jinworks
Effect of adding the `samples` metadata column on inferred communications
Hi and thank you for developing and maintaining CellChat!
I noticed that adding the metadata column samples to the CellChat object affects the number of detected cell–cell communications even when no filtering is being done using the min.samples parameter of the filterCommunication function.
After adding the samples column, the inferred communication network changes, with fewer interactions detected compared to the same data without the samplesmetadata column.
Could you clarify how adding the samples metadata column affects communication inference?
The documentation states:
A column named samples should be provided for spatial transcriptomics analysis, which is useful for analyzing cell-cell communication by aggregating multiple samples/replicates.
I am working with spatial transcriptomics data from a patient cohort and currently define each patient as a “sample.” Does this approach make sense if my goal is to identify common communication signals across patients?
Thank you for your guidance!
@AlbertoFabbri93 Yes, please use the min.samples parameter when identifying common communication signals across patients.
It is a bit confusing that adding the metadata column samples to the CellChat object affects the inference. Except for adjusting the min.samples parameter of the filterCommunication function, I do not think the results will change.
