Is it possible to specify sample?

[alexvpickering]

Thanks for the great package,

Is it possible to account for sample identity in comparison analyses? Results seem to be steered by individual samples. For example this is the setup I have:

1. cellchat.test: 2 disease samples
2. cellchat.ctrl: 20 control samples

When I perform comparison analyses as described in the vignette many of the results are driven by a a single disease sample and the ligand/receptor pairs are not consistently differentially expressed between groups. Is there any way to account for sample identity to avoid this?

[sqjin]

@alexvpickering CellChar does not support this. One way is that you can run cellchat on each of the disease samples, and then find the intersection between the two samples.