How to merge bedGraph records which are next to each other and have the same score?

Question

[James Ashmore]

Is anyone aware of software to merge bedGraph records if the score is the same? I have a bedGraph file calculated at single base pair resolution and I would like to decrease the file size by merging records next to each other which have the same score.

[Devon Ryan]

awk?

Answer 1

[GouthamAtla]

cat test.bdg

chrY    1   2   10
chrY    2   3   10
chrY    3   4   11
chrY    4   5   12
chrY    5   6   12
chrY    6   7   13
chrY    7   8   14
chrY    8   9   14
chrY    9   10  12

.

cat test.bdg  | groupBy -g 1,4 -c 2,3 -o min,max | awk -v OFS="\t" '{ print $1,$3,$4,$2}'

output:

chrY    1   3   10
chrY    3   4   11
chrY    4   6   12
chrY    6   7   13
chrY    7   9   14
chrY    9   10  12

Bedtools groupBy

[James Ashmore]

Thank you for the reply, however I think this will only work on small bedGraph files. I got the following error when I tried on my base-pair resolution bedGraph file:

terminate called after throwing an instance of 'std::bad_alloc'
  what():  std::bad_alloc
/home/s1437643/conda/bin/groupBy: line 2: 12111 Aborted                 ${0%/*}/bedtools groupby "$@"

[GouthamAtla]

Did it work for small file and failed for large file ?

[James Ashmore]

Actually it ended up being an error in the bedtools installation I had, specifically the groupBy command. I've compiled from the last source code and it works perfectly. Thank you!

Answer 2

[Alex Reynolds]

Via BEDOPS, bash and GNU core utilities:

$ SORTED_BEDGRAPH=in.bedGraph
$ while read -r score; do awk -v s=$score '$4==s' ${SORTED_BEDGRAPH} | bedops --merge - | awk -v s=$score '{print $0"\t"s}'; done < <(cut -f4 ${SORTED_BEDGRAPH} | sort | uniq) | sort-bed - > answer.bed

This should perform decently and scale to large inputs.

Use sort-bed if you first need to sort the bedGraph file, so that merging works correctly.

For others working with BED, instead of bedGraph, the score data are usually in the fifth column, which would need adjusting of the two awk and cut -f4 statements.