This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Bedtools: Merging Many Bed Files

I am using the algorithm CookHLA for my research. As part of its preparation, I need to feed it a bed file representing at least 100 of my samples.

I have made the bed files for 500 samples using samtools and bedtools in a pipeline:

samtools view -@ CPUs -b --reference $REF_FILE $FILE_NAME.cram chr6:29000000-34000000 | bedtools bamtobed -i stdin > ${FILE_NAME}.bed

I then sorted these bed files with the following:

for FILE in * ;  
    do 
    echo $FILE
    FILE_NAME=$(echo $FILE | rev | cut -c5- | rev)
    bedtools sort -i $FILE > $FILE_NAME.sorted.bed
    done

Now I want to merge all these sorted bed files into one. I have read into the mergeBed function and intersect offered by bedtools. When running these, I am unfortunately getting the following error:

Error: Sorted input specified, but the file cat_beds.bed has the following out of order record
chr6    28999852    29000003    A00266:357:HFKFMDSXY:4:1449:1127:5791/1 60

Clearly there is a sequence outside of the bounds I specified through samtools. Anyone have advice?

bedtools cookhla
• 6,013 views • link
written by J • J

As an addendum, I got that output after a pipeline where I used cat to combine the sorted bed files, then attempted mergeBed

• 0 views • link
written by J • J

2 answers

I got that output after a pipeline where I used cat to combine the sorted bed files

The input to mergeBed shoud be sorted so I think you should do something like this:

cat *.bed | sortBed | mergeBed > merged.bed

where *.bed are the output files from bamtobed (there so no need to sort these files individually).

• 0 views • link
written by dariober dariober

Thanks, I got to that solution as well after looking at it further.

• 0 views • link
written by J • J

Here's another option to merge N files at once:

$ bedops --merge fileA.bed fileB.bed ... fileN.bed > answer.bed

If you have unsorted files and sufficient memory:

$ bedops --merge <(sort-bed fileA.bed) <(sort-bed fileB.bed) ... <(sort-bed fileN.bed) > answer.bed

Or sort them in a loop, depending on system memory, and merge at the end:

$ for fn in `ls *.bed`; do sort-bed ${fn} > ${fn%.bed}.sorted.bed; done; bedops --merge *.sorted.bed > answer.bed
• 0 views • link
written by Alex Reynolds Alex Reynolds

Log in to answer this question.

More posts like this