Estimation of genetically homogenous clusters within Europe

Abstract

The most powerful designs and methods in genetics for the analysis of complex diseases are performed at large scale, involving many countries. These studies face the challenge of including genotypes collected in different latitudes what can potentially produce false positives if genetic stratification is present. Here, genome-wide patterns of SNP among 4416 individuals clustered in 2121 families and inhabitants of 15 different cities across Europe are studied. Principal Component Analysis and the cluster method implemented by FRAPPE are used to estimate the presence of a hidden structure in the European population. Correlated measures are challenging for both statistical techniques and different approaches are then proposed to cope with the problem. From the analysis, three clusters were identified, the axes of variation in the sample data show two clear patterns: northsouth and east-west. Tampere, Finland was unambiguously identified as a homogeneous population by all analysis, wh