Tutorial notebooks

• Quick Start: Train MintFlow on a Single Tissue Section
  • 1. Download the anndata object
  • 1. Read the defualt configurations
  • 2. Customise the 4 configurations
  • 3. Verify and post-process the four configurations
  • 4. Setup the Data/Model/Trainer
  • 5. Train the Model
  • 8. Select the best checkpoint
  • 9. Use MintFlow predictions for analysis
• Quick Start: Generate Expression Data for a Customised Tissue
  • 1 Overview
  • 2. Download a sample anndata object and a sample MintFlow checkpoint
  • 3. Load the MintFlow checkpoint
  • 4. Make a tissue with customised cell type labels and cell spatial locations
  • 5. Generate expression data for the specified tissue
• Quick Start: Train MintFlow on Multiple Tissue Sections
  • 1. Overview
  • 2. Download the raw anndata objects
  • 3. Process the anndata files
  • 4. Create the four configuration objects
  • 5. Verify and post-process the four configurations
  • 6. Setup the Data/Model/Trainer
  • 7. Train the Model
  • 8. Select the best checkpoint and perform the analysis
• Quick Start: In silico Tissue Perturbation
  • 1 Overview
  • 2. Download a sample anndata object and a sample MintFlow checkpoint
  • 3. Load the MintFlow checkpoint
  • 4. Make the original tissue
  • 5. Make a modified (i.e. perturbed) version of the tissue
  • 6. Create neighbourhood graphs for both tissues
  • 7. Generate expression data for each tissue
  • 8. Put the generated expression data in .obsm field of anndata objects
  • 9. Idnetify cells whose microenvironment cell type composition (MCC) is changed
  • 10. Perform differential expression analysis