SPACE-GM

This is an example code base for SPACE-GM: geometric deep learning of disease-associated microenvironments from multiplex spatial protein profiles.

See Example.ipynb for a tutorial of using Graph Neural Networks (GNNs) to perform predictions on spatial cellular graphs.

See Microenvironment_Clustering.ipynb for an example analysis pipeline that uses a trained GNN to define microenvironment clusters.

Requirements

• numpy
• scipy
• pandas
• networkx
• matplotlib
• scikit-learn
• geovoronoi
• pytorch >= 1.13.0
• pytorch_geometric >= 2.2.0
• CUDA 11.7 (if with GPU)