Clearly there is significant noise with regard to allele assignment For example the FF treated as FH should have had no paternal (H) alleles, yet it has several, and they appear to be consistent with the patterns we see in the F1s. We could either mask these problem loci or, better yet, adjust the counts to compensate for the errant ratios.

There is some problem with the "FF_6136(F)". It looks more like an F1 than an inbred.