find overlaps/common in multiple bed file

[epi]

Does anyway have experience to find commonalies (or overlaps) between bed files. I have 4 bed files and want to find what intervals are common to all.

And I don't have any rigid criteria for overlap, any intersection will do.

Appreciate any answers.

[dpryan]

Have a look at bedtools, specifically the intersectBed utility.

[GenoMax]

http://pypi.python.org/pypi/pybedtools is a python extension of the bedtools that was mentioned in the post #2 by dpryan.

[husamia]

Originally Posted by epi

Does anyway have experience to find commonalies (or overlaps) between bed files. I have 4 bed files and want to find what intervals are common to all.

And I don't have any rigid criteria for overlap, any intersection will do.

Appreciate any answers.

I mentioned the wrong function I meant to mention Galaxy with the Intersect function in the operate on genomic intervals functions
here is example


http://main.g2.bx.psu.edu/

[epi]

Originally Posted by dpryan

Have a look at bedtools, specifically the intersectBed utility.

Thanks for response everyone !

I could implement intersectBed as pairwise. And to find common among 4, I can do multiple pairwise, but it seems there is some chance for false negatives.

I dint get chance to look at the python script so far, I wonder if that adresses this issue.

[husamia]

I mentioned the wrong function, I modified my response above to reflect the correct reference function. Sorry for confusion

[dpryan]

Originally Posted by epi

Thanks for response everyone !

I could implement intersectBed as pairwise. And to find common among 4, I can do multiple pairwise, but it seems there is some chance for false negatives.

I dint get chance to look at the python script so far, I wonder if that adresses this issue.

The python interface isn't very different from direct command line usage and I would suspect produces the same results. I also don't see how you'd get a false negative, provided you actually want regions existing in all of the bed files.

[epi]

Originally Posted by dpryan

The python interface isn't very different from direct command line usage and I would suspect produces the same results. I also don't see how you'd get a false negative, provided you actually want regions existing in all of the bed files.

Thanks again for responding. I realize i did not state my objective well enough.
This is chip-seq analysis for which I have bed files (peaks). There could be a situation when peak 1 intersect peak 3 at 5' and peak 2 intersect peak 3 at 3`. but peak 1 and peak 2 do not intersect.
Read Peak1, Peak2 and Peak3 coming form Samples 1,2 and 3 please.
intersectBed will not reveal these peaks on my first paiwise comparison (peak1 and peak2) so it will be gone. Logically, it seems they come from same region so I was wondering if there is a tool that could capture those.

I must mention that this is not a real example, just i theoretical possibility which crossed my mind. May b I am just too obsessed over it

[mgogol]

You want multiIntersectBed...

http://biostar.stackexchange.com/que...iple-bed-files

[arvid]

Originally Posted by epi

Read Peak1, Peak2 and Peak3 coming form Samples 1,2 and 3 please.
intersectBed will not reveal these peaks on my first paiwise comparison (peak1 and peak2) so it will be gone. Logically, it seems they come from same region so I was wondering if there is a tool that could capture those.

You can include both A and B regions in the output from intersectBed (e.g. with -wo), merge them with mergeBed and then intersect the merged output to the next sample (to "grow" the overlapping regions).

[epi]

Originally Posted by mgogol

You want multiIntersectBed...

http://biostar.stackexchange.com/que...iple-bed-files

bullseye !!

Looks like can the job, will try it out ...
appreciate your response

[sjneph]

If you prefer a much more scalable solution that can do this simple intersection (and any other set-like operation) on any number of bedfiles at once, check out BEDOPS.

http://code.google.com/p/bedops/