Cut off for DE genes

Question

[divya.nandakumar]

Hi,

I am using RNA Seq to identify genes that are down-regulated in a mutant cell line for a viral transcription factor compared to WT. The cell line has been infected with a virus and I am specifically interested in viral transcript down regulation.

In the mutant cell line, almost all viral genes are down-regulated although to varying degrees. What would be an acceptable cut off to count as being "down-regulated" in the absence of the factor? Could I use the median of the Log2FC as a cut off or is some other statistic more appropriate?

[h.mon]

I hope you are using edgeR, DESeq2 or other specialized package for the differential gene expression. I would use just the fdr (corrected p-value) to delimit "significant" from "non-significant", with either an fdr < 0.05 or fdr < 0.01, depending on how strict you want to be.

[divya.nandakumar]

I am using DESeq2 for determining the log2FC and the fdr. Most of the viral genes that are down regulated are significant as seen by the fdr.

Sorry if my questions wasnt clear but my question was of the ones that are significantly down-regulated, is there a statistic that I can use to say that one set are more strongly down-regulated than others? Such as taking the mean of the log2FC and then looking at one or 2 standard deviations from it to say some of them are more down regulated than others?

Answer 1

[waqaskhokhar999]

log2 fold change can give you better insight of the picture. For more details you can check these links RNA-Seq workflow for differential gene expression

From reads to genes to pathways: differential expression analysis of RNA-Seq experiments using Rsubread and the edgeR quasi-likelihood pipeline