search

jianhuupenn / SpaGCN

SpaGCN: Integrating gene expression, spatial location and histology to identify spatial domains and spatially variable genes by graph convolutional network
MIT License
192 stars 59 forks source link

Using SpaGCN without image #11

Closed Yao-14 closed 2 years ago

Yao-14 commented 2 years ago

Hallo,

I am very interested in SpaGCN, but I don’t have a stained image. Can I run SpaGCN only with gene expression data and spatial location data, and get a good result?

jianhuupenn commented 2 years ago

Thanks for your interest in SpaGCN. Yes, SpaGCN can be applied to ST data without histology images. The only difference is the way of calculating the adjacent matrix. For more details and guidance, please refer to the tutorial https://github.com/jianhuupenn/SpaGCN/blob/master/tutorial/tutorial.md Section 4. Integrate gene expression and histology into a Graph

#If histlogy image is not available, SpaGCN can calculate the adjacent matrix using the function below
adj=calculate_adj_matrix(x=x_pixel,y=y_pixel, histology=False)

Jian

Yao-14 commented 2 years ago

Well, thank you very much for your reply, which is very useful to me. I have a new confusion. I found that SpaGCN integrates the process of performing PCA, computing neighbor graphs, and clustering, but I now have multiple batches of data that I want to integrate, and then cluster. At the same time, the suggestion given by Scanpy is to perform PCA first, then Use harmony to integrate different batches, and then compute the neighbor graph. How can I implement this process in SpaGCN ?

jianhuupenn commented 2 years ago

The goal of SpaGCN is to identify the spatial domain in a single section and does not have any steps for batch effect removal. If you are interested in analyzing multiple sections from different tissue samples, you can refer to my suggestions in: https://github.com/jianhuupenn/SpaGCN/issues/6 If the tissue sections are from the same tissue and are spatially adjacent to each other, you can refer to my tutorial part

  1. Multiple tissue sections analysis https://github.com/jianhuupenn/SpaGCN/blob/master/tutorial/tutorial.md#8-multiple-tissue-sections-analysis