Graph Contrastive Learning of Subcellular-resolution Spatial Transcriptomics Improves Cell Type Annotation and Reveals Critical Molecular Pathways

requirements:

• torch (=1.11.0)
• networkx (>=2.6.3)
• scikit-learn (>=1.0.2)
• torch-scatter (>=2.0.9)
• torch-sparse (>=0.6.16)
• torch-cluster (>=1.6.0)
• torch-geometric (>=2.1.0)

Dataset

The offical raw FOCUS dataset is avaiable:

• [CosMx SMI ]
• [MERFISH MOP ]
• [Xenium DCIS ] Please download and put into data folder.

Preprocessing

1. knn-based gene neighborhood network generation

python focus/data/knn_data_cosmx_lung.py # take CosMx lung data as an example.

2. data split and generation

python focus/data/build_graph_datasets_lung.py 

Training

torchrun --nnodes=1 --nproc_per_node=4 run_all.py