RNA-SEQ: Pathway analyses correlation to pathway activity

Question

[Muha0216]

hi i have a question regarding pathway analysis from RNA-SEQ data.

From what i understand, if the data states that a particular pathway is enriched, it i suggesting that there is increased/decreased activity of that pathway since the genes (proteins) making up the pathway cascade is differentially expressed. But from my understanding, generally when one talk about pathway activation, it mainly involves phosphorylation levels of the proteins in that pathway that alters activity. That is whether a pathway is activated or not mainly involves phosphorylation of proteins, not so much about the expression level of the protein. How reliable is it then that pathway analysis from rna-seq data can be used to suggest the activity of a particular pathway?

Answer 1

[jared.andrews07]

While phosphorylation is certainly a factor for certain pathways, it doesn't play a part if said proteins are not expressed. This is particularly true for pathway enrichment, given that it makes sense that genes involved in the same biological processes would be transcriptionally co-regulated.

Additionally, this is why experimental validation is so key to really make a strong case for conclusions derived from pathway enrichment analyses. Saying X samples have higher expression of genes in Y pathway, implicating Z is one thing, but saying this enrichment causes Z is a much tougher claim to make. This is why overexpression, knockdown/KO experiments are so common when trying to support hypothesis generated from pathway enrichment analyses (and RNA-seq as a whole).

[Muha0216]

Good comment but all proteins are generally expressed in a cell, at baseline levels. i needed a fresh perspective since i mainly work on signalling pathways at phosphorylation levels.

[jared.andrews07]

To some extent, sure, but there are plenty of proteins that aren't expressed at any meaningful level. If many copies of a protein are necessary for measurable function, a few copies aren't going to yield any functional difference from a complete knockout, regardless of any post-translational modifications that contribute to their activity. If you knockdown BTK or SYK, but they're phosphorylated as normal, you'll still see downstream effects, right? Pathway enrichment isn't much different from that, particularly since the way you're presumably identifying the genes is by comparing group of (likely phenotypically) different samples.

And as I said previously, there are many proteins for which phosphorylation isn't so crucial (or necessary at all) for their activity.