| Maize is one of the most important staple crops and it is cultivated all over the world.The production of maize has exceeded rice and wheat.Maize kernel size and weight are important contributors to its yield and the dissection of their genetic basis is helpful for maize yield improvement.In this study,we combined linkage mapping and association mapping in 10 RIL populations to dissect the genetic basis of maize kernel size and weight.Then,we used comparative genomic analysis and candidate gene association analysis to gain knowledgements about the conserved genetic basis for seed size and weight in rice and maize.The main results were:1.We identified the maize ortholog of rice GS5(GRMZM2G123815)and found that there is a synteny between the region on which GS5 was located of rice genome and the region on which Zm GS5 was located of maize genome;26 paralogs of Zm GS5 were identified and expression GWAS of Zm GS5 and 15 paralogs shown that Zm GS5 and other3 paralogs were regulated by a same trans-element Zm BAK1-7.Zm GS5,3 paralogs of Zm GS5 and Zm BAK1-7 were highly co-expressed.Through candidate gene association analysis,we found that Zm GS5 was significantly associated with kernel length,width and hundred-kernel weight in Hainan,2011;Zm BAK1-7 was significantly associated with kernel length in 4 environments and kernel width,thickness and hundred-kernel weight in at least one environment.The expression GWAS results of 28,769 genes shown that Zm BAK1-7 could regulate the expression of 882 genes.Zm GS5 was located in a QTL(R2=8.50%)for kernel test weight identified in Teosinte×Mo17 BC2F6 population;Zm BAK1-7 was located in a QTL(R2=10.60%and 9.80%)for kernel width and hundred-kernel weight identified in YU87-1×BK population.2.We performed QTL mapping separately(SLM)in 10 RIL populations with composite interval mapping for hundred-kernel weight,kernel test weight,kernel length,width and thickness and we identified 373 QTLs in total,90,70,61,89 and 63 for each trait,respectively.Joint linkage mapping(JLM)and GWAS was conducted through combining 10 RIL populations.In joint linkage mapping,300 QTLs were identified with56,59,55,68 and 62 QTLs for hundred-kernel weight,kernel test weight,kernel length,width and thickness,respectively.In GWAS,754 significant SNPs were identified with194,136,138,163 and 123 SNPs for hundred-kernel weight,kernel test weight,kernel length,width and thickness,respectively.After backward regression analysis of 754significant SNPs to reduce redundancy,we finally identified 135 candidate SNPs with 30,22,26,32 and 25 SNPs for hundred-kernel weight,kernel test weight,kernel length,width and thickness,respectively.3.With comparative analysis,we identified 30 maize genes which were orthologs of18 rice cloned genes for seed size or weight and found that 26 of them were located in the candidate region of QTLs mapped with at least one method mentioned before and 5 with both three methods.Candidate gene association analysis between these 30 genes and 5kernel traits identified significant associations between 5 of them(Zm SLG,Zm GLW7-1,Zm GL2,Zm GW7-2,Zm SRS1-2)and at least one kernel traits.Expression GWAS of 5genes which were significantly associated with at least one kernel traits found that expression of Zm GL2 was regulated by trans-elements and its expression was significantly correlated with kernel width,while expression of Zm SLG was regulated by cis-elements near this gene and its expression was significantly correlated with kernel width.4.We also fine-mapped a major QTL for both kernel with and hundred-kernel weight identified in ZHENG58×SK RIL population.The candidate region was narrow down to~600 Kb and we proposed a candidate gene,q HKW1-9,for this QTL based on whether genes in this region expressed in kernel and the annotation of functions.We identified two indepent Mu-insertion lines for q HKW1-9.We also constrcted q HKW1-9:YFP infusion gene over-expression vector and q HKW1-9 CRISPR-Cas9 gene editing vector and successfully transformed them into maize. |