Can I run cellchat on multiple seurat objects?

[Chiranjit1504]

Hi,

Firstly, thank you so much for the great package. It makes analysis easier.

I have an annotated sc-RNA seq object for CD45+ cells but this object has broad annotations (T,B,myeloid cells, etc). I have sub-clustered and annotated each cell type separately so I have a separate object for T cells, B cells, and so on, each with very high-resolution annotations (Cd4 T, Th1, Th17, memory B, etc). I wanted to run cellchat at the fine resolutions but across the cell types. Since these are all separate seurat objects and in my knowledge I cannot feed in multiple seurat objects to the cellchat input, I would like to get some advice.

Should I merge the sub-clustered objects into one and pass it in as input? I am worried as this would result in changing the scale.data slot and the merged seurat again has to be re-scaled and pass through the pre-processing steps. Or is it not necessary since cellchat would only need the normalized counts and the annotations which wont change after merging?

Thank You

[hong-hui-zhao]

hello , do you slove this problems , i have meet same problem when i want to merge six dataest by running cellchat.

[jzheng25]

Not a member from CellChat team. You can merge the seurat objs (put all cell types in one that you want them to be interact). You will NOT merge the seurat objs where you want to see the changes across conditions. Example, you can Condition-A-T cells, Condition-B-T cells, Condition-A-Mye cells, Condition-B-Mye cells.

• You will merge (Condition-A-T cells+Condition-A-Mye cells) and pass to CellChat (CellChat-resA)
• You will merge (Condition-B-T cells+Condition-B-Mye cells) and pass to CellChat (CellChat-resB)
• You will follow the multi-sample tutorial to merge and compare CellChat-resA and CellChat-resB.

Hope it helps