Raw Data Preprocessing¶
This page is for users whose input data does not yet contain spliced and unspliced layers.
If your .h5ad file already includes these layers, you can skip preprocessing and go directly to the Quick Start workflow.
Purpose¶
A practical challenge in spatial RNA velocity analysis is generating spliced and unspliced count matrices from raw sequencing data. The STEER repository includes preprocessing resources to support this step before velocity inference.
These resources are organized under:
tutorials/raw_data_processing/
Available platform routes¶
The repository currently provides preprocessing routes for major spatial transcriptomics platforms, including:
- Slide-seq
- 10x Visium
- Stereo-seq
Platform notes¶
Slide-seq¶
This route provides an end-to-end workflow for generating spliced and unspliced matrices, including handling barcode mismatch issues and running velocyto.
10x Visium¶
This route provides a practical workflow that connects velocyto-based processing with Visium and Seurat-style outputs.
Stereo-seq¶
This route provides a preprocessing path based on the official stereopy ecosystem and Stereo-seq-style data organization.
Repository location¶
GitHub directory:
Recommended usage path¶
Choose your starting point as follows:
- already have
spliced/unsplicedlayers → go to Quick Start - raw Slide-seq data → use the Slide-seq preprocessing route
- raw 10x Visium data → use the 10x Visium preprocessing route
- raw Stereo-seq data → use the Stereo-seq preprocessing route
After preprocessing¶
Once spliced and unspliced matrices are available in your processed input object, continue with:
- Quick Start