Ability of RFLPs to uniquely identify and to show associations among 160 elite lines used within a proprietary breeding program during 1930-1990

--J. S. C. Smith, O.S. Smith, S.L. Bowen, R.A. Tenborg, S. Wall, and D.N. Duvick

Inbred lines of key importance in the proprietary U.S. maize breeding program of Pioneer Hi-Bred International, Inc. from 1930 through to 1990 are being profiled for 100 RFLP probes; preliminary data for 160 lines and 37 probes are presented herein. RFLP probes were selected on the basis that they were well dispersed throughout the genome, that they revealed a high degree of polymorphism among previous screens of elite U.S. maize germplasm (Smith et al., Theor. Appl. Genet., 80:in press, 1990), and that banding profiles were readily scorable in terms of discrete variants. Distances were calculated between all pairs of lines according to Nei and Li (1979) and associations among lines were revealed by cluster analysis. All lines gave unique profiles. Two lines that were 94% related on the basis of pedigree (Malecot's Coefficient of Parentage) had 98.5% of their variants in common and were the most similar according to RFLP data, as would have been expected from their pedigrees. The most dissimilar lines, on the basis of RFLPs, had 22% common variants. Seventy percent of the lines clustered at a level of >30% profile dissimilarity between lines. Two hundred and eighty-six variants (mean of 7.73 variants per probe) were scored. Collectively the data show an abundant breadth of genetic diversity among these elite inbred lines. Cluster analysis of the RFLP data resulted in seven large groups of lines in broad agreement as would be expected from pedigrees. Correlation of RFLP distances between lines versus distances calculated from pedigree records (1 - Malecot's Coefficient of Similarity) was r = 0.58. This value was lower than that (r2 = 0.81) reported previously (Smith et al., Theor. Appl. Genet., 80:in press, 1990) for 37 elite Corn Belt lines. However, among these 160 lines there were no connections by pedigree breeding for 66% of the pairs of lines. For the majority of pairs, therefore, there was no information upon which to compute a reliable estimate of distance other than to use RFLPs (or other data such as F1 yield or heterosis that are extremely resource-consuming to obtain). RFLPs can be very useful to show associations among lines on the basis of their genetic constitutions. Studies that track genotypes through generations of breeding are one means by which an index of agronomic worth could be assigned to RFLP profiles. Such information could assist in further objective utilization of available genetic resources by breeders.

Return to the MNL 65 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage