spatialGE

spatialGE quantifies and visualizes spatial heterogeneity in tumor microenvironments using spatial transcriptomics to relate gene expression patterns to clinical data.

Key Features:

• Gene expression surfaces: Generates visualizations and quantifications of spatial gene expression patterns across tissue sections using gene expression surfaces.
• Spatial heterogeneity statistics: Computes statistical measures of spatial heterogeneity that can be correlated with clinical information.
• Spot-level cell deconvolution: Performs deconvolution of cell types at individual spatial transcriptomics spots to estimate cellular composition.
• Spatially informed clustering: Implements clustering analyses that incorporate spatial information to identify distinct microenvironmental niches.
• Integrated data object: Stores raw spatial transcriptomics data and resulting analyses together within a single data object.

Scientific Applications:

• Cancer prognosis and biomarker discovery: Relates spatial heterogeneity metrics to patient outcomes to support prognosis and biomarker identification.
• Targeted therapy development: Identifies spatial microenvironmental niches and patterns associated with therapy resistance or response.
• Clinical integration: Correlates spatial transcriptomics-derived statistics with clinical data to study disease progression and treatment efficacy.
• Demonstration and benchmarking: Applied to published spatial transcriptomics datasets such as the Thrane study for method demonstration and benchmarking.

Methodology:

Leverages spatially resolved transcriptomics data to generate gene expression surfaces and statistical analyses, performs spot-level cell deconvolution, computes spatial heterogeneity statistics, and applies spatially informed clustering.