DaVinci: Deconvolving latent Variables for integrated niche cluster identification
Tissue function and pathology are shaped by spatially organized cellular microenvironments (niches) that are remodeled across development, perturbations, and disease. Spatial omics enables genome-wide measurement of molecular state across multiple modalities in situ, yet most computational methods for spatial data remain limited to single-modality analyses, section-by-section processing, and offer limited biological interpretability. Here we present DaVinci, an interpretable framework for joint multi-omics, multi-section spatial analysis that defines niches as structured combinations of molecular programs and spatial context. DaVinci provides a flexible, interpretable solution for integrating multi-section, multi-modal spatial data, enabling comparative, mechanistic analysis of how niches are organized and rewired across biological contexts.
remotes::install_github("FunctionLab/DaVinci")
We provide detailed installation instructions for Linux, macOS and Windows.
We provide step-by-step tutorials for applying DaVinci to a variety of spatial omics datasets. Detailed documentation is available here.