Drought-tolerant Identification And Genome-wide Association Study Of Soybean Germplasm From Yangtze And Huaihe Valleys At Seedling And Adult Stage

Soybean[Glycine max(L.)Merr.]originated in China,is an important source of plant proteins and edible oils.Soybean occupies the important position in agricultural production as oil crops.The Yangtze and Huaihe River Valleys(YHRV)is an important area of soybean production in China.However,seasonal drought occurs frequently in this region,thus it is necessary to explore new drought tolerance materials and genes in the region.In the present study,210 accessions of soybean germplasm,including 159 new breeding lines developed from YHRV and 51 parental lines,were planted in plastic pots under well-watered(WW)and water-stressed(WS)conditions in both 2015 and 2016.The drought tolerance coefficients of shoot dry weight(SDW),plant height(PH),tap root length(TRL),and root dry weight(RDW)at soybean seedling stage and the drought tolerance coefficients of shoot dry weight(SDW),plant height(PH),node number on main stem(NMSS),single-plant grain number(SPGN),hundred grain weight(HGW)at adult stage were used as drought tolerant indices.Principle component analysis,the subordinate function value method and cluster analysis were jointly applied for comprehensive evaluation of drought tolerance for the tested genotypes.A total of 63,513 SNPs of soybean were used for genetic diversity analysis,linkage disequilibrium(LD)estimation and population structure analysis.The mixed linear model(PCA+K)was used for genome-wide association studies for nine traits and their drought tolerance coefficient(DTC)under soybean seedling stage and adult stage.The main results of this research were as follows:1.Identification of drought tolerance soybean germplasm from YHRV at seedling and adult StageAt the seedling stage,the results showed that the four traits were significantly decreased under the water-stressed condition compared with those under the well-watered condition.The averages for SDW,RDW,PH and TRL were reduced by 54%,42%,39%and 15%,respectively.The analysis of variance(ANOVA)showed that there had very significant differences between the two water treatments among the tested genotypes for all the traits;Moreover,the first order interactions of the three factors,i.e.genotype,water treatment and year,for the PH and RDW were very significant.There had significant positive correlation relationship for drought-tolerant coefficients between SDW and PH,SDW and RDW,and TRL and RDW,which reflected the inherent relations among the original drought tolerant indicators.The first three independent principal components extracted from principal component analysis accounted for 83.61%of the total variability and could adequately replace the original four indices which had overlapped information.To obtain the drought tolerance comprehensive evaluation D-value,the principal component scores for all the genotypes were then standardized through subordinate function and further average with corresponding eigenvalue as weight.All accessions were divided into five classes,i.e.high drought tolerance,drought tolerance,medium drought tolerance,drought susceptibility and high drought susceptibility,using cluster analysis based on D-value.Five accessions(IA2077(L160),YC4H/NN88-31//NN73-935(L077),M8108(L001),NN88-48/NN86-4(L078)and NN88-48/D76-1609(L079)were identified as high drought tolerance and 57 lines were identified as drought tolerance among the total samples.27 lines(32.53%out of 83 genotypes in this region)from southern YHRV and 19 lines(25.00%out of 76 genotypes in this region)from northern YHRV were identified as high drought tolerance or drought tolerance.At the adult stage,the five traits were significantly decreased under the water-stressed condition compared with those under the well-watered condition.The averages for SDW,PH,NMSS,SPGN and HGW were reduced by 39%,22%,6%,45%and 7%,respectively.ANOVA showed that there had significant differences between the two water treatments among the tested genotypes for all the traits;Moreover,the first and the second order interactions of the three factors,i.e.genotype,water treatment and year were extremely significant.There had significant positive correlation relationship for drought-tolerant coefficients between SDW and PH,SDW and NMSS,SDW and SPGN,SDW and HGW,which reflected the inherent relations among the original drought tolerant indicators.The first three independent principal components extracted from principal component analysis accounted for 80.42%of the total variability and could adequately replace the original five indices which had overlapped information.To obtain the drought tolerance comprehensive evaluation D-value,the principal component scores for all the genotypes were then standardized through subordinate function and further average with corresponding eigenvalue as weight.All accessions also were divided into five classes,using cluster analysis based on D-value.Two accessions(JL31(L168)and FD51(L039))were identified as high drought tolerance and 27 lines were identified as drought tolerance in whole sample.14 lines(20.89%out of 67 genotypes in this region)from northern YHRV and 7 lines(9.86%out of 71 genotypes in this region)in southern YHRV were identified as high drought tolerance or drought tolerance.2.Genetic diversity analysis,linkage disequilibrium estimation and population structure analysis of SNPThe average major allele frequency was 0.788,the average polymorphism information content(PIC)was 0.247,and the average nucleotide diversity was 0.301.The analysis of genetic diversity indicated that there was abundant genetic variation in soybean germplasm from YHRV.The decay of linkage disequilibrium(LD)was approximately 400 kb.A total of 197 lines were divided into three subpopulations according to neighbor-joining tree.In subpopulation I(including 76 accessions),most soybean accessions were derived from the main parent NN86-4.In subpopulation ?(including 35 accessions),most soybean accessions were descended from the parent TongZhouDou and Liu Feng.In subpopulation?(including 86 accessions),most soybean accessions were derived from the main parent NN86-48.The Principal component analysis showed consistent results with the neighbor-joining tree analysis.3.Genome-wide association studies of soybean germplasm from YHRV at seedling and adult stageAt the seedling trial:twenty-one SNPs were identified for 8 traitxtreatment combinations and their drought tolerance coefficient(WS:9 SNPs;WW:1 SNP;DTC:11?).One of these SNPs was co-localized to QTL regions with previous studies reported.A total of 17 drought tolerance candidate genes were annotated.Three of these genes(Glyma03g32500,Glyma03g32440 and Glyma03g32290)were related to PH,Glyma08g45560 was related to TRL.At the adult trial:a total of 32 SNPs were identified for 10 traitxtreatment combinations and their drought tolerance coefficient(WS:5 SNPs;WW:16 SNPs;DTC:11?).A total of 57 drought tolerance candidate genes were annotated.Most of drought tolerance genes were related to HGW.