Entering edit mode
Hello!
I am analyzing gene expression with DESeq2 and was wondering if there is a way to look at genes that are constituently expressed across samples, but not differential. So far, I identified genes with a p-value of >0.05 (I understand this comes with the caveat of that gene simply having no evidence for differential expression) and then subsetted those genes from the TPM count matrix. I took the mean TPM across each replicate group and marked the gene as "expressed" if mean TPM >= 3. I was wondering if this approach seems reasonable, or if there is a better method for identifying constituently expressed genes?
Thank you so much for your help!
Hello! Thank you so much for your response. I have actually done the resLA <- results(dds, lfcThreshold=.5, altHypothesis="lessAbs") method before--as I understand, this is a statistical way to look at genes that do not undergo expression differences by changing the hypothesis; now the p-value indicates the probability that a gene is NOT deferentially expressed in a contrast. When I did this, I noticed a very small number of significant genes that we have confidence are not deferentially expressed. I have taken note of this gene list.
However, my question remains--is there a way to look at genes that are expressed regardless of our ability to estimate whether they are expressed differentially or not? In addition to looking at whether a gene is statistically significant in their expression pattern, I am also interested in knowing whether a gene is expressed at all in a tissue. Do you know of any ways to do this?
Thank you so much!
If I wanted that sort of information, I would probably just get the data from GTEx. But if you want to use your own data, you will probably have to use something ad hoc like your TPM >= 3.
Thank you, I appreciate it!