anndata

Store and manipulate annotated single-cell data matrices with observation and variable metadata.

4|Updated Mar 2, 2026
One-click install
npx skills add https://github.com/shushuzn/Rairos --skill anndata-shushuzn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: anndata
Source: https://github.com/shushuzn/Rairos/tree/main/skills/anndata
Command: npx skills add https://github.com/shushuzn/Rairos --skill anndata-shushuzn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

AnnData provides a standard, efficient container for annotated experimental matrices so you can store measurements together with observation and feature metadata and reliably move data through single-cell workflows.

Core Features & Use Cases

  • Unified annotated data model: Represents the main matrix X plus obs (observations), var (variables/features), and additional fields like layers, embeddings (obsm), pairwise graphs (obsp), and unstructured metadata (uns).
  • Fast I/O for research datasets: Load and save common genomics formats (especially .h5ad and zarr) and convert between them for downstream tools.
  • Practical single-cell operations: Subset, filter, concatenate across batches/modalities, transpose axes when needed, and follow memory-friendly practices for large or sparse data.
  • Ecosystem interoperability: Acts as the foundational object for scverse tools such as Scanpy, and supports integrations for multimodal data and model training workflows.

Quick Start

Use the anndata skill to load a single-cell dataset from an .h5ad file named data.h5ad into an AnnData object for further analysis.

Frequently Asked Questions about anndata

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I load and work with annotated single-cell matrices in .h5ad format?

You can load .h5ad files into an AnnData object to store the main data matrix alongside observation (obs) and variable (var) metadata for single-cell analysis. This provides a unified container to reliably move data through workflows.

What is the best way to concatenate single-cell data across multiple batches or modalities?

AnnData provides built-in operations to subset, filter, and concatenate annotated data matrices across different batches or modalities. This allows you to merge datasets while maintaining consistent indices and preserving auxiliary layers.

Can I use backed mode or sparse matrices to handle large single-cell datasets efficiently?

Yes, AnnData supports efficient large-data handling through backed mode and memory-aware practices using sparse matrices and categoricals. This enables you to process experimental matrices that exceed available memory.

Does anndata work with scverse tools like Scanpy for downstream analysis?

AnnData acts as the foundational object for scverse tools such as Scanpy and supports integrations for multimodal data. You can pass the annotated matrix directly into these tools for model training and downstream workflows.

How do I manage embeddings and auxiliary layers when processing annotated matrices?

AnnData lets you manage embeddings in obsm, pairwise graphs in obsp, and unstructured metadata in uns alongside the main matrix. You can also transpose axes when needed to align data for specific operations.

What are the limitations when concatenating .h5ad files with inconsistent indices?

Concatenation and subsetting require a well-formed AnnData object with consistent indices to function correctly. If indices are misaligned or malformed, operations will fail or produce invalid merged matrices.