bam-readcount usage

[ElenaN]

Hi, could someone tell me what is the proper way to run bam-readcount (for later use with the VarScan2 fpfilter.pl?)
Specifically, what should be the "list of regions to report readcounts within" (-l parameter): the SNPs found (from .vcf file)? the exome .bed file?
And how is the "optional region specificaton" different?
An example of a command-line call to bam-readcount would also be useful.
Thank you!
- Elena

[ernfrid]

Hi Elena,

You should use the SNPs found for your import to -l. The file should be tab separated, and contain chr start stop in 1-based (as in VCF not BED) coordinates for both start and stop. This is similar to the optional region specification except that it handles multiple sites at once and that the format for the optional region specification is the same as that for samtools view.

This command would filter out bases below BQ of 15 and report counts for each site in the varscan.positions file.

bam-readcount -b 15 -f reference.fa -l varscan.positions some.bam

This command prints the counts for position 10000 on chromosome 1 of the reference sequence.

bam-readcount -b 15 -f reference.fa some.bam 1:10000-10000

This command prints counts for all of chromosome 1.

bam-readcount -b 15 -f reference.fa some.bam 1

This command prints counts from position 10000 of chromosome 1 to its end.

bam-readcount -b 15 -f reference.fa some.bam 1:10000

Hope that helps,

Dave

[ElenaN]

Thanks.
So the -b value should be what the user considers to be a reasonable cutoff?

[ernfrid]

Yes, I would recommend that users choose -q and -b cutoffs that they consider reasonable. As this depends on how the data is being used, I have not set defaults on the program itself.

[himanshu04]

Error while using bam-readcount

I have tumor-normal sample and I have high confidence somatic calls from Varscan which I then filtered using Somatic filter. I wish to use Bam-read count on them and then use the fpfilter to bring it down to most confident set of calls.
I am not able to understand which bam file should I give since this is a combination of normal-tumor.
Also, I tried to use the tumor file but it gave me an error saying single quality mapping not found. Check if SM group present in the bam file.
What should I be doing for making it work?.
I am a newbie to varscan and any help will be much appreciated.
Also, I must tell you that I am running it on RNA seq data from Mice models.
Thanks,
Himanshu

[ernfrid]

Hi Himanshu,

You should use the tumor BAM file only. You can also safely ignore the errors about the single end mapping quality, it will not affect your filtering results. There is a fix to report this only once in the current source, but we have not yet released a new stable version.