Delin Li, Zhan Li, Shiyu Guo, Qi Wang, Xinkang Feng, Hao Zhang, James C. Schnable, YingHui Li, LiJuan Qiu

Molecular Plant; 2026; IF: 32.70

DOI: 10.1016/j.molp.2026.09.011

Abstract

Protein abundance is the ultimate functional layer that translates genetic information into phenotype, yet its genetic regulation remains largely unexplored in plants, leaving a critical layer of post-transcriptional regulation unresolved in the genotype-to-phenotype map. Here, we report a population-scale pQTL map and a direct proteome-wide association study (PWAS) framework in soybean. Using 200 diverse accessions, a pan-protein reference database, and data-independent acquisition mass spectrometry, we quantified 19,700 proteins encoded by 17,341 genes. Genome-wide association analysis identified 3,190 pQTL for 2,550 proteins encoded by 2,280 genes, of which 60.2% were cis-pQTL. Comparison with transcriptome-wide eQTL revealed 444 proteins with pQTL but without eQTL and 634 genes with cis-pQTL but without cis-eQTL, indicating extensive post-transcriptional regulation. We further validated the protein-folding factor HSP90, in which the non-synonymous variant p.Ala478Ser presumably altered protein abundance by affecting protein stability rather than transcript level. We then applied direct PWAS to qualitative and quantitative traits and precisely pinpointed known causal genes, including T for pubescence color and Rhg1 and NSF07 for soybean cyst nematode resistance. Notably, the DNA, RNA, and protein layers captured unique and complementary associations. Thus, our data demonstrate that protein abundance is a direct, heritable mediator of agronomic traits, and that pQTL mapping and PWAS provide a protein-level framework for dissecting complex traits in plants.