Population Stratification Analysis in Genome-Wide Association Studies - Mathematical Approaches to Polymer Sequence Analysis and Related Problems - page 187

Fig. 9.3 Summary plot of estimates of Q (estimated population ancestries) in UCI sample for
three clusters: ( a ) clustering obtained with a single ethnic group shows the homogeneity of each
group ( b ) clustering obtained with the entire sample shows the genetic divergence between groups
With STRUCTURE, we analysed each population sample separately (for instance,
in the UCI sample, Caucasians, Africans and Asians separately) as well as the en-
tire subset, using a subset of 400 selected markers. We observed that each single
population is homogeneous (Fig. 9.3 a), even though it differs compared to the other
populations (Fig. 9.3 b): if we consider a single ethnic group, we observe that the
individuals belong to a unique cluster (Fig. 9.3 a), while for the entire sample the
software is able to distinguish the differences between the different ethnic groups
(supplementary Fig. 9.3 b) as in the PCA plot (Fig. 9.1 ) with a limited number of
well-selected SNPs describing the genetic variance between different groups.