Identity by Descent

We calculated two types of IBD summary metrics. First, we calculated the average amount of total IBD shared between an individual of one cluster and an individual in a second cluster. We refer to this as the "mean IBD of iLASH reported," as iLASH only reports IBD that is present and that is greater than 3cM. This metric is the same as many other group IBD averages reported in the literature.

We defined a second metric, which we call the "mean IBD of all possible pairs." For every cluster, there might be many pairs of people who do not share any IBD, and thus, would not be reported by iLASH. This observation bias means that the first metric could be biased upwards and not be an accurate representation of the actual amount of relatedness between two clusters. Therefore, we find all possible pairs of individuals between two clusters. We set their IBD sharing to be the total IBD estimated by iLASH. If the pair had no estimated sharing, we assign that pair a 0. We then take the average of this vector for our second metric.

The "proportion of pairs sharing IBD" refers to the proportion of all possible pairs of individuals between two clusters that share any IBD detected by iLASH (greater than 3cM).