Anndata

If you're working with single-cell genomics data, this is the data structure you'll be using whether you realize it or not. It's the foundation under scanpy, scvi-tools, and most of the scverse ecosystem. Think of it as a specialized container that bundles your expression matrix with all the metadata about cells and genes, plus support for sparse matrices and lazy loading when files get huge. The skill covers creating and manipulating these objects, reading h5ad and zarr files, concatenating datasets from multiple batches, and staying memory-efficient with backed mode. One thing to note is that for actual 10X file reading you'll want the scanpy skill instead, since those convenience functions live there now.

Install to Claude Code

npx -y skills add k-dense-ai/scientific-agent-skills --skill anndata --agent claude-code