Fig. 3. Genetic relatedness and geographic distances in the UK Biobank dataset.

For each of the 432,968 UK Biobank samples with available geographic data, we detected the individual sharing the largest total amount (in cM) of genome IBD within the past 10 generations (referred to as closest individual). a For each value x of total shared genome (in cM) on the X-axis, we report the percentage of UK Biobank samples (Y-axis) that share x or more with their closest individual. b For each value x of total shared genome (in cM) on the X-axis, we report the median distance (km, computed every 10 cM) for all pairs of (sample, closest individual) who shared at least x. Vertical dashed lines indicate the expected value of the total IBD sharing for kth degree cousins, computed using 2G(1/2)^2(k+1), where G = 7247.14 is the total diploid genome size (in cM) and k represents the degree of cousin relationship (e.g. k = 2 for second degree cousins, separated by 2(k + 1) generations)¹¹. The value of 45 observed when no sharing cutoff is considered (i.e. when the x value approaches 0) reflects the median prediction error for a random individual, regardless of how much IBD they share with the closest individual.