Optimal plant density increases cotton yield potential through high light use efficiency supported by regulating gene expression at middle and upper canopy

Appropriate plant density could optimize light distribution within canopy and improve light use efficiency (LUE). However, most cotton studies have merely focused on whole canopy, and lacking in-depth exploration of the underlying mechanisms by which plant density regulates LUE as well as the gene expression. A two-year field experiment was conducted to assess the effect of plant density (low density, LD; moderate density, MD; high density, HD) on canopy light distribution, photosynthesis and its relevant gene expression, LUE, and cotton yield in upper, middle and lower canopy. The results showed that cotton responded to an increase in plant density by enhancing leaf area index at upper canopy, which subsequently led to higher fractional interception of photosynthetically active radiation (FIPAR) and a decrease of FIPAR at middle and lower canopy both in 2024 and 2025. Furthermore, HD caused a decline in photosynthetic rate ( A N ), chlorophyll content, leaf area, and specific leaf weight, which were significantly lower than those under MD. More importantly, LD and MD consistently achieved the maximum biomass and yield in 2024 and 2025, and biomass-based light use efficiency (LUE B ) in each canopy under MD was higher than that under LD and HD. Meanwhile, random forest model showed that FIPAR played a dominant role in regulating the variation in LUE B , with A N coming second, and these variations were supported by regulating photosynthetic genes expression at different canopy layers under different densities. The partial least squares model indicated that yield is mainly influenced by LUE B and leaf photosynthesis. Overall, MD optimized canopy light distribution, fully exploited the photosynthetic productivity at middle and upper canopy, thereby enhancing LUE and yield potential. • Divide canopy into three layers by height to observe their responses to density. • Combine physiological measurements with gene expression level. • Random forest model and partial least squares model were used. • Canopy light distribution is key to regulating light use efficiency.