Abstract

PENA MALAVERA, A; GUTIERREZ, L  and  BALZARINI, M. Principal components in associative mapping. BAG, J. basic appl. genet. [online]. 2014, vol.25, n.2, pp.32-40. ISSN 1852-6233.

Association mapping (or linkage disequilibrium mapping) is used to find specific parts of the genome associated with phenotypic trait variation. It is a widely used in plant breeding because it allows the use of populations that do not come from specific experimental designs. If the population of individuals used in association mapping is genetically structured, the number of false positives, in the marker-trait association, increases. Several strategies can be used to model associations taken into account the underlying genetic structure. The principal components analysis can be used to identify the structure and express it in a reduced number of principal components (PCs). Then, these PCs can be incorporated as covariates in the association model. Different models strategies can be used to account for genetic structure in association mapping. The aim of this paper is to estimate expected false positive rates in association mapping performed by three different statistical models, under genetically structured populations. Compared models were M1: without correction for structure, M2: including PCs, as covariates of fixed effects, and M3: including PCs as random effects within a linear mixed model. Model comparison was performed using both, real and simulated data, for self-pollinated specie. The results suggested that the use of PCs as random covariates decreases the false positive rate in the inference of marker-trait associations.

Keywords : Mixed linear models; Principal components analysis; Genetic structure.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )