PlantST: Unraveling plant tissue heterogeneity and developmental trajectories via multimodal graph contrastive...

Spatial transcriptomics (ST) enables the analysis of spatial heterogeneity and developmental progression within intact tissues. In plants, however, its application is constrained by complex tissue architectures, such as the concentric vascular organization of stems and the irregular geometries of floral organs, which challenge computational methods developed primarily for animal tissues. Here, we present PlantST, a computational framework tailored to plant spatial omics. By integrating spatial and transcriptional information, PlantST identifies complex spatial domains and reconstructs continuous developmental trajectories consistent with physical growth patterns. We evaluated PlantST in established developmental systems, including secondary vascular growth in Populus stem nodes and morphogenesis in orchid floral organs. Compared with existing methods, PlantST achieved higher biological resolution, resolved narrow functional boundaries such as the cambium-xylem-phloem continuum, and more accurately reconstructed radial and floral developmental gradients. These results show that PlantST captures biologically meaningful spatial organization and developmental dynamics in structurally complex plant tissues. PlantST therefore provides a useful analytical framework for plant ST datasets with complex spatial organization and offers a methodological basis for future high-resolution plant spatial atlas construction.