Cell-GPS (COSTE Windows Application): Cophenetic Spatial Topology Embedding for multiscale tissue architecture analysis in spatial omics

This record provides Cell-GPS, the standalone Windows application of COSTE (Cophenetic Spatial Topology Embedding), a computational framework for quantifying multiscale tissue architecture from spatial omics data. COSTE captures global spatial organization beyond local neighborhood analysis by embedding cell-type or transcript spatial relationships into a hierarchical metric space and deriving the Spatial Separation Score (SSS) to measure spatial separation between populations.

As described in the associated manuscript, COSTE is parameter-free, density-independent, and can operate at both the cell-type level and the single-transcript level without requiring cell segmentation. It has been demonstrated on spatial transcriptomics datasets (e.g., pulmonary fibrosis and triple-negative breast cancer), where it delineates tissue structures, identifies spatially defined subpopulations, and reveals disease-related architectural patterns.

Cell-GPS is a GUI no-code workflow that can be used directly without programming experience and is designed to support interactive spatial omics analysis.

File: CellGPS.exe (~445.77 MB) Checksum (MD5): 4624fda684bbbc1c21586c1e34f01289

For questions/support, please open an issue on the project repository or contact the corresponding author.