Re: [Vcftools-help] HELP: How is QUAL score recalculated by vcf-merge?

[Petr D. <pd...@sa...>]

Hi Peter and Mark,

On Mon, 2012-10-08 at 12:57 +1000, Peter Johansson wrote:
> Isn't the QUAL defined as "Phred-scaled probability of all samples being 
> homozygous reference"? If that is correct, I think it follows that we 
> merge two data sets with QUAL = Q1 and Q2, respectively, the merged QUAL 
> will be greater than both Q1 and Q2, since
> P(all sample are ref/ref) <= P(all samples in a subset are ref/ref)

for biallelic sites, yes. Note that the specification says that QUAL is
defined as probability of the assertion made in ALT is wrong. Therefore
it is not just probability of all samples being hom ref, but generally
any other combination of alleles not listed in ALT as well.

> Thanks,
> Peter
> 
> 
> 
> On 10/05/2012 11:30 PM, Mark Aquino wrote: 
>
> > Thanks Petr, so you
> recommend using the geometric mean of the quality scores where
> available and to omit no calls from the calculation?

I would not go so far as to make such a recommendation. It's just one of
the possible ways to do it and not necessarily the best one. Certainly
one of the simplest ones.

Petr




> > Best,
> > Mark
> > Sent from my iPhone
> >
> > On Oct 5, 2012, at 3:33 AM, "Petr Danecek"<pd...@sa...>  wrote:
> >
> >> Hi Mark,
> >>
> >> the current implementation is somewhat naive. It is similar to geometric
> >> mean but weighted by the number of samples in the VCF, so that when a
> >> VCF with one sample is being merged with a VCF with many samples, the
> >> quality from the first does not override the latter. Better would be to
> >> use the individual quality scores, but this is not always available.
> >>
> >> Petr
> >>
> >> On Thu, 2012-10-04 at 21:03 +0000, Mark Aquino wrote:
> >>> When multiple quality scores are combined from different VCF files
> >>> (via vcf-merge), how is the NEW quality score in the multi-sample vcf
> >>> calculated using the individual quality scores from the single-sample
> >>> vcfs? I've tried everything from geometric means to converting back to
> >>> p-values and re-calculating the Phred score and nothing gives me the
> >>> same value that VCFTools calculates!
> >>>
> >>>
> >>>
> >>>
> >>> Thanks,
> >>> Mark
> >>
> >>
> >>
> >> -- 
> >> The Wellcome Trust Sanger Institute is operated by Genome Research
> >> Limited, a charity registered in England with number 1021457 and a
> >> company registered in England with number 2742969, whose registered
> >> office is 215 Euston Road, London, NW1 2BE.
> >>
> >
> > ------------------------------------------------------------------------------
> > Don't let slow site performance ruin your business. Deploy New Relic APM
> > Deploy New Relic app performance management and know exactly
> > what is happening inside your Ruby, Python, PHP, Java, and .NET app
> > Try New Relic at no cost today and get our sweet Data Nerd shirt too!
> > http://p.sf.net/sfu/newrelic-dev2dev
> > _______________________________________________
> > Vcftools-help mailing list
> > Vcf...@li...
> > https://lists.sourceforge.net/lists/listinfo/vcftools-help
> 
> 
> ------------------------------------------------------------------------------
> Don't let slow site performance ruin your business. Deploy New Relic APM
> Deploy New Relic app performance management and know exactly
> what is happening inside your Ruby, Python, PHP, Java, and .NET app
> Try New Relic at no cost today and get our sweet Data Nerd shirt too!
> http://p.sf.net/sfu/newrelic-dev2dev
> _______________________________________________
> Vcftools-help mailing list
> Vcf...@li...
> https://lists.sourceforge.net/lists/listinfo/vcftools-help




-- 
 The Wellcome Trust Sanger Institute is operated by Genome Research 
 Limited, a charity registered in England with number 1021457 and a 
 company registered in England with number 2742969, whose registered 
 office is 215 Euston Road, London, NW1 2BE.