Data from: Are molecular markers useful predictors of adaptive potential?

Estimates of molecular genetic variation are often used as a cheap and
simple surrogate for a population's adaptive potential, yet empirical
evidence suggests they are unlikely to be a valid proxy. However, this
evidence is based on molecular genetic variation poorly predicting
estimates of adaptive potential rather than how well it predicts true
values. As a consequence, the relationship has been systematically
underestimated and the precision with which it could be measured severely
overstated. By collating a large database, and using suitable statistical
methods, we obtain a 95% upper bound of 0.26 for the proportion of
variance in quantitative genetic variation explained by molecular
diversity. The relationship is probably too weak to be useful, but this
conclusion must be taken as provisional: less noisy estimates of
quantitative genetic variation are required. In contrast, and perhaps
surprisingly, current sampling strategies appear sufficient for
characterising a population's molecular genetic variation at
comparable markers.