1. ¹G I Genética y Fisiología Forestal, Universidad Politécnica de Madrid, Madrid, Spain
2. ²Unidad Mixta de Genética y Ecofisiología Forestal INIA-UPM, Universidad Politécnica de Madrid, Madrid, Spain
3. ³Département de Biologie des populations, Unité Mixte de Recherche (UPR) 5175, Centre d'Ecologie Fonctionnelle et Evolutive, CNRS, Montpellier Cedex, France
4. ⁴INRA, UMR1202 BIOGECO, Cestas, France
5. ⁵Université de Bordeaux, UMR1202 BIOGECO, Cestas, France

Correspondence: Dr Á Soto, GI Genética y Fisiología Forestal, ETSI Montes, Universidad Politécnica de Madrid, Ciudad Universitaria s/n, Madrid 28040, Spain. E-mail: alvaro.soto.deviana@upm.es

⁶Current address: Departamento de Sistemas y Recursos Forestales, Instituto Nacional para la Investigación Agraria y Alimentaria (CIFOR-INIA), Madrid, Spain.

Received 15 July 2008; Revised 23 December 2008; Accepted 7 January 2009; Published online 25 February 2009.

Abstract

Powerful and accurate detection of first-generation (F1) hybrids and backcrosses in nature is needed to achieve a better understanding of the function and dynamics of introgression. To document the frequency of ongoing interspecific gene exchange between two Mediterranean evergreen oaks, the cork oak (Quercus suber) and the holm oak (Q. ilex), we analyzed 1487 individuals originating from across the range of the two species using eight microsatellite loci and two Bayesian clustering approaches (implemented in the programs STRUCTURE and NEWHYBRIDS). Simulated data were used to assess the differences between the two clustering methods and to back up the choice of the threshold value for the posterior probability to discriminate admixed from pure individuals. We found that the use of STRUCTURE resulted in the highest power to detect hybrids, whereas NEWHYBRIDS provided the highest accuracy. Irrespective of the approach, the two species were clearly distinguished as independent genetic entities without any prior information. In contrast with previous reports, we found no evidence for unidirectional introgression. The overall hybridization rate was very low (<2% of introgressed individuals). Only two individuals were identified as F1 hybrids and five as early backcrosses. This work shows that the combined application of the two complementary Bayesian approaches and their systematic validation with simulations, fit for the case at hand, helps gain resolution in the identification of admixed individuals.