| Isoflavones are a group of phenolic secondary metabolites which are relatively abundant in soybean and some other legumes.Total of 12 kinds of components are isolated from soybean seed,they can be grouped into three categories based on the difference of benzene ligands:daidzin group,genistin group and glycitin group.Soybean isoflavones have special health care functions on the human body,mainly in cancer prevention,menopausal syndrome remission,adjunctive treatment to cardiovascular disease and neurological disorders.Different components of isoflavones also have their own specific health care efficacies,such as daidzin group can adjust the level of male and female hormones in the human body,genistin group can effectively prevent cancers,and glycitin group can treat bone loose disease and has antioxidant activity,and so on.Dedicated variety breeding of total soybean isoflavone content and component content may bacome an important direction for modern soybean breeding of quality traits.The germplasm resources of plants are very abundant in China.The landraces were collected originally from farmers rather than the released cultivars from breeders,and therefore accumulated genetic variation through farmers' selection generation after generation.The genetic diversity of Chinese soybean germplasm provides rich source of material and gene bank for the modern breeding.The necessary prerequisites for the practical implementation of isoflavones breeding are understanding the genetic variation of isoflavone in germplasm resources,determining the genetic structure for controlling isoflavone content related traits,and accurately detecting the major gene loci related to isoflavones.The major process of plant breeding is to dig out the best recombinants from potential germplasm through optimal cross design and accurate progeny selection.Molecular marker technology has provided tools in genetic dissection of the germplasm for optimal cross design,from which marker-assisted progeny selection can be practiced.The GWAS(genome-wide association studies)with SNP(single nucleotide polymorphism)provides a way for genome-wide genetic dissection of breeding populations.But its bi-allelic property does not fit the populations with multiple alleles,and the long LD decay distance causes false positives in mapping QTL in inbreeding populations.A marker type suitable to GWAS for natural population and the resulted QTL-allele matrix of the population provides the compact dataset in prediction of the optimal crosses,which can fulfill the plant breeders'dream for incorporating molecular techniques into conventional breeding procedures.Based on previous researches,a representative population of 366 Chinese soybean landraces(CSLRP)derived from 24 provinces in six eco-regions was assessed,four isoflavone content traits were determined:TISF(total isoflavone content),TD(total daidzin group content),TG(total genistin group content)and TGL(total glycitin group content),and the phenotypic data obtained in 4 environments were used to analyse genetic variation characteristics of different eco-regions,and screen a batch of excellent variety resources for breeding.A set of SNPLDBs(SNP linkage disequilibrium blocks)derived from genome-wide SNPs was established according to the LD(linkage disequilibrium)level of CSLRP which had extensive genetic variation,and its genome-wide distribution characteristics was analysed.Through comparing the utility of the SNP and SNPLDB markers using in GWAS of isoflavone content traits,it suggested that SNPLDBs is more effective in GWAS.Using SNPLDB markers and "two steps”method to perform GWAS of isoflavone traits in CSLRP population,the genetic structures of isoflavone traits were identified in order to construct QTL-allele matrices,from which the difference of genetic structure among different eco-regions could be understand.The excellent alleles was digged out in order to provide parental information and predict crosses for subsequent isoflavone breeding.Finally,the commonality and specificity of genetic systems of 4 isoflavone traits were found and genetic information of TISF trait was used to assist the cross design of content traits.The main results are as follows:1 The isoflavone traits had great genetic variation in CSLRP,genetic differences existed among eco-regions,elite germplasms was selected in every eco-regions.TISF trait was showed wide phenotypic variation in four environment in CSLRP,the range of environmental average data were 423.8?6,579.6?g g-1,phenotype in individual environment was up to 8,514.0 ?g g-1,genetic variation coefficient GCV was 27.1%,and the broad sense heritability were high(92.5%).Combined analysis of variance showed that although the interaction between accessions and environments were significant,it were far less than that of genotypes,the influence of environment on the phenotypic variation was small.Genetic variation characteristics of TD,TG and TGL were similar to TISF,with variation range of 116.7?2,984.9 ?g g-1,319.4?4,040.0 ?g g-1 and 8.5-1,208.0 ?g g-1,respectively,The percentage of the three components to the total content were approximately 35.2%,51.9%and 2.9%,respectively,variation of TGL was most widely,while the highest content was in TG trait.The GCV were 34.5%,32.1%and 35.0%respectively,the genetic variation degree of all three component content were greater than the total isoflavone content.The broad sense heritability of TD,TG and TGL were also high,with 93.0%,93.4%and 92.9%respectively,it showed that the isoflavone traits had relatively high breeding selection efficiency.Four traits in accessions in six eco-regions exhibited different degrees of genetic variation.For TISF,the lowest phenotypic mean and the highest GCV existed in eco-?,indicating that the isoflavone content of accessions from eco-? was generally not high but its genetic variation was abundant,while the highest mean values and the lowest GCVs were in eco-? and eco-?.The phenotypic variation features of TD and TG were similar with TISF,the lowest phenotypic means were derived from eco-? and the highest ones were from eco-? region and eco-? also had the most extensive variation.The lowest mean TGL also came from eco-?,but the highest mean and the most extensive variation were from eco-?.In addition to the highest TG content was in eco-?,the highest content of the other three traits were all in eco-?.There were 6,7,10 and 9 elite gerrnplasms selected based on the criterias of TISF>6,000 ?g g-1,TD>2,400?g g-1,TG>3,200 ?g g-1 and TGL>900?g g-1,the total number was 24,from all the six eco-regions,the most material was from eco ? of 13,while the least was from the eco ? of only ? accession.The accessions with highest TISF,TD,TG and TGL were N23587(6,579.6 ?g g-1,eco-?),N24452(2,984.9?g g-1,eco-?),N24603(3,894.8?g g-1,eco-?)and N24296(1,053.4 ?g g-1,eco-?),respectively.There are many accessions with advantages of multiple traits,for example,the N23587 from Henan in eco ? had both high TD(2,605.1?g g-1),TG(3,251.2?g g-1)and TISF contents;the N23576 from Shanxi in eco ? was specific germplasm of TD(2,572.0 ?g g-1),TGL(933.4 ?g g-1)and TISF(6,497.4?g g-1).The selected specific germplasms could be as parents directly for breeding,and also could be used as intermediate material.2 Establish SNPLDB marker suitable for GWAS and subsequent breeding in germplasm.In CSLRP,a total of 116,769 SNPs on the whole genome,which then were grouped under D'(the standardized disequilibrium coefficient)>0.7 within the windows of 200 kb length into 29,121 SNP linkage disequilibrium blocks(SNPLDB),each composed of 1?128 SNPs with physical length 1 bp-200 kb and 2-12 haplotypes/alleles.The SNPLDB GWAS results showed that SNPLDBs have obvious advantages over SNPs in that:multiple alleles fitting the natural population;increase of mapping accuracy due to shortened LD decay distance;covering built-in additional non-associated SNPs;reducing hypothesis-testing times and GWAS Type I errors;and as a genome segment naturally close to the target QTL/genes.The SNPLDB markers can be effectively applied to GWAS in complex traits and subsequent molecular assisted breeding.3 The GWAS genetic constitution of isoflavone traits in CSLRP and regional differentiation.The genetic structure of 4 isoflavone target traits were dissecting in 366 accessions of CSLRP using "two steps" association analysis method.The genetic part of TISF explained 92.9%of phenotypic variation,the total phenotypic variance contribution rate of six "major QTLs"(contribution rate being more than 3.0%)was 31.2%,the total contribution rate of the remaining 38 "small QTLs"(contribution rate being less than 3.0%)was 41.0%,and the unmapped minor QTLs explained the remaining 20.7%of the variation.For the three components of TD,TG and TGL,phenotypic variation explanation rates of genetic part were also higher,which were 93.3%,93.6%and 93.6%,respectively;they had 8,7 and 5"major QTLs",respectively,the phenotypic variation explanation rates were slightly higher than that of TISF,39.4%,36.4%and 38.2%,respectively,that of TD was the highest one;the explanation rates of their "small QTLs" were lower than that of TISF,which were 34.3%,37.1%and 34.3%,respectively;the explanation rates of minor QTLs were 19.6%,20.1%and 21.1%,respectively,and that of TGL was the highest one.The 44 SNPLDB associated loci were obtained for TISF,distributed on 16 chromosomes,Gm03 has the most QTLs(5),and the locus qTisf-a-03-5 with the highest phenotypic rate(8.2%)was also in 29.2 Mb of this chromosome.The three component traits of TD,TG and TGL were obtained 50,42 and 37 QTLs,respectively.The genetic structure between component traits and the total content had both generality and discrepancy,the identical or similar loci between TD,TG,TGL and TISF were located at 14,12 and 12 chromosome intervals.The most TD loci were on Gm04(7),but the locus qTd-a-05-3 with the highest phenotypic rate(7.4%)located on Gm05 near 38.5 Mb.In addition to TISF,Gm03 also had the largest number of TG and TGL QTLs(both 6),the TG locus qTg-a-03-2 with the highest phenotypic rate(8.3%)was also in in 29.2 Mb of Gm03.The TGL locus qTgl-a-11-1 on 8.3 Mb of Gm11 was the only one QTL with phenotypic rate more than 10%(up to 16.8%).Each locus of TISF had 2-12 alleles with the mean of 4.5,the accumulative total alleles were 199,and for the three components of TD,TG and TGL,each QTL alleles were 2-10,2-10 and 2-11,with the mean of 4.2,4.7 and 4.6,respectively,the accumulative total alleles were 212,198 and 171,respectively.The difference among different traits was not significant for the allele number of single locus,but according to differences in QTL number,he accumulative total alleles of TD was the most and the least alleles was for TGL.According to the distribution of allele effect in CSLRP,four QTL-allele matrices were established for 4 target traits,respectively,as a compact genome dataset.The matrix size was QTL number plus 366 accessions.All the materials,superior or inferior,were found to contain both positive and negative alleles,indicating great potential of recombination for breakthrough segregants hidden in the population.Using the six eco-regions of Chinese soybean as units,the QTL-allele matrix in CSLRP population can be decomposed into six eco-regional QTL-allele matrices.The QTL structure in different eco-regions of all the target traits was inconsistent,which showed that there was some difference existed in the distribution of allelic variation in genetic material from different eco-regions,it was also the intrinsic reason for the phenotypic differences among eco-regions.Most of the differential loci had more allele types.Genetic diversity in six eco-regions were differences,the diversity of all the 4 traits in eco-? and IV were higher,while that of TISF and TGL in eco-V and of TD and TG ineco-VI was lower.The accessions from the same eco-regions were used for parental prediction,as well as that from inter eco-regions,a number of potential excellent cross combinations were obtained,which were 34,49,42 and 38 representatively for TISF,TD,TG and TGL.The potential of selecting offspring with higher phenotype values than parents existed for every traits.The ratio between the predicted value of progenies and the maximum phenotypic value of parents could up to 1.28,2.08,1.61 and 1.50 times for TISF,TD,TG and TGL,respectively,and TD had the highest potential in breeding.4 The genetic system commonality of isoflavone traits can help to realize comprehensive optimization breeding by design.A total of 173 QTLs of all the 4 target isoflavone traits were detected,distributed in all 20 chromosomes of the soybean genome.About three quarters of total loci in isoflavones genetic system could control two or more traits.A total of 37 chromosome intervals(about 5 Mb)contained QTLs belong to more than two traits,involving 130 assocaited QTLs,distributed on 17 chromosomes except Gm07,Gm10 and Gm12.Gm03 had the largest number of intervals(5).The QTLs of all four target traits were detected on 36.6-38.5 Mb interval of Gm05 which could be used as a focus segment for the study of genetic commonality of isoflavones.The remaining 43 loci reflected the specificity of genetic system which was also the genetic basis of differences between the component contents.Because of the generally significant correlation between different traits of isoflavones,the breeding of every trait would affect other traits.Combining with the total isoflavone content trait,it was expected to obtain progenies with increasing of both the absolute and relative contents of the three isoflavone content traits,total of 32,24 and 13 comprehensive optimization corsses were predicted based on materials in different eco-regions. |
PDF Full Text Request