On the use of large marker panels to estimate inbreeding and relatedness: empirical and simulation studies of a pedigreed zebra finch population typed at 771 SNPs

In recent years there has been a dramatic increase in the availability of high density genetic marker data for both model and non-model organisms. A potential application of these data is to infer relatedness in the absence of a complete pedigree. Using a marker panel of 771 SNPs genotyped in three generations of an extensive zebra finch pedigree, correlations between pedigree relatedness and seven marker-based estimates of relatedness were examined, as was the relationship between heterozygosity and inbreeding. Although marker-based and pedigree relatedness were highly correlated, the variance in estimated relatedness was high. Further, the correlation between heterozygosity and inbreeding was weak, even though mean inbreeding coefficient is typical of that seen in wild vertebrate pedigrees; the weak relationship was in part due to the small variance in inbreeding in the pedigree. Our data suggest that using marker information to reconstruct the pedigree, and then calculating relatedness from the pedigree, is likely to give more accurate relatedness estimates than using marker-based estimators directly.