Hi folks, I've recently started a new job in a new country and one of my first tasks is going to be to validate a mouse cell model (ME3), primarily using RNA-Seq. We're working with adipocyte differentiation, and we've run a time-course RNA-Seq of the differentiation procedure (About 7 days). I have found plenty of GEO datasets which include a similar cell culture RNA-Seq procedure, but from primary mouse pre-adipocytes, and it is against these that I plan to validate our model.
There are plenty of great ideas regarding validating models via static-time RNA-Seq experiments, and this is definitely where I will begin. My thought is that the first half of the process will be graphing gene expressions to show similarity or dissimilarity between our day 0 and day 7 data and those day 0 and day 7 sets I have found at GEO. Hopefully, this will show that at least our pre-adipocytes taste like pre-adipocytes, our adipocytes taste like adipocytes, our snozberries taste like... I am also planning on clustering all of our expression sets together on as many common genes as I can, to get a good dendrogram visualisation of how this model compares to primary cell culture, primary tissue extract and other, more trusted cell models (3T3).
My question is this, after I've done this initial validation (And hopefully found enough reason to continue), is there a lot of scope to then validate further based on the time-course dynamics of the sets. Is the best way to do this to simply compare all of the intermediate days (Day 1 to day 1, day 2 to day 2, etc.)?
One of my considerations is to use WGCNA to form co-expression clusters from each data set, and then test these for module preservation (Are genes being allocated into the same/similar modules in each set?) and for network similarity (Are the intra-module connections between the genes the same/similar in each set?).
If anyone has any ideas that might be useful, I'd really appreciate it.