Skip to content

Prepared Inputs

Quick Start

Use this page when your data are already prepared for STEER. The recommended starting point is the Quick Start Notebook, which walks through the standard workflow on processed input data.

Open Notebook See Tutorials

Before you begin

STEER is designed for RNA velocity analysis and therefore requires input data in which spliced and unspliced molecules can be quantified reliably. In practice, this usually means sequencing-based transcriptomic data generated by workflows that preserve enough transcript structure information to distinguish mature and nascent RNA states.

Typical compatible inputs include:

  • standard scRNA-seq or snRNA-seq datasets that have already been processed to include spliced and unspliced layers
  • sequencing-based spatial transcriptomics datasets from untargeted or whole-transcriptome workflows, provided that the preprocessing pipeline preserves the information needed for spliced and unspliced quantification
  • spatial datasets that also include X_spatial, which is required for the spatial workflow used in the quick start notebook

Examples of potentially suitable upstream data sources include standard single-cell RNA-seq pipelines and sequencing-based spatial assays such as fresh-frozen whole-transcriptome workflows. However, compatibility ultimately depends on the final processed object rather than the platform name alone: your input .h5ad file must contain validated spliced and unspliced layers, as well as X_spatial for the spatial workflow.

By contrast, some spatial transcriptomics technologies are generally not suitable for direct RNA velocity analysis. Imaging-based assays usually do not provide sequencing reads that allow intronic and exonic signal to be separated in the way required for unspliced quantification. Likewise, probe-based or targeted sequencing assays, including many workflows commonly used for FFPE samples, often measure only predefined transcript regions and typically do not capture the intronic information required for reliable RNA velocity inference.

If your data do not already contain the required layers, or if your upstream technology does not support reliable spliced and unspliced quantification, please start from the raw-data preprocessing tutorials instead.

Quick start

For most users, the recommended order is:

  1. Complete the installation
  2. Open the Quick Start Notebook
  3. Prepare or load an input .h5ad
  4. Run the core STEER workflow
  5. Explore downstream velocity-related analyses

If you want to see the same core workflow on real datasets rather than the demo-style quick start example, you can also open the Spatial Mouse Placentation Run or the scRNA Mouse Erythroid Run. These notebooks provide compact end-to-end runs for an actual spatial mouse placentation dataset and a single-cell mouse erythroid lineage dataset.

  • Primary notebook

    Run the guided STEER workflow on processed inputs.

    Quick Start Notebook

  • Real data example

    See the workflow on an actual spatial mouse placentation dataset.

    Spatial Mouse Placentation Run

  • Single-cell example

    Run STEER on a mouse erythroid lineage scRNA-seq dataset.

    scRNA Mouse Erythroid Run

  • Demo data

    Browse the example files used in the tutorial workflow.

    tutorials/demo_data/

  • Need preprocessing?

    Start here if your data do not yet contain validated spliced and unspliced layers.

    Raw Data Preprocessing

If you prefer to browse the original repository files:

Next steps

After completing the quick start, you can continue with: