Abstract Divergence processes in crop-wild fruit tree complexes in pivotal regions for plant domestication such as the Caucasus and Iran remain little studied. We investigated anthropogenic and natural divergence processes in apples in these regionsusing 26 microsatellite markers amplified in 550 wild and cultivated samples. We found two genetically distinct cultivated populations in Iran that are differentiated from Malus domestica, the standard cultivated apple worldwide. Coalescent-based inferences showed that these two cultivated populations originated from specific domestication events of Malus orientalis in Iran. We found evidence of substantial wild-crop and crop-crop gene flow in the Caucasus and Iran, as has been described in apple in Europe. In addition, we identified seven genetically differentiated populations of wild apple (M. orientalis), not introgressed by the cultivated apple. Niche modeling combined with genetic diversity estimates indicated that these wild populations likely resulted from range changes during past glaciations. This study identifies Iran as a key region in the domestication of apple and M. orientalis as an additional contributor to the cultivated apple gene pool. Domestication of the apple tree therefore involved multiple origins of domestication in different geographic locations and substantial crop-wild hybridization, as found in other fruit trees. This study also highlights the impact of climate change on the natural divergence of a wild fruit tree and provides a starting point for apple conservation and breeding programs in the Caucasus and Iran. Methods Microsatellite genotyping data for M. orientalis, M. sieversii from Kazakhstan, M. domestica and M. baccata were from previously published studies, including: 207 M. orientalis individuals from Turkey, Armenia and Russia (23 sites, Tables S1 and S2, (Cornille et al., 2013; Cornille et al., 2012)), four apple cultivars from Armenia, 40 “pure” European cultivated M. domestica individuals (i.e., not introgressed by M. sylvestris) (Tables S1 and S2, (Cornille et al., 2013, 2012)), 20 M. sieversii individuals from Kazakhstan (Cornille et al., 2013) and 22 M. baccata individuals from Russia (Cornille et al., 2012). We collected 257 new samples in 2017 in Iran and in 2018 in Kyrgyzstan (M. sieversii) for this study: 167 M. orientalis individuals from the Hyrcanian Forests and the Zagros region in Iran (Table S1), 48 local Iranian apple cultivars from the Seed and Plant Improvement Institute (Karaj, Iran) (Table S1) and 42 M. sieversii individuals from Kyrgyzstan. Note that for 18 of the 48 local Iranian samples, fruit size was measured (Table S1). Collections meet the requirements of the recently enacted Nagoya protocol on access to genetic resources and the fair and equitable sharing of benefits. Thus, a total of 550 individuals were analyzed, comprising 374 M. orientalis, 48 Iranian and four Armenian apple cultivars, 40 European apple cultivars belonging to M. domestica, 62 M. sieversii (from Kyrgyzstan and Kazakhstan) and 22 M. baccata individuals (details are provided in Table S1). DNA from the new samples (N = 257) was extracted from dried leaves with the NucleoSpin plant II DNA extraction kit (Macherey & Nagel, Düren, Germanyâ) following the manufacturer’s instructions. Multiplex microsatellite PCR amplifications were performed with a multiplex PCR kit (Qiagen Inc.â) for 26 microsatellite markers as previously described (Cornille et al., 2012; Patocchi, Frei, Frey, & Kellerhals, 2009). Note that on each DNA plate, we included three controls, i.e., one sample of M. orientalis, one of M. sieversii and one of M. domestica for which data were already available (Cornille et al., 2013). Genotypes of the controls were compared with the 2013 dataset. We retained only multilocus genotypes for which < 20 % of the data was missing. The suitability of these markers for population genetics analyses has been demonstrated in previous studies (Cornille et al., 2013; Cornille, Gladieux, & Giraud, 2013; Cornille et al., 2012). Software used -STRUCTURE https://web.stanford.edu/group/pritchardlab/structure.html -ABCtoolbox: http://cmpg.unibe.ch/software/ABCtoolbox/ -fastsimcoal2 http://cmpg.unibe.ch/software/fastsimcoal27/ -Genepop: https://genepop.curtin.edu.au -R software

We thank the Franco-Iranian Campus France program « Gundhishapur » 2016-2018, the Institut Diversité Écologie et Évolution du Vivant (IDEEV), ATIP-Avenir for funding, and the ANR JCJC "PLEASURE". We thank Bolotbek Tagaev (Sustainable Livelihoods Coordinator of FFI-Kyrgyzstan) for sampling and prospection, Fauna & Flora International and more specifically the Global Trees Campaign (GTC) Program. We also thank Adrien Falce, Olivier Langella and Benoit Johannet for help and support on the INRAE-Génétique Quantitative et Evolution- Le Moulon lab cluster and the genotyping platform GENTYANE INRA UMR 1095. We thank the INRAE MIGALE bioinformatics platform (http://migale.jouy.inra.fr) for providing help and support, in particular Véronique Martin, Eric Montaubon and Valentin Loux. We also thanks Céline Bellard for her advices for ecological niche modeling analyses.