Spatial autocorrelation analysis and ecological niche modelling allows inference of range dynamics driving the population genetic structure of a Neotropical savanna tree

Aim Spatial autocorrelation analysis of genetic diversity was combined with ecological niche modelling (ENM) to better infer how ecological and evolutionary processes underlie population structure in Eugenia dysenterica, a widely distributed tree in the ,'Cerrado' region of Central Brazil. Location 'Cerrado' region, Central Brazil. Methods Data were derived from 11 microsatellite loci in 23 populations of E. dysenterica, totalling 249 allele frequencies. The expected heterozygosity (He) within populations and the first principal coordinates extracted from pairwise F-ST and from the difference between R-ST and F-ST among populations were correlated with shifts in suitability from ENM. Frequencies were then analysed using a spatial autocorrelation analysis based on Moran's I and Mantel tests to contrast population differentiation for mean allele frequencies, allele size and shifts in suitability since the Last Glacial Maximum inferred from ENM. Results Spatial correlograms based on Moran's I and Mantel tests showed a linear decrease in autocorrelation with distance, which revealed north-westsouth-east gradients in allele frequencies, genetic diversity and differences between RST and FST. These spatial patterns varied among loci and alleles, and the strongest spatial patterns were found for more common alleles with higher levels of differentiation among populations and for those correlated with shifts in ENM suitability. Main conclusions Current genetic diversity and population structure in E. dysenterica can be explained by geographical range shifts associated with Quaternary climate dynamics, thus demonstrating the value of applying spatial analyses to study the ecological and evolutionary processes underlying differentiation even within populations possessing a continuous distribution.