Hi to all community!!
First of all thank everyone that helped me!
Have a question: I have analyzed RNAseq data using TopHat, then CuffLinks and finally parsing results to R and analyzing via Limma Package. I obtained following the tutorial the TopTable with the: Name of the Gene, log2FC, Average Expression, adjusted p value, among others (t, B,...); and also computed the FPKM and FC and added to the table.
Here you can find a preview of my data analyzed:
My objective at this point is to retrieve the full list of genes (of nearly 23,000) that appear to be UP or DOWN according to fold change.
The filtering that usually I do is to set:
- adj.P.Val < 0.05 (FDR)
- FC > 1.4 (UP) or < -1.4 (DOWN)
But only applying this criteria I reach too many genes UP or DOWN (3000 up and 4000 down). Considering previous experiments in our lab I expect to find close to 2000 up/down genes.
My question is about the FPKM and if I should use it to be more stringent in my filters: do you have experience with FPKM filtering ? which threshold do you use in your experiments? What can you recommend me?
I read some literature about this and found some papers which use an FPKM > 1 together with FDR and FC, and others not use the same FPKM or even use the RPKM > 0.1 and I don't know exactly how to proceed to set my filters with more statistically significant limits for genes UP/DOWN.
Thanks to all community! Any help will be kindly appreciated! Dani.