Gene expression and expression quantitative trait loci analyses uncover natural variations underlying the improvement of important agronomic traits during modern maize breeding

SUMMARY Maize ( Zea mays L.) is a major staple crop worldwide, and during modern maize breeding, cultivars with increased tolerance to high‐density planting and higher yield per plant have contributed significantly to the increased yield per unit land area. Systematically identifying key agronomic traits and their associated genomic changes during modern maize breeding remains a significant challenge because of the complexity of genetic regulation and the interactions of the various agronomic traits, with most of them being controlled by numerous small‐effect quantitative trait loci (QTLs). Here, we performed phenotypic and gene expression analyses for a set of 137 elite inbred lines of maize from different breeding eras in China. We found four yield‐related traits are significantly improved during modern maize breeding. Through gene‐clustering analyses, we identified four groups of expressed genes with distinct trends of expression pattern change across the historical breeding eras. In combination with weighted gene co‐expression network analysis, we identified several candidate genes regulating various plant architecture‐ and yield‐related agronomic traits, such as ZmARF16 , ZmARF34 , ZmTCP40 , ZmPIN7 , ZmPYL10 , ZmJMJ10 , ZmARF1 , ZmSWEET15b , ZmGLN6 and Zm00001d019150 . Further, by combining expression quantitative trait loci (eQTLs) analyses, correlation coefficient analyses and population genetics, we identified a set of candidate genes that might have been under selection and contributed to the genetic improvement of various agronomic traits during modern maize breeding, including a number of known key regulators of plant architecture, flowering time and yield‐related traits, such as ZmPIF3.3 , ZAG1 , ZFL2 and ZmBES1. Lastly, we validated the functional variations in GL15 , ZmPHYB2 and ZmPYL10 that influence kernel row number, flowering time, plant height and ear height, respectively. Our results demonstrates the effectiveness of our combined approaches for uncovering key candidate regulatory genes and functional variation underlying the improvement of important agronomic traits during modern maize breeding, and provide a valuable genetic resource for the molecular breeding of maize cultivars with tolerance for high‐density planting.