how to split BED file according to chromsome

[sunsnow86]

Does anyone know a program which can split BED file according to the chromosome? I have generate a BED file which contains the data for all chromosome, but it is not sorted. When I did sorting using BedSort， the output was not ordered according the numeric order, it always give chr10 on the top and then followed chr11, up to chr19. It seems I have to do the sorting for each chr respectively, I wonder whether there is a program which can split BED file according to the chromosome. Thanks

[zee]

You could try the following with your bed file:

sort -k 1V,1 -k 2n,2 file.bed -o file.sorted.bed

if you want to split your bed file you could do with bash:

mkdir -p split_results
for chr in `cut -f 1 file.bed | sort | uniq`; do
                grep -w $chr file.bed > split_results/$chr.output.bed

done

[adamdeluca]

An alternative:

awk '{close(f);f=$1}{print > f".bed"}'

[quinlana]

Similar to adamdeluca's suggestion, here is another simple awk solution. Note that the ">>" creates and appends to files named CHROM.bed, where CHROM is column 1 of the bed input bed file (in this case, example.bed).

So, in plain English, the awk command prints each entire line ($0) from example.bed to distinct files that are each named by the chrom field ($1).

This strategy is useful in many other cases where you want to do a context-based "grep", and route the results to distinct files.

$ awk '{print $0 >> $1".bed"}' example.bed

$ ls -1 *.bed
chr1.bed
chr2.bed
... (snip)
chrY.bed
example.bed

arq

[sunsnow86]

Thank you !

Many thanks to you guys! I have worked it out.