VarScan for heterozygous calls

[Irina Pulyakhina]

Dear all,

I'm running VarScan:
java -jar ./VarScan.v2.2.8.jar mpileup2snp --min-var-freq 0.1 --output-vcf --variants input.mpileup

And I get two output files, with SNPs and indels. I have three questions:
1) how do I identify heterozygous SNPs? I thought I should see something like "A,G" or "A/G", but I see only single nucleotides in the corresponding column.
2) Can I get raw counts for the allele frequencies (not only one, DP, as I get it now for homozygous calls)
3) what does "str10" in the field "QUAL" mean?

Thank you in advance!
Irina

This is how the file look like:

#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT Sample1
chr1 20729 C T . PASS DP=10 GT:GQP 0/1:2:10
chr1 73935 G A . PASS DP=43 GT:GQP 0/1:3:43
chr1 73939 A C . PASS DP=44 GT:GQP 1/1:6:44
chr1 73978 A G . PASS DP=119 GT:GQP 0/1:12:119
chr1 73996 T C . PASS DP=66 GT:GQP 0/1:2:66
chr1 74008 A C . PASS DP=47 GT:GQP 0/1:3:47
chr1 74010 A C . PASS DP=42 GT:GQP 0/1:5:42
chr1 74020 A C . PASS DP=34 GT:GQP 0/1:3:34

[brofallon]

Take at look at the last column. The bit before the first colon reads 0/1 or 1/1. 0/1 implies that heterozygous is the most likely genotype at the position, 1/1 means that homozygous non-reference is most likely. (The / means the calls are unphased, you may see a | if you've phased the variants somehow).
HTH
B

[dkoboldt]

I strongly recommend updating to the latest release of VarScan (v2.3.3), which outputs more extensive information into VCF format including the read counts supporting each allele. Otherwise, you could get that information using your current version if you didn't output to VCF format, but used VarScan's native output format instead.

[yangjr]

Same problem here

I think Irina actually means she's expecting something like this:

#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT Sample1
chr1 20729 C T,A . PASS DP=10 GT:GQP 0/1:2:10
chr1 73935 G A,T . PASS DP=43 GT:GQP 0/1:3:43

Note that it is heterozygous calls, not heterozygous sites. I hand-pick some pretty good heterzygous SNP (DP>30 for either allele, and again, either allele is the reference allele, no quality problems), but VarScan seem to just cut it down to the most likely alternative allele, but throw the other alternative allele.

Could someone recommend other tools that could pick up the secondary alternative allele? I know that mpileup could do it, but it just ignores some other obvious regular SNPs in my case and VarScan seems better on that issue. So I just wish VarScan could deal with the 2ndary alternative allele too.

jianrong