Genetic Dissection Of Kernel-Related Traits And Fine Mapping Of A Major QTL, QKL1.07, For Kernel Length In Maize

Kernel size and weight are major determinants of grain yield. For understanding molecular mechanisms of kernel related traits, QTL for kernel related traits were detected. Further, by the combining strategy of linkage and assocaition approach, a kernel-length related QTL in maize, q KL1.07, was fine mapped. Meanwhile, the candidate gene was also supposed by genome sequence analysis. The main research results are as follows:1. QTL mapping for kernel related traitsThe average kernel length of HZS was about 0.86 cm and the average kernel length of Mo17 was about 1.02 cm. In population of RIL, the date of kernel related traits fitted normal distribution. QTL for kernel related traits were detected by stepwise regression using Windows QTL ICI-Mapping software in five environments. In totally, 30 QTL were detected in single environment on chromosome 1、2、4、6、7、8、10 with the range of variation of explaining phenotypic variance 6.28%-27.63% and 11 QTL were detected in joint environment on chromosome 1、2、4、6、9、10 with the range of variation of explaining phenotypic variance 6.01%-26.79%.2. Fine mapping of a major QTL q KL1.07 for kernel lengthUsing the Ici Mapping V3.3, the q KL1.07 was mapped to the marker interval of umc2237 and ML93 in BC3F2 population(Fig. 3A). To further narrow down the target region, five new In Del markers were developed applying to genotype the BC3F3 population. As a result, q KL1.07 was narrowed down to the marker interval of ML61 and ML93, explaining ~11 % of phenotypic variation. In order to fine map q KL1.07, a large BC3F4 population(2341 plants) derived from homozygous plants of BC3F3 was evaluated in the last season. With much more recombinant events occurred, the q KL1.07 was detected within a region flanked by the two markers of ML74 and ML162(Fig. 3A). Furthermore, using a new marker ML194 within the narrowed region, a total of 156 homozygous recombinants were divided into twelve recombinants lines in the BC3F4 population. The results indicated that recombinants, harboring the introgressed segment of Mo17 flanked by ML194 and ML162, showed significantly longer KL than those lines with the HZS genotype in the corresponding interval(Fig. 3B). Finally, the region of q KL1.07 was fine mapped to a 1.6Mb interval with the flanked marker of ML194 and ML162.3. Candidate gene association analysis for kernel-length QTL q KL1.07Diverse KL variation was observed for the association panel(AP) with 627 maize inbred lines in this study. According to the mapping results of linkage analysis, 425 SNP markers within the target QTL region flanked by the markers umc2237 and ML186 were selected to process a regional association analysis. Considering the family relatedness within the population(the first 3 PCs), the association analysis was conducted with a general linear model(GLM) by the TASSEL V4.2.1 software for the AP. As a result, four significantly associated SNPs(significant level, P<0.00001) were detected to be associated with KL. For the peak signal, the marker PZE-101164750 showed the strongest association with KL(P = 2.36E-09), which could explain 5.49% of the total variance for the used AP in our study. The averaged LD decay for the target QTL region was 64 kb in the AP. Then, four candidate genes could be identified by the window of 64 kb around those significantly associated SNPs(GRMZM2G122064, GRMZM2G122126, GRMZM2G438749, and GRMZM2G363004)(Fig. 6). Meanwhile, an obvious LD region with about 212 kb distance could be detected around the most associated SNP(PZE-101164750). Within this LD, another candidate gene, GRMZM2G348452(Zm CKX10), exactly located at the 98 kb upstream of PZE-101164750, which was also treated as one of the candidate gene for q KL1.07.