Genetic Dissection Of Fusarium Ear Rot Resistance In Maize And Genomic Prediction Analysis

Fusarium ear rot(FER)is a destructive maize fungal disease worldwide.In this study,three Fusarium ear rot(FER)is a destructive maize fungal disease worldwide,which affects the yield and quality of corn seriously.The Pathogens caused corn ear are different between regions and years,and it can infect ear alone or in combination,so it has complex pathogenesis.Some studies show that FER resistance exists in maize is controlled by minimal multigenes,and there is no immune material identified.It is difficult to clarify the genetic mechanism of resistance because of the undesirable genetic material for FER resistance,the complex traits and identification methods,so the study of QTL fine positioning and verification about FER resistance goes slowly,which hinders the breeding process about FER resistant exist in maize.The major QTLs for disease resistance were identificated using linkage populations and the SNPs associated with significant disease resistance were identificated using related populations are considered to be effective ways to analyze disease resistance mechanisms.GS for improving FER resistance in maize can polymerizate minimal multigenes fast and improve breeding efficiency.In this study,three tropical maize populations consisting of 874 inbred lines were used to perform association mapping and genomic prediction analyses of FER resistance,three biparental populations consisting of 462 double haploid lines were used to perform linkage mapping analyses.these results are used to analyze the genetic mechanism of FER resistance,explore how to use whole-genome selective breeding technology to improve FER resistance and improve breeding efficiency.The main results are as follows:(1)There are rich FER resistance variations in the 3 related populations and the 3 linkage populations,so it should be used for further genetic analyses: All the 1336 inbred lines were inoculated in four to six environments with a nail punch/sponge method and evaluated their FER performance.Broad phenotypic variation and high heritability for FER were observed,heritability of FER ranged from 0.53 to 0.91 in the individual environment analyses,and 0.54 to 0.84 in the combined environments analyses.Multiple environments analyses indicated that both the genotypic variance and genotype-by-environment interaction variance are significant at P value of 0.01.The results showed that FER resistance was controlled by genetic factors,and highly influenced by large genotype-by-environment interactions.Phenotypic data estimated from the combined environments analyses should be used for further genetic analyses.(2)The Significant SNPs of FER resistance in maize were detected in the three association analysis populations and all 874 inbred populations,but there were few overlapping significant SNPs: At the P-value threshold of 1 × 10-5,association mapping with mixed linear model(MLM)identified 12,5,22,19 single-nucleotide polymorphisms(SNPs)significantly associated with FER resistance in the populations of CML,DTMA AM,SYN＿DH,and all the 874 inbred lines,respectively.The average PVE value of these SNPs was 7.01%,5.01%,4.88%,and 1.60%,respectively.A few stable genomic regions conferring FER resistance were identified,which located in bins 3.04/05,7.02/04,9.00/01,9.04,9.06/07,and 10.03/04.The genomic regions in bins 9.00/01 and 9.04 are new.More SNPs associated with FER resistance were observed in the populations of SYN＿DH and all the 874 inbred lines.The SNPs associated with FER resistance in the SYN＿DH population has smaller P values and MAF values.Overlapping SNP associations across populations were rare.These results indicated that FER resistance is a genetic complex trait,it is controlled by minor QTL with small effects,and highly influenced by the genetic background of the populations studied.The statistical power and the mapping precision in association mapping could be improved by using the phenotypic data from combined environments analyses,bigger population size,and multiple-parental population with unique phenotypic and genotypic diversity.(3)4 QTL were detected using Linkage mapping analyses in three bi-parental populations: Linkage mapping analyses in three bi-parental populations detected four QTL at the LOD threshold of 3.0,the PVE of these QTL ranged from 11.84% to 21.19%.The physical position of the QTL on chromosomes of 2,3,4,and 5 was in 21.78 Mb to 26.27 Mb,1.87 Mb to 2.55 Mb、167.17 Mb to 175.21 Mb,and 42.70 Mb to 57.76 Mb.(4)The overlapping genomic regions were detected in joint linkage-association mapping analyses: The QTL on chromosomes 4 and 5 were consistent with the significant associated SNPs detected from association mapping analyses.These results implied that the genomic regions detected in joint linkage-association mapping analyses are stable across environments and populations.(5)Candidate genes related to FER resistance: 39 candidate genes related to FER resistance were detected in this study of MLM association analysis,but there was only one candidate gene was repeatedly detected among different populations,it is GRMZM2G146020,and also is Transcription factor Pos F21,it was detected in the SYN＿DH population and the population containing 874 copies of all inbred lines.In addition,the gene function annotations of a few candidate genes are related to disease resistance,such as GRMZM5G879570 in zone 7.03 、 GRMZM2G025997 in zone 8.03 、 GRMZM2G061314 in zone 9.07 and GRMZM2G127416 in zone 10.03,and so on.(6)The whole genome prediction using the Significant SNPs of FER resistance in maize were detected in the three association analysis populations and all 874 inbred populations by GLM method can improve the prediction accuracy: Genomic prediction produced moderate accuracies around 0.50 with the phenotypic data from combined environment analyses and the genome-wide markers,they are higher than the prediction accuracies estimated with the phenotypic data from individual environment analyses.The prediction accuracies estimated with the significantly associated SNPs detected from the association mapping are relatively high and ranged from 0.55 to 0.74 across populations.Moderate prediction accuracies,i.e.,0.45,were observed when the training and validation sets were closely related.Increasing the size of the training populations by including more genetic less related materials does not improve the prediction accuracy.The prediction accuracy could be improved by using the phenotypic data from combined environments analyses,larger size of training population including closer related materials,and incorporating SNP associations detected from association mapping.Genomic prediction and genomic selection are promising tools for improving FER resistance in maize,it can replace part of the phenotyping work and improve breeding efficiency.