search

ventolab / CellphoneDB

CellPhoneDB can be used to search for a particular ligand/receptor, or interrogate your own HUMAN single-cell transcriptomics data.
https://www.cellphonedb.org/
MIT License
305 stars 52 forks source link

Missing interactions post-method2 analysis #135

Closed jl1693 closed 7 months ago

jl1693 commented 11 months ago

Hi CellphoneDB team,

I have recently analysed my single cell seq data of some in vitro cells on CellPhoneDB v4.1.0 but am getting strange results. My data contained 6 leiden clusters of cells of which i selected two clusters to perform cell-cell interaction analyses on. The data was normalised and log transformed. On comparing the two clusters, I obtained 143 interactions. I then repeated the analysis but this time including all the 6 leiden clusters but came into 2 key issues:

  1. Both sets of analyses yielded 143 interactions (of which these interactions were very similar between cell type pairs)
  2. The mean values I am getting from these analyses seem to all be <1 majority of the time.

Do you know what I can do to fix this? I know that there are a lot more previously validated interactions that are being missed out.

ktroule commented 11 months ago

Hi.

It's hard to judge what might be happening based on your comment. Check that the inputs of the two runs are correct, they seem to be the same based on the 143 interactions you are getting.

Kind regards