Is it ok to use DREME to predict motifs from RNA-Seq data?

Question

[Harumi]

Is it ok to use DREME to predict motifs from RNA-Seq data?

I saw that DREME is used to predict motifs from large datasets provided by CHIP-Seq experiments, however I have some diferentially expressed genes from RNA-Seq data that I want to understand regulation by motif prediction. I ran MEME analyses but I've got only one motif. On DREME I've got several. Are these results reliable?

Thanks

[h.mon]

I do not have much experience on motif finding, but for what I've read, DREME is used to predict motifs when there is some reason to believe the data is enriched for some motif, like protein-DNA capture, or protein-RNA. Why would you expect a set of differentially expressed genes to share motifs?

[Harumi]

I'm sorry I couldn't express myself properly. Since these differentialy expressed genes were upregulated (and the other set downreglated) in response to a stress condition, maybe their promoters could share motifs. But I'm a not sure because I didn't do protein-DNA capture or protein-RNA. Thank you for your reply.

[ytzhao]

Hello, I guess that you mean you now have several genes, in fasta format. you wanna find motif from the promoter region. but what to do next ? I suggest that you could have a try to find transcription factor using homologous genes that have been analysed with Chip-seq or Clip-seq. May it clear ?

[ATpoint]

You can try that but note that gene regulation is much based on distal regulatory elements so scanning only promoters might (and probably will) underestimate the complexity.