Genetic differentiation of wild and managed Coffea arabica populations in Afromontane rainforest fragments in SW Ethiopia

Abstract

*1.* Coffea arabica L. has its origin and centre of diversity in the Afromontane rainforests of southwestern Ethiopia. Deforestation, habitat fragmentation and forest degradation threaten this wild coffee gene pool, which is vital to global coffee cultivation.*2.* We investigated genetic variation and diversity within and between eleven coffee stands across a gradient of increasing forest management intensity (‘forest coffee’, unmanaged large continuous forest vs. ‘semi-forest coffee’, managed, small fragments) using two multiplex panels of five microsatellite markers each. We used analysis of molecular variance (AMOVA) to partition total genetic diversity and applied a Bayesian model-based cluster method to infer population structure.*3.* Genetic variability was larger within (72%) than between (21%) stands. There was a significant differentiation between three wild coffee genotypes in the least disturbed forest and a single genotype prevailing in the semi-forest coffee system. Managed populations were closely allied to local coffee cultivars indicating an effect of local selection, but measures of neutral genetic diversity (Ne, I, He, MV) did not differ significantly between management systems.*4.* Despite severe forest fragmentation, we found no evidence for genetic erosion in C. arabica, possibly because traditional Ethiopian coffee cultivation involves collecting, exchanging and replanting coffee. Nevertheless, this poses a threat to wild coffee populations because they are replaced by (or exposed to genetic exchange with) genotypes imported from other areas and that are preferred by local farmers due to specific characteristics. In-situ conservation efforts must therefore focus on avoiding intensification of coffee production in forest coffee systems.