Effects of life history traits and sampling strategies on genetic diversity estimates obtained with RAPD markers in plants

A compilation of studies using RAPD markers for evaluating population differentiation resulted in 78 estimates of AMOVA-derived Φ_ST and 31 estimates of Nei's G_ST, as well as in 41 estimates of Nei's within-population diversity. In outcrossing taxa, estimates of between-population diversity were closely correlated with maximum geographic distance between sampled populations. A corresponding association was not found in selfing taxa. These results suggest that RAPD can be a sensitive method for detection of genetic structuring according to the isolation-by-distance model. However, it also means that sampling strategies, as applied in individual studies, can seriously influence the resulting estimates of between-population diversity. Other sampling strategies, like number of plants per population and number of scored polymorphic markers, do not seem to impart any serious artefacts. As previously verified with allozyme data, RAPD markers showed that long-lived, outcrossing, late successional taxa retain most of their genetic variability within populations. By contrast, annual, selfing and/or early successional taxa allocate most of the genetic variability among populations. Estimates for between- and within-population diversity, respectively, proved to be negatively correlated, as previously reported for allozyme data. The only major discrepancy between allozymes and RAPD markers concerns geographic range; within-population diversity was strongly affected by distributional range of the investigated species in the allozyme data but not in the RAPD data. Moreover, RAPD-based values for between-population diversity increased with increasing distributional range whereas the opposite has been reported in a large allozyme data compilation. Contrary to allozymes, RAPD marker-derived within-population diversity is probably therefore not a very good predictor of total species genetic diversity.