AN EVALUATION OF THE AFLP FINGERPRINTING TECHNIQUE FOR THE ANALYSIS OF PATERNITY IN NATURAL-POPULATIONS OF PERSOONIA-MOLLIS (PROTEACEAE)

The accurate assignment of paternity in natural plant populations is r equired to address important issues in evolutionary biology, such as t he factors that affect reproductive success. Newly developed molecular fingerprinting techniques offer the potential to address these aims. Here, we evaluate the utility of a new PCR-based multi-locus fingerpri nting technique called Amplified Fragment Length Polymorphism (AFLP) f or paternity studies in Persoonia mollis (Proteaceae). AFLPs were init ially scored for five individuals from three taxonomic levels for 64 p rimer pairs: between species (P. mollis and P, levis), between subspec ies (P. mollis subsp. nectens and subsp. livens), between individuals within a single population of P. mollis, as well as for a naturally po llinated seed from a single P. mollis subsp. nectens plant. Overall, 1 164 fragments (24.6% of all fragments) were polymorphic between specie s, 743 (16.5%) between subspecies, 371 (8.6%) between individuals with in a single population, and 265 (6.2%) between a plant and its seed. W ithin a single P. mollis population of 14 plants, 42 polymorphic fragm ents were scored from profiles generated by a single AFLP primer pair. The mean frequency of the recessive allele (q) over these 42 loci was 0.773. Based on these observations, it will be feasible to generate w ell over 100 polymorphic AFLP loci with as few as three AFLP primer pa irs. This level of polymorphism is sufficient to assign paternity unam biguously to more than 99% of all seed in experiments involving Small, known paternity pools. More generally, the AFLP procedure is well sui ted to molecular ecological studies, because it produces more polymorp hism than allozymes or RAPDs but, unlike conventionally developed micr osatellite loci, it requires no prior sequence knowledge and minimal d evelopment time.