hi i really need advise, kind of in a dilemma.
my RNA-seq data has shown about 430 DEGs based on cut off of 2 fold change, and p-value <0.05. That seems like a reasonable and adequate amount of DEGs to study on further but i am in a dilemma when it comes to presenting the GO or pathway analyses. None of the GO/pathway terms are significantly enriched based on Q value. I am trying to reduce fold change to 1.5 fold hoping more DEGs would get me higher chance of enrichment in GO/Pathway analyses so that remains to be seen.
In the event that none of these terms are enriched even upon lowering cutoff, what is the best approach to take in presenting the RNA-SEQ data? I cant even say that for example ''genes involved in metastasis process'' is signicantly enriched because i have some DEGs scattered across the terms in bits and pieces.
The only way i can think of presenting the data is to just put a table indicating all the DEGs and their fold change/P values, followed by a validation of some of the genes by RT-PCR. Any better ideas on the direction i should take from here?
Hi, please do not double post. Your other thread is here: RNA-SEQ: Lowering fold change cutoff from 2 to 1.5.
Closing this. You can edit the text of your original post, if you wish.
Kevin
I don't think these are the same question, Kevin.
yes Ram is correct. this is not a double post. i just needed advise on direction to take post rna-seq results.
Not sure. Your other question is virtually contained within this question (above), with some extra details added. You could have edited your previous question.
You could have acknowledged, in addition, the contribution of Grant's comment in your other question prior to opening another question (I was the one who up-voted his comment, as I deemed it helpful).
Best wishes, Kevin
Kevin, IMO this is a follow up to OP's previous question. OP should definitely have responded to posts there, and this follow up is as weak as it gets - especially because it shows OP is fine with tweaking data to show more dramatic results than those actually obtained. But this does not preclude this question from being a follow up. I think a new post was better than editing their original post, given their original post is a lot more relevant than this one.
How is that tweaking data? Isint it common practice to lower down cut off to a minimum of 1.5 if no substantial data can be jnterpreted from a more stringent cutoff? It would be an entirely different story if i also lower stringency of p value which clearly i am not.
Also im not sure if you are a biologist. But even a fold change increase of 1.5 can be very significant, for receptors or ligands or even kinases in terms of changing biological processes. So i dont think im going overboard on this one.
Since we can't see your data we can't comment on what conclusions are appropriate for that dataset. Downstream analysis of data is an exercise in exploration and you have to decide what "story" makes sense with the data you have at hand and what cutoff's (which are artificial) are appropriate.
Ultimately you are going to have to validate the hypotheses experimentally (as you already indicated above). You are already following the right path. I don't think there is any other/better direction.
Looks like I read your post wrong, sorry.
Not sure, Ram - we are all correct and we are all incorrect at the same time. Hope all is well in NYC. I could have been there right now!
Cool! I've moved to Houston though :-)