Parents Progeny Expected ratio χ2

Locus ISSR banding phenotype ISSR genotype Phenotype (genotype)

1 (+/+ or +/-) 0 (-/-)

UBC812-0640 1 × 0 (+/+ or +/- × -/-) 15z (8)x 19 (4) 1:1 0.53

UBC825-0760 1 × 1 (+/+ or +/- × +/+ or +/-) 22 (8) 12 (4) 3:1 2.17

UBC825-0696 1 × 0 (+/+ or +/- × -/-) 13 (4) 21 (8) 1:1 2.13

UBC841-1075 1 × 0 (+/+ or +/- × -/-) 19 (7) 15 (5) 1:1 0.53

UBC841-0480 1 × 0 (+/+ or +/- × -/-) 17 (5) 17 (7) 1:1 0.00

UBC853-1930 1 × 0 (+/+ or +/- × -/-) 18 (7) 16 (5) 1:1 0.13

UBC860-0875 1 × 0 (+/+ or +/- × -/-) 13 (4) 21 (8) 1:1 2.13

UBC860-0940 1 × 0 (+/+ or +/- × -/-) 17 (8) 17 (4) 1:1 0.00

UBC873-0480 1 × 0 (+/+ or +/- × -/-) 18 (7) 16 (5) 1:1 0.13

UBC873-0800 1 × 1 (+/+ or +/- × +/+ or +/-) 21 (8) 13 (4) 3:1 3.60

UBC880-1150 1 × 1 (+/+ or +/- × +/+ or +/-) 23 (8) 11 (4) 3:1 1.11

10 mm Tris-HCl, pH 9, 50 mm KCl, 0.1% Triton X-100, 200 μM dNTPs, 2 mm MgCl2, 200 nm primer, 1.5 units of Taq polymerase, and 2 ng DNA template in 20 μL total volume. DNA was amplified using a GeneAmp PCR system 9700 (Perkin Elmer Applied Biosystems, Calif.). The program consisted of an initial period of 94 °C for 5 min, followed by 45 cycles of 45 s at 94 °C, 1 min at 50 °C, and 2 min at 72 °C with a final extension period of 10 min at 72 °C, followed by storage at 4 °C until use. The amplified products were electrophoresed at 4.5 V·cm-1 through a 1.4% agarose gel using a 1-kb ladder as the molecular weight standard (Promega Co., Madison, Wis.). The gels were then stained using ethidium bromide and visualized under UV light. Photographs were taken using a Kodak digital camera and digital photos were analyzed using Kodak Digital Science ID software (Eastman Kodak Company, Rochester, N.Y.).

MARKER SCORING AND DATA ANALYSIS. ISSR primers (University of British Columbia, Canada, microsatellite set #9) were screened against 34 progeny of a controlled cross and parents to identify segregating polymorphisms. Segregating ISSR markers were identified by the manufacturer primer code corresponding to the primer responsible for their amplification, followed by a four digit number indicating product size in base pairs (Table 2). ISSR phenotypes were scored as being either present (1) or absent (0) and the corresponding genotypes were assigned as +/+ or +/- and -/-, respectively (Table 2). Genotypes that had a polymorphic band in at least two of three gels were scored as +/+ or +/-. Chi-square tests were used to determine the goodness-of-fit of these segregating polymorphisms to their expected Mendelian ratios (n = 34). Those with a χ2 value of ≥ 3.84 were rejected.

Cultivars were screened for the presence of ISSR Mendelian markers that were identified in the segregating population (Table 2). Three replicate gels for cultivars were evaluated and scored for markers being either present (1) or absent (0). Cultivars having a specific Mendelian marker in at least two of three replicate gels were scored as having the marker present. For cultivars that had Mendelian markers in only one of three gels, an additional three gels were scored and genotypes with the marker in at least two of six gels were scored as having the marker present. Allele frequencies were indirectly estimated using the homozygous null genotypes (−/−) and corrected for dominance according to Lynch and Milligan (1994). Genetic statistics were estimated, corrected for small sample size (Nei, 1978), and used to determine the percentage of polymorphic loci (P) (95% criterion) and expected heterozygosity (He) (Miller, 1997). The level of genetic similarity among cultivars was determined by Nei's genetic distance (Nei, 1978). A dendrogram was constructed based on the matrix of the distances using unweighted pair-group mean analysis (UPGMA). Scores were entered in a matrix and analyzed with NTSYSpc software, version 2.02i (Exeter Software, Setauket, N.Y.). A similarity matrix was generated using the Dice coefficient, S = 2Nxy/(Nx + Ny), where Nx and Ny are the numbers of bands observed in clones X and Y, respectively, and Nxy is the number of bands common to both clones (Dice, 1945). The Dice values were then used to perform UPGMA cluster analysis and generate a dendrogram.

percentage of polymorphic loci (P) (95% criterion) and expected heterozygosity (He) (Miller, 1997). The level of genetic similarity among cultivars was determined by Nei's genetic distance (Nei, 1978). A dendrogram was constructed based on the matrix of the distances using unweighted pair-group mean analysis (UPGMA). Scores were entered in a matrix and analyzed with NTSYSpc software, version 2.02i (Exeter Software, Setauket, N.Y.). A similarity matrix was generated using the Dice coefficient, S = 2Nxy/(Nx + Ny), where Nx and Ny are the numbers of bands observed in clones X and Y, respectively, and Nxy is the number of bands common to both clones (Dice, 1945). The Dice values were then used to perform UPGMA cluster analysis and generate a dendrogram.

Eighty ISSR primers were screened that yielded a total of 234 amplified products. Of these 234 products, 30 polymorphic markers were found, and 11 of these segregated in a simple Mendelian fashion, as confirmed by chi-square analysis (Table 2). The 11 ISSR loci confirmed for Mendelian inheritance were used to assess genetic diversity in 19 pawpaw cultivars (Table 3). Marker UBC873-0800 had unreliable amplification and was not included in genetic diversity or cultivar separation analysis. Marker UBC860-0940 was present in all cultivars and UBC825-0760 was present in all but one cultivar (Sunflower), while the 8 remaining markers showed a high degree of polymorphism (Table 3). Isozyme marker variation analysis in cultivated pawpaw suggested genetic diversity components that are comparable to those of other long-lived temperate woody perennials of widespread geographic range with insect-pollinated outcrossing breeding systems, secondary asexual reproduction and animal-dispersed seed (Huang et al. 1997, 1998) (Table 4). Using the 10 ISSR markers, estimates of genetic diversity (P = 80% and He = 0.358) were higher than those based on isozymes (P = 44% and He = 0.172) for cultivated pawpaw, for RAPD markers for wild pawpaw accessions (P = 64% and He = 0.249) and cultivated pawpaw (He = 0.285) by Huang et al. (1997, 2000, 2003) (Table 4). However, care must be taken in comparing the ISSR genetic diversity estimates to those of Huang et al. (1997, 2000). Six additional pawpaw cultivars were included in this study that were not included in the isozyme study of Huang et al. (1997);