Question about use of the population.size parameter in computeCommunProb for spatial datasets

[kathleenabadie]

Hello, Thank you for making and maintaining CellChat! I have been using it to analyze cell-cell interactions in a 10x Xenium spatial dataset. I am running computeCommunProb using the parameters below. My questions is whether the population.size parameter should be set to FALSE (default) or TRUE.

cellchat <- computeCommunProb(cellchat,
    type = "truncatedMean", trim = 0.05, # or 0.1 
    distance.use = TRUE, interaction.range = 100,
    scale.distance = 1,
    contact.dependent = TRUE, contact.range = 10,
    population.size = FALSE,        # or true                           
)

Conceptually, it seems to me that it should be set to TRUE, since the nature of spatial data accurately captures population fractions (e.g. there is no sorting step that interferes with cell proportions). However, in the spatial vignettes, it seems that this parameter is kept at the default of FALSE. I have tried running it both ways, and the results are different as expected, where the lowly represented cells in my dataset have higher signaling scores when population.size=FALSE.

Any insight or guidance you can provide here would be much appreciated.

Thank you.

[AlbertoFabbri93]

Yes, I have the same exact doubt. I am running an analysis on CosMx data and I have population.size = TRUE.