I would like to know what the general consensus is regarding cut-off values for gene expression fold changes (is it mainly >2 up and down-regulated?). Also, is this cut-off applied together with the cut-off for p-value which is p
I think the general consensus is > and < than 2-fold, however, we should all justify our rationale for using 2-fold. In our specific case, a difference up or down of 2-fold was our cutoff comparing RNA-seq from two different samples. Of the 8,935 genes we were analyzing the 2-fold cutoff put 95% of the genes less than this difference p
Agree with Raul's comment. Only cutoff for p-value makes sense. Could be
In my experience, it is also helpful to include general expression values (RPKM, quantile etc.) when scoring differential expression. I used Microarrays and qPCR to verify differential expression discovered by RNA-Seq and found out that it works good for highly and moderately expressed genes, but not for weak expression, even if fold change is >2 and p
According to my experience , for RNA-seq analysis most important criteria is RPKM. After taking this cutoff you can apply fold change cut-off and this generally depends on what king of experiment you are doing. Generally people go for 2 fold cutof as it does suffice their purpose. Also P
Agree with Michal Okoniewski's comment. I don't think there are a "consensus" value in all of researches. It's depend on you.
To my experience, for low-expressed genes, the fold change is much more sensitive and should be very careful to say that a low-expressed gene is a signature or not.
- Badges
- Science method