A New Nonparametric Approach to Delineating Spatial Population Genomic Variation

Abstract Text: The plot of the first two principal coordinates as often derived from the analysis of large-scale SNP data is essential for identifying the population stratification that may lead to high false positive rate in genome-wide association studies. The current use of bivariate plot is for visualization without confidence regions (CRs) needed to justify the observed population structures. Parametric approaches are often used but sampling distributions of parameters estimated from SNPs are unknown. The objective of this paper is to describe a new nonparametric method for constructing CRs without knowing the sampling distributions. The statistical properties of the new method are evaluated using simulation data and its application is illustrated with the analysis of a human SNP data. Several parametric and nonparametric methods of constructing CRs are also compared with the new method in both simulation and real data analysis. 

Keywords: Confidence region; Nonparametric statistics; Population genomics