Data Readiness
Knowledge BaseLife-sciences R&D
93 profiles ·
practical standards guide

Reference architecture · AnnData 0.13.2 · MuData 0.3.10 · CELLxGENE schema 7.1.0 · SpatialData 0.8.0

Single-cell and Spatial Data Ecosystem

Maintained by scverse · Chan Zuckerberg Initiative

What it helps you do

AnnData ecosystem supports annotated single-cell matrices, multimodal containers, repository metadata requirements, ontology-bound annotations, and spatial images, labels, points, shapes, tables, and coordinate systems.

  • Omics
  • Bioinformatics
  • Imaging
  • AI / ML
PlanAcquireHarmonizeExchangeLearn + reuse

01

Where it fits and where it does not

Use these four checks before committing implementation time.

Use it when
An implementable analysis and publication stack for single-cell, multi-omic, and spatial-omics data when the exact component, encoding, schema, and ontology releases can be pinned.
Limits
This is a fast-moving software ecosystem rather than a single governed standard. Container validity does not guarantee consistent preprocessing, cell annotation, biological comparability, or lossless conversion to Seurat or other ecosystems.
Best for
Omics and Bioinformatics and Imaging and AI / ML teams working across Acquire → Harmonize → Exchange → Learn + reuse.
Maturity
ScalingUsable now, but adoption or tooling is still developing. Pilot the exact stack first.

02

See it in the workflow

This view shows the input, the change the standard introduces, and the resulting output.

  1. InputWhat starts

    Omics and Bioinformatics and Imaging and AI / ML source data, metadata, and local mappings

  2. AnnData ecosystemWhat changes

    Use AnnData ecosystem as a pinned reference architecture across Acquire → Harmonize → Exchange → Learn + reuse

  3. OutputWhat becomes possible

    A handoff the next system or team can validate against the same release

Readiness gateThis is a fast-moving software ecosystem rather than a single governed standard. Container validity does not guarantee consistent preprocessing, cell annotation, biological comparability, or lossless conversion to Seurat or other ecosystems.

03

A concrete example

A study stores RNA counts and cell annotations in AnnData, links RNA and ATAC modalities in MuData, validates publication metadata against a pinned CELLxGENE schema, and stores tissue images, segmentations, transcript points, and coordinate transforms in SpatialData.

Why it matters: Common matrices, annotations, spatial coordinates, and loaders support scalable model development, but raw-count provenance, reference builds, QC, batch effects, label uncertainty, privacy, leakage, and split policy remain separate evidence.

04

What it fits with

AnnData represents annotated matrices; MuData groups modalities as linked AnnData objects; the CELLxGENE Discover schema constrains AnnData metadata and pinned ontologies; SpatialData adds spatial elements, transformations, and OME-NGFF-based storage.

05

Implementation starter

Start with one bounded handoff. Pin, test, and review it before scaling.

  1. Define one handoff, its accountable owner, and the decision AnnData ecosystem must support.

  2. Pin the exact version and companion artifacts: AnnData 0.13.2 · MuData 0.3.10 · CELLxGENE schema 7.1.0 · SpatialData 0.8.0.

  3. Map one representative input to the required reference architecture artifacts.

  4. Test the result against the canonical source and record every exception.

  5. Preserve the source data, mappings, and review evidence before scaling.

06

Test the main limitation

Risk

This is a fast-moving software ecosystem rather than a single governed standard. Container validity does not guarantee consistent preprocessing, cell annotation, biological comparability, or lossless conversion to Seurat or other ecosystems.

Test

Run one representative end-to-end pilot and record exactly where AnnData ecosystem loses context, needs an extension, or depends on another standard.

Risk

Machine-readable output may still be unfit for analysis or ML.

Test

Test the output for missing context, provenance, terminology alignment, time leakage, and the intended downstream decision. Common matrices, annotations, spatial coordinates, and loaders support scalable model development, but raw-count provenance, reference builds, QC, batch effects, label uncertainty, privacy, leakage, and split policy remain separate evidence.

07

Official resources

Specifications, diagrams, examples, and guides from the organizations that maintain them.

  • Primary sourceComponents version independently

    scverse foundational data structures

    The ecosystem catalogue positioning AnnData, MuData, and SpatialData as interoperable foundational structures.

    Publisher
    scverse
    Open official source
  • Specification0.13.2

    AnnData documentation

    The annotated matrix model, on-disk encodings, APIs, and release documentation used across the Python single-cell ecosystem.

    Publisher
    scverse
    Open official source
  • Specification7.1.0

    CELLxGENE Discover schema

    The versioned metadata and ontology contract applied to annotated single-cell datasets published through CELLxGENE Discover.

    Publisher
    Chan Zuckerberg Initiative
    Open official source
  • Specification0.8.0

    SpatialData documentation

    The data framework, schema, and serialization guidance for spatial images, labels, points, shapes, tables, and coordinate transforms.

    Publisher
    scverse
    Open official source

Next action

Put this profile in context

Compare its role with adjacent standards or place it inside an end-to-end data pathway before choosing an implementation.