Figure 4.

The effect of incomplete data on error reporting. In this pedigree two female F1 offspring of female 175162 by male 175216 have been wrongly re-assigned to sire 175184 (in both A and B the two affected sisters and their relations have been highlighted in yellow.) Note that the dam ID 175162 has been duplicated in the dam row of the sandwich because it is now mother to two families, and is connected by an arc underneath the dam row. In (A) the genotype data for the F1 generation is complete, and these two offspring report both multiple failures to inherit from the (incorrect) father and also the occurrence of novel alleles inherited from neither asserted parent. In (B) 50% of F1 genotypes have been erased prior to analysis, and because the algorithm infers incorrect F1 genotypes from the (wrong) parent, errors are revealed between these inferred genotypes and the (correctly identified and genotyped) F2 progeny. Thus the true cause of the errors is obfuscated and the errors are reported in the progeny of the wrongly assigned sisters, rather than in the erroneous sisters themselves.

Paterson et al. BMC Bioinformatics 2012 13(Suppl 8):S5   doi:10.1186/1471-2105-13-S8-S5