| High and stable yield are always the main objectives of maize breeding. Investigating the genetic basis of grain yield and yield-related traits, and dissecting the genotype×environment interactions thoroughly are of great importance for the genetic improvement of grain yield in maize. Two F2:3 populations and their testcross populations derived from the foundation inbred lines used in maize breeding in China were materials in the present study. Molecular marker techniques combined with phenotyping in multiple years and locations were used to investigate the genetic basis of eight traits in maize, i.e. grain yield per plant (GYPP), kernel number per plant (KNPP), 100-kernel weight (KWEI), kernel density (KDEN), 100-kernel volume (KVOL), 10-kernel length (KLEN), 10-kernel width (KWID) and 10-kernel thickness (KTHI). The main results were as follows:1. The genetic ralationships between grain yield and yield components in maize at phenotypic and QTL level were studied.An F2:3 population (Q/H) including 230 families derived from two foundation lines in maize breeding of China, i.e. Qi319 and Huanzgaosi, was used to conduct conditional analysis and QTL mapping. The genetic relationships between GYPP and yield components (KNPP and KWEI) were investigated. The results showed that the heritability of KWEI was higher than those of GYPP and KNPP. Both KNPP and KWEI were positively significantly correlated with GYPP, and the correlation coefficient between KNPP and GYPP was higher than that between KWEI and GYPP. A total of five QTLs and five pairs of epistatic loci were detected for GYPP by the unconditional QTL mapping. Of these, four QTLs were found with pleiotropic effects on three traits (GYPP, KNPP and KWEI), while another one QTL controlled two traits (GYPP and KNPP). All of the five pairs of epistatic loci were related to both KNPP and KWEI. Consequently, either KNPP or KWEI was closely related with GYPP at phenotypic level, suggesting that the improvement of KNPP and KWEI can efficiently contribute to high grain yield. The conditional QTL mapping demonstrated the strong genetic relationship between GYPP and two yield components (KNPP and KWEI) at the individual QTL level.2. QTL-by-environment interactions and stability of QTLs across multi-environemnts with normal irrigation for grain yield and yield-related traits in maize were investigated.Two F2:3 populations derived from the crosses of Qi319×Huangzaosi (Q/H) and Ye478×Huangzaosi (Y/H), respectively, and their parents were evaluated under 6 environments including Henan, Beijing and Xinjiang in the years of 2007 and 2008. Correlation analysis and the hypergeometric probability function showed the dependence of yield components on kernel-related traits. The stepwise joint QTL mapping procedures were used to identify quantitative trait loci (QTLs). About 70% of the QTLs for Q/H and 90% of the QTLs for Y/H did not show significant QTL×environment interactions in the joint analysis across all environments. In contrast to the QTLs for GYPP and KNPP, the QTLs for KWEI, KVOL, KLEN, KWID and KTHI were quite stable across different environments. Seven constitutive QTLs (QTLs with R2>10% in at least one environment and also detected in more than four environments based on single environment analysis) affecting more than one trait were identified on chr. 1, chr. 4, chr. 6, chr. 7, chr. 9 and chr. 10 in the two populations, among which two were for each of KLEN and KWID, one for each of KNPP, KDEN and KTHI. Moreover, the QTLs on chr. 1, chr. 4 and chr. 9 could be detected in both of the two populations.3. QTL-by-environment interactions and stability of QTLs across different water regimes for grain yield and yield-related traits in maize were investigated.Two F2:3 populations derived from the crosses of Qi319×Huangzaosi and Ye478×Huangzaosi, were used to investigate the genetic basis of grain yield and yield-related traits under different regimes in Xinjiang (including well-water and water-stress environments). The results showed that above 70% of the QTLs for grain yield and yield-related traits expressed stably under the same water regime across the 2 years. The QTLs detected in water-stress environments were less stable than those detected in well-water environments across the 2 years in Xinjiang. The joint analysis combining data of all environments indicated that the stability of the QTLs for all traits decreased, but above 60% of them still expressed stably. A total of 11 constitutive QTLs (QTLs with R2>10% in at least one environment and also detected in more than two environments based on single environment analysis) distributed on bin1.10, 2.00, 4.09, 7.02, 9.02, 10.04 and 10.07 were detected in the two populations, and all of them except bin10.04 were stable across all environments. Consequently, most of the QTLs for grain yield and yield-related traits stably expressed under the same water regime across different years, and even under different water regimes in Xinjiang.4. The genetic correlation between line per se performance (LP) and their testcross performance (TP) for grain yield and yield-related traits across multi-environment with normal irrigation were studied and four common QTLs between LP and TP with high stability across environments were detected.Two F2:3 populations derived from the crosses of Qi319×Huangzaosi (Q/H) and Ye478×Huangzaosi (Y/H), respectively, and their testcross populations derived from testcross the two F2:3 populations with Mo17 (unrelated tester) were used to further investigate the genetic basis of grain yield and yield-related traits by the methods of molecular marker technique under three locations within one year. Genotypic correlation analysis showed that the genotypic correlation coefficient of line per se performance (LP) with testcross performance (TP) for GYPP and KNPP were low, and rg(LP,TP) within 0-0.16. While for the traits with additive effects as the main gene action, such as KWEI, KDEN, KVOL, KLEN, KWID and KTHI, the genotypic correlation coefficients of LP with TP were relatively higher, and rg(LP,TP) within 0.32-0.69. Consequently, selection of LP for KWEI, KDEN, KVOL, KLEN, KWID and KTHI to improve the correspondent trait in TP could be efficient. The genotypic correlations for GYPP in Q/H and KNPP in Y/H between observed TP and its prediction based on QTL positions and effects for LP, rg(MLP,YTP), were both of zero, and marker assisted selection (MAS) for QTLs detected in LP were no gains in the improvement in TP. Besides those traits, for other traits in each population could receive certain genetic gains.A total of 16 common QTLs between LP and TP in both populations were distributed in 10 genomic regions. i.e., bin1.10, 3.09, 4.03, 4.05, 4.06, 6.00, 6.07, 9.02, 9.03 and 10.04. Among those 10 genomic regions, 4 were detected with constitutive QTL. Especially for bin4.05 (QTL for KLEN), which was detected common QTL between LP and TP in each population, explained for 12.7% (Q/H) and 17.2% (Y/H) phenotypic variations for LP and 6.0% (Q/H) and 5.3% (Y/H) phenotypic variations for TP, furthermore with high frequency in 1000 times CV/G.
|
PDF Full Text Request