Morphological, molecular and statistical tools to identify Castanea species and their hybrids

The aim of this study was to investigate, in relation to reference samples, the genetic structure of several natural Castanea populations of unknown origin and of hybrids from the clone collection of the Lourizan Forest Research Center. A total of 115 individuals sampled from four Castanea sativa stands located in the northwest of Spain, 61 Castanea crenata individuals and 27 Castanea mollissima individuals were classified on the basis of four morphological traits, and genotyped with 11 microsatellite loci to define a set of reference samples. The data analyzed with the program STRUCTURE detected four clusters: the two Asiatic species and two clusters in C. sativa. From these reference samples, pure individuals and hybrids of known genealogy were simulated to determine the efficiency with which STRUCTURE and NEWHYBRIDS assigned them to pure or hybrid groups and to a specific genealogical class, respectively. As expected, the discrimination and assignment of the simulated individuals improved with increasing F-st value. The two clusters identified within C. sativa may correspond to gene pools with different adaptive characteristics previously identified in provenance tests; pure and admixed populations of both C. sativa gene pools were identified.