Comparison of QTL mapping for plant height in F2:3 and F6:7 generations
--D.F. Austin and M. Lee

The first objective of our study was to locate and characterize genetic factors controlling quantitative traits in recombinant inbred (RI) progeny of an elite maize population. The second objective was to compare the RFLP linkage map and detection of quantitative trait loci (QTL) in F2:3 and RI (F6:7) lines of the same population grown at the same location in different years. With the RI population, we expect to detect QTL with smaller phenotypic effects due to increased replication of the homozygous parental marker classes. Also, the near homozygous state of RI progeny should result in a higher precision of trait measurement. The population was derived from a cross between inbred lines Mo17 and H99, which differ for several traits, including insect resistance, kernel size, grain yield, ear length, plant height, and flowering traits. From the original cross, 186 unselected F6:7 lines were developed. The lines were grown in a 14x14 lattice design with two replications near Ames, IA in 1993. A total of 14 traits were evaluated including ear height, plant height, ear-to-tassel interval, anthesis, silk emergence, silk delay, grain yield, ear number/plant, ear length, ear diameter, cob diameter, kernel row number, kernel depth, and 300 kernel weight.

A linkage map of 101 loci (100 RFLP and 1 morphological) was generated with MAPMAKER/EXP 3.0 using the "ri self" setting. Ten well characterized linkage groups were observed with an average interval between loci of 15.4 cM. Using 150 F2:3 lines from the same population, Veldboom et al. (Theor. Appl. Gen. 88:7-16, 1994) produced a linkage map with 104 loci (103 RFLP and 1 morphological). The total map distances for the F2:3 and RI maps using the 87 common marker loci are nearly identical at 1419 cM and 1420 cM, respectively. Map order is identical except for two loci on the end of the long arm of chromosome 9 linked by 2 cM in the RI map. These loci, npi209 and bnl14.28, are present in the opposite order in the F2:3 map.

Single-factor analyses of variance were conducted for all pair-wise marker loci and quantitative trait combinations. Trait data consisted of adjusted entry means from the lattice analysis. A significant (P<0.05) contrast between the homozygous parental genotypic classes was interpreted as evidence for linkage between a QTL and a marker locus. From the total number of significant loci for a trait, a subset of loci was selected to represent the total number of QTL detected by selecting the loci with the highest significance from a cluster of closely linked significant loci. If multiple loci on the same chromosome were significant when evaluated simultaneously in a model, they were included in the subset of significant regions as distinct QTL explaining unique portions of the trait variation.

For plant height, 31 loci and 7 chromosomes were significantly associated with trait variation. Twelve QTL were detected accounting individually for 2.3 to 11.8% of the phenotypic variation. Collectively, the additive effects of the 12 QTL explained 47.2% of the phenotypic and 52.4% of the genotypic variation as determined in a multiple model. QTL on 1S, 1L (2 QTL), 2S, 2L, 4L, 5L, and 8L have alleles from Mo17 (the taller parent) contributing to increased plant height. QTL on 3L, 4S, and 7L (2 QTL) have alleles from H99 contributing to increased plant height. Also, six of the seven QTL with the largest additive effects have Mo17 contributing increased plant height, but the second largest QTL is associated with H99 alleles. The QTL with the largest R2 value, identified by umc37 on 1L, has an additive effect of 6.7 cm, representing a difference of 13.4 cm in RI lines polymorphic at this locus.

Veldboom et al. (Theor. Appl. Gen. 88:7-16, 1994) identified five plant height QTL in the F2:3 generation of this population using interval mapping with MAPMAKER QTL. Four of the five regions were also identified with RIs. QTL were detected in both generations at umc37 (1L) and umc34 (2S). QTL were detected in both generations on 4S and 7L but at loci differing in location by 17 cM and 11 cM (based on F2:3 map), respectively. For all four common QTL, the parental alleles contributing to increased plant height were the same for the two generations. A QTL was detected in the F2:3 on 6L, but the region was not significantly associated with plant height variation in the F6:7. The F2:3 and F6:7 generations were both grown at the same locations, but the environmental conditions were very different in 1989 and 1993. Conditions in 1989 were the seventh driest on record while 1993 was one of the wettest growing seasons on record. Despite this difference in generations and environments, most regions associated with plant height variation in the F2:3 were also detected in the F6:7.

In the F2:3 generation, a large significant region on chromosome 1 was attributed to a single QTL identified by umc37 with the genetic effect for increased plant height derived from Mo17. The region could not be further resolved with the F2:3 generation. This region has been resolved into three distinct QTL in the F6:7 identified by P1, umc37, and umc86A on the basis of evaluation in multiple models (Figure 1). Genetic effects for increased plant height for all three QTL are derived from Mo17. This suggests that the increased recombination fraction and precision of RIs allowed the separation of one large linkage group of significant loci into multiple linked QTL.

The remaining eleven traits evaluated with the RI population in 1993 have also been analyzed in the same manner. The experiment was repeated in 1994 for an additional environment. Also, the F2:3 experiment has been repeated a second year by Veldboom et al. (in review). Additional comparisons will be made to determine consistency of QTL detection across generations, environments, and independent samples of this population.

Figure 1. Comparison of plant height QTL detection on chromosome 1 in F2:3 and F6:7 generations of a single-cross population of inbreds H99 (short) and Mo17 (tall). Names of RFLP marker loci are on the left. In the F2:3, a LOD threshold of 2.0 was used. Significance levels in the F6:7 analysis are noted by *, **, and *** representing 0.05, 0.01, and 0.001, respectively

Please Note: Notes submitted to the Maize Genetics Cooperation Newsletter may be cited only with consent of the authors

Return to the MNL 69 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page