Empirical assessment of allozyme and RAPD variation in Pinus sylvestris (L.) using haploid tissue analysis

We analysed 20 allozyme and 22 putative random amplified polymorphic DNA (RAPD) loci in two populations of Pinus sylvestris (L.) from northern Sweden. Genotypes for individual allozyme and RAPD loci were inferred from segregation patterns in haploid macrogametophytes. Therefore, it was possible to distinguish between homo- and heterozygotes carrying a RAPD fragment and to estimate directly the frequencies of RAPD fragments. The percentage of polymorphic loci and the expected and observed heterozygosity were lower for allozymes than for RAPDs. Average fixation indices for both types of markers were negative indicating a heterozygote excess over panmictic expectations. The apportionment of genetic variation within and among the investigated populations was similar for allozymes and RAPDs and showed that most of the variation resided within populations. RAPD genotypes inferred from haploid material were subsequently converted to diploid phenotypes and used to estimate indirectly the frequencies of RAPD fragments. Gene diversity measurements derived from indirectly estimated RAPD frequencies were distinctly lower than those based on directly estimated RAPD frequencies. This result was caused by the absence of the null homozygote at many loci which appeared as monomorphic in the diploid data set. Population differentiation coefficients based on the indirectly estimated RAPD frequencies were not concordant with those based on directly estimated RAPD frequencies. Our present results indicate that when complete genotype information can be obtained, RAPD analysis provides genetic information similar to that revealed by analysis of allozyme variation. On the other hand, our results are concordant with theoretical results suggesting that analysis of RAPD variation in diploid material can produce unreliable estimates of population-genetic parameters.