A comprehensive study of graph deep learning enabled spatial domains discrimination technologies for spatial transcriptomics
Graph deep learning has been regarded as a promising methodology to address genetic transcriptomes and spatial locations in spatial omics data. To this end, a comprehensive analytical toolbox, GRAPHDeep, is presented to aggregate two graph deep learning modules (i.e., Variational Graph Auto-Encoder and Deep Graph Infomax) and twenty graph neural networks for spatial domains discrimination. Towards spatial omics data with various modalities and scales, the best integration of graph deep learning module and graph neural network is determined. Consequently, this built framework can be regarded as desirable guidance for choosing an appropriate graph neural network for heterogeneous spatial data.
- Python (>=3.8)
- Scanpy
- Squidpy
- Pytorch_pyG
- pandas
- numpy
- sklearn
- seaborn
- matplotlib
- torch_geometric
- Selecting GNNs for spatial clustering task in DGI module
running DGI Example_DLPFC.ipynb to see the simulation results step by step
- Selecting GNNs for spatial clustering task in VGAE module
running VGAE Example_DLPFC.ipynb to see the simulation results step by step
Benchmarking methods used in this paper include:
- Spatial protein: 4i
- Spatial protein: MIBI-TOF
- 10X Xenium
- Human DLPFC
- MERFISH
- Mouse brain
- SeqFISH
- Slide-seqV2
- STAR-map
Distributed under the MIT License. See LICENSE.txt for more information.
-
data files
Please down load the spatial transcriptomics data from the provided links. -
Porch_pyg
Please follow the instruction to install pyG and geometric packages.