Refining the functional dynamics of deoxynivalenol in wheat Fusarium head blight via MALDI-TOF-MSI and deep learning.

Wheat grains face widespread contamination by deoxynivalenol (DON), a major contaminating mycotoxin produced by Fusarium graminearum. However, DON's role in mediating F. graminearum infection in Fusarium head blight (FHB) remains contentious, largely due to challenges in capturing its holistic spatial accumulation dynamics. Here, we established a systematic MALDI-TOF-MSI-based workflow enabling untargeted, tissue-scale in situ visualization of DON distribution in wheat. By identifying 1,6-Diphenyl-1,3,5-hexatriene (DPH) as the optimal matrix, we achieved accurate DON detection in negative ion mode. Consequently, we settled the controversy regarding DON as a virulence factor in coleoptiles and mapped its translocation patterns in rachises. Ultimately, visual observation identified the germ and inner crease as low accumulation zones for DON in wheat grains, with its full distribution pattern being further revealed by deep learning-assisted 3D imaging. Visualizing DON in situ and refining its functional roles offer more valuable strategies for wheat processing and breeding.