Skip to main content
. 2015 Nov 14;16:940. doi: 10.1186/s12864-015-2181-1

Fig. 4.

Fig. 4

Proportion of assigned individuals and misassignment rate in assignment tests based on supervised clustering. The 2250 individuals from 45 public populations of the world reference dataset were assigned against the world reference dataset, using 32,966 (33K) SNP, 10K SNP and 1K SNP, with different values for the EM algorithm’s ε stopping criterion (0.01, 0.1 and 1). The proportion of assigned individuals a and the misassignment rate b were plotted against ancestry thresholds (0–1)