Hi all,
I have analyzed an RNA-seq with 2 conditions. And I have DEG using Deseq2. Now I would like to compare this to a table (actually maybe 2 or 3 tables) of DEG between 2 groups of cells (actually this come from single cell) and I'd like to know if the differences in my first DEG after treatment accounts for the recruitment of one population or the other. I'm checking the best genes (either by log2FC or by adj p-Val). And I see a tendency but I'm just checking genes one by one from the text file. Now I want to have a real global assessment that would tell me how much these correlate or anti-correlate (or don't correlate with anything). So I want to see if globally most genes that are up-regulated in my treatment are also mostly up-regulated in one population or the other and this is globally significant. And if possible get a list of genes that do correlate or not, or have a view such as heatmaps to compare. Could you tell me what is best to try and do this type of analysis and represent it?
Edit: I have thought of just doing venn-diagrams but I don't think this gives me a good enough view of the real correlation between data, although if the results are clear-cut this could be enough.
thanks a lot for your help, best, Xavier
You could just perform correlation of log2FC in the two analysis, preferably using Spearman's correlation, instead of Pearson's. Also, if you perform a Wilcoxon test between the 2 log2FC distributions you can test if they look the same. You can also do the same with the padj distributions.
thanks for your reply. Sorry for my lack of knowledge, I'm still beginner in some of these analyses. I guess I have to do these type of correlations but I do not know how to do it. Is it in R, I have to create a data table and there is a simple function to use or is there another way of doing that? Could you either tell me how to do or direct me by telling me which tool is the best and I can search more online? thanks again
edit: basically I have 2 tables, I can put the genes in the same order in each and I have a value (up or down-regulated log2FC) I want to check that globally when it goes up in one, it mostly goes up in the other. (I keep only those with adjusted p-value < 0.05) edit2: I will try this: in 1 data.table the 2 columns of the log2FC with the same order of genes (actually here I will have to run a quick script to remove genes that are not in common first) and do spearman.test (x, y) x and y being my 2 columns. then spearman.test(x,-y) to test the anti-correlation