Natural variation of GhSI7 increases seed index in cotton.

qSI07.1, a major QTL for seed index in cotton, was fine-mapped to a 17.45-kb region, and the candidate gene GhSI7 was verified in transgenic plants. Improving production to meet human needs is a vital objective in cotton breeding. The yield-related trait seed index is a complex quantitative trait, but few candidate genes for seed index have been characterized. Here, a major QTL for seed index qSI07.1 was fine-mapped to a 17.45-kb region by linkage analysis and substitutional mapping. Only GhSI7, encoding the transcriptional regulator STERILE APETALA, was contained in the candidate region. Association test and genetic analysis indicated that an 845-bp-deletion in its intron was responsible for the seed index variation. Origin analysis revealed that this variation was unique in Gossypium hirsutum and originated from race accessions. Overexpression of GhSI7 (haplotype 2) significantly increased the seed index and organ size in cotton plants. Our findings provided a diagnostic marker for breeding and selecting cotton varieties with high seed index, and laid a foundation for further studies to understand the molecular mechanism of cotton seed morphogenesis.