Deconvolution of bulk RNA sequencing PBMC: different results for different methods

Question

[MatStat]

Hi all,

I am running deconvolution on a PBMC bulk RNA sequencing data in R. The problem is that I get very different results when using different methods: quantiseq CIBERSORT EPIC ABIS

Besides EPIC it seems that most methods are having difficulty characterizing the cells fractions. I would expect that a well established method would be able to recognize a PBMC data and therefor I won't have a large fraction of cells which are uncharacterized when using appropriate reference matrix.

How can I tell which method yields the best result?

Does anyone have an experience with one of the methods on PBMC/in general?

All the best