I have a question regarding the interpretation of CibersortX results to assess the abundance of different cell populations in xenograft samples.
- I have prepared the single cell reference matrix files (from a publicly available scRNAseq dataset) and the mixture files (in-house bulk RNAseq) as instructed: both in linear space, same normalization between scRNAseq and mixture sample, S-mode batch correction to adjust for different platform variation.
However, we are unsure of whether all the cell populations which are normally present inside the single cell reference file are also present inside our mixture sample.
- Is CibersortX stable under these conditions? In other words, will the inference of cell fractions not be biased by the possible lack of one or more cell populations in our mixture file compared to the single cell reference?
I thank you very much in advance for your kind help.