Genome-wide Association Study And Identification Of Candidate Genes For Fiber Traits In G.hirsutum Races

G.hirsutum,a natural textile fiber and oil crop contributes 95%of the total world cotton production.Genome-wide association study(GWAS)has gained wide spread popularity and become a prominent alternative tool to traditional linkage mapping in understanding the functional roles of genetic variants for complex traits.Herein,we performed GWAS for five fiber quality traits:fiber elongation(FE),fiber length(FL),fiber micronaire(FM),fiber strength(FS)and fiber uniformity(FU)using high density 80 K Illumina Infinium single nucleotide polymorphism(SNP)array in diverse G.hirsutum races or also known as semi-wild cottons.The semi-wild cottons were evaluated at two locations for three years.Fiber quality traits among the semi-wild cottons displayed wide range of variations.The variations ranged from 2.00 to 8.90%with mean 6.06%and CV%17.32 for FE,19.61 to32.80 mm with mean 24.48 mm and CV%8.63 for FL,2.10 to 6.50 with mean 4.33 and CV%16.65 for FM,18.23 to 34.00 cN/tex with mean 24.24 cN/tex and CV%8.17 for FS,and 74.10 to 87.70%with mean 81.35%and CV%2.63 for FU.Comparison of correlation coefficients showed positive correlations between FL and FU,FL and FS,FU and FS,FM and FE,and FE and FS,whereas significant negative correlations were observed between FL and FM,FL and FE,FU and FE,FU and FM,and FM and FS.Moderate to high broad sense heritability was observed,with the highest for FL(0.95),while the lowest for FE(0.45).After filtering SNP data set,51841 high quality polymorphic SNPs,covering all the 26cotton chromosomes with an average marker density of 1 SNP in every 0.04 Mb were used for subsequent analyses.We observed discrepancies among the geographical origins of the semi-wild cottons and the grouping patterns.In the whole genome,LD decayed to r~2=0.1at about 2.72 Mb.In the At and Dt subgenomes LD decayed to r~2=0.1 at distances of approximately 2.88 Mb and 2.58 Mb,respectively.Using the mixed linear model(MLM)approach that takes into account population structure as fixed effects and pair-wise kinship(K)as random effects,in total,we detected 63significant SNPs,representing 23 for FE,15 for FL,3 for FM,14 for FS and 10 for FU on24 of the 26 cotton chromosomes,explaining 8.86-27.47%of phenotypic variation(PV).TM6231 showed significant association with both FE and FU that suggests pleiotropy of an underlying gene simultaneously affected both the traits.Of the total significant SNPs,17(26.15%)were repeatedly detected in two different environments.We also conducted RNA sequencing(RNA-seq)and identified 111 candidate genes between closely identified significant SNPs,of which,Gh_A08G0898 and Gh_A08G0890for FE,Gh_A07G0087 for FL,Gh_D12G0815 for FS,and Gh_A10G1564 for FU continuously showed higher expression in three wild cottons.The identified SNP loci and candidate genes provide important insights into the understanding of the genetic basis of the traits in G.hirsutum races.Further validation and manipulation of the predominantly expressed candidate genes will be also great interest.