Skip to content

Single-cell data in ODM: Getting started

Upload → Transform → Index → Search

This page walks you through the single-cell HDF5 workflow in ODM (uploading a file, running a transformation to produce indexed objects, confirming indexing, and exploring your data with search and analytics queries) and links to the notebooks for each stage.

Who this is for

This guide is for users who want to try ODM's single-cell functionality on prepared datasets.

Prerequisites

  • ODM instance URL: <HOST>
  • API token: <TOKEN>. See Authentication and tokens.
  • An environment set up to run Jupyter notebooks.

Step 1: Upload and transform a single dataset

The goal here is to walk through the full workflow on one HDF5 file and verify that analysis-ready objects appear in ODM.

  1. Create a study with an HDF5 file as an attachment.
  2. Run a transformation job to convert it into ODM-indexed single-cell objects.
  3. Verify that objects were created, linked, and indexed correctly.

Notebook: Single-Cell RNA-Seq: Data Transformation and Upload to ODM

This notebook covers: uploading the HDF5 file, creating a configuration, running a dry run, and checking job status and outputs.

Step 2: Load curated public datasets

The goal of this optional step is to populate ODM with a ready-made catalogue of curated public single-cell studies so you can test cross-study search without preparing your own data.

  1. Load the public dataset template.
  2. Load the curated datasets (HDF5 attachments included).
    • Ready-to-run import commands: Curated public datasets, per-dataset copy-paste commands with placeholders for server, token, and template.

Step 3: Transform curated datasets

The goal is to transform the curated datasets into fully indexed objects with harmonised metadata.

  1. Using the transformation notebook from Step 1 as a reference, run the transformation for each curated dataset.
  2. Use the provided configurations to ensure consistent curation. You can skip the dry-run step, as the configurations are pre-tested.
  3. Monitor transformation jobs until all complete successfully.
  4. Confirm that the expected objects are present: Cell Group, Expression Group, and metadata objects.

Prepared configurations: Curated public datasets, the configuration to use for each dataset.

Step 4: Confirm indexing

Make sure all datasets are marked as indexed and ready to query.

  • Each transformed dataset shows the Indexed label in the ODM Metadata Editor.
  • All indexing tasks show Done status in Task Manager.

A completed transformation job does not mean data is immediately searchable. ODM automatically triggers indexing after ingestion, but data is only available for querying once indexing finishes.

Step 5: Query and analyse single-cell data

Use ODM's search and analytics notebooks to explore your indexed datasets.

Notebook: Single-Cell RNA-Seq: Cohort Selection and Data Retrieval

This notebook covers: cross-study search examples, filtering by curated attributes, example analytical queries, and result inspection.

Next steps