SCP for data generators

This article describes the AnnData format (.h5ad) and SCP-required metadata, to help inform data management cores and other data generators who want their single cell datasets to be explorable using Single Cell Portal and its reusable visualization tools.

 

How to process single cell datasets for use in Single Cell Portal:

• Step 1: Generate primary data (FASTQs)
• Step 2: Generate BAMs and raw count matrix, using tools like Optimus
• Step 3: Generate SCP-ready AnnData files
  • Fill in obsm frames in your AnnData with sample-level values for SCP required metadata (e.g. organ ID, organ__ontology_label)
  • Use tools like Scanpy or Seurat to:
    • Apply normalization, filtering, etc., generating an processed expression matrix from raw count matrix
    • Generate (X, Y) clustering coordinates for your cells via dimensionality reduction methods like UMAP, t-SNE, or PCA
  • Use tools like scANVI or Cellarium CAS to annotate cell types, and put the Cell Ontology values in the cell_type ID and cell_type__ontology_label frames in obsm -- these are optional fields in the SCP metadata schema

Our articles on setting up a study on SCP give more detailed instructions on how to create, upload, and manage your dataset on Single Cell Portal.

Questions?  Contact us at [e-mail].