Somatic calling comparing differnt platform BAMs

[Mamun]

Hi guys,

I am trying to make some somatic calling using ( varscan / mpileup ) a tumor
sample from whole exome sequencing data and Normal sample from whole
genome data.

Size of the BAM files

Whole exome tumor ( 6.7 gb )
Whole genome tumor ( 18 gb )

I have ~6 million calls from this procedure.

I was wondering is it appropriate to perform somatic calling with two samples from
different platform.

thanks in advance,
Mamun

[adaptivegenome]

If this is the only data you have, then you can certainly try. However, it is probably a bad idea for many reasons. Tools are often optimized to call genotypes and with the big differences in coverage (exome is usually ~100X; genome is usually ~30X) this could skew results especially when you are assessing changes that arise in the tumor.

[sdvie]

just one thought:

The exome sample must have been submitted to some enrichment in the experimental part of the workflow. Usually you would have a .bed file containing the enriched regions. So one thing to do, would be to select from the whole genome data only the regions that were enriched for the exome. Then, the coverage might still be different between the two bam files.... you might consider adjusting the amount of reads from the exome experiment by downsampling the number of reads in the exome bam file, as discussed here.