Tutorials

Welcome to the SpaHDmap tutorials. These tutorials will guide you through various aspects of using SpaHDmap for spatial transcriptomics data analysis.

Each tutorial provides step-by-step instructions and code examples to help you get started with SpaHDmap for your specific use case.

We tested the tutorials on the following system:

• Ubuntu 22.04 LTS

• Python 3.11.9

• CPU: Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz

• GPU: NVIDIA A40 (48GB)

• RAM: 512GB

It takes about 1~2 hours to complete each tutorial. If you encounter any issues, please open an issue on the SpaHDmap GitHub repository.

• Step-by-step guide for analysis of H&E-stained imaging data

Provides a complete, step-by-step walkthrough of the SpaHDmap workflow using a 10X Visium H&E stained dataset, from data loading to downstream analysis.

• Analysis of IHC-stained imaging data

Focuses on applying SpaHDmap to spatial transcriptomics data with IHC-stained images, covering data preparation and analysis for this image data.

• Analysis of multi-section ST data

Explains how to analyze multiple tissue sections together, including how to handle batch effects between different sections for an integrated analysis.

• Running SpaHDmap from the command line

Details how to run the entire SpaHDmap pipeline using the command-line interface, including how to set up the necessary JSON configuration file.

• Select the optimal dimension of SpaHDmap

Guides you through selecting the optimal rank (number of components) for SpaHDmap analysis using cophenetic correlation to ensure stable and meaningful results.

• Transfer a pre-trained model to new data

Demonstrates how to apply a pre-trained SpaHDmap model to a new dataset, allowing for rapid analysis without retraining from scratch.

• Spatial domain refinement

Shows how to refine the analysis by focusing on a specific region of interest, extracting high-score spots and re-running the pipeline for a more detailed view.

• Correct image batch effects using color normalization

Illustrates how to use color normalization to reduce image-based batch effects in multi-section H&E stained datasets, improving consistency across sections.

• Differential expression and GO enrichment analysis

Show how to perform differential expression analysis and GO enrichment analysis for the results of multi-section data analysis

• How to download and prepare data

Illustrates how to download and prepare for use with SpaHDmap, including using scanpy and manually downloading from 10X Genomics or Google Drive.