Of the 90 RAPD primers used to amplify DNA from the seventeen lines, 73 (81%) revealed polymorphisms among the seventeen parents. These 73 primers generated a total of 453 non-redundant polymorphic bands with an average of 6.2 and a range of 3 RAPD profiles. Among the 453 RAPD variants, 10 (2.2%) were present in only one of the seventeen parents. For AFLP analysis, sixteen primer combinations were used to assay the seventeen inbreds. These primer combinations revealed approximately 1038 selectively amplified DNA fragments ranging in size from 80 to 900 bp nucleotides. Among 1038 AFLP variants, 621 (59.8%) were polymorphic bands with an average of 38.8 and a range of 30 to 59 per AFLP primer combination, 75 (7.2%) were detected in only one of the seventeen parents. These results suggest that AFLPs can reveal a large number of polymorphisms in a more efficient way compared with RAPDs.

Nei's genetic distance (GDs) were computed for all 136 combinations of the seventeen parents based on AFLP and RAPD markers. GDs based on AFLP data among the seventeen parents ranged from 0.19 to 0.56, with an average of 0.43 across all 136 pairs. GDs based on RAPD data among the seventeen parents ranged from 0.09 to 0.67, with a mean of 0.59 across all 136 pairs. These results indicate that GDs based on AFLP data were significantly different from GDs based on RAPD data used in this study.

Cluster analysis based on the AFLP data and RAPD data resolved the seventeen parental lines into five major groups that were consistent with pedigree information, GD values among groups were significantly greater than that within group. Dendrograms were also constructed on the basis of either 21 AFLP or 453 RAPD variants and compared with the dendrogram generated from the entire data set of AFLPs and RAPDs. The AFLP-based dendrogram had quite similar clustering structure as the dendrogram developed from the RAPD data. The only discrepancy was that 1301 and P167 (1301 and P167 were synthetical-bred inbreds) were in different groups. These results suggest that AFLP and RAPD markers used in this study can assign genotype to different heterotic or subspecific groups.

The correlations of genetic distance with the F1 yield (F1Y)/mid-parent heterosis (MPH)/specific combining ability (SCA) of the 136 hybrids are presented in Table 1.

Table 1. Correlations of genetic distance (GD) based on AFLP and RAPD data respectively, with F1 yield (F1Y)/mid-parent heterosis (MPH)/specific combining ability (SCA) of grain yield for 136 crosses.
 

Variables	F1Y	MPH	SCA
GD-AFLP	0.4352**	0.3453**	0.4732**
GD-RAPD	0.4018**	0.3247**	0.4217**

**Significant at 0.01 probability level

As shown in Table 1, genetic distances based on AFLP and RAPD data were significantly correlated with F1 yield. The correlation coefficient(r) was 0.4352 for AFLPs and 0.4018 for RAPDs. The correlation coefficients of GDs calculated for AFLP and RAPD data with mid-parent heterosis for 136 hybrids were highly significant. Genetic distance based on AFLP and RAPD data were also correlated with specific combining ability (SCA)(P<0.01). Finally, it was worth noting that correlation between GD computed from AFLP data and F1Y/MPH/SCA were higher than those based on RAPD data.

In summary, the results from the current study indicate that AFLP and RAPD offer a reliable and effective means of assessing genetic variation and of assigning maize inbred lines into different heterotic groups and thus reduce the field work associated with making cross and hybrid field testing. In particular, AFLP technique may allow maize breeders to predict combinations of lines that result in high-yielding, single-cross hybrids.

Return to the MNL 74 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage