Evaluation Of Linkage Disequilibrium And Genome-wide Association Studies Of Four Agronomic Traits In Brassica Napus

In order to reduce the complexity of B.napus genome,we used double-digested restriction-site associated DNA sequencing(ddRAD-seq)technology to conduct high-throughput SNP genotyping for 189 B.napus inbred lines.The developed SNPs were used to comprehensively analyze the genetic diversity,linkage disequilibrium(LD)and haplotype block(HB)across the genome for the panel.And we used resequencing to develop ultra high-density SNPs/InDels for the panel to perform genome-wide association studies(GWASs)for flowering time(FT),oil content(OC),erucic acid content(EAC)and glucosinolate content(GLSC).Meanwhile the markers were used to detect lots of selective regions associated with agronomically important traits.The main results were as follows.1.Evaluation of LD and HB patterns in B.napus using SNPs developed by ddRAD-seqIn order to reduce the genome complexity,we sequenced the association panel containing 189 B.napus inbred lines using ddRAD-seq with 90 bp pair-end reads to enrich the sequences associated with restriction-site based on the Illumina high-throughput sequencing platform.The reads for each inbred line ranged from 113 to 633 millions,with an average of 268 millions.We genotyped 19,327 tags using the modified Universal Network Enabled Analysis Kit(UNEAK)pipeline which mainly contains three steps:tag extraction,network building and allelic tags discrimination.We randomly selected one SNP in each tag to represent it which was considered as one marker.With the LD-based mapping method,a total of 15,921 SNPs were assigned to a published genetic linkage map.Among these SNPs,11,326(71.1%)SNPs were uniquely mapped to the reference genome Darmor-bzh.There were 4,708 and 6,618 SNPs aligned to A and C subgenomes,respectively.We used 10,343 SNPs with MAF>0.05 to estimate the LD level for the panel.The distance of LD decay was 1,214 kb across the genome at the background level(r2=0.26),with the distances of LD decay being 405 kb and 2,111 kb in the A and C subgenomes,respectively.And differences were also observed for LD decay distances among all chromosomes.A total of 361 HBs with length>100 kb were identified,which covered 145 Mb of the entire genome.It suggests that the breeding with the goal of low GLSC has resulted in the selective regions arcoss the genome and subsequently the long HB formed.We found that the C subgenome had longer HBs than the A subgenome(112 Mb vs 33 Mb).This discrepancy may be correlated with the difference selection and contribution of two subgenomes to the breeding.Based on Structure analysis,the panel could be classified into two groups,P1 and P2,which represented the semi-winter type mainly from China and the spring type mainly from Europe and Canada.In P1 the PIC value of A subgenome was higher than that of C subgenome(0.255 vs 0.214),whilst in P2 the PIC-value of C subgenome was higher than that of A subgenome(0.275 vs 0.234).The hybridization between B.napus and Chinese B.rapa,which introduced the fragments of B.rapa chromosomes to B.napus,may lead to a substantial increase of genetic diversity of A subgenome in P1.We detected 241 common HBs with total length of 91.2 Mb for two groups.And 24 HBs with total length of 5.3 Mb had the frequency difference larger than 0.4 between two groups for the main haplotype of these common HBs,suggesting that some experienced distinct direction in two groups among all common selection regions.Meanwhile a large number of group-specific HBs were identified.P1-specific HBs in A and C subgenomes were 16 and 38 Mb,and P2-specific HBs in A and C subgenomes were 19 and 31 Mb,respectively.Maybe it resulted from adapting their different growth habitats or experiencing distinct selections in breeding programs for the inbred lines in different groups.2.GWASs of four agronomic traits and detection of selective regions in B.napusWe resequenced the 189 inbred lines and subsequently developed 3.82 million high-quality SNPs/InDels.And we performed GWASs for FT,OC,EAC and GLSC using these markers.Based on MLM,109 significant loci were detected to be associated with four traits.Among them,56 loci were repeatedly detected in at least two years,and some were commonly detected in the previous studies,suggesting that the results of GWAS in this study were reliability.Two peak signals for EAC and GLSC were landed to the corresponding target genes BnaC03g65980D(BnA8.FAEl)and BnaC03g65980D(BnA8.FAE1),respectively,suggesting the high mapping resolution of GWAS.In addition,for other 49 loci,the candidate genes can also found with the distance of 300 kb flanking to the peak signals.Based on all subsets regression,all peak signals of significant loci of each year can explain phenotypic variation ranging from 22.8%to 95.0%,with an average of 61.8%,for four agronomic traits.The difference for the explained phenotypic variations may be correlated with the different genetic basis of the traits.Lower GLSC and higher OC were usually regarded as the breeding goals for B.napus,therefore the alleles controlling favorable characters(the favorable alleles)of the traits experienced selection in the process of breeding.We observed that the inbred lines with lower GLSC and higher OC usually had more favorable alleles,and high correlation were found for GLSC(R2=0.88)and OC(R2=0.51)between the phenotypic values of traits and the number of favorable alleles,suggesting that we can make it straight forward for breeding through pyramiding multiple favorable alleles.Liho and Bronowski were previously reported as the materials of breeding which had the favorable alleles that reduced EAC and GLSC,respectively.We estimated the identity by descent(IBD)regions between them and the inbred lines and found that some regions covering the favorable alleles were under selection,such as the loci containing BnA8.FAE1和 BnA9.MYB28.The population in this study can be divided into two groups P1 and P2 which represented different ecotypes.And 111 selective regions with total length of 39.4 Mb were uncovered for Chinese semi-winter type oilseed rape through performing XP-CLR by regarding the genome of the spring type oilseed rape as the reference.Among these regions,three loci were overlapped to the significant loci associated with agronomic traits detected by GWAS.And 63 genes correlated with FT were observed in the selective regions,inferring that these genes may contribute to the differentiation of different ecotypes.Meanwhile gene ontology(GO)analysis reveals lots of significantly enrichment GO terms associated with important traits,such as the biosynthesis and storage of lipid,the defense to biotic and abiotic stress and the function of maintaining normal mitosis.To sum up,the present study detected lots of significant loci associated with agronomic traits and provided insight into the footprints of selective regions during breeding of oilseed rape,especially for Chinese semi-winter type.These results provide the foundations for genetic improvement of oilseed rape in the future.