Hi,
So maybe this is too far-fetched question but I have a question. I was thinking about whether or not this is possible...Let's say you are studying disease X, where people have discovered multiple regulators from loss-of-function screening in animals or cells. This is a common occurrence, where after the screening, people identify "novel" genes for that are regulators. And sometimes, it turns out that those "novel" genes were studied in other areas but no one suspected them to play a role in this context.
I guess my question is, statistically or computationally speaking, is this sort of "screening" feasible given just the available information (literature, data sets, etc)? That is, if you take a core list of known regulators for disease X, can you discover "novel" regulators? I know this has been done but I feel like those hits are never surprising - they are very biased towards more studied genes so the hits that are not as well-known (which turn out to be the ones that people actually discover) do not show up.
Just a thought.
Thanks for the links of the papers although first two were not as useful to my question. I get what you mean by looking at the motif and pathway analysis based on experimental data (IPA, MetaCore, etc.) but is it still possible to do this:
Say I have two genes A and B that both regulate (not necessarily transcriptionally - it could be a number of different ways) gene C, and upstream of gene A in one tissue is a gene D. Now, based on the known experimental condition and prediction, could we make an association between gene A and B (through their both involvement with gene C) and somehow identify gene D as a "potential" regulator of gene B as well? Obviously, this will depend on tissues, but I'm just wondering whether this is possible because known computational search will not pick up gene D as a regulator of gene B because this could be conditional based on gene A. Hope that makes sense...