The use of RFLP markers has proven useful in the rapid construction of detailed linkage maps in several crop species and made possible the dissection of quantitative traits into Mendelian factors. The identification and examination of individual quantitative genes should provide information about the organization of genomes and insight into the relative contributions of quantitative genes to continuous variation. In this respect, genetic markers linked to factors associated with metric traits have been advanced in the literature to study quantitative inheritance.
The objective of this research was to identify RFLP loci associated with QTLs affecting expression of yield and other agronomic traits in a cross between two maize inbred lines, B73 and A7. From this cross 294 F3 lines were developed through two selfing generations with each F3 line tracing back to a different F2 plant. In the 1990 breeding nursery at Bergamo, Italy, two sets of testcrosses for each of the 294 F3 lines were produced: one with the tester inbred line A1, and the other with Mo17.
For each kind of testcross, the materials were subdivided into 3 sets and evaluated in simple 10x10 lattices at two locations, Bergamo and Brescia, Italy, in 1992. A total of 75 genomic maize clones were selected from collections of mapped clones available from Brookhaven National Laboratory and the University of Missouri to provide a uniform coverage of the genome.
Mapping of QTLs and estimation of their genetic effects were performed according to the method of interval mapping described by Lander and Botstein (Genetics 121:185-199, 1989) using the computer package MAPMAKER/QTL (Lincoln and Lander, Whitehead Inst. for Biomed. Res., Tech. Rep., Cambridge, MA, 1990). Presence of a putative QTL in a given genomic region was declared when the LOD scores of the additive model exceeded 2.5, corresponding to a probability P<0.05 that a false positive occurs somewhere in the genome. The total variation accounted by significant QTLs and the total LOD score were obtained by fitting a model including all putative QTLs for the respective trait simultaneously.
For grain yield in the testcross to Mo17, the long arm of chromosome 4 and the short arm of chromosome 6 showed highly significant effects with LOD scores of 3.2 and 6.1, respectively. In the testcross to A1, the short arm of chromosome 6 and the long arm of chromosomes 9 and 10 showed highly significant effect with LOD scores of 2.9, 5.4, and 2.9, respectively. Thus, the testcross to Mo17 showed at least two QTLs which collectively accounted for 21.7% of the variation for grain yield, while the backcross to A1 showed at least three QTLs which collectively accounted for 25.2% of the variation for grain yield. In the combined analyses for means over testcrosses, three genomic regions located on 4L, 6S, and 10L were found to significantly affect grain yield. LOD scores at peaks of QTL likelihood maps ranged from 2.6 to 7.4 for the genomic regions on 4L and 6S, respectively. The multiple QTL model indicated that these QTL, collectively, accounted for 35.4% of the variation for grain yield. It was interesting to note that the QTL on the short arm of chromosome 6 (LOD 7.4) accounted for 24.5% of the total phenotypic variation.
A total of three QTLs influencing grain dry matter content were detected. Analysis of Mo17 testcross data revealed two factors on chromosomes 1 and 2 with LOD scores of 4.8 and 4.9, respectively. The two loci together accounted for 22.7% of the phenotypic variance. Analysis of A1 testcross data confirmed the presence of the QTL on chromosome 2 between umc134 and umc131 markers, and suggested a second locus on chromosome 8. Collectively they accounted for 10.7% of the phenotypic variation for grain dry matter. The loci on chromosome 1 and 2 influencing grain dry matter content were confirmed in the combined analysis across testers. LOD scores at peak of QTL likelihood maps were 3.6 and 8.9, respectively. In total, 26.4% of the phenotypic variance was explained by the two QTLs.
Only a single QTL influencing plant height was detected in the testcross to Mo17. This QTL mapped on chromosome 3 and had a LOD score exceeding 2.7, which accounted for 16% of the phenotypic variation for plant height. In the A1 testcross, chromosomes 3, 5, 9, and 10 showed LOD scores exceeding 2.7, which altogether accounted for 28.8% of the phenotypic variation. QTLs found on chromosome 3 in the two testcrosses had extremely large, and overlapping, support intervals (>50 cM). Hence they were considered as identical loci even if they map between different flanking markers. It was also evident that for the QTL on chromosome 3 the B73 allele performed better than the A7 allele; the reverse was true for all the other QTLs detected.
Most QTLs found for the traits evaluated in our study were consistent across locations, although variations were observed in the LOD score levels, indicating that expression of genes controlling these traits was mainly independent of the environments. Only QTLs with larger effect were consistent across testcrosses suggesting that genetic background may contribute to the identification of the QTLs in a specific fashion. It is conceivable that data averaged over more than one testcross should be used for QTL identification. Obviously, further experiments will be required before sufficient evidence is available to verify this effect.
Loci for grain yield found on chromosomes 9 and 10 in our study are likely to have overlapping support intervals with QTLs for grain yield found in the cross B73xMo17 by Stuber et al. (Genetics 132;823-839, 1992). Moreover, all the loci for plant height found in our study mapped in chromosomal regions where the previous authors have found QTLs for plant height, although only the QTLs located on chromosome 3 have overlapping confidence intervals. A further observation which originates from our data is that on chromosomes 9 and 10, the likelihood peaks for the putative QTLs for plant height and grain yield fell in the same marker intervals. The direction of the effects of allele substitution was also consistent. The A7 allele increased both plant height and grain yield, suggesting evidence of an interrelationship of the genes regulating the two traits in this genomic region. The phenomenon of significant association of molecular markers with more than one trait has also been observed by others.
In conclusion, although further investigations will be required to establish the consistency of the detected effects in other genetic backgrounds, our results demonstrated the value of this type of investigation for identifying and localizing genetic factors (QTLs or specific genomic regions). This approach should be useful for marker-facilitated improvement programs, including intrapopulation selection or transfer of desired factors to other germplasms. Research involving facilitated breeding approaches is currently being addressed in our laboratory.
Return to the MNL 68 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage