Scanvi Tools

Configuration File: remote_tools/scanvi_tools.json Tool Type: Remote Tools Count: 1

This page contains all tools defined in the scanvi_tools.json configuration file.

Available Tools

run_scanvi_annotate (Type: RemoteTool)

Transfer cell-type labels to unlabeled single cells with scANVI (semi-supervised single-cell anno…

run_scanvi_annotate tool specification

Tool Information:

  • Name: run_scanvi_annotate

  • Type: RemoteTool

  • Description: Transfer cell-type labels to unlabeled single cells with scANVI (semi-supervised single-cell annotation; Xu 2021 / scvi-tools, Gayoso 2022). Pretrains scVI on a raw-count matrix, refines it semi-supervised on the known labels, then predicts a cell type for every cell and returns the predictions for the previously-unlabeled cells. Input is a server-accessible .h5ad of raw UMI counts whose labels_key obs column holds known cell types for some cells and the unlabeled_category sentinel (default ‘Unknown’) for the cells to annotate.

Parameters:

  • adata_path (string) (required) Server-accessible path or URL to an .h5ad AnnData of RAW UMI counts.

  • labels_key (string) (required) obs column holding known cell-type labels for some cells and the unlabeled_category sentinel for the cells to annotate.

  • unlabeled_category (string) (optional) Value in labels_key marking unlabeled cells (default ‘Unknown’).

  • scvi_epochs (integer) (optional) Unsupervised scVI pretraining epochs (default 100).

  • scanvi_epochs (integer) (optional) Semi-supervised scANVI refinement epochs (default 20).

Example Usage:

query = {
    "name": "run_scanvi_annotate",
    "arguments": {
        "adata_path": "example_value",
        "labels_key": "example_value"
    }
}
result = tu.run(query)