Identification photosynthesis-related genes in peanut (Arachis hypogaea L.) by hyperspectral phenomics, GWAS and RNA-seq

Peanut (Arachis hypogaea L.) contributes importantly to the global food supply, and its yield potential is closely linked to leaf photosynthetic efficiency. However, conventional measurement of leaf net-photosynthetic-rate (Leaf Pn) is time- and labor-intensive, limiting its application to large breeding populations. We established a phenomics-driven framework integrating hyperspectral reflectance and genome-wide association studies for high-throughput evaluation and genetic dissection of Leaf Pn. Machine-learning models were trained on preprocessed spectral data from 1,200 leaf samples collected in four environments. Random forest performed best after Savitzky-Golay (S-G) smoothing and first-derivative transformation (R2 = 0.81, root mean square error [RMSE] = 1.50), outperforming the vegetation index-based model (R2 = 0.66). Applying this model to a panel of 173 peanut genotypes identified two quantitative trait loci, qPn5.1 and qPn13.1, encompassing seven candidate genes. Integrated analyses prioritized Arahy.K5L3ZW, which encodes protein kinase 2B. Silencing this gene increased Leaf Pn, whereas transient overexpression decreased it, indicating that Arahy.K5L3ZW negatively regulates Leaf Pn. This framework enables high-throughput genetic dissection of seedling photosynthetic capacity and identifies a candidate target for improving photosynthetic efficiency in peanut.