This is a test version of Biostars. For the public version, visit https://www.biostars.org.
GFOLD value cutoff

Dear all,

I have two groups of RNA-seq data, one of them are for treated sample and another one is for control. Unfortunately they are without any replicates (I know experimental design without any biological replicate is not a good thing). so I have used GFOLD calculated GFOLD value for each transcripts. then I ranked the value got more than 5900 transcripts (GFOLD value>1) out of more than 56000 assembled sequences. here I wanna set a cutoff for following enrichment analysis, but I do not know how to set a reasonable one, could you give me some suggestions?

gfold
• 3,594 views • link
updated by Ram Ram • written by Kurban Kurban

1 answer

There is no gold standard cutoff. To understand this, just consider gfold as a reliable log2 fold change. The only thing hold is that large absolute gfold value corresponds to large express changes.For enrichment analysis, just take the top genes, 1000, 500, 2000, etc. The biological conclusion should be consistent.

• 0 views • link
written by jianxing.tongji • jianxing.tongji

Hi, Jianxing, I have a further question. Kurban set GFOLD value > 1. That will select up-regulated genes. Then how to set GFOLD value to select down-regulated genes using the same criteria as selecting up-regulated genes? Is it GFOLD value < -1? Thank you very much!

• 1 views • link
written by biolab biolab

The GFOLD values are symmetric around 0. For up-regulated genes, you can take GFOLD>x and for down-regulated ones, you can take GFOLD<-x.

• 1 views • link
written by jianxing.tongji • jianxing.tongji

Thank you very much. It's much helpful.

• 1 views • link
written by biolab biolab

Log in to answer this question.

More posts like this