Methods Of Kernel Traits Measurement And QTL Mapping And Analysis Of Association

Maize is not only a major crop for food and feed, but also constitutes important resource for industryand bio-energy. Besides it plays major role in providing food, feed and bio-energy, maize is also animportant material for biological research. Constantly high yield and superior quality are the two majorbreeding goals of maize, both of which focus on maize kernels. Grain kernel can be into two types,physiological and quality. Physiological traits include kernel color, kernel shape (flint, dent, sweet, waxyetc.), kernel size (length, width, and thickness), kernel architecture (length/width, length/thickness,width/thickness), kernel weight, kernel specific gravity, kernel volume, kernel weight per unit volume,kernel dry-down rate, kernel water content, and kernel density,. Quality traits refer to kernel protein,starch, fat and mineral contents and compositions. Currently, there are numbers of reports on QTLmapping of kernel quality traits; however, similar research on kernel physiological traits except for kernelweight has lagged far behind. Since kernel physiological traits are correlated with grain yield, they aregiven more and more attention recently. Researches on kernel physiological traits will provide insightfulinformation for future maize breeding. The purposes of this study were:(1) to establish an efficientmethod to measure maize kernel size and volume;(2) to map quantitative trait loci (QTL) or genesunderlying maize grain yield by utilizing both linkage and association analyses. In this study, six kernelphysiological traits: kernel length (KL), kernel width (KWH), kernel thickness (KT), kernel weight (KW),kernel volume (KV) and kernel density (KD) ware focused in a recombinant inbred line (RIL) populationderived from inbred lines Zong3and87-1under two years field testing. Meanwhile, an associationpanel of121maize inbred lines was also used for measuring the same six traits, and a maize gene,ZmDEP1, a homolog of rice OsDEP1gene, was taken as candidate gene to investigate for its contributingto those six traits. Major findings are as follows:1. An efficient method to measure kernel length, kernel width and kernel thickness was established.This method includes two major steps: taking digital pictures of maize kernels and converting pixels toactual size measurement by comparing to a standard pixel/size conversion rate using ImageJ. Using thisnew method and a standard method using vernier caliper, we measured kernel size in a panel of70maizehybrid lines which showed a wide range of variation:9.50～13.50×10-3m for kernel length,8.0～10.0×10-3m for kernel width, and3.50～5.50×10-3m for kernel thickness. High and significant correlation wasobserved for measurements from the two methods:0.9942,0.9945and0.9945for kernel length, kernelwidth and kernel thickness respectively. 2. A simple volume-measurement instrument was made based on the “water displacement” principleusing common materials. The accuracy of this instrument was confirmed by comparing with steel ballsfor which volume is known. We applied this method to measure30hybrids and compare themeasurement with a traditional method, graduated cylinder. Single seed volume among those30hybridsvaries between0.091and0.325cm3, and the correlation between these two methods was very high,0.9996.This provided an efficient means to measure kernel volume and kernel density.3. Using the established method from1and2, we measured KL, KWH, KT, KW, KV and KD in panelof294RILs derived from Yuyu22in two years field testing. Correlation analysis was performed. We alsomapped QTLs for those traits by using Composite interval mapping method. A total of33QTLs wasmapped distributed on all chromosomes except chromosome3. Those QTLs were not evenly distributed,with chromosomes1and5having the most. QTLs for different traits co-localized with each other. Forexample, on chromosome1between bnlg1025and umc2029, QTLs were detected for KW, KT, KD andKV; on chromosome5between umc1155and bnlg1237, QTLs were detected for KW, KT, KW, KV andKD; on chromosome7between mmc0411and bnlg339, QTLs were detected for both KW and KV. Wefound that QTLs for highly correlated traits tend to localize at similar or closely related regions, implyingthat the correlation at trait levels might be due to the correlation at QTL levels.4. We obtained and confirmed the full-length genomic sequences of ZmDEP1gene in maize based onthe sequences of its rice homolog, OsDEP1, which is a gene controlling rice grain yield. We found thatZmDEP1co-localized with two KW QTLs: q300k20and qgrwt6. We resequenced the3’ end of ZmDEP1in a panel of121elite maize inbred lines and found a substantial number of polymorphisms in this region.Base on Association analysis performed between those polymorphisms and kernel traits measured in thesame panels and in two different years field testing, we found a SNP site within3’ UTR significantlyassociated with KW and KV in both two environments. This same SNP is also significantly associated withKW and KT in one environment. Together, those results suggested that ZmDEP1might be a geneaffecting KW and KV, indicating functional conservation of its rice homolog.