from it, send an email to macs-announcem...@googlegroups.com.
To post to this group, send email to macs-ann...@googlegroups.com.
Visit this group at http://groups.google.com/group/macs-announcement.
For more options, visit https://groups.google.com/groups/opt_out.

[Ian Donaldson]

MEME (online) takes a maximum of 600 sequences that 100bp in length.  I have usually avoided MEME because of this limitation in favour of Weeder that does not have such a conservative limit.  However, I recently used MEME and got some very nice results, using the top 600 sequences from the MACS list filtered to remove regions with fold enrichment <5 (ordered by q-value or p-value).

Ian

[francis.poulat]

MEME is not very efficient but everybody use it, then people reading publications use it…..

I have experienced  "peaks motifs" on RSAT website where you can  scan all the peak you want directly on the website without downloading the program, installing it on your computer….I've scan up to 10000 peaks without any problems. Moreover you can inject the database of matrices that you want directly online.

Another interesting tool is matrix-scan which allow to scan your peak with your favorite transcription factors matrices, it's much more efficient than MEME!

you can find all these tools here : http://rsat01.biologie.ens.fr/rsa-tools/index.html

and here : http://rsat.ulb.ac.be/rsat/  (there is 8 mirors)

good luck

francis

[Vince]

Hi,

I think it is better to use all peaks to make sure that the motifs are found in a range of possible qualities.  There are a number of tools that allow you to do this for large datasets quickly.  I like to use the peak position plus (an arbitrary) 50 bp on either side, rather than the whole called region.  Here are some tools

Homer
http://biowhat.ucsd.edu/homer/chipseq/peakMotifs.html

meme-chip and DREME
http://meme.nbcr.net/meme/

chipmunk
http://bioinformatics.oxfordjournals.org/content/26/20/2622.full

chipseeqer
http://physiology.med.cornell.edu/faculty/elemento/lab/chipseq.shtml

rsat
http://nar.oxfordjournals.org/content/early/2011/12/08/nar.gkr1104.full

motifRG
http://www.bioconductor.org/packages/2.12/bioc/html/motifRG.html

posmo
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3326300/

Vince

[TRypdal]

Thanks for your help. I had some interesting results using the MEME-chip suite and the following input:

-pick peak summits from MACS2 summits.bed file

-sort by p-value and obtain the first 1500 peaks

-extend the coordinate of each peak summit by 100bp upstream and downstream

The interesting bit is that MEME-chip will run a suite of tools (Centrimo, TomTom) and will not limit the number of input sequences.