How extract the longest overlap?

Question

[star]

I have two files that run intersect and find overlap between them using bedtools intersect as bellow: bedtools intersect -a A.bed -b B.bed -wao > intersect.bed . but i would like to extract the longest overlap from my data. Would you please let me know if there is any solution for that.

A.bed

chr   start  end
1     200   250
1     240   300
2     100   120
4     300   360
4     310   400

B.bed

chr   start   end 
1      180   220 
1      210   260  
1      213   348
4      305  352 
4      310   370
4      315  382
4      350  400

Th output of bedtools intersect:

chr start end  chr start end   overlaps.bd
 1  200    250  1  180  220        20
 1  200    250  1  210  260        40 
 1  200    250  1  213  348        37
 1  240    300  1  180  220         0
 1  240    300  1  210  260        20
 1  240    300  1  213  348        60
 4  300    360  4  305  352        47
 4  300    360  4  310  370        50
 4  300    360  4  315  382        45
 4  300    360  4  350  400        10
4  310     400  4  305  352        42
4  310     400  4  310  370        60
4  310     400  4  315  382        67 
4  310     400  4  350  400        50

Expected.bed

chr start  end    chr start end   overlaps.bp 
 1  200    250     1    210  260      40
 1  240   300      1    213  348     60
 4  300    360     4    310  370     50 
 4  310   400      4    315  382     67

[cpad0112]

with datamash, output:

$ datamash  -sH -g 1,2,3 -f max 7 < test.txt | cut -f 1-7 
chr start   end chr start   end overlaps.bd
1   200 250 1   210 260 40
1   240 300 1   213 348 60
4   300 360 4   310 370 50
4   310 400 4   315 382 67

input:

$ cat test.txt 
chr start   end chr start   end overlaps.bd
1   200 250 1   180 220 20
1   200 250 1   210 260 40  
1   200 250 1   213 348 37
1   240 300 1   180 220 0
1   240 300 1   210 260 20
1   240 300 1   213 348 60
4   300 360 4   305 352 47
4   300 360 4   310 370 50
4   300 360 4   315 382 45
4   300 360 4   350 400 10
4   310 400 4   305 352 42
4   310 400 4   310 370 60
4   310 400 4   315 382 67  
4   310 400 4   350 400 50

[Friederike]

did you check bedtools sort that allows you to sort by region size (you would have to modify your original intersect command to only return the overlap instead of -wa and -wb)

[star]

yes I have checked it but it was not useful in my case. I want to extract the largest overlap while the same coordinate in file A have several overlaps with coordinates of B file.

Answer 1

[GouthamAtla]

test.bed

1   200 250 1   180 220 20
1   200 250 1   210 260 40
1   200 250 1   213 348 37
1   240 300 1   180 220 0
1   240 300 1   210 260 20
1   240 300 1   213 348 60
4   300 360 4   305 352 47
4   300 360 4   310 370 50
4   300 360 4   315 382 45
4   300 360 4   350 400 10
4   310 400 4   305 352 42
4   310 400 4   310 370 60
4   310 400 4   315 382 67
4   310 400 4   350 400 50

cat test.bed | sort -k1,1 -k2,2n -k7,7nr | groupBy -g 1,2,3,4 -c 7 -o first -full | cut -f1-7

1   200 250 1   210 260 40
1   240 300 1   213 348 60
4   300 360 4   310 370 50
4   310 400 4   315 382 67

Ps: groupBy is bedtools groupBy

[Friederike]

didn't know about groupby, that's an elegant solution (that doesn't even need the sort if the information from b are not needed)

 bedtools intersect -a test1.bed -b test2.bed  -wao | bedtools groupby -grp 1-3 -c 7 -o max

[star]

Thanks for your solution. what is the -o first -full and what is it difference with -o max -full?

[GouthamAtla]

-o max -full provides the first full line irrespective if its max or not. Sorting makes sure the 'first' is the actual full overlap.

[h.mon]

Did you try geek_y suggestion and did it solve your problem? Please provide feedback by accepting and / or up-voting helpful / correct answers.