1
$\begingroup$

I'm performing differential gene expression analysis on rnaseq of tumor samples that might have different amounts of immune infiltrate. Tumors are extracted from patients that respond/not respond to immunotherapy

In my lists I found several immune response related genes significantly modulated but i suspect that this could be related to the different amount of immune infiltrate. In order to highlight the genes that are induced by the tumors and not by the different amount of immune cells... it could be appropriate to normalize the raw reads counts (on which is based my analysis) on the basis of CD3 gene?

$\endgroup$
1
$\begingroup$

Consider that while a fair portion of your immune infiltrate will be T cells, that you can also have dendritic cells, macrophages, B cells, NK and NKT cells, and other cell types. I like to use this figure to demonstrate that fact, and also make note that the relative abundance of immune mRNA by cell type differs by indication (figure 1).

enter image description here

Source: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4852857/

Also consider that for T cell markers like CD4 or CD8, these are expressed by antigen-presenting cells, as well.

$\endgroup$
  • $\begingroup$ Thank you very much You are right . If i correctly understand it is not easy to separate the mRNA that came from the immune fraction from the mRNA that came from tumors also because some genes are expressed from apc as well. $\endgroup$ – GuidoL Jul 3 '18 at 7:24
  • $\begingroup$ the best that i might do is estimating the fraction of immune cells present in my samples (with cibersort) and evalute if the amount of immune cells is comparable between samples $\endgroup$ – GuidoL Jul 3 '18 at 7:30
  • $\begingroup$ @GuidoL I'm not 100% sure the method to delineate your gene signatures but a common issue I'd run into with mac/mono contamination, for example, is you'd see this huge p < 0.005 IL-8 expression, but the common cell that produces a lot of the IL-8 are macrophages. And IL-8 is also secreted by other cells, so you can't link the expression back to a single type of cell, either. $\endgroup$ – CKM Jul 3 '18 at 13:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.