Re: [Vcftools-help] filter/erase samples based on sample attributes

[Sébastien L. <seb...@in...>]

Great, that fit my needs.

I saw that 'FORMAT' is not recognized, only 'FMT'.
I just mention it, because the doc page says that we could use both.

I could make my complex filter.

Thanks,
Sébastien L.


Le 20-05-2014 13:15, Petr Danecek a écrit :
> Ah, I am sorry, I pointed you in a wrong direction. It is the "filter"
> command which can set the genotypes to missing
> 
> bcftools filter -S. -e'EXPRESSION'
> 
> Petr
> 
> 
> On Tue, 2014-05-20 at 11:27 +0200, Sébastien Letort wrote:
>> Thanks Petr for your information.
>> My version of bcftools was outdated.
>> 
>> Nevertheless, I didn't succeed.
>> 
>> `bcftools annotate -x FORMAT/DP` seems to be designed to remove all DP
>> format values.
>> 
>> I tried 'FORMAT/DP<100' to remove DP when DP < 100 and that's not the
>> way do do this.
>> 
>> I need something clause to :
>> for s in samples
>> do
>>   if( s.DP<100 || s.SRR/s.SRF <0.8 )
>>   then
>>    s = '.'
>>   fi
>> done
>> 
>> Maybe what I want is to complex, should I try to write a plugin ?
>> I did not find documentation on it.
>> 
>> Best,
>> Sébastien L.
>> 
>> Le 19-05-2014 09:31, Petr Danecek a écrit :
>> > Hi Sebastien,
>> >
>> > `bcftools annotate` can do what you want via the -x option
>> >
>> > Best,
>> > Petr
>> >
>> >
>> > On Fri, 2014-05-16 at 10:47 +0200, Sébastien Letort wrote:
>> >> Hi,
>> >>
>> >> I need to remove some sample genotypes of my VCF file regarded their
>> >> attributes.
>> >>
>> >> Here is a row starting with FORMAT field.
>> >>
>> >> GT:FAO:FSAR:SAR:AO:SRF:FSRF:SAF:DP:GQ:FRO:SRR:RO:FDP:FSAF:FSRR  .
>> >> 0/1:7:6:4:5:20:21:1:32:12:27:5:25:34:1:6    .   .
>> >>
>> >> I need to change sample value with '.' when one attribute doesn't
>> >> reach
>> >> a threshold for example "AO > 10"
>> >> or when a more complicated test fail like : "0.9 < SRR/SRF && SRR/SRF
>> >> <
>> >> 1.1"
>> >>
>> >> I check vcf-annotate, but it doesn't seem to manage the complexe test.
>> >>
>> >> Can you confirm me that nothing exist to do this task ?
>> >> I can use Vcf.pm if nothing else already exist.
>> >>
>> >>
>> >> regards,

-- 
Sébastien Letort - bio-informaticien
Inserm U1078, Génétique, Génomique fonctionnelle et Biotechnologies
46 rue Félix Le Dantec
CS 51819
29218 BREST CEDEX 2
Tel. 02.98.22.34.08 (EFS)
Tel. 02.98.01.82.37 (UBO)