phase_common sometimes creates Mendelian errors for missing genotypes when phasing with --pedigree

[kkellysci]

I am phasing a sample of parent/offspring duos/trios/quads using phase_common with the --pedigree flag.

When I ran checks on the output phased data, I found that it contained Mendelian errors that were not present in the original data.

I checked some of affected people/loci in the original data, and found these are all sites where at least one parent had missing genotype data. I think phase_common may be ignoring the pedigree data when it imputes the missing genotypes, and sometimes creating Mendelian errors as a result.