| Soybean is cultivated all over the world widely as a important soure of protein and oil for humans.As the origin of soybean,there are abundant germplasm resources in China.In the past introduction and dissemination process,ecological environments adaption,the soybean materials in different regions have formed a serious genetic bottleneck,which led to different genetic background in different soybean materials planted in different regions.Nowdays,the soybean reference genome Williams82 is derived from only one US soybean material which is derived from multiple soybean materials with complex genomic information,and more and more studies have suggested that a single genome cannot represent the genetic information of a specie,including structural variations,as in the Arabidopsis,rice and maize,more and more genomes of different materials are assembled comprehensively.The genetic linkage map construction based on Chinese soybean landrace materials is benefit to soybean genomic structure analysis and the accuracy improvement of soybean molecular breeding in China.In addition,most of agronomic and seed quality traits are quantitative traits with complex genetic bases which are controlled by multiple genes.With the explosive increase of high-throughput sequencing data,the molecular markers have become more and more plenty,so plant biologists and breeders have moved their sights from a few functional genes to the comprehensive understanding of genetic system in the quantitative traits gradually.Linkage mapping methods are popular in the QTL mapping of bi-parental population,which are all based on inteval mapping method.The computational burden is happened with the increase in molecular marker number,and these methods can only detect a few major QTLs which contribute a part of phenotypic genetic variation or clustered QTLs which cause the overflowing genetic contribution.Association mapping as another method of QTL mapping,has the advantage of high precision which is usually used in the natural population.The population structure is clear in a bi-parental population,and the application of association mapping method in a bi-parental population is benefit for the comprehensive genetic analysis of the quantitative trait,and understanding the controlling of the genetic structure of the trait and make it possible to exploring their comprehensive genetic system.The flowering date trait is an important ecological trait for soybean,which is related with maturity period,yeild and so on.Isoflavone is a class of secondary metabolites which are unique to legume species such as soybean.Because of their health care effects on human body and variouse relationships with the disease resistance of plants and environment interaction,seed isoflavone content will be an important target trait for the selection and breeding of special soybean variety in the modern soybean quality breeding programs.The genetic analysis of flowering date and seed isoflavone related traits and the gene regulatory networks in the seed isoflavone triats provide a necessary theoretical basis for the development in the soybean breeding.In this study,427 recombinant inbred lines(RILs)deriving from Kefeng No.l and NN1138-2 which were selected from pure lines of Chinese soybean landrace materials were merged as a population named as "NJRIKY".From the different genotype detection pipelines,including PCR(Polymerase Chain Reaction)and WGR(whole genome sequencing),different types of molecular marker,including SSRs(simple sequence repeats),BINs(recombinant breakpoint intervals)and SNPLDBs(single nucleotide polymorphism linkage disequilibrium blocks),were applied to reveal the characteristics of genome structure in NJRIKY population and the variation of recombination rate on the whole soybean genome.Different type of markers and QTL methods were applied to select the best maker and QTL procedure for QTL system detection.The QTL mapping for different traits was applied to detect the genetic foundation of flowering date,seed isoflavone related traits,explore the candidate gene and gene network in the genetic system.The main results are as follows:1 Characterization of different genetic linkage maps with different marker ty pes and genomic structure changes of Williams82 from the Chinese landrace genomeIn the NIRIKY RIL population which derived from the Kefeng No.1 from a landrace material in the Huang-Huai-Hai Double Cropping Spring and Summer Planting Varietal Eco-Region and the NN1138-2 from a landrace material in the Middle and Lower Changjiang Valley Double Cropping Spring and Summer Planting Varietal Eco-Region,three different types of molecular markers were applied for genetic linkage map construction,including 834 SSRs,4,737 BINs and 3,683 SNPLDBs.SSR is non-genomic marker,the physical positions of 628 SSR markers were determined by sequence alignment with the reference genome Williams82.BIN and SNPLDB markers as genomic markers with known physical position were used for genetic map construction.In the construction process,some markers were removed to ensure the accuracy of genetic map,the final genetic linkage maps were constructed by 4,737 BINs and 3,310 SNPLDBs.The rank correlation of marker order between genetic map and physical map was low(0.51:0.75),and the correlation in heterochromatin is much lower than that in euchromatin(0.58:0.15 vs.0.91:0.45)in the SSR-map and SNPLDB-map,respectively.The results indicate that the SSR-map with non-genomic markers and SNPLDB-map with genomic marker were shown significant differences between genetic map and physical map in the marker order.After the three genetic maps information combination and detail local comparison on different chromosomes,183 structural variations were identified on 20 chromosomes which were accounted for 137 Mb of genome region with four tpyes of variation,including 82 inversion regions,32 intrachromosommal translocations,85 complex rearrangements and 13 interchromosomal translocations.192 reported QTLs associated with seed oil,seed protein,disease resistance,plant type and yield related traits were located in these rearrange genome regions,which suggests these genomic structural variation regions play an important role in different traits.Although the total length of the detected structural variations contributed only～14%of whole soybean genome,but it also indicates that there are some differences in the genomes among different soybean materials under various regions.A single reference genome of a single material can no longer meet the need of molecular studies and breeding programs on soybean in the future.In order to apply the genomic information in the future better,we need to analyze more genome information of different soybean materials in various regions of the world for the soybean molecular breeding program.As a strict self-pollination crop,soybean has significantly different recombination rate in the different chromosomal regions.The SNPLDB-map was used to analysis the recombination rate of different regions on the whole genome.The linear relationship between the genetic map and physical map was calculated and shown that the average recombination rate in euchromatin(6.99 cM/Mb)was more higher than that in heterochromatin(0.36 cM/Mb),the recombination rate of different chromosomes were different,which indicate the recombination rate in the soybean genome is significantly different in different regions.From the refinement of chromatin regions of soybean genome determined by SNPLDB-map,it was found that the euchromatin regions were increased 31.2 Mb compared to the public genetic linkage map,while the maximum change on chromosome 15 increased by 6.9 Mb,the minimum change on chromosome 17 increase by 1.1 Mb.Among many previous reported QTLs,seven cqQTL(confirmed QTL)associated with seed oil and protein,100-seed weight and cyst nemotodes resistance trait and El affected growth period were located in heterochromatin with low recombination rate,so this study suggests the genetic linkage map can be applied in the molecular marker breeding design which will increase the accuracy of progeny prediction.2 The RTM-GWAS procedure with SNPLDB markers is efficient for QTL mapping in bi-parental populationIn order to analysis the genetic foundation of quantitative trait in the bi-parental population comprehensively,the flowering date trait as an example was applied for QTL mapping in three types of markers(SSRs,BINs and SNPLDBs)and three different QTL mapping procedures(CIM,MLM-GWAS and RTM-GWAS).The RTM-GWAS with SNPLDB showed more powerful(86 QTLs),efficient and reasonable(89.92%phenotypic variation contribution)than the others(23 QTLs in CIM and 14 QTLs in MLM-GWAS),including five known E series QTLs,J locus and Dtl locus.These results indicate that the innovations in RTM-GWAS procedure,including molecular markers(SNPLDB marker)and QTL model(two-stage multi-locus mapping model),make it is suitable for QTL mapping for bi-parental population and improve the QTL detection power.In addtion,SNPLDB marker has several advantages for QTL mapping in a bi-parental population as follows:it contains the complete genome sequence information and distributes on the whole genome evenly,including in the SNP scarce region;The block size of SNPLDB marker is natural depending on the genome sequence LD property,not on a fixed sliding window;As an extension,its characteristic of multi-allelic variation makes itself having the potential in detecting QTLs with multiple alleles in different kinds of populations,including various bi-parental populations with different parents shared a same set of SNPLDB markers,although it is not required in the represent bi-parental population.From the above results,the RTM-GWAS procedure was the best efficient mapping procedure in NJRIKY population,so it was used to detect the genome-wide QTL system of flowering date and isoflavone related traits in the following text.3 Genetic architecture of flowering date trait in NJRIKY populationUsing the 3,683 SNPLDB markers under RTM-GWAS procedure,a total of 86 QTLs were detected on 20 chromosomes,accounting for 89.92%of the phenotypic variation(PV),of which 5 QTLs were large-contrubution(LC)major QTLs PVE>3%),81 were small-contribution(SC)major QTLs,unmapped QTLs contributed the remaining 8.28%of PV.29 QTLs were novel in NJRIKY population.A comprehensive analysis of the candidate gene system revealed that two LC major QTLs on Gm11 and Gm12(PVE=9.92%and 10.89%)were located in the repeat region after the second replication event of soybean genome～13 Myr ago.The two paralogous candidate genes,Glymallg15580 and Glyma12g07861 were related with photoperiodism and can be regarded as important new genes in the genetic system of flowering date in future.In addition,24 of 54 candidate genes were predicted into three protein protein interaction(PPI)networks,including a large PPI network which contained E1,E2,E9 and J from two LC major QTLs and two SC major QTLs and other 14 candidate genes.These results indicate one gene is a member of the whole candidate gene system,these genes work together to affect the trait expression.4 Genetic architecture of seed isoflavone component traits in NJRIKY populationA total of 86,88,56 and 49 QTLs were detected on 20,20,20 and 16 chromosomes,which accounting for 78.1%,81.4%,65.0%and 62.6%PV in the total isoflavone content(TISF),total daidzein content(TD),total genistin content(TG)and total glycitin content(TGL)traits,respectively.Among them,45,37,25 and 26 QTLs of TISF,TD,TG and TGL respectively were reported previously,the remaining 41,51,31 and 23 QTLs were new QTLs in NJRIKY population.The phenotypic correlation relationship among these traits indicates that there are commonality and difference in the genetic foundation of these traits.The correlation among TD,TG and TISF were higher(0.83,0.93 and 0.93),while the TGL with TD,TG or TISF was lower(0.35,0.46 and 0.38,respectively).These results suggest that the genetic variation and synthesis between TD and TG are similar,and they are main components of soybean seed isoflavone content,while the synthesis pathway of TGL may has a certain specificity.Further comparing the positions of QTLs,55,29 and 31 QTLs were specific QTLs in TD,TG and TGL,respectively,other 98 QTLs affected the synthesis of isoflavones in other ways,altogether 213 QTLs were detected in NJRIKY population.From the gene expression data of two parents and the correlations between the phenotype and gene expression,a total of 161 candidate genes were detected,no candidated genes were detected in the remaining 52 QTLs.Among them,36,25 and 23 candidate genes were unique to TD,TG and TGL,respectively,and the total PVEs were 18.6%,15.8%and 58.4%,respectively.The prediction of protein and protein interaction network among 161 candidate genes indicates that 55 candidate genes were involved in the protein protein interaction(PPI)n.etworks,especially 10 known key enzyme proteins and 42 new candidate genes were involved into one PPI network.The above results indicate that these candidate genes in the gene system are not independent,the interaction and cooperation among these genes completed the expression of the seed isoflavone traits,5 The consistency between the expression patterns and the pathway position of candidate genes in the genetic system of seed isoflavone traitsThe development of various omics techniques has enabled a large number of quantitative and qualitative assays for rapidly quantify and improve the development of basic systematic biology.The combination between omic data and forward genetic results would accelerate to explore the important genes associated with target trait.According to the clustering results of 161 candidate genes and 44 enzyme genes expression data,the expression patterns of these candidate genes and enzyme genes in the seed development tissues were different,93 candidate genes and 21 enzyme genes are positive regulatory genes,and the remaining 68 candidate genes and 23 enzyme genes are negative regulatory genes,indicating that these negative regulatory genes need to be suppressed in the molecular breeding process,the overexpression of positive regulatory genes will increased the seed isoflavone content.In order to insight the position of each candidate gene in the isoflavone synthesis pathway,the weighted gene co-expression network(WGCNA)analysis was applied for 161 candidate genes and 44 reported enzyme genes.These genes were separated into two co-expression modules.A total of 76 genes(including 16 enzyme genes)were grouped together as model 1,and almost all enzyme genes involved with isoflavone synthesis directly were in this module,which indicated that these candidate genes may be directly,while 93 genes(including 17 enzyme genes)were involved in model 2,and the enzyme genes in this model were locate in the upstream of isoflavone synthesis pathway,which suggested that these genes may regulate the main part of phenylpropanoid metabolism pathway.Throughout the relationship between the expression patterns of candidate genes and their position in the isoflavone synthesis pathway,58 of the 76 genes in module 1(76.3%)expressed positively,and among the 93 genes in module 2,58 genes(62.4%)was negatively regulated,and the remaining 35 genes(37.6%)were positively regulated.From the above results,the positively regulated expression pattern in the module 1 were directly involved in the synthesis of isoflavones predominately,while the negative regulation genes in the module 2 were involved in the upstream pathway of isoflavone synthesis predominately.In conclusion,the candidate genes which are co-expressed with the enzyme genes involved in the isoflavone synthesis directly are most positive regulated during the seed development stages,these genes are more directly beneficial for the accumulation of seed isoflavone content. |
