search

huBioinfo / CytoCommunity

A spatial omics data analysis tool that enables both unsupervised and supervised discovery of complex tissue cellular neighborhoods from cell phenotypes.
MIT License
16 stars 7 forks source link

[CytoCommunity] Using visium data #7

Open jpark27 opened 4 months ago

jpark27 commented 4 months ago

Dear authors,

Hi, all! Thank you so much sharing great tool to community. I happen to come across you used DLPFC 10x visium data on CytoCommunity as well but can't find any toy example how you converted raw spaceranger output into current cytocommunity input files. Can you guide us with tutorial or explanation how one can do this?

many thanks in advance, J

huBioinfo commented 4 months ago

Hi Jun,

I'm very sorry for replying late. Yes, CytoCommunity can be used for spot-based spatial omics like 10x visium data, but needs an additional procedure, that is, cell-type deconvolution. First, you can apply cell-type deconvolution methods (e.g. CARD, by Ma Y, Nature Biotechnology, 2022) to 10x visium data (usually single-cell transcriptomics data as reference) to obtain cell type fractions for each spot. Then, you can construct a spot-spot proximity graph and use the cell type fraction as the attribute vector of each spot. So the Step1.py should be revised since there will be no CellTypeLabel.txt. If you need further assistance on how to revise the code in Step1.py, please contact me via email: huyuxuan@xidian.edu.cn

Best, Yuxuan