Association Analysis And Genome-Wide Selection Of Maize Kernel Dry-Down Related Traits

Mechanized harvesting has become a common means of increasing maize production in many countries to meet the growing global food demand.For efficient mechanical harvesting,low grain moisture content at harvest time is essential.From the perspective of maize breeding,maize varieties are required to have a faster dry-down rate after entering the physiological maturity stage,Drydown rate(DR),which refers to the reduction in grain moisture content after the plants enter physiological maturity,is one of the main factors affecting the amount of moisture in the kernels.Dry-down rate is estimated using kernel moisture content at physiological maturity and at harvest time;however,measuring kernel water content at physiological maturity,which is sometimes referred as kernel water content at black layer formation(BWC),is time consuming and resource demanding.At present,the genetic basis of maize dry-down rate after maturity is still unclear.Therefore,inferring BWC from other correlated and easier to measure traits could improve the efficiency of breeding efforts for dry down related traits.This study used a natural population of 397 maize inbred lines with extensive polymorphisms,sequenced their genotypes,and measured grain moisture at harvest,grain moisture at black layer appearance,dry-down rate,and flowering time for phenotypic investigations.The genetic basis of grain dry-down related traits was analyzed by genome-wide association analysis,genome-wide selection methods,and a multi-trait genomic best linear unbiased prediction model was used to estimate the grain moisture content and harvest time when the black layer appeared,when maize entered the mature stage.Genetic correlations between grain water content and traits at flowering time was estimated,and multitrait Genomic best linear unbiased prediction analysis was carried out using a multi-trait Genomic best linear unbiased prediction model and a variety of cross validation methods to explore the most suitable genome-wide breeding methods for maize dry-down related traits predictive models and breeding strategies.The results of the study are as follows:1.Heritability,phenotype analysis and genome-wide association analysis.When the black layer appeared,the grain moisture content of maize entering the mature stage,the grain moisture content at harvest and the traits and flowering time showed moderate to high heritability(0.22-069).Moreover,the phenotypic variation among different varieties was extensive,and the phenotypes were basically in a normal distribution with good continuity,which laid a phenotypic basis for the subsequent analysis.In the genome-wide association analysis study of 3 dry-down related traits in multiple location and years,a total of 79 significant SNP loci were found,and 11 candidate genes in four categories were identified,of which the first category was related to coding regulation of plant growth.candidate genes for development-related proteins;the second category is candidate genes related to plant hormones;the third category is candidate genes related to the cell wall structure of plant structural substances;the fourth category is candidate genes related to drought response,combined with RNA-seq database comparison should focus on the expression of Zm00001d038988,Zm00001d021316 and Zm00001d047860 genes in the corresponding traits in the follow-up research.2.Genome-wide selection analysis.In a genome-wide selection analysis of grain water content at black layer appearance,grain water content at harvest,and dry-down rate,four models of Bayesian ridge regression,Bayes B,Bayes C,and Bayes LASSO were used to perform genome-wide selection for target traits analysis,multi-environment,multi-year and multi-model comparative analysis was carried out on the obtained prediction accuracy,combined with multi-year multi-location and multi-model full-gene selection,the average prediction accuracy of grain water content when black layer appeared was 0.3,and the average prediction accuracy of grain water content when the black layer appeared was 0.3.The average prediction accuracy of grain moisture content was 0.19,and the average prediction accuracy of dehydration rate was 0.16.3.Multi-traits genomic prediction.Combined with the genetic correlation of the target traits,in the Multi-traits genomic prediction analysis of the target traits,four cross-validation methods were evaluated to explore the most suitable genome-wide prediction models for maize dry-down related traits,as well as breeding strategies and genetic basis analysis,in most cross-validation methods,the use of the multi-trait genome-wide prediction model can significantly improve the prediction accuracy(compared with the standard 5-fold cross-validation compared with the single-trait genome-wide test model The prediction accuracy is improved from 0.3 to 0.56).In addition,even if the target trait was not measured in all field trials,including historical data related to the genetics of the target trait in the breeding database in the model also improved the prediction accuracy(average prediction accuracy was better than 0.7).The obtained results clearly demonstrate that the use of genetically related traits such as grain water content at harvest and flowering time in a multi-trait genomic prediction model can be compared with phenotypic measurements of grain water content at the appearance of black layer.The prediction accuracy of target traits is significantly improved,and the prediction accuracy of target traits with high cost of phenotypic measurement and high resource consumption,such as grain water content when black layer appears,can be reduced,and resource consumption and manpower and material resources can be saved.In multi-trait genomic breeding strategies that measure phenotypic traits in only a subset of families in a population,high prediction accuracy is obtained,but the need for phenotypic measurements for genetically related traits can limit increased selection intensity to some extent.Including phenotypic information of genetically related traits in genomic prediction models using multiple traits can greatly improve prediction accuracy and improve breeding efficiency.Molecularly assisted breeding strategies for dehydration-related traits are explored to help breeders select the most efficient breeding strategy for grain dry-down related traits.