Dissecting the molecular basis of spike traits by integrating gene regulatory network and genetic variation in wheat.

Spike architecture influences both grain weight and grain number per spike, which are the two major components of grain yield in bread wheat (Triticum aestivum L.). However, the complex wheat genome and the influence of various environmental factors pose challenges in mapping the causal genes affecting the spike traits. Here, we systematically identify genes involved in spike trait formation by integrating the information of genomic variation and gene regulatory network (GRN) controlling young spike development in wheat. We obtained 170 loci that are responsible for variations of spike length (SL), spikelet number per spike (SNS) and grain number per spike (GNS) through genome wide association study (GWAS) and meta-QTL analysis. The GRNs for young inflorescence at double ridge stage (DRS) and floret primordium stage (FPS) in which spikelet meristem and floret meristem are predominant, respectively, were constructed by integrating transcriptome, histone modifications, chromatin accessibility, eQTL, and protein-protein interactome. Strikingly, we identified 169 hub genes, whose polymorphisms are significantly associated with variation of spike trait from the network and they are located in 76 of the 170 QTL regions. In addition, the functions of TaZF-B1, VRT-B2 and TaSPL15-A/D in establishing wheat spike architecture were verified. The study provides valuable molecular resources to understand spike traits and the approach of combining genetic analysis and developmental regulatory network is robust to dissect complex trait.