Tutorials

List of Tutorials

• Using the AdataDict Class
  • Build an AdataDict
  • Manipulate the Hierarchy of an AdataDict
  • Set the hierarchy
  • Iterate Over an AdataDict
  • Write and Read AdataDict
• Annotating with AnnDictionary
  • Automated Cell Type Marker Gene Scores
  • Cell Type Annotation with Large Language Models (LLMs)
  • Transfer Labels Using an Sklearn Classifier
  • Automated Annotation of Spatial Transcriptomics

Some background information

For reference, and to get you started, here’s the rationale behind the package’s core modules.

AdataDict

When analyzing large amounts of single cell sequencing data, we frequently wanted to do things like: - do the same analysis separately on each cell type - then, separate by tissue and try again

Manually looping became too clunky, so we wrote this package to simplify the syntax. We also built in multithreading and smart argument broadcasting.

• See AdataDict Core documentation

Annotate

When analyzing many anndata, there’s the additional bottleneck of cell type annotation. So, we built a suite of functions to annotate cell types with LLMs, with label transfer models, etc., and functions to benchmark and compare results.

• See Annotate Module documentation

Automated Label Management

Finally, there’s a major friction point–managing all these labels (for example, cell type labels that differ only typographically). So, we built functions that use LLMs to automatically clean categorical columns.

• See Automated Label Management documentation