bedops wig2bed issue

Question

[rbronste]

Hi using the following command and either getting an empty file or an empty STDIN through sub:

convert2bed --input=WIG [--output=BED] < GSM1724233_mm9_liver_ERWT_ER_E2_RMR-32-42b.wig > GSM1724233_mm9_liver_ERWT_ER_E2_RMR-32-42b.bed

Any ideas? Thanks!

Rob.

[Alex Reynolds]

How large a file is GSM1724233_mm9_liver_ERWT_ER_E2_RMR-32-42b.wig?

[rbronste]

Just under 2gb. I saw the memory options but wasn't sure about how it would handle large wig files....

I assign 24gb memory through qsub on the cluster but was just an empty STDIN.

[Alex Reynolds]

What happens if you run $ convert2bed --input=wig --do-not-sort < in.wig | head?

[rbronste]

I get the following:

$ convert2bed --input=WIG [--output=BED] --do-not-sort < GSM1724233_mm9_liver_ERWT_ER_E2_RMR-32-42b.wig | head
chr1    3000370 3000380 id-1    2.000000
chr1    3000380 3000390 id-2    5.000000
chr1    3000390 3000400 id-3    5.000000
chr1    3000400 3000410 id-4    5.000000
chr1    3000410 3000420 id-5    5.000000
chr1    3000420 3000430 id-6    5.000000
chr1    3000430 3000440 id-7    5.000000
chr1    3000440 3000450 id-8    5.000000
chr1    3000450 3000460 id-9    5.000000
chr1    3000460 3000470 id-10   5.000000

[Alex Reynolds]

Conversion looks good. Looks like it's a temporary file issue, probably. Memory allocation for your conversion job doesn't need to be much higher than 2 GB. A 3 GB allocation would more than handle conversion and sorting, but you might need to specify a different temporary directory if the compute node your job goes on gets its /tmp folder filled up. See here for a previous discussion and options.