Number entry in VCF header

[GuilhemSempere]

Hi

Version 0.8.1 of Platypus generates the following header lines:

##FORMAT=<ID=GT,Number=1,Type=String,Description="Unphased genotypes">
##FORMAT=<ID=GQ,Number=.,Type=Integer,Description="Genotype quality as phred score">
##FORMAT=<ID=GOF,Number=.,Type=Float,Description="Goodness of fit value">
##FORMAT=<ID=NR,Number=.,Type=Integer,Description="Number of reads covering variant location in this sample">
##FORMAT=<ID=GL,Number=.,Type=Float,Description="Genotype log10-likelihoods for AA,AB and BB genotypes, where A = ref and B = variant. Only applicable for bi-allelic sites">
##FORMAT=<ID=NV,Number=.,Type=Integer,Description="Number of reads containing variant in this sample">

I agree that according to VCF specs, the number entry has to be "." for GL since it is a list of values. However wouldn't it be more accurate to set it to "1" for GQ, GOF, NR and NV, just as for GT?

G.

[GuilhemSempere]

Sorry my question only makes sense for GOF and GQ since NR and NV contain one value per allele

[tdelhomme]

I think that following the VCF specifications, this should be an "A" (one value per alternate allele).
At the moment this creates a bug if using vt for example, as it uses this number/value to know into how many fields it should divide the site (decompose mode).