STdeconvolve

STdeconvolve performs reference-free deconvolution of multi-cellular spatial transcriptomics datasets to recover cell-type transcriptional profiles and per-pixel cell-type proportions.

Key Features:

• Reference-free deconvolution: Deconvolves spatial transcriptomics data without requiring external single-cell transcriptomics references.
• Multi-cellular pixel resolution handling: Applies to data from multi-cellular-pixel spatial technologies including Spatial Transcriptomics, 10X Visium, DBiT-seq, and Slide-seq.
• Recovery of cell-type transcriptional profiles: Recovers transcriptional profiles specific to inferred cell types and estimates their proportional representation within each pixel.
• Performance relative to reference-based methods: Delivers performance comparable to reference-based methods when suitable single-cell references exist and can outperform them when such references are absent.

Scientific Applications:

• Developmental biology: Resolves spatial organization and cell-type composition during tissue development.
• Cancer research: Characterizes intratumoral cellular heterogeneity and spatially localized cell-type distributions.
• Immunology: Maps immune cell types and their spatial interactions within tissues.
• Tissue engineering: Assesses spatial cell-type composition and distribution in engineered tissues and constructs.

Methodology:

Performs reference-free computational deconvolution to infer cell-type-specific transcriptional profiles and estimate per-pixel cell-type proportions from multi-cellular spatial transcriptomics data.