The first objective of our study was to identify RFLP loci associated with agronomic traits in inbred progeny of an elite maize population. The second was to determine the minimum number of QTL controlling each trait and the proportion of the total phenotypic variation explained by each locus. The recombinant inbred (RI) population was derived from a cross between elite inbred lines Mo17 and H99, which differ for several traits including insect resistance, kernel size, grain yield, ear length, plant height, and flowering date. From the original cross, 186 unselected F6:7 lines were developed. By using RIs, we expect to detect smaller phenotypic effects because of increased replication of the homozygous parental marker classes. The power of detecting significant differences between the homozygous marker classes in a RI mapping population is increased over an F2 derived population of equivalent size. This is due to the reduction of the expected frequency of heterozygous individuals from 50% in the F2:3 to 3% in the F6:7 at any given locus.

A linkage map consisting of 100 RFLP loci and 1 morphological marker was developed using Mapmaker. The map consisted of ten well characterized linkage groups with a total map length of 1408 cM and an average interval between loci of 15.4 cM. Veldboom et al. (Theor. Appl. Genet., 1994) produced an RFLP map in the same population using 150 F2:3 lines with 103 RFLP loci and 1 morphological marker. Their total map length was 1419 cM with an average interval length of 15.0 cM. Between the two studies, 84 loci are in common. Marker order within chromosomes is conserved between maps with one exception. At the end of the long arm of chromosome 9, the order of two loci, separated by 2 cM on the RI map, is reversed. Burr et al. (Genetics 118:519-526, 1988) discussed the expectation of a two-fold expansion of the map for closely linked markers when mapping with RIs. To analyze this expectation, we compared common loci intervals which were less than 15 cM in the F2:3 map. Of the 37 intervals, 17 are larger and 21 smaller in the RI map. The average expansion of the 17 intervals which are larger in the RI map is 1.3X.

Single-factor analysis of variance was conducted on all locus-trait combinations to test for significant differences between the homozygous marker classes. For plant height, 32 marker loci were significant (.05) and were located on 7 chromosomes. Significant regions include 1L, 2S, 2L, 3L, 4S, 4L, 5L, 7L, and 8L. The proportion of phenotypic variance explained by marker classes for individual markers ranged from 2.1% to 12.3%. In three of the regions (3L, 4S, and 7L) H99 alleles are associated with increased plant height. The remaining regions are associated with Mo17, the taller parent, contributing the positive effects. Veldboom et al. identified putative QTL using Mapmaker QTL for plant height on 1L, 2S, 4S, 6L, and 7L. Four of the five regions were also identified in our study. Loci umc37 (1L), umc34 (2S), and umc35 (7L) were identified in the previous study and were significant (.01) in the present study. Both studies identified region 4S but different loci were indicated. In all four regions, the parent contributing the positive effect is consistent between studies.

Several additional traits have been measured on the RI population and will be analyzed in a similar manner. Additional traits to be analyzed include ear height, silking date, anthesis date, silk delay, yield and yield components. 

Return to the MNL 68 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage