Resequencing of cv CRI-12 family reveals haplotype block inheritance and recombination of agronomically important genes in artificial selection

Abstract

Although efforts have been taken to exploit diversity for yield and quality improvements, limited progress on using beneficial alleles in domesticated and undomesticated cotton varieties is limited. Given the complexity and limited amount of genomic information since the completion of four cotton genomes, characterizing significant variations and haplotype block inheritance under artificial selection has been challenging. Here we sequenced Gossypium hirsutum L. cv CRI-12 (the cotton variety with the largest acreage in China), its parental cultivars, and progeny cultivars, which were bred by the different institutes in China. In total, 3.3 million SNPs were identified and 118, 126 and 176 genes were remarkably correlated with Verticillium wilt, salinity and drought tolerance in CRI-12, respectively. Transcriptome-wide analyses of gene expression, and functional annotations, have provided support for the identification of genes tied to these tolerances. We totally discovered 58 116 haplotype blocks, among which 23 752 may be inherited and 1029 may be recombined under artificial selection. This survey of genetic diversity identified loci that may have been subject to artificial selection and documented the haplotype block inheritance and recombination, shedding light on the genetic mechanism of artificial selection and guiding breeding efforts for the genetic improvement of cotton.

Keywords: CRI-12 family; SNPs; artificial selection; haplotype block; inheritance and recombination.

Figure 1

Genetic relationships and genome‐wide SNP variation analysis in eight Upland cottons. (a) among these eight accessions, CRI‐12 was the focal variety. Xingtai6871 comes from China and Uganda4 comes from Uganda. Yumian2067, Lumianyan16, Yumian11, Jinmian33 and Jinmian20 were formed using CRI‐12 as one of the two parents. Twelve elite varieties were also included in the accompanying analyses. (b) Upland cotton (Gossypium hirsutum L.) contains 26 chromosomes, including A subgenome (A01‐13) and D subgenome (D01‐13). The figure showed the distribution of whole‐genome wide variation on each chromosome. The outermost layer (①) represents chromosome names and other tracks, from the second outside circle (②) to inside, show gene density (window size = 1 Mb, sliding window = 1 Mb, nonoverlapping), Uganda4 (③, green), Xingtai6871 (④, blue), CRI‐12 (⑤, red), Yumian2067 (⑥, brilliant green), Lumianyan16 (⑦, blue grey), Yumian11 (⑧, grey), Jinmian33 (⑨, brilliant blue) and Jinmian20 (⑩, orange). The highest value of SNP, Indel and SV were 6000, 800 and 100, respectively. The deeper the colour is, the bigger the density.

Figure 2

Whole‐genome selective sweep analysis for tolerance‐related genes in cotton. Genome‐wide distribution of F _st and ln (π_R/π_S) values (sliding window = 1 Mb). Orange and Green colours delineate 26 chromosomes in Gossypium hirsutum L. Magnified regions show two genes for each trait out of the three traits examined that showed strong selective signals.

Figure 3

The distribution of SNP haplotype blocks in the eight CRI‐12 family varieties. (a) SNP haplotype blocks identified in CRI‐12 family varieties. Each line represents a SNP haplotype block, with the CRI‐12 family varieties listed along the X axis. When a haplotype block matched that of CRI‐12, we coloured it red (A). If it matched a second, different haplotype block, we applied another colour (B) and so on with the subsequent haplotypes and their associated colours (C, D, E, F, G, H). Lines may be different from each other. We found 31 894 SNP haplotype blocks to be shared among the CRI‐12 family of cultivars. Only 100 haplotype blocks with significant difference were shown here. (b) 3‐D principal component analysis. The horizontal axis represents principal component one and the two vertical axes represent the other two principal components. Different colours represent different cotton varieties. We divided these eight varieties into two groups. (c) A Venn diagram showing unique and shared SNP haplotype blocks in CRI‐12 and its two parents.

Figure 4

A schematic diagram of haplotype blocks inheritance and recombination in CRI‐12 family. The haplotype blocks in Uganda4 and Xingtai6871 could be recombined into one haplotype in CRI‐12, and these recombined haplotype blocks in CRI‐12 were inherited to its progeny (Yumian2067, Lumianyan16, Yumian11, Jinmian33 and Jinmian20). Purple modules represent haplotype blocks shared by eight varieties. Blue modules represent haplotype blocks inherited only Uganda4 and black modules represent haplotype blocks inherited from Xingtai6871. The recombined modules represent haplotype blocks recombined from Uganda4 and Xingtai6871. Each type of haplotype blocks were listed once in the figure. Not all haplotypes are visualized here. The numbers 23 752, 1029, 1003 and 3420 represent sums of each type of haplotype blocks.

Figure 5

The distribution of haplotypes shared by eight cotton varieties in At and Dt subgenome. (a) represents the distribution of haplotypes shared by eight cotton varieties in At subgenome; (b) represents the distribution of haplotypes shared by eight cotton varieties in Dt subgenome. Red bars mean haplotype density, more red bars, higher the haplotype density.