MACS output format (NAME_peaks.bed)

[rebrendi]

Hi all,

I am analyzing a ChIP-Seq output file with TF enrichment peaks from the published dataset. It looks like the data are in the MACS output format. The files are named "NAME_peaks.bed" and the content looks like this:

chr1 2011287 2011686 . 34 +

I guess the columns, left to right, are the chromosome name, peak start, peak end, the last one is the strand, but I don't know what is in between. Do you have any ideas what is in the 4th-5th columns? Also, do you have any ideas why the distance between the start and end is so large, and how one would extract the transcription factor binding site location from this broad peak?

Thanks!

[dpryan]

It's a BED file, so just read the spec. From the MACS readme file, the 5th column is the "-10*log10pvalue of peak region".

[rebrendi]

dpryan, thank you.

concerning my second question: is it safe to assume that the TF binding site is located in the middle of the identified broad peak (start+end)/2 ?

[dpryan]

I wouldn't assume that. You could use something like meme for motif discovery.

[rebrendi]

dpryan, I am not looking for motifs, just need the coordinates for each TF binding site (assuming that each peak contains one binding site). There is no single consensus motif for this TF, that is known. Do you think there is a simple way to use these peak.bed files to detect the coordinates of the binding sites?

[mudshark]

basic question: you sure your file is a MACS output file?

[rebrendi]

basic question: you sure your file is a MACS output file?

nope. it is provided as is by the authors

[mudshark]

so it has been published, any description on how the peak calling was performed? did you ask the authors?
are the raw data also published? maybe you better do the peak calling yourself then

[rebrendi]

so it has been published, any description on how the peak calling was performed? did you ask the authors?
are the raw data also published? maybe you better do the peak calling yourself then

you are right, have to contact the authors...

[anurag.gautam]

HI,
Once we are done with identifying the peaks for chip seq data, how can we validate those results in wetlab. Since these all are bioinformatics predictions, are there any techniques other than PCR to validate the peaks we called??

[mudshark]

that all depends on your target, model organism etc etc.
reporter gen assays are probably most straight forward in any case.