Consider option to calculate log-likelihoods ignoring dosage for VCF output

[timothymillar]

This can be achieved by treating an allelic phenotype as a genotype (i.e. ploidy = number of unique alleles).
Calculating this over all possible genotypes would give genotype likelihoods that are invariant to dosage bias.
This would not be used in the haplotype assembly, only as a VCF output for downstream analysis and potential dosage correction.