| Most important agronomic and quality traits are quantitatively inherited,controlled by multiple genes and interactions between genes and environments.Linkage mapping and association mapping are two commonly used procedures in dissecting the genetic architecture of complex traits.Linkage mapping based on bi-parental populations often identifies only the polymorphic loci that are segregated between two parents.Genome-wide association study(GWAS)based on natural population can analyze the multiple alleles at a given locus synchronously.Linkage mapping and association mapping complement each other.Nested association mapping(NAM)population derived from a multiple-cross mating design sharing one common parent,which exploited multiple recombinant inbred lines(RIL)populations derived from crosses between a common founder line and a set of diverse founders.As NAM design takes advantage of multi-parental lines with multiple alleles per locus and large population size,and it has emerged as a promising multi-parental mapping design for complex trait genetic dissection.The present study constructed four soybean RIL populations(LM,ZM,MT and MW)sharing a common parent(M8206)as an NAM population(designated LZTWM),which was genotyped through the RAD-seq(restriction-site-associated DNA sequencing)and tested under multiple field environments between 2012 and 2014.The genetic architecture of the five parents(Linhe,Zhengyang,Tongshan,WSB and M8206)entirely covered the phenotypic variance of 623 lines in the LZTWM population.The LZTWM population was used for revealing the genetic structure conferring flowering date,100-seed weight,oil content and protein content of the five parents through QTL mapping of the NAM soybean population and to explore the recombination potentials among the five parents.The major research results are summarized below.1 RTM-GWAS was the efficient QTL mapping procedure for LZTWM populationTo choose a suitable mapping procedure for the soybean LZTWM population,this study compared the QTL results of flowering date trait using five mapping procedure among procedures with available software.The restricted two-stage multi-locus GWAS(RTM-GWAS)with SNP linkage disequilibrium block(SNPLDB)as multi-allele genomic markers performed the best among the five available software mapping procedures and identified the greatest number of quantitative trait loci(QTL)(139)and alleles(496)on 20 chromosomes covering almost all the QTL detected by the other four mapping procedures(29 QTL in total),except one additional QTL by CIM and MCIM,one additional QTL by JICIM and one additional QTL by JICIM and MLM-GWAS.It provided the highest genetic contribution without overflowing and missing heritability problems(81.7%genetic contribution vs.h2=97.6%)and matched the population property with 2-5 alleles per locus.From the above results,RTM-GWAS was the best-identified mapping procedure for the soybean LZTWM population,and used to detect the genome-wide QTL-allele system of flowering date,100-seed weight,oil content and protein content in the following text.2 Genetic architecture of flowering date in LZTWM population and breeding potentials among five parentsIn the LZTWM population,a total of 55,936 SNPs were organized into 6,137 SNPLDBs to fit the property of multiple alleles/haplotypes per locus for the five parents of the NAM population.The two-stage association analysis procedure RTM-GWAS with SNPLDB as multi-allele markers was used for dissecting the genetic architecture of flowering date in LZTWM population.The combined analysis of variance under multi-environments showed that the hereditability of flowering date was estimated as 97.6%.The 139 loci on 20 chromosomes were identified with the QTL number from 2(Chr.09)to 12(Chr.06,Chr.10 and Chr.13),and the genetic contribution per locus ranged from 0.03%to 18.27%,with a sum of 81.7%of the total phenotypic variance.Of those,the nine large-contribution major loci each contributed more than 1.0%phenotypic variance,for a total of 53.2%phenotypic variance,and the remaining 130 small-contribution major loci each contributed less than 1.0%phenotypic variance for a total of 28.5%phenotypic variance.With the heritability being 97.6%,the remaining heritability(15.9%)must reflect the unmapped minor QTL.The QTL × environment interaction and experimental error explained 3.4%of the phenotypic variance.The 139 QTL with 496 alleles of 623 lines and that of five parents were organized into QTL-allele matrices of the NAM population and the five parents,showing the corresponding flowering date genetic architecture,and 91.53%of the allele effects were between-2 d?+2 d.All lines and parents comprised both positive and negative alleles,implying a great potential of recombination for early and late flowering date improvement.From the detected QTL system,126 candidate genes were annotated as a flowering date candidate gene system accounting for 79.44%of the phenotypic variance,which indicated that the trait is a typical quantitative trait involving a complex of biological processes rather than only a handful of major genes.3 Genetic architecture of 100-seed weight in LZTWM population and breeding potentials among five parentsThe two-stage association analysis procedure RTM-GWAS with 6,137 SNPLDB as multi-allele markers was used for dissecting the genetic architecture of 100-seed weight in LZTWM population.The combined analysis of variance under multi-environments showed that the hereditability of 100-seed weight was estimated as 97.7%.The 119 loci on 20 chromosomes were identified with the QTL number from 1(Chr.14 and Chr.17)to 13(Chr.06),and the genetic contribution per locus ranged from 0.06%to 7.06%,with a sum of 84.6%of the total phenotypic variance.Of those,the 18 large-contribution major loci each contributed more than 1.0%phenotypic variance,for a total of 49.4%phenotypic variance,and the remaining 101 small-contribution major loci each contributed less than 1.0%phenotypic variance for a total of 35.2%phenotypic variance.With the heritability being 97.7%,the remaining heritability(13.1%)must reflect the unmapped minor QTL.The QTL × environment interaction and experimental error explained 2.3%of the phenotypic variance.The 119 QTL with 407 alleles of 623 lines and that of five parents were organized into QTL-allele matrices of the NAM population and the five parents,showing the corresponding 100-seed weight genetic architecture,and 93.86%of the allele effects were between-2 g?+2 g.All lines and parents comprised both positive and negative alleles,implying a great potential of recombination for small and big grain soybean improvement.From the detected QTL system,92 candidate genes were annotated as a 100-seed weight candidate gene system accounting for 65.82%of the phenotypic variance,which indicated that the trait is a complex trait involving a complex of biological processes.4 Genetic architecture of oil content in LZTWM population and breeding potentials among five parentsThe two-stage association analysis procedure RTM-GWAS with 6,137 SNPLDB as multi-allele markers was used for dissecting the genetic architecture of oil content in LZTWM population.The combined analysis of variance under multi-environments showed that the hereditability of oil content was estimated as 90.2%.The 128 loci on 20 chromosomes were identified with the QTL number from 2(Chr.12)to 17(Chr.17),and the genetic contribution per locus ranged from 0.02%to 6.16%,with a sum of 58.1%of the total phenotypic variance.Of those,the 13 large-contribution major loci each contributed more than 1.0%phenotypic variance,for a total of 29.6%phenotypic variance,and the remaining 115 small-contribution major loci each contributed less than 1.0%phenotypic variance for a total of 28,5%phenotypic variance.With the heritability being 97.6%,the remaining heritability(32.1%)must reflect the unmapped minor QTL.The QTL × environment interaction and experimental error explained 9.8%of the phenotypic variance.The 128 QTL with 422 alleles of 623 lines and that of five parents were organized into QTL-allele matrices of the NAM population and the five parents,showing the corresponding oil content genetic architecture,and 91.53%of the allele effects were between-0.5%?+0.5%.All lines and parents comprised both positive and negative alleles,implying a great potential of recombination for oil content improvement.From the detected QTL system,106 candidate genes were annotated as a oil content candidate gene system accounting for 51.57%of the phenotypic variance,which indicated that the trait is a complex trait involving a complex of biological processes.5 Genetic architecture of protein content in LZTWM population and breeding potentials among five parentsThe two-stage association analysis procedure RTM-GWAS with 6,137 SNPLDB as multi-allele markers was used for dissecting the genetic architecture of protein content in LZTWM population.The combined analysis of variance under multi-environments showed that the hereditability of protein content was estimated as 85.0%.The 112 loci on 20 chromosomes were identified with the QTL number from 1(Chr.09,Chr.12 and Chr.17)to 14(Chr.06),and the genetic contribution per locus ranged from 0.02%to 3.94%,with a sum of 46.9%of the total phenotypic variance.Of those,the 10 large-contribution major loci each contributed more than 1.0%phenotypic variance,for a total of 19.4%phenotypic variance,and the remaining 102 small-contribution major loci each contributed less than 1.0%phenotypic variance for a total of 27.4%phenotypic variance.With the heritability being 85.0%,the remaining heritability(38.1%)must reflect the unmapped minor QTL.The QTL x environment interaction and experimental error explained 15.0%of the phenotypic variance.The 112 QTL with 371 alleles of 623 lines and that of five parents were organized into QTL-allele matrices of the NAM population and the five parents,showing the corresponding protein content genetic architecture,and 90.84%of the allele effects were between-1.0%?+1.0%.All lines and parents comprised both positive and negative alleles,implying a great potential of recombination for protein content improvement.From the detected QTL system,94 candidate genes were annotated as a protein content candidate gene system accounting for 42.46%of the phenotypic variance,which indicated that the trait is a complex trait involving a complex of biological processes. |
PDF Full Text Request