Inferring and outlining past population declines with linked microsatellites: a case study in two spruce species

Abstract

Estimating demographic parameters is fundamental for conservation programs. Inferences are generally derived from the variation of unlinked nuclear genomic regions, which are often unavailable for non-model species. These limitations can be circumvented using universal polymorphic markers that can be easily transferred across taxa, such as cytoplasmic single sequence repeats (SSRs). These markers are sensitive to population expansions, but no formal test has been conducted to explore if they can be used to infer and distinguish between competing bottleneck scenarios. Herein, we simulated the evolution of ten linked haploid SSRs in populations submitted to different bottleneck regimes (θ 1 = 1, 10, 50 and 90 % of θ 0) at different times (τ = 0.5, 1 and 10). The variation of these markers, as compiled with six summary statistics, allowed to detect severe population collapses independently of τ, and as long as populations kept roughly constant effective sizes after the size reduction. Mild declines became difficult to infer as τ increased, and small bottlenecks were virtually undetectable with these markers. More complex frameworks, such as bottlenecks followed by expansions, were also difficult to infer. Comparisons with chloroplast SSR variation in the Mexican relict Picea mexicana and the eastern North American Picea rubens suggested that these species went through bottlenecks of different intensities. While P. mexicana suffered a severe population decline that could be dated back to the last interglacial, P. rubens went through a more recent (i.e. late Pleistocene) and milder bottleneck. These results indicate that linked SSRs can be used as proxies to infer basic parameters related to strong population declines in species that lack adequate genomic resources.