Something to share regarding multiple samples with CellChat

[NPTL1201]

Based on my experience, I suggest running CellChat on separate samples when having multiple biological replicates. I found that even the most frequent interactions observed across samples in a group often appear only in specific samples. Running CellChat on a merged dataset may overestimate the importance of these interactions.

[sqjin]

@NPTL1201 You can set the parameter 'min.samples' when running filterCommunication, leading to robust identification across multiple replicates.