genome. This low proportion of families with elevated homozygosity is consistent with low reported rates of consanguinity in the AGRE collection. Nonetheless, in the few outlier families, rates of homozygosity are far higher than generally observed in individuals whose parents have no common ancestry (≤1.6%), and overlap or exceed in some cases the predicted range of homozygosity expected in offspring of first cousin parents (6.25%) [22] (Figure 1A). The sizes of homozygous blocks in probands from these outlier families ranged from ∼5–19 cM on average (Figure 1B), suggesting ancient shared ancestry in these families compared to larger blocks of homozygosity seen in consanguineous families (≥20 cM) [22]. Since the AGRE dataset provides no specific information about shared ancestry or consanguinity between parents, we explored the level of shared ancestry between parents, by performing tests to estimate relatedness between individuals based on identical-by-state (IBS) and identical-by-descent (IBD) genotype information [23], [24]. We find that for 16 families where probands had the largest amount of homozygosity in their genomes, some of the parental pairs were more closely related than average (Figure S1), but that parental relatedness by itself, as analyzed by these methods, did not always predict the degree of homozygosity in
