This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Query regarding single cell rna seq annotation using marker genes

In one study, their dataset contains NORMAL tissue while my dataset contains CANCER tissue of the same organ. Using canonical markers they have annotated the cell types(epithelial, b cell, t cell etc.). But for annotating cell states/cell subtypes they have mentioned for- CD4+ T cells included naive T cells (SELL), T helper (TH) and TH-like (IL7R, CCR6,CCL20), T effector memory (TEM) (LMNA) and Tregulatory (Treg) (FOXP3,CTLA4) cells.But in methods they have mentioned that they used deferentially expressed genes and pathway enrichment analysis to annotate the cell sates. so can i annotate the cell states in my dataset using the cell state genes they have mention for their dataset, especially for epithelial cells as these are highly altered in cancer tissue(my dataset) in comparison to normal tissue(their dataset)?

scrna-seq
• 886 views • link
updated by Ram Ram • written by TNF_alpha • TNF_alpha

1 answer

I would say that using just DEG for cell type annotation can be unreliable sometimes. You could perform differential gene expression using Wilcoxon-rank sum and then perform GO pathway analysis and compare the geneset/pathways you get with available literature to see if the markers make sense.

If you are just trying to get a preliminary overview of the dataset then I would suggest using canonical markers mentioned in the paper. Immune cell types are tricky to characterize and results can vary, hence using canonical markers would be a better and safer option in my opinion.

I would refer you to this poster on the ABCAM website that is quite informative: https://www.abcam.com/en-de/technical-resources/pathways/immune-cell-markers-poster

• 0 views • link
written by bio_info • bio_info

Log in to answer this question.

More posts like this