VarScan pileup2snp VarFreq calculation

[petter]

I have a question about how VarFreq is calculated in pileup2snp.

I’ve got these results:

Chrom 1
Position 260300
Ref G
Cons R
Reads1 3
Reads2 6
VarFreq 50%
Strands1 1
Strands2 2
Qual1 42
Qual2 22
Pvalue 0.004525
MapQual1 1
MapQual2 1
Reads1Plus 3
Reads1Minus 0
Reads2Plus 4
Reads2Minus 2
VarAllele A


Shouldn’t VarFreq be #Reads2/(#Reads1 + #Reads2)=3/(3+6)=1/3?

and at another position I've got:

Reads1 0
Reads2 8
VarFreq 80%

What am I missing here?

Also, when I look at the bam file in tablet there are more than 100 reads. Why only 9 reads in output from pileup2snp?

Kind Regards
Petter

[petter]

snp freq calculation

Can anyone comment or answer my question above? What is the formula for VarFreq? It has to involve some other measurements (e.g quality scores) than just read counts. Otherwise the output is slightly wrong for many of the variants.

regards Petter

[Jane M]

I do not use the pileuptosnp command so I cannot help on your VarFreq problem.
But concerning

Also, when I look at the bam file in tablet there are more than 100 reads. Why only 9 reads in output from pileup2snp?

Did you run mpileup with -A and especially -B options to generate your pileups? (you can find several topics about this issue on seqanswers)

[petter]

mpileup2snp

Thanks Jane,

The -B and -A option using mpileup helped a lot.

Still though I can't understand how the varfreq (freq column in mpileup2snp) is calulated? Anyone who can answer this?