getting rid of 0 entries in bedTools genomeCoverageBed tool

[efoss]

I am trying to use genomeCoverageBed in the bedTools suite to find coverage of each base in the human genome from an RNA seq bam file that has been generated with TopHat. The output file gets enormous because all of the positions with 0 hits are included. Is there a way I can get the output file to include only chromosome-coordinate pairs that have non-zero hits? This is the command that I use to generate the output file:


genomeCoverageBed -ibam bamfile.bam -g human.hg19.genome -d -split > genome_coverage_split_071711_1.txt

Thanks.

Eric

[biznatch]

So this is basically making a .wig file? Is it a fixedStep (one column) or variableStep (two column) file? If you have a lot of 0's it is probably better to use variableStep. I haven't used bedTools so I don't know if that is an option within that program.

[dariober]

I am trying to use genomeCoverageBed in the bedTools suite to find coverage of each base in the human genome from an RNA seq bam file that has been generated with TopHat. The output file gets enormous because all of the positions with 0 hits are included. Is there a way I can get the output file to include only chromosome-coordinate pairs that have non-zero hits? This is the command that I use to generate the output file:


genomeCoverageBed -ibam bamfile.bam -g human.hg19.genome -d -split > genome_coverage_split_071711_1.txt

Thanks.

Eric

Hi,
Assuming the BED output looks something like this (where line2 is to be filtered out):

chr1	1	10	line1	1	+
chr2	100	1000	line2	0	-
chr3	200	210	line3	10	-

You could parse the output of genomeCoverageBed with grep:

genomeCoverageBed -ibam bamfile.bam -g human.hg19.genome -d -split | grep -P -v .+?[[:blank:]].+?[[:blank:]].+?[[:blank:]]0[[:blank:]].+? > genome_coverage_split_071711_1.txt

(Untested... Make sure it does what you want!)

Dario

[idonaldson]

Were you using the -d flag (reports coverage for every single base)? I once used:

awk '$3>0' < in.file > out.file

This ensures the bedTools output only has coverage greater then 0.

[efoss]

Thanks very much, all. I'll look in to these solutions.