Entering edit mode
6.9 years ago
Edalat
▴
30
Hello all,
I have some RNA-SEQ data and do some analysis by metaseq package for meta-analysis,for choosing genes I calculate q-value and local FDR,but after using local FDR<0.1 for upper and lower of fisher and stouffer, I don't have any genes for S$Lower and F$Lower,because value of local FDR for all of my analysis are 1 or .99, so it means that every time I do not have down-regulate genes in every analysis!while I have genes for S$Upper and F$Upper.
Besides, for threshold q-value<0.1,I have some genes not only up-regulated DEGs but also down-regulated DEGs.
can I use q-value?and what is difference between q-value and local FDR in term of statistical accuracy?
