search

dviraran / xCell

Cell types enrichment analysis
172 stars 60 forks source link

Proteomics data, number of genes #83

Open anuar-makhmut opened 5 months ago

anuar-makhmut commented 5 months ago

Hello,

I am trying to use xCell with Proteomics data, before using it I normalized and imputed the data. I also checked the number of genes, the tool requires minimum 5000 genes, my file contains 5294, which I confirmed by checking unique rows in R. However when I run the tool, it says I only have 3850 genes, so somehow the tool treats some of the rows being non-unique. Do you know is there any way I can fix this issue?

With best regards Anuar

dviraran commented 5 months ago

Hi,

The reason is that xCell has a list of 10K genes that it uses, and the 5K limit means that the intersection between your genes and the 10K genes should be more than 5K.

Anyhow, I wouldn't rely on such analysis. We know well that proteomics is not well correlated with transcriptomics.

Best, Dvir

fengqlin commented 3 weeks ago

Hello.I am currently working on proteomics, and I am wondering whether it is possible to convert the results of his ssgsea step, for example, into PCA or Z-score scores, and then proceed with the transformation of cell type scores. I hope to receive a relevant reply, Thank you.

dviraran commented 3 weeks ago

Hi,

ssGSEA step takes a signature of genes and ranking (just ranking) of gene expression and calculates a score, so I don't see how your suggestions may be helpful.

Regarding analyzing proteomics data - I would not recommend using the xCell signatures as they were not trained for proteomics.

Best, Dvir