I understand that pathway gene lists in databases also include the inhibiting genes.
Suppose I have a sorted list of genes, from high differentially expressed genes (hyper expression) through neutral genes to negatively expressed genes (hypo expression). When searching for pathways, should I submit one list with both hyper and hypo subsets? Are there insights to be gained when submitting each subset individually?
What kind of pathways are you talking about? Please define. A metabolic pathway? Kinase pathway? As you can see already with my two examples, pathways can be very different and therefore differently regulated. In metabolic pathways, you'll have kinetics with enzymes, more expression of a gene might correlate with more of that particular enzymatic reaction, however some enzymes may also inhibit a reaction (so down regulation of some members of the pathway could be expected). So I would say, both up- and down regulated genes would be interesting is such a pathway. However for a kinase pathway, that is regulated by post-translational modifications, looking at up- or down genes wouldn't really make sense. So hard to give one answer, it depends on your definition of pathway.
I was referring more to metabolic type of mechanisms. However, is it correct to say that for post-translational modifications, I should still submit both subsets but separately or that the hypo list cannot provide any relevant information?
What I am trying to say is that up/down regulation of genes is not a simple (one to one) down/up regulation of a pathway. There is more to a pathway than only gene expression. So my advice is to only use pathway analysis as a tool to get more insight in your data, you can try up or down separately and both: up and down as well.
I'd say it depends on biological question you need to address. If you just want to see pathways that are deregulated (whcih is not really informative), you can submit everything together.
And (obviously) if you submit for example only up-regulated genes, you will know for sure the a corresponding pathway is activated/up-regulated, which I think is more valuable information than in previous approach.
I am making the distinction between hyper/hypo sets of genes versus pathway enhancing/inhibiting sets of genes. From RNA-Seq and DE analysis I only know which genes are overly expressed in my treated sample (hyper) and which genes are under-expressed (hypo). But both sets may have a regulation effect. So I am not sure why submitting both sets would show deregulated (aka not regulated) pathways.
By 'deregulated' I havent't meant 'not regulated', but rather 'altered' without distinction if it is inhibition or activation. BTW, which methods are you planning to use for pathway analysis? This might also matter for deciding the strategy of the analysis.
I was referring more to metabolic type of mechanisms. However, is it correct to say that for post-translational modifications, I should still submit both subsets but separately or that the hypo list cannot provide any relevant information?
What I am trying to say is that up/down regulation of genes is not a simple (one to one) down/up regulation of a pathway. There is more to a pathway than only gene expression. So my advice is to only use pathway analysis as a tool to get more insight in your data, you can try up or down separately and both: up and down as well.