Re: Digest for bedops-discuss@googlegroups.com - 4 updates in 1 topic

8 views

Skip to first unread message

Shane Neph

unread,

Sep 21, 2015, 9:05:00 PM9/21/15

to bedops-...@googlegroups.com

How about:

bedmap --echo --echo-map regions.bed main.bed

For each row of regions.bed, you will get a semicolon separated list of regions from main.bed that overlap. Add --skip-unmapped if you don't want an output for the case where an element of regions.bed has no overlapping element in main.bed. Here, I included --echo which will output information from regions.bed too.

Example output:

chr11 13202 14981 2|chr11 13302 13303 1;chr11 13327 13328 1;chr11 13980 13981 1

chr11 13980 41480 2|chr11 13980 13981 1;chr11 30923 30924 1

chr11 30923 51480 2|chr11 30923 30924 1;chr11 51476 51477 1;chr11 51479 51480 1

You can use --delim to change the '|' or --multidelim to change the ';'. Follow-on cut or awk statements would likely follow. I'd argue that you should try to keep everything in one file rather than multiple - it requires a bit more coding but scales much better if regions.bed has a lot of entries.

Shane

On Mon, Sep 21, 2015 at 4:13 PM, <bedops-...@googlegroups.com> wrote:

bedops-...@googlegroups.com Google Groups

Topic digest
View all topics

output overlap for each region in a file - 4 Updates

output overlap for each region in a file

francy....@gmail.com: Sep 20 08:40PM -0700

Dear Experts,

I have a sorted large BED file (main.bed) and several regions defined in a
file (regions.bed) which can be overlapping with each other. I am trying to
find a quick way to output the overlaps from the main BED file with the
regions defined in regions.bed, and create a separate BED output file for
each of these regions.

"bedops --everything" seems maybe would work, but is there a way to output
different intersects between main.bed and regions.bed in a different file
for each line of regions.bed?

Thank you very much for any suggestions,
Fra

Alex Reynolds <alexpr...@gmail.com>: Sep 20 10:50PM -0700

You could pipe the results of BEDOPS *bedmap* to *split*:

$ bedmap --echo --echo-map regions.bed main.bed | split -l 1 -a 5 -
bedmap_result_

For each line in *regions.bed*, you would get files called
*bedmap_result_aaaaa*, *bedmap_result_aaaab*, and so on.

See *man split* for more information on the *-l* and *-a* options used in
this example.

francesca casalino <francy....@gmail.com>: Sep 21 09:04AM -0400

This is not giving me what I was looking for though, it gives me many
output files with 1 line indicating the overlap between the two bed files,
but what I was hoping to get is, in each output file, the subset of
overlaps in main.bed that are included in each region defined by
regions.bed. So I am looking to obtain one output file for each region
which should contain many entries since main.bed is based on a 1 base pair
while regions is many base pairs. This will also mean that several entries
in the output files will be duplicated with other output files since the
regions can be overlapping... Is this possible to do?

One example is:

cat main.bed

chr11 13302 13303 1

chr11 13327 13328 1

chr11 13980 13981 1

chr11 30923 30924 1

chr11 51476 51477 1

chr11 51479 51480 1

cat regions.bed:

chr11 13202 14981 2

chr11 13980 41480 2

chr11 30923 51480 2

And the output should be:

cat res.region1

chr11 13302 13303 1

chr11 13327 13328 1

chr11 13980 13981 1

cat res.region2

chr11 13980 13981 1

chr11 30923 30924 1

cat res.region3

chr11 30923 30924 1

chr11 51476 51477 1

chr11 51479 51480 1

francesca casalino <francy....@gmail.com>: Sep 21 09:04AM -0400

Thank you Alex!

But this is not giving me what I was looking for though, it gives me many
output files with 1 line indicating the overlap between the two bed files,
but what I was hoping to get is, in each output file, the subset of
overlaps in main.bed that are included in each region defined by
regions.bed. So I am looking to obtain one output file for each region
which should contain many entries since main.bed is based on a 1 base pair
while regions is many base pairs. This will also mean that several entries
in the output files will be duplicated with other output files since the
regions can be overlapping... Is this possible to do?

One example is:

cat main.bed

chr11 13302 13303 1

chr11 13327 13328 1

chr11 13980 13981 1

chr11 30923 30924 1

chr11 51476 51477 1

chr11 51479 51480 1

cat regions.bed:

chr11 13202 14981 2

chr11 13980 41480 2

chr11 30923 51480 2

And the output should be:

cat res.region1

chr11 13302 13303 1

chr11 13327 13328 1

chr11 13980 13981 1

cat res.region2

chr11 13980 13981 1

chr11 30923 30924 1

cat res.region3

chr11 30923 30924 1

chr11 51476 51477 1

chr11 51479 51480 1

Back to top

You received this digest because you're subscribed to updates for this group. You can change your settings on the group membership page.
To unsubscribe from this group and stop receiving emails from it send an email to bedops-discus...@googlegroups.com.

Shane Neph

unread,

Sep 22, 2015, 6:53:34 AM9/22/15

to bedops-...@googlegroups.com

Consider:

bedmap --echo-ref-name --echo-map regions.bed main.bed \

| awk -F"|" 'BEGIN {OFS="\t"} ; { lng=split($2, arr, ";"); for(i=1; i<=lng; ++i) { print arr[i]"\t"$1 } }' \

| sort-bed -

The output is all in 1 file and otherwise duplicate entries from main.bed have a column that shows to which element in regions.bed it belongs to.

Example output:

chr11 13302 13303 1 chr11:13202-14981

chr11 13327 13328 1 chr11:13202-14981

chr11 13980 13981 1 chr11:13202-14981

chr11 13980 13981 1 chr11:13980-41480

chr11 30923 30924 1 chr11:13980-41480

chr11 30923 30924 1 chr11:30923-51480

chr11 51476 51477 1 chr11:30923-51480

chr11 51479 51480 1 chr11:30923-51480

Shane

Reply all

Reply to author

Forward

0 new messages