Upgraded genome assembly reveals the key importance of TEs, in-dels, SNPs and point mutations during the evolution of spinnable cotton fibres

It is generally believed that the ancestors of Gossypium arboreum (A2 genome) or G. herbaceum (A1 genome) together with that of G. raimondii (D5 genome) provided the genetic basis for the modern, cultivated allotetraploid cotton species G. hirsutum (AtDt genome). Here we upgraded the A2 genome assembly by integrating SMRT sequencing and Hi-C technologies. We further resequenced 230 G. arboreum and 13 G. herbaceum accessions (average 6x depth) in China to generate a map of genome variation including ~18 million SNPs and ~2 million indels. Independent analysis suggested that Chinese G. arboreum originated in South China and was subsequently introduced to the Yangtze River and Yellow River regions. GWAS for 11 agronomically-important traits in G. arboreum identified a total of 98 significant peak associations. We found that a burst of transposable elements (TEs) contributed significantly to the 2-fold increase in the size of the G. arboreum genome, when compared to that of the D genome. Comparative genomics analysis suggested that an Ile/Val substitution in conserved catalytic motif DDVAE of 4 gossypol biosynthesis CDN1 genes renders the ability to produce gossypols only in cotton plants. Sequence alignments of the promoters of ACO gene, which is key to ethylene biosynthesis in cotton, revealed that specific deletions in both G. arboreum and in A-subgenome of G. hirsutum caused the loss of several cis-elements, including MYB binding sites, and may be responsible for the inactivation of ACO gene transcription in G. arboreum ovules. This reduced ACO gene expression correlated to the short-fibered phenotype in G. arboreum, whereas very high levels of ACO transcriptions, with a resultant ethylene burst, in G. raimondii ovules seems to force an early fiber senescence phenotype. A Cysteine/Arginine substitution of GaKASIII seems to have conferred significant alteration of the fatty acid composition (C16:0 and C16:1) in cotton seed. Also, gain of Fusarium wilt disease resistance in YZ and YR accessions is associated with GaGSTF9, whose expression is highly inducible upon fungal inoculation. The GaGSTF9-silenced cotton line shows sensitivity to fungal inoculation, whereas overexpression of cotton GaGSTF9 in A. thaliana displays resistance to Fusarium wilt disease compared to wild-type.