Modeling Spatial Omics and Cellular Niches with Graph Neural Networks

Explore cutting-edge techniques for modeling spatial omics and cellular niches using graph neural networks in this comprehensive lecture from the Computational Genomics Summer Institute. Delve into single-cell analysis, spatial proteomics, and imaging methods for measuring spatial omics. Learn about the Space GM framework, representation learning, and message passing techniques for capturing 2D slices and clustering spatial data. Examine a detailed case study demonstrating the application of these methods to analyze spatial clusters and cellular coherence. Discover how to train and generalize models for subcellular morphologies and generate synthetic data. Gain insights from related research papers on characterizing tumor microenvironments, generating in silico CODEX data, and integrating spatial gene expression with tumor morphology through deep learning approaches.

Introduction
Single cell analysis
Spatial proteomics
Measuring spatial omics
Imaging spatial omics
Modeling spatial omics
Space GM
Representation Learning
Message Passing Walkthrough
Capturing 2D Slices
Clustering
Case Study
Spatial Clusters
Coherence
Overall Framework
Training the Model
Generalizing the Model
Workflow Summary
Subcellular Morphologies
Generating Synthetic Data
Summary