spatial-preprocess

What problem does it solve?

This Skill automates the complex and repetitive process of loading, quality controlling, normalizing, and preparing spatial transcriptomics data for downstream analysis, ensuring consistency and reproducibility.

Core Features & Use Cases

• Multi-platform Data Loading: Supports various spatial transcriptomics formats (Visium, Xenium, MERFISH, Slide-seq, generic h5ad).
• Automated QC and Filtering: Applies standard metrics to filter low-quality cells and genes.
• Normalization and Feature Selection: Performs library-size normalization, log transformation, and highly variable gene selection.
• Dimensionality Reduction and Clustering: Computes PCA, UMAP, and Leiden clustering for exploratory analysis.
• Use Case: You have raw Visium data from a new experiment. Use this Skill to load it, perform QC, normalize counts, select important genes, and generate UMAP and Leiden clusters, all ready for further biological interpretation.

Quick Start

Run the spatial-preprocess skill on the provided Visium dataset located at '/path/to/visium_data' to generate a processed AnnData object and analysis report.