Comparison of AFLP and pedigree-based genetic diversity assessment methods using wheat cultivars from the Pacific Northwest

Comparing results of different genetic diversity estimation methods may be indicative of their utility as parental selection tools for plant breeders. Forty-three spring and winter wheat lines (Triticum aestivum L.) from the Pacific Northwest (PNW) were used to compare pedigree and amplified fragment length polymorphism (AFLP)-based genetic diversity estimates (GDE(PED) and GDE(AFLP), respectively). The mean of the 903 GDE(PED) pairwise comparisons was 0.96, and 89% of the values were >0.90. In contrast, GDE(AFLP) values were normally distributed (mean = 0.54), suggesting that this diversity measure was better able to identify pairs of genotypes representing the entire range of possible GDEs. Thus, GDE(AFLP) may have more utility than GDE(PED) for identifying parental combinations with maximum allelic variation. Despite the substantially different means and distributions of the two diversity measures, a moderate rank correlation (r(s) = 0.42, P < 0.001) was detected between the two GDE matrices. AFLP fragments from hypomethylated portions of the genome (generated with Pst I:Mse I) were more highly associated with GDE(PED), than were fragments generated with the methylation insensitive combination Eco RI:Mse I (r(s) = 0.44 and 0.28, respectively). Pedigree and AFLP-based GDEs detected a similar hierarchical pattern of genetic diversity among the 43 cultivars evaluated. Although GDE(PED), was not well suited for identifying degree of relatedness among individuals, it may be adequate for assessing overall patterns of genetic variation among regionally adapted germplasm. It remains to be determined which method, if either, provides predictive estimates of heterosis or genetic variance among progeny from specific parental combinations.