Exploring population genetic structure from microsatellites via spatial statistical modeling

BRUNO, C.; MACCHIAVELLI, R.; BALZARINI, M.

Sanya, Hainan.

Conferencia; 2nd International Conference on Plant Molecular Breeding (ICPMB); 2007

Generation Challenge Program (GCP) y China Association of Agricultural Science Societies

Polymerase chain reaction amplification of microsatellites in many species showed that they are highly polymorphic, somatically stable and inherited in a co-dominant Mendelian manner. They increased significantly the number of alleles and loci for genetic studies. The abundance and amount of information derived from these multiallele and multilocus co-dominant markers, make them ideal tools for their use in plant breeding, especially for varietal identification and population genetic studies. Studies of the spatial population genetic structure in plants, mainly open-pollinated, permit to understand the patterns of local breeding. Plant populations will exhibit local genetic structure when gene flow is restricted (Wright, 1943). Therefore, the genetic distances among samples in a meta-population may correlate with their geographic distances. The spatial autocorrelation of genetic data is expected to decline as field distance increases. In a microspatial scale, the spatial structure could be weak, and rarely consistent across loci (Heywood,1991). Efficient statistical approaches are needed to detect spatial autocorrelation among observation in microgeographic scales and fully understand contemporary dispersion patterns. Unfortunately, most population genetic analyses from microsatellite data are limited to a single-statistic based on single locus allele frequencies, providing only a relative ´snapshot´ of the spatial autocorrelation. Other approaches proposed to account for the multivariate nature of microsatellite data analyze multiallele-multilocus genetic distances but in a discrete set of between sample distances. Our objective is to assess spatial autocorrelation among multiallele-multilocus genetic distances of samples taken in a microgeographic scale by the use of geostatistical models which yield direct fit of genetic distances over a continuum of geographic distances. Modeling of multivariate genetic distances as function of spatial distances to analyze spatial structure in this scale should be more powerful than separate analysis based on single-locus allele frequencies or multivariate analysis for discrete between sample distances.