Figure 4

Illustration of the formation of the initial partitioning by density-based mean clustering on the first two principal components of the Pima-Surui data with 100 randomly selected markers without missing genotypes. The numbers (1 and 2 here) over the triangles indicate the order of cluster starting points identified. Points with black plus signs indicate the points identified as the first cluster by Gap statistics. Points with red crosses indicate the points identified by Gap statistics as the second cluster. Blue circles represent the original points in the two-dimensional space after singular value decomposition. Downward triangles indicate the starting points for each cluster. Points with red crosses indicate the points identified by Gap statistics as the second cluster.