Genomic signatures and candidate genes of lint yield and fibre quality improvement in Upland cotton in Xinjiang

Abstract

Xinjiang has been the largest and highest yield cotton production region not only in China, but also in the world. Improvements in Upland cotton cultivars in Xinjiang have occurred via pedigree selection and/or crossing of elite alleles from the former Soviet Union and other cotton producing regions of China. But it is unclear how genomic constitutions from foundation parents have been selected and inherited. Here, we deep-sequenced seven historic foundation parents, comprising four cultivars introduced from the former Soviet Union (108Ф, C1470, 611Б and KK1543) and three from United States and Africa (DPL15, STV2B and UGDM), and re-sequenced sixty-nine Xinjiang modern cultivars. Phylogenetic analysis of more than 2 million high-quality single nucleotide polymorphisms allowed their classification two groups, suggesting that Xinjiang Upland cotton cultivars were not only spawned from 108Ф, C1470, 611Б and KK1543, but also had a close kinship with DPL15, STV2B and UGDM. Notably, identity-by-descent (IBD) tracking demonstrated that the former Soviet Union cultivars have made a huge contribution to modern cultivar improvement in Xinjiang. A total of 156 selective sweeps were identified. Among them, apoptosis-antagonizing transcription factor gene (GhAATF1) and mitochondrial transcription termination factor family protein gene (GhmTERF1) were highly involved in the determination of lint percentage. Additionally, the auxin response factor gene (GhARF3) located in inherited IBD segments from 108Ф and 611Б was highly correlated with fibre quality. These results provide an insight into the genomics of artificial selection for improving cotton production and facilitate next-generation precision breeding of cotton and other crops.

Keywords: Gossypium hirsutum; Xinjiang cotton improvement; identity by descent; resequencing.

Figure 1

Single nucleotide polymorphisms (SNPs) annotation, phylogenetic tree, genetic structure and principal component analysis (PCA) of the 76 accessions. (a) Summary of SNP annotation. The upper pie chart shows the distribution of SNPs, while the lower pie chart shows the detailed distribution of SNPs in gene coding regions. (b) Phylogenetic analysis of 76 accessions. The neighbour‐joining tree was constructed using whole‐genome SNP data. The cotton samples were divided into clade 1 (orange) and clade 2 (green). (c) Population structure analysis of all accessions. The accessions were divided into 2 groups when K = 2. The y‐axis quantifies cluster membership, and the x‐axis represents the different accessions. (d) PCA plot of the first three components. The left plot shows PC1 and PC2, and the right shows PC1 and PC3. Group 1 and group 2 are orange and green, respectively.

Figure 2

Genome‐wide screen of artificial selection sweeps. Whole‐genome analysis of the selective sweeps through the comparison of parents and cultivars. The genome‐wide thresholds of 3.5721 and 0.09157 were defined by the top 5% of the π_parents/π_cultivars and Fst values. The arrows indicate the sweeps that contained 31 genes with nonsynonymous SNPs.

Figure 3

Identification of candidate genes related to trait development under artificial selection. (a) Selection signals from 3.7 to 4.2 Mb on chromosome A13. The upper line plot shows Fst in red, and the lower line plot shows π_parents/π_cultivars in blue. (b) Gene structure and polymorphisms of GhAATF1 and GhmTERF1. (c) Lint percentage (%) analyses of accessions with CC and TT genotypes of GhAATF1 (left). Lint percentage (%) analyses of accessions with AA and CC genotypes of GhmTERF1 (right). Centre line, median; box limits, upper and lower quartiles; and whiskers, 1.5x the interquartile range (**P < 0.01, two‐sided t‐test). (d) Transcriptomic patterns of GhAATF1 (upper) and GhmTERF1 (lower) in distinct tissues, based on the number of fragments per kilobase of the exon model per million mapped reads (FPKM) in a single experiment, including root, stem and leaf tissues during ovule and fibre development stages.

Figure 4

Distribution of modern cultivar identity‐by‐descents (IBDs) inherited from foundation parents. (a) Distribution of sixty‐nine modern cultivar IBDs inherited from seven foundation parents. The genetic constitutions of these sixty‐nine cultivars were identified from 108Ф, 611Б, DPL15, KK1543, C1470, STV2B and UGDM. Different colours represent different parents, and the corresponding colour of each parent is noted on the legend (top right). (b) The contribution of various genetic pools from different areas (the former Soviet Union and America or Africa) in modern cultivars.

Figure 5

Identification of candidate gene GhARF3. (a) Identification of GhARF3 gene structure and location in identity‐by‐descent (IBD) regions. The IBD regions inherited from 108Ф and 611Б on chr.A10, a nonsynonymous SNP (T‐to‐A) resulted in a change from isoleucine to asparagine. (b) Transcriptomic patterns of GhARF3 in distinct tissues, including root, stem and leaf, in ovule and fibre development stages based on FPKM in a single experiment. (c) Fibre length (upper) and fibre strength (lower) analyses of accessions with AA and TT genotypes of GhARF3. Centre line, median; box limits, upper and lower quartiles; and whiskers, 1.5× the interquartile range (*P < 0.05, **P < 0.01, two‐sided t‐test).