DOMINO: diffusion-optimised graph learning identifies domain structures with enhanced accuracy and scalability

Spatial transcriptomics enables molecular profiling in situ, and identifying spatial domains facilitates our understanding of functional compartments of tissues. However, most existing domain calling methods do not scale to rapidly increasing data sizes and focus only on local structure while missing the global view of the tissue. Here, we present DOMINO, a diffusion-optimised contrastive learning framework for spatial domain detection. DOMINO utilises graph diffusion convolution to propagate information beyond immediate neighbours, and jointly optimises local and global graph structure via contrastive learning. This novel framework yields biologically interpretable domains with clearer boundaries and scales to large datasets, outperforming state-of-the-art methods across healthy and malignant benchmark datasets. Applying DOMINO to a newly generated spatial dataset from endometriosis-associated ovarian cancers, which failed to be processed by existing domain detection methods due to its large size, we found that DOMINO reveals a shared epithelial continuum along a proliferation associated axis spanning FOXJ1 high proliferative to CD55 high non-proliferative domains, consistently coupled to coordinated stromal remodelling. DOMINO also identified subtype specific spatial domains, including spatially restricted WNT beta catenin activation in endometriod and discrete NDRG1 high metabolic domains in clear cell ovarian cancers, features that are not recovered by standard expression based clustering workflows.