SPCS

SPCS applies two-factor smoothing that combines transcriptome and spatial relationships to reduce noise and improve interpretability of spatial transcriptomics (ST) data.

Key Features:

• Two-factor smoothing: Integrates both transcriptome and spatial relationships rather than relying solely on single-factor transcriptomic smoothing.
• k-nearest neighbor technique: Employs a k-nearest neighbor approach to incorporate local transcriptomic similarity and spatial proximity into smoothing.
• Noise and drop-out mitigation: Reduces technical noise and drop-out events present in ST datasets.
• Improved separability and partition accuracy: Enhances separability of molecular patterns and accuracy of spatial partitioning compared to one-factor methods used in scRNA-seq analyses.
• Biological interpretability: Increases the clarity of tissue-specific gene expression patterns for downstream biological interpretation.
• Demonstrated datasets: Applied and evaluated on pancreatic ductal adenocarcinoma (PDAC), dorsolateral prefrontal cortex (DLPFC), and simulated high-grade serous ovarian cancer (HGSOC) ST datasets.

Scientific Applications:

• Spatial transcriptomics denoising: Reducing technical noise and drop-outs in ST data to recover underlying gene expression signals.
• Tissue region identification: Improving partitioning and separability of spatial domains within tissue sections.
• Biological interpretation of spatial patterns: Enhancing detection and interpretation of tissue-specific gene expression patterns in PDAC, DLPFC, and HGSOC contexts.
• Method comparison and benchmarking: Serving as a comparative approach against one-factor smoothing methods common in scRNA-seq analyses.

Methodology:

SPCS employs a k-nearest neighbor technique to perform two-factor smoothing by jointly considering gene expression patterns and spatial proximity.