Data from: Effect of microsatellite selection on individual and population genetic inferences: an empirical study using cross-specific and species-specific amplifications

Although whole-genome sequencing is becoming more accessible and feasible
for nonmodel organisms, microsatellites have remained the markers of
choice for various population and conservation genetic studies. However,
the criteria for choosing microsatellites are still controversial due to
ascertainment bias that may be introduced into the genetic inference. An
empirical study of red deer (Cervus elaphus) populations, in which
cross-specific and species-specific microsatellites developed through
pyrosequencing of enriched libraries, was performed for this study. Two
different strategies were used to select the species-specific panels:
randomly vs. highly polymorphic markers. The results suggest that reliable
and accurate estimations of genetic diversity can be obtained using random
microsatellites distributed throughout the genome. In addition, the
results reinforce previous evidence that selecting the most polymorphic
markers leads to an ascertainment bias in estimates of genetic diversity,
when compared with randomly selected microsatellites. Analyses of
population differentiation and clustering seem less influenced by the
approach of microsatellite selection, whereas assigning individuals to
populations might be affected by a random selection of a small number of
microsatellites. Individual multilocus heterozygosity measures produced
various discordant results, which in turn had impacts on the
heterozygosity-fitness correlation test. Finally, we argue that picking
the appropriate microsatellite set should primarily take into account the
ecological and evolutionary questions studied. Selecting the most
polymorphic markers will generally overestimate genetic diversity
parameters, leading to misinterpretations of the real genetic diversity,
which is particularly important in managed and threatened populations.