What problem does it solve?
This Skill helps you manage and analyze annotated data matrices by organizing large experimental measurements together with observation and feature metadata, reducing the pain of juggling separate files and mismatched indices.
Core Features & Use Cases
- AnnData data model guidance: Build and reason about the core structure (X, obs, var, layers, obsm/varm, obsp/varp, uns, raw) used across single-cell genomics and related workflows.
- I/O for common genomics formats: Load and save datasets (especially h5ad) and convert between practical formats, including backed mode for large files.
- Key operations for analysis workflows: Subset, concatenate batches/modalities, transform data, and apply best practices for memory efficiency and reproducibility.
- Integration into scverse ecosystem: Use AnnData as the backbone for Scanpy workflows and multimodal setups (e.g., via Muon), keeping embeddings and metadata consistent.
Quick Start
Ask the assistant: "Read the AnnData skill and tell me the safest way to load a large h5ad file in backed mode, subset by cell quality stored in obs, and then convert to memory for downstream analysis."
