Allele frequencies have long been studied by biologists interested in evolution and speciation. More recently, with the application of molecular markers in human DNA profiling we have also seen the need for reliable population allele frequency estimates for making probabilistic inferences. There is now interest in applying the same DNA profiling technology to identification of plant varieties. HortResearch maintains a large germplasm of horticultural plant species. It is becoming evident that accurate identification of these accessions through DNA fingerprinting is essential for effective utilisation and maintenance of this germplasm. Microsatellites are the markers of choice for this fingerprinting. However, such markers do not reveal the dosage of alleles in a polyploid. Polyploidy is common amongst horticultural plants. Estimating allele frequencies in a polyploid population is, therefore, complicated because of some marker genotypes being phenotypically indistinguishable. For example, in a tetraploid, with four alleles at a locus showing polysomic inheritance, although 35 genotypes are possible, these will fall into only 15 marker phenotypic classes. Furthermore 'null' individuals are rarely detected in polyploids. Furthermore, some polyploids can be cryptic exhibiting disomy, instead of the polysomic inheritance. We will discuss the implications of these factors and present an EM-type algorithm for estimating allele frequencies of a polyploid population under certain patterns of inheritance. The method will be demonstrated on simulated data. We also discuss the nature of some of the additional problems that may be encountered with estimating allele frequencies in polyploids for which other solutions still need to be developed. Heredity (2005) 95, 327-334.

Flowers of diploid Actinidia chinensis (kiwifruit) were hand-pollinated with pollen from either hexaploid A. deliciosa or diploid A. chinensis males and the subsequent fruit were evaluated. Following pollination with A. deliciosa pollen, fruit set, fresh weight, dry matter content, and seed weight and number were reduced. However, the most striking effect was on fruit flesh colour: the proportion of seedlings expressing red pigmentation, the intensity of pigmentation and the anthocyanin concentration were greatly reduced. The effects on maternal fruit tissues were probably indirect consequences of a reduction in the number of fertilized ovules due to partial pollen incompatibility. Effects on seed development could be explained largely by the ploidy difference between the seed and pollen parents. Growers should be cautious about using A. deliciosa pollen to pollinate diploid A. chinensis females, especially red-fleshed cultivars.