Rapid Mapping And Evaluation Of Breeding Potential Of Functional Genes In Soybean Mutants

The mutant derived from mutagenesis are widely used in gene function research and variety breeding of soybean[Glycine max（L.）Merr.].At present,there are three aspects that need to be improved.Firstly,for the classical method of causal mutation mapping,such as bulked segregent analysis（BSA）,the mutant material needs to undergo self-crossing or hybridization for several generations,which greatly prolonged the cycle of causal mutation mapping.Secondly,there is a lack of prior reference databases for research on the functional mechanisms of soybean mutant genes.Thirdly,the causal mutations of mutants often lead to the growth and development defects in soybean,thus they are not suitable for direct application in soybean breeding.Therefore,this study established a new method for rapid mapping of causal mutations of mutants,which greatly shortened the cloning cycle of functional genes of soybean mutants.The gene co-expression network of soybean was constructed,which provided a direct clue for the functional mechanism analysis of target genes and prior information for subsequent molecular experiment verification.It was proved that genomic selection based on haplotype was effective in high generation breeding population of soybean,which laid a foundation for introducing elite alleles of target functional genes into genomic selection model in the following study.The main progress of this study is as follows:（1）Development and application of a new method for rapid mapping causal mutations of soybean mutants.In this study,a BSA method called M2-seq was developed,which only needed to pool samples from soybean mutant population in M₂ generation for the whole genome sequencing and bioinformatic analysis to map the causal mutation.The method mainly comprised the two improvements:1)Mutagenic variations could be screened out directly without genotype of wild type plants.2)The absolute value of delta SNP index was applied to draw the mapping curve,which effectively eliminated the interference of phase repulsion between mutant alleles of M₂ population to mutation mapping.Among the 10 independent soybean M₂ populations involved in this study,the M2-seq method successfully mapped all genomic regions harboring candidate causal mutations.The average mapping interval of 10 populations was 7.7 Mb.In addition,candidate causal mutations were successfully identified in 9 of the 10 populations.We selected two of the mutants to analyze their phenotypic characteristics and the function of their candidate causal mutant genes,and proved that the candidate causal mutations were related to their phenotypic abnormalities.These results demonstrated the effectiveness of M2-seq for mutagenic gene cloning.Compared with the reported methods,M2-seq shortens the time required for causal mutation mapping by at least one generation and improves the speed of functional mutations mining for mutants.（2）Construction of soybean gene weight co-expression network and regulation mechanism analysis of causal mutagenic genes.A gene weight co-expression network covering gene expression characteristics of seven major organs in soybean was constructed using 531 transcriptome data from a public database.The network consisted of 41 co-expression modules covering 22,760genes.All the 9 candidate causal mutant genes identified by M2-seq could be detected in specific modules of the network.The functional mechanisms of the causal mutant genes from two mutants were analyzed.The analysis showed that the causal mutant gene Glyma.08G193200（Gm RHD3）from mutant Mut01 might be regulated by transcription factor TGA.Gm RHD3 might not only regulate phenotypic morphology such as epidermis hair density of soybean,but also affect growth,development and stress response of soybean by regulating the expression of transcription factors such as Gm ERF5.The causal mutant gene Glyma.10G251500（Gm THI1）from mutant Mut07,as a thiamine synthase,occupied a hub node in the regulatory network.The analysis also revealed that chaperonins and Gm THI1 might be co-involved in the regulation of stress resistance of soybean.These results have demonstrated that the gene regulatory network constructed based on transcriptome data from a public database could effectively predict the regulatory mechanisms of the causal mutagenic genes of mutants and provide clues for further molecular experiments.（3）The validity of haplotype based genomic selection in breeding population of soybean was proved.In this study,we didn’t find the 9 candidate causal mutations from mutants in the natural population,but the corresponding mutagenic genes were abundant in other variations and haplotypes in the natural population.Through genome wide association study,it was found that there was no significant association between 9 mutagenic genes and the two target traits of soybean oil content and protein content.Nevertheless,the feasibility of using haplotypes for genomic selection was demonstrated in this study,which would lay a foundation for introducing elite alleles of other mutagenic functional genes from natural populations into the genomic selection model in the future.Haplotype based and SNP based genomic selection analyses were performed in a F₄ population composed of 688 progenies and a breeding population composed of 953 high generation varieties.The results showed that:1)The haplotype diversity in the soybean F₄ population was low,and the highest prediction accuracy based on SNPs for oil content and protein content reached 0.39 and 0.54,respectively.The prediction accuracy based on haplotypes had no significant advantage;2)The haplotype diversity in the soybean breeding population was high,and the highest prediction accuracy based on the short haplotypes consisting of five consecutive SNPs for oil content and protein content was 0.50 and 0.48,respectively,which was superior to prediction accuracy based on the haplotypes consisting of 25 consecutive SNPs（0.44for oil content and 0.42 for protein content）;3)On the premise that the training population size exceeded 60%of the whole population,the prediction accuracy based on haplotypes composed of five consecutive SNPs was superior to the prediction accuracy based on SNPs（oil content 0.50 vs 0.45;protein content 0.48 vs 0.44）.Therefore,haplotype based genomic selection will have greater application potential in high generation breeding populations of soybean.