I came across the PSEA package which is used for computational deconvolution of microarray expression datasets. Their paper demonstrates its usage on human brain microarray data.

The paper and reference manual mention microarray, but I wonder if the same method is applicable to RNA-seq data? Considering the model is different, I'd initially suspect it isn't. But does someone know an alternative? The majority of the deconvolution methods I've seen are optimized/tested for impure tumors, which may or may not a representative system for the complexity in celltype composition in the case of neurodegeneration.

Thanks in advance!