Why is there a celining in the addPeakMatrix function?

[piyushjo15]

Hi,
My question is why is there a ceiling parameter in addPeakMatrix() function? Does it cause highly enriched CRE regions to be under-counted? When there is normalization of the peaks which uses the ceiling parameter, is it some sort of scaling? Can I make the ceiling as NULL?
What is an appropriate ceiling for a dataset with 400K cells? I am assume that the default ceiling=4 is for a typical sample having 3-5k cells, or am I wrong?