Statistics behind variant/consensus calling

[ZHIHUA.LI]

Hi fellow NGSers,

I've been doing NGS data analysis for a while. But lately when I was asked what's the statistics behind the variant/consensus calling of, say, SAMtools, I realized that I never thought of assigning a p value to a variance call. SAMtools document does not talk about statistics except for association studies. Is it because there's no statistical test involved in its variance calling? Do people know of any statistical methods being used in such type of tasks?

Thanks a lot!

[ZHIHUA.LI]

Looks like I was wrong. In SAMtools' FAQ it says:
By default, SNPs are called with a Bayesian model identical to the one used in MAQ. A simplified SOAPsnp model is implemented, too. Indels are called with a simple Bayesian model.

So there is statistics behind the variant calling of SAMtools. But all these models assume a diploid genome, and my sample is haploid. I saw some posts using SAMtools to call SNPs from haploid genome anyway. Maybe I can try that.....