Read group and mpileup

[adrian]

Hi:

Why is it important to have a read group. Is it to keep track of
samples belonging to a family or a phenotype etc. OR is it useful to
optimize the variant calls among samples due to coverage issues.

What I mean:

For instance I have a family of 4 (Mother, father, 2 children) and 5
samples from this family (4 germ line from members and 1 tissue sample
from affected child).

Likewise I have 4 different families. Now using exome sequencing data,
I might not have equal coverage among a family and across families for
a locus. In a family that locus might be interesting , but due to
coverage in affected child, it is difficult to call that locus as a
significant call. In such case, If I specify read group, based on RG
info, will mpileup optimize the SNV at that locus and call it
significant?

Is this why we need a read group? Is mpileup works this way. Sorry for
asking a naive question.

thanks
Adrian

[n00c]

I don't know enough about mpileup, but in my case I found setting the RG tag useful because it allowed me to mark reads as belonging into two separate libraries for each of the individual genomes I worked with. Because the reads from the two libraries were separated thanks to the read group data, tools such as Picard MarkDuplicates and Dindel were able to create separate distributions of insert sizes for each of the libraries (otherwise all reads were lumped together into one distribution, at the expense of precision).