RNA seq, secreted protein

Question

[Rob]

Hi all, I have about 200 differentially expressed genes. How should I know which genes make the secreted protein?

[Joe]

Short answer: unless the function of the genes is already known, you can't learn this from RNAseq.

[Matthias Zepper]

Wouldn't it be possible to screen for export signals in the translated peptide sequences? Certainly not the most promising approach (which is why I would use annotations where available), but a viable alternative - or am I missing something here?

[Joe]

Sure, but my point was really that that's a waste of RNAseq when you could just do that from a genome annotation as you mention.

I think a better approach would be to use this for pathway enrichment and interrogate which pathways are associated with secretion?

[Maka]

Hi, I am trying to use this approach with my list of DE genes with GO annotations. How can I verify which pathways are associated with secretion? Do you have any suggestions? thank you

[Joe]

You should find certain GO terms that are reasonably obvious, but otherwise you can feed the data to tools like STRING which will give you some idea of whether secretion features as a significant component of the dataset.

Though the real answer here is: time to hit the literature. Find the genes that are most different/prominent, and look at what people have found out about them.

[Matthias Zepper]

Is your organism annotated? In that case, see for example this and this response.