| Rice is one of the most important crops in the world, and has become the model crop in plant for genetics and genomic studies. Yield componens and related traits are the most important agronomical traits and are controlled by quantitative trait loci (QTL) in rice. In recent two decades, with constructing of high-density molecular linkage map, QTL mapping of rice had got great achievement for rice yield, quality, resistance and some other important agronomic traits. However, the study for QTL mapping needs to be furthered because of the diversity of global germplasm resources and speciality of QTLs. The goal of this study was to investigate some important agronomic traits including plant height, heading time, yield of plant and yield components by using the methods of variance analysis, correlation analysis, stepwise regression analysis and path analysis with a set of 241 RIL population (Fg) derived from an elite hybrid cross of 'Zhenshan 97 x Minghui 63'. A molecular linkage map of 221 markers was constructed and QTLs, epistasis and QTL x evironment interaction for these traits were analyzed by using the mixed-model-based composite interval mapping with the 2 year's data (1999 and 2000). The results are as follows:1. The significant transgressive segregation for the traits studied was found in the population and the frequency of these traits was approximately normally distributed. Significant differences between the parents were detected for the traits except for panicle number per plant. The correlation coefficient between grain weight per plant and filled spikelets per panicle, seed setting rate, panicle number per plant, 1000-grain weight, spikelets per panicle, plant length, heading date, plant height were 0.801, 0.377, 0.266, 0.203, 0.136, -0.090, 0.050, -0.045 , respectively. The four traits including filled spikelets per panicle, seed setting rate, panicle number per plant, 1000-grain weight contributed 88.7% of genetic variation of the grain weight per plant. The complex correlations of these trait phenotypes could be well explained with regression analysis and path analysis. The path coeffiecint and correlation coefficient between grain weight per plant and panicle number per plant were opposite. The main reason was the significant positive correlations that leaded to the indirect effects from filled spikelets per panicle and seed setting rate to grain weight per plant were more than 0.514. So it covered the real effect from panicle number to grain weight per plant.2. A total of 45 quantitative trait loci (QTLs) and 47 pairs of interaction were found for these traits in 1999. The contribution rates of single QTL were 2.83-28.46%. The QTLs detected for heading date, panicle number per plant, plant length, plant height, spikelets per panicle, filled spikelets per panicle, seed setting rate, 1000-grain weight and grain weight per plant were 3, 4, 6, 6, 7, 2, 6, 8 and 3, respectively. The C1023-R1440 in chromosome 7 was the most active flanking marker and held the pleiotropism, which was simultaneously associated with the same QTL in 5 traits and the 3 pairs of significant epistatic loci in 3 traits. There were significant correlations among these traits.3. A total of 65 QTLs detected in 1999 and in 2000, which includes 45 QTLs in 1999, 35 QTLs in 2000. Among these QTLs, 15 QTLs were all been found in both years. These QTLs were in each of 12 rice chromosomes except for the forth one. But the epastatsis were detected in the forth chromosome. The genetic effects in 2000 were 2.83-14.99 and were lower than those in 1999. The expressing of QTLs could be affected by the different environments.4. The results indicated that 8 QTLs held significant QTLx environment interaction effects. One QTL for panicle length, heading date and 1000-grain weight, respectively, two QTLs for spikelets per panicle and three QTLs for seed setting performed significant QTL x environment interaction effects. The contributions from interactions were slightly larger than those from the QTLs inv...
|
PDF Full Text Request