False discovery rate control in association mapping under genetically structured populations

PEÑA MALAVERA, A.; BRUNO, C.; BALZARINI, M.

Marrakech

Congreso; 61st World Statistics Congress; 2017

The association tests between molecular markers and phenotypic traits are crucial for the Quantitative Trait Loci (QTL) identification. Biotechnological advances increased the molecular marker information; consequently, the number of genotype-phenotype association tests required incremented too. The multiple statistical inferences (multiplicity) demand corrections of the p-values obtained for each comparison in order to keep limited the error rates for the family of association tests. However, classic statistical correction methods such as Bonferroni, False Discovery Rate (FDR) and the Effective Number of Independent Test (Meff) were developed in the context of independent data. Wherever, when there are population genetic structure the data are not longer independent. In this paper, we propose a method of correction for multiplicity based on estimation of the effective number of tests from a models that adjust for the underlying correlation structure. We evaluate the performance of the proposed procedure in the estimation of p-values for a set of simulated QTL. The results suggest that the proposed method provides control of FDR and has more power than other methods for multiplicity correction commonly used in association mapping.