At a given locus in the reference genome, VCF records phased and non-phased alleles by separating numeric references to the alleles with either | (phased) or / (unphased). For example, "0/1" might be found in an unphased diploid sample.
My question has to do with the case where you have polyploidy with 3 or more alleles. With VCF's notation, it is possible to represent a combination of phased and unphased alleles at a single locus, e.g., "0|1/1". But is this actually possible in practice? From what I understand of software that computes phasing, this shouldn't happen, so "0|1|1" or "0/1/1" might be reported at a given locus, but not "0|1/1" or "0/1|1".
Have I got this right?
My question has to do with the case where you have polyploidy with 3 or more alleles. With VCF's notation, it is possible to represent a combination of phased and unphased alleles at a single locus, e.g., "0|1/1". But is this actually possible in practice? From what I understand of software that computes phasing, this shouldn't happen, so "0|1|1" or "0/1/1" might be reported at a given locus, but not "0|1/1" or "0/1|1".
Have I got this right?
