how to intersect multiple bed files and get results in uncollapsed genomic intervals

Question

[Alewa]

Is there anyway I can get the results on intersection of multiple bed files in uncollapsed format? i.e., the result should have interval/width of 1.

Here's the result I get in collapses form

(base) bash-4.2$ bedtools multiinter -header -names gr1 gr2 gr3 -i gr1.bed gr2.bed gr3.bed > collapsed_multiIntersect.bed

(base) bash-4.2$ cat collapsed_multiIntersect.bed
chrom   start   end     num     list    gr1     gr2     gr3
chr1    0       1       1       gr2     0       1       0
chr1    1       6       2       gr1,gr2 1       1       0
chr1    6       7       1       gr1     1       0       0
chr2    0       2       2       gr1,gr3 1       0       1
chr2    2       6       3       gr1,gr2,gr3     1       1       1
chr2    6       8       2       gr2,gr3 0       1       1
chr2    8       11      1       gr3     0       0       1

Here's example of data

(base) bash-4.2$ cat gr1.bed
chr1    2       2       1       *       0.7276748239528388      67.1151612419635
chr1    3       3       1       *       0.6065833754837513      49.11124410573393
chr1    4       4       1       *       0.17338064685463905     89.53636873047799
chr1    5       5       1       *       0.34515149844810367     44.35192057862878
chr1    6       6       1       *       0.656820923788473       91.98388278018683
chr2    1       1       1       *       0.21105075324885547     37.15807143598795
chr2    2       2       1       *       0.24114407738670707     49.36463311314583
chr2    3       3       1       *       0.8360638627782464      92.60302006732672
chr2    4       4       1       *       0.8508221041411161      44.008928118273616
chr2    5       5       1       *       0.8572670503053814      3.721454436890781

(base) bash-4.2$ cat gr2.bed
chr1    1       1       1       *       0.6819672391284257      97.34485428780317
chr1    2       2       1       *       0.481787727214396       30.3413461195305
chr1    3       3       1       *       0.3573118313215673      64.61741852108389
chr1    4       4       1       *       0.51849395618774        27.47685764916241
chr1    5       5       1       *       0.8251491808332503      83.9870688971132
chr2    3       3       1       *       0.42395976139232516     1.6344564035534859
chr2    4       4       1       *       0.6377549672033638      37.02820355538279
chr2    5       5       1       *       0.27647473104298115     87.21347157843411
chr2    6       6       1       *       0.6640123652759939      55.96593914087862
chr2    7       7       1       *       0.8544245520606637      64.64025254826993

(base) bash-4.2$ cat gr3.bed
chr2    1       1       1       *       0.9348035349976271      52.820321382023394
chr2    2       2       1       *       0.5684890144038945      92.6313241943717
chr2    3       3       1       *       0.47866105823777616     7.479210733436048
chr2    4       4       1       *       0.4196885135024786      95.4758029896766
chr2    5       5       1       *       0.4240405154414475      39.72786704543978
chr2    6       6       1       *       0.6786996794398874      60.717192641459405
chr2    7       7       1       *       0.7414848282933235      9.92186910007149
chr2    8       8       1       *       0.6762841297313571      56.700556515716016
chr2    9       9       1       *       0.19594529480673373     44.58376122638583
chr2    10      10      1       *       0.9014762006700039      95.27598922140896

Answer 1

[Alex Reynolds]

Generally, you could use bedops --chop with some piping and process substitutions with bedmap to get count/ID/score data:

bedops --chop 1 some_regions.bed | bedmap --echo --count --echo-map-id - <(bedops --everything signal1.bed signal2.bed ... signalN.bed) > answer.bed

The file some_regions.bed could come from a simple merge operation upstream, e.g.:

bedops --merge signal1.bed signal2.bed ... signalN.bed | bedops --chop 1 - | bedmap --echo --count --echo-map-id - <(bedops --everything signal1.bed signal2.bed ... signalN.bed) > answer.bed

Files signal1.bed through signalN.bed would have a unique identifier in the ID (fourth) column ("gr1", "gr2", whatever).

Note: Your gr1.bed, gr2.bed and gr3.bed do not have correct BED coordinates. You'll need to run them through awk or similar to fix them, e.g. awk -vFS="\t" -vOFS="\t" '{ $2 -= 1; print $0; }' in.bed > out.bed or similar.

[Alewa]

thank you so much Alex!