Bucci G*, Vendramin GG
IMGPF-CNR, v. Atto Vannucci 13 – 50134 Firenze (FI)
Collocazione: c2.2.1 – Tipo Comunicazione: Presentazione orale
2° Congresso SISEF
Sessione 2: “Biodiversità nei sistemi forestali: dalle indagini di base alle prospettive applicative”
Abstract: We applied geostatistics to previously reported cpSSR haplotype-frequency data from 95 European Norway spruce populations to provide preliminary evidence about the following issues: 1) delineation of genetically homogeneous regions (’breeding zones’); 2) prediction of their haplotype frequencies and definition of related criteria to be applied for provenance identification and certification of seedlots; 3) identification of putative allochthonous stands within the natural range; 4) construction of a continental-scale ’availability map’ of the intraspecific biodiversity for Norway spruce. A direct evidence of large-scale geographic structure of cpSSR haplotypes over the European natural range was obtained. Experimental variograms showed the existence of both geographic clines, expected to occur when a gradient of haplotype frequency is present in the dataset, and stationary patterns, detected when haplotype-frequency surface has marked peak andor troughs within the interpolated area. The increase of the mean genetic divergence by geographic distance (up to about 1,800 km apart) provided a strong hint that geographic distance is a major factor of population differentiation in Norway spruce. Haplotype-frequency surfaces were obtained by applying ordinary kriging to haplotype-frequency sampling data. Cluster analysis carried out on haplotype-frequency surfaces revealed a fair discrimination among 16 genetically homogeneous regions called ’genetic zones’, with an accuracy of 0.916. Dendrogram analysis (using a restricted maximum likelihood method) carried out on the predicted mean haplotype frequency confirmed a fairly good separability of the genetic zones detected. Discriminant analysis carried out on the observed haplotype frequencies assigning sampling stands to genetic zones based on their geographic location revealed a fairly high proportion of correctly reassigned stands (0.909), confirming the robustness of the genetic subdivision obtained. Combining the information from discriminat analysis and cross-validation of sampling points based on kriging predictions, a subset of 3 populations that are likely to be of non-local origin was identified. Application of geostatistical analysis to genetic marker data, combined with the use of GIS applications, is discussed in relation to breeding activities (certification of seedlots and provenance identification) and to formulation of appropriate conservation strategies for this species (identification of suitable seed sources andor gene reserves).
Citazione: Bucci G, Vendramin GG (1999). Statistiche spaziali applicate allo studio della biodiversità: Identificazione di ’breeding zones’ in specie forestali . 2° Congresso Nazionale SISEF, Bologna, 20 – 22 Ott 1999, Contributo no. #c2.2.1