CellsFromSpace (CFS)

CellsFromSpace (CFS) deconvolves and annotates spatially distributed omics data without requiring a single-cell reference dataset by deploying independent component analysis (ICA) to decompose spatial transcriptomics data into interpretable components associated with distinct cell types or activities.

Key Features:

• Reference-free deconvolution: Analyzes spatial transcriptomics data without relying on a single-cell reference dataset.
• Independent component analysis (ICA): Decomposes spatial transcriptomics data into interpretable components associated with distinct cell types or activities.
• Noise and artifact reduction: Enables noise or artifact reduction through component selection.
• Subset analysis: Allows subset analysis of cell types of interest through component selection.
• Cross-technology compatibility: Processes datasets from Visium, Slide-seq, MERSCOPE, and CosMX technologies.
• Multisample processing: Handles complex, even multisample datasets with speed and scalability.
• Quantitative deconvolution: Quantitatively deconvolutes spatial transcriptomics datasets.

Scientific Applications:

• Spatial cell type identification: Identifies spatially distributed cells as well as rare diffuse cells in real-world samples.
• Cell crosstalk analysis: Supports analysis of cell crosstalk in healthy and diseased organs by capturing transcriptomic profiles within spatial contexts.
• Component annotation and interpretation: Annotates and interprets components based on spatial distribution and contributor genes.
• Full downstream analysis: Performs complete downstream analysis of spatial transcriptomics data.

Methodology:

CellsFromSpace applies independent component analysis (ICA) to decompose spatial transcriptomics data into interpretable components.