Spatial genetic structure of Persian walnut (Juglans regia L.) populations across its native range

Persian walnut is highly economically important species cultivated throughout the temperate regions of the world for its high quality wood and edible nuts. Anthropized species often suffered genetic erosion. Nevertheless, an extensive and accurate evaluation of worldwide genetic resources of walnut has not been carried out. Establishing the center(s) of diversity for the species, understanding its post glacial expansion events and shedding light to its history of domestication and dispersal by humans, which began about 2000 yr. ago in Eurasia, can predict how walnut species will respond to possible future climate change scenarios. In the framework of WALNET agreement, we will provide valuable information on how landscape features influenced gene flow and promoted walnut local adaptation. The general objective is to apply landscape genetic analysis with support of GIS technologies and spatial statistic tools to identify and quantify the effects of refugia during Last Glacial Maximum, post-glacial expansion, physical and climatic barriers, human domestication events, anthropological factors and connectivity between rural human communities on genetic differentiation of walnut populations in Eurasia. In this study, we determined the spatial genetic structure of 29 J. regia populations (number of adult trees = 703) collected in the native range of the species (China, Uzbekistan, Kyrgyzstan, Tajikistan, Pakistan, Iran and Georgia) using 14 neutral nuclear SSR markers. The level of allelic richness was relatively high within each population and the AMOVA revealed that 84.67% of molecular variance was partitioned within individuals, while the 15.33% was distributed among populations. Evidence for human-mediated domestication bottleneck events was detected in only 10.3% of populations. A Bayesian approach, divided 703 walnut samples into three clusters: (1) Kyrgyzstan genotypes, (2) Chinese genotypes and (3) trees from South-Western Asia. The combination of geostatistical analysis (ArcGIS 9.3 software) such as Kriging interpolation of Q-membership (clustering surface maps) and Monmonier's maximum difference algorithm analysis allowed to detect spatial coincidence of landscape features with genetic discontinuities (overlay approach). Our data confirmed the presence of multiple glacial refugia hypothesis in the native range of J. regia and the presence of secondary contact zone (hybridization) between populations originating from two glacial refugia in Uzbekistan.