What problem does it solve? Spatial transcriptomics platforms like Visium and Xenium measure gene expression at spots or cells without known cell type labels. This Skill performs cell type deconvolution with RCTD (spacexr) using scRNA-seq references, so you can assign cell types, detect doublets, and integrate results into Seurat objects. ## Core Features & Use Cases - Reference Preparation: Downsample scRNA-seq references to 300-500 cells per type and build RCTD Reference objects with cleaned cell type names. - Single and Multi-Sample RCTD: Run deconvolution in doublet or full mode with parallel processing, then normalize weights and extract per-cell predictions. - Seurat Integration & Reporting: Add predicted cell types and weights to Seurat metadata, export cell groups for Xenium Explorer, and generate FGCZ-compliant R Markdown reports. - Use Case: You have a Xenium breast cancer dataset and a published scRNA-seq reference. Use this Skill to validate gene overlap, run RCTD across all samples, filter low-confidence cells, and produce spatial cell type maps with summary statistics. ## Quick Start Ask the agent to annotate your Xenium or Visium Seurat object with RCTD using your scRNA-seq reference and generate the spatial cell type plots and summary report.
