Combining Large Recombinant Inbred Lines Population And Ultra-high Density Molecular Markers To Identify QTL For Important Agronomic Traits In Maize

Genetic study of agronomic traits is of great importance for maize molecular breeding. Generally, maize agronomic traits are controlled by Quantitative Trait Loci (QTL). Population size and marker density are two important factors affecting the resolution and accuracy of QTL mapping. Here, we constructed two population with1147and887Recombinant Inbred Lines (RILs) derived from super maize hybrid Zhengdan958and Nongda108respectively, penotyped them for more than15traits at3locations, genotyped them by Genotyping By Sequence (GBS) method, constructed high resolution bin maps by sliding windows approach, conducted QTL mapping by R/qtl software, and verified4QTL by PCR based markers. The results are as followed:1. Analysis of the phenotype data, we found Plant Height (PH), Ear Height (EH), relative ear height (EH/PH), Leaf Length (LL), Upper Leaf Number (ULN), Tassel Length (TL), Tassel Branch Number (TBN), Ear Length (EL), Ear Diameter (ED), Row Number (RN), Ear Weight (EW) were in accordance with normal distribution in the two populations, but the Silk Color (SC) in Zhengdan958RILs, and Cob Color (CC), Leaf Width (LW) and Maize Rough Drawf Disease (MRDD) in Nongda108RELs were not. Several traits showed strong correlation (r2>0.5), including PH and EH, EH and EH/PH, ED and RN, ED and EW. Meanwhile, PH and TL, EH/PH and ULN, TL and EL, EL and EW, RN and EW had moderate correlation (r2-0.4) within each other.2. After strictly filtering, only826and670RILs for Zhengdan958and Nongda108were used for sliding window analysis for bin map construction. The final bin maps contain24,322and23,843bins respectively, with average of87kb and91kb between adjacent bins. Via QTL mapping for SC trait in Zhengdan958RILs, we got4QTL, with four maize anthocyanin biosynthesis related genes P1,pll, bzl and r1located in these QTL intervals. QTL mapping for CC and MRDD in Nongda108RILs resulted PI and qMrddl being mapped respectively. These results demonstrated our map quality.3. Via QTL analysis for12agronomic traits, we got120and67QTL in Zhengda958and Nongda108RILs population respectively. Almost all of these QTL were minor QTL with no more than10%of phenotypic variation explained, expect qLW4_N in Nongda108RILs with effect value of about25%. These traits with signal of correlation were found always with co-located or close linked QTL. There were only3QTL responsible for the same trait overlapped between the two populations, reflecting the complexity of agronomic traits genetic background.4. The median of QTL intervals in Zhengda958and Nongda108population were3.108Mb and2.870Mb, and many QTLs were overlapped with previous reports using smaller size of population or markers, but our physical intervals were remarkably reduced. Several genes with known function in maize were shown to be located within the QTL intervals, including ral and kn1for tassel branch number and br1, brd1, nal for ear height or relative ear height. Additional candidates near QTL peaks, inferred from homologous genes of known function, included zapl for upper leaf number, GA20OX2and ZmELDl for ear height, ZmALT5for tassel branch number. These finding demonstrated the accuracy and precision of our result, and implied the posibility of using large RILs and high-density markers to map QTL to sufficiently small intervals to identify candidate genes.5. We used PCR based markers to verify qPHla_Z, qEL3_Z, qEL8_Z in Zhengdan958population, and qLW4_N in Nongda108population, demonstrated their authenticity, and validated our results again.