Genetic Dissection Of Disease Resistance To Ear Rot In Maize By Linkage Mapping And Association Mapping

Ear rot is one of the most destructive maize diseases worldwide.In recent years,the incidence rate of ear rot in China has increased year by year.As one of the most important dominant pathogens of maize ear rot,Fusarium verticillioides not only causes Fusarium ear rot（FER）,but also produces Fumonisin,which poses a serious health threat to human beings and livestock.As the resistance to maize ear rot is a complex quantitative trait,which is controlled by many minor quantitative trait locus,it is greatly influenced by pathogens,environmental factor and genetic background.It is of great significance to analyze the genetic basis of maize ear rot and mine the resistant genes for breeding and improvement of disease-resistant varieties.A set of 183 F_6:7recombinant inbred lines（RILs）of KA105×KB024 and 168 diverse lines of natural populations were used to carry out multi-environment ear rot resistance QTL mapping and genome-wide association study.Besides,genomic prediction was carried out using the RILs.In order to improve FER resistance of inbreds in Shann A and Shann B heterotic group and improve the efficiency of disease-resistant breeding by using genome-wide selection prediction technology.The main research results are as follows:1.Under the six single environments,the differences in the FER phenotype values of the parents KA105 and KB024 reached a very significant level,the FER of the KA105 line was 7.7%-12.5%,and the FER of the KB024 line was 46.3%-76.6%.The FER phenotype of the RIL population conforms to the basic characteristics of quantitative traits and is suitable for QTL mapping.Analysis of variance showed that the differences between genotypes,environments,and intercation also reached a very significant level.The broad-sense heritability was 80.4%.2.The population of recombinant inbred lines was genotyped with the Maize6H-60K chip.A genetic linkage map was constructed using 6193 high-quality SNP markers after quality control,with a total length of 4540.58 c M and an average marker density of 0.73c M.Combining the FER phenotypes and BLUP values in 6 single environments,a total of18 FER resistance QTL were detected.Most of them are minor QTLs.The phenotypic variation explained by a single QTL is 4.66%-12.41%.Among them,q FER1-1 located in bin1.02（which can explain 4.69-7.49%of phenotypic variation,and the resistance allele comes from KB024）,q FER3-1 located in bin3.03（which can explain 5.15-9.97%of phenotypic variation,and the resistance allele comes from KA105）,and q FER4-2 located in bin4.03（which can explain 5.11-7.82%of phenotypic variation,and the resistance allele comes from KA105）,and q FER9-1（which can explain 4.88-8.39%of phenotypic variation,and the resistance allele comes from KA105）were stable QTL in multiple environments.3.The content of fumonisins in kernels of RIL families in two single environment were determined.The content of fumonisins in KA105 ranged from 38.9～194.9 ppm,and that in KB024 ranged from 131.8～630.9 ppm.The results of variance analysis showed that genotype,environments and the interaction between environments and genotype reached extremely significant level,and the broad-sense heritability was 78.0%.A total of 17 FUM resistance QTLs were detected by combining genotype with two single environmental fumonisin content phenotypes.q FUM4-1,located in bin4.05,and the disease resistance allele came from KA105,which explained the phenotypic variation of 6.93-11.97%.q FUM3-1 was a stable QTL in multi-environment,located in bin3.03,and the disease resistance allele comes from KA105,which can explain 5.98-9.16%of phenotypic variation.Among them,q FER3-1 and q FUM3-1 were detected in bin3.03 segment,and there might be candidate genes for improving ear rot resistance and reducing toxin pollution.4.Using the offspring of residual heterozygous lines（RHLs）in RIL population,the effect verification and fine mapping of three stable QTLs were carried out.q FER1-1 could reduce the FER phenotype by 17.1%-25.8%.q FER1-1 was located between 21.5 and 23.1Mb on chromosome 1 by progeny-test mapping strategy,and the physical position interval was about 1.6 Mb.The effects of q FER4-2 and q FER3-1 were verified by the offspring of the RHLs,and it was clear that the alleles from KA105 at these two loci could significantly improve FER resistance.q FER4-2 was finally mapped between 15.6 and 17.5 MB on chromosome 4.5.The kernels of KA105 and KB024 were inoculated with Fusarium verticillioides 15days after pollination,and the samples of 2 hpi and 6 hpi were sequenced by transcriptome.Mock treatment and inoculation treatment of KA105 were compared,and 1669differentially expressed genes were screened.Mock treatment and inoculation treatment of KB024 were compared,1849 differentially expressed genes were screened.According to KEGG’s analysis results,differentially expressed genes mainly focus on disease defense pathway,vitamin metabolism pathway,asparagine metabolism,signal transduction mediated by plant hormones and so on.There were 38 genes in the effect segment of q FER1-1.According to the sequencing results of transcriptome,there are three differentially expressed genes（Zm00001d028074,Zm00001d028075,Zm00001d028086）.These genes may be key candidate genes for q FER1-1.In addition,the analysis of transcriptome data showed that there were significant differences in the expression of Zm00001d028074,Zm00001d028075 and Zm00001d039648 in the effect interval of q FER3-1,and these three genes might be the key candidate genes for q FER3-1.Zm00001d049155 and Zm00001d049166 are differentially expressed in the q FER4-2 effect interval,and these two genes may be important candidate genes for q FER4-2.6.Based on the Farm CPU model,168 natural groups were used to carry out genome-wide association study.As a result,a total of 34 significant SNPs were identified,and 13 SNPs were detected in the 21SY and 21PC environments.Based on the marker’s physical position（B73 Ref Gen＿v3）,we compared the significant SNPs detected by GWAS with linkage analysis results,and found that a total of 17 SNPs fell within the QTL interval.Among them,five superior alleles were mined.A total of 13 genes were found within 200kb interval of these co-localized SNPs.According to the results of transcriptome sequencing,Zm00001d039647 on chromosome 3 and Zm00001d049094 on chromosome 4were differentially expressed.7.In order to effectively improve the genetic gain of FER trait through genome-wide prediction,and thus provide reference for ear rot resistance breeding of Shaan A and Shaan B groups.In this study,the whole genome prediction model of FER phenotype of 183 RIL families was screened and optimized.Based on GBLUP model,the interaction between marker and environment was taken into account.The results showed that the cross-environment model performed better than the single-environment model,and the prediction accuracy of the marker-environment interaction model was the best.Six resistant families were screened based on genomic prediction,and then hybridized with four elite inbred lines respectively.It was found that compared with the control varieties（SD650）,the FER resistance of hybrid combinations PH6WC×RIL-50 and Zheng58×RIL-149 were better in both environments and the yield were not significantly affected.