Trying to figure out reason for mutant phenotype using RNA seq

Question

[robertkmaeda]

Hello,

Sorry everybody, but I am a complete beginner here. Our lab recently performed an RNA-seq experiment to try to understand the phenotype of a particular mutant in Drosophila. I now have a list of up and down-regulated genes (analysis performed by our genomics facility). I have been staring at this list for a while now, trying to put together groups of genes that might account for the phenotype. Being a fly geneticist, with a long history of working on individual genes with 100% penetrant visible phenotypes, I have been just looking at the list an trying to find trends, then checking those trends using immunostaining, or whatever other method is appropriate. This is tedious and has led to only minor breakthroughs. Is there a way I am supposed to go about this? Of course I have read hundreds of papers where the authors have displayed Venn diagrams of GO analysis, but I do not know how to do it. I mean I can get GO Venn diagrams using Panther or other programs, but I am never sure how to tell if these are meaningful or represent just a random set of genes. Also, Panther does not seem to look at the level of misregulation. Is there a better program to use?

In any case, I am hoping somebody here can point me in the right direction as to where I should start. Some good references and programs?

Thanks,
Rob

[Anima Mundi]

GO analysis sure helps but personally I do not think this is going to be the "meat" of your pipeline. In your shoes, I would somehow define a list of differentially expressed genes (say, the best 50) and perform histology (ISH/immunostaining) on them (both on mutant and wild-type). This will define a narrow subset of genes with remarkable/remarkably changed expression patterns. On this subset I would perform functional studies (namely, mutants for each selected gene, resque experiments etc.). You will end up, aside from the wide set of differentially expressed genes provided by the RNA Seq, with a few robust connections between your mutated locus and some effectors. By the way, did you perform RNA Seq both in a loss-of-function and gain-of-function mutants? This would have improved the quality of you initial "top-50" list, making the tedious downstream wet lab work more valuable.

Answer 1

[h.mon]

Gene set enrichment analysis is one possible avenue to explore. Drosophila melanogaster has a ton of online resources, you could try to define gene sets based on these resources and see if your DGE list is enriched for something. As a more concrete example: you could define gene sets based on tissue of preferential expression (taken from FlyAtlas), and use GAGE to see if any of these custom-defined gene sets is significantly enriched for your DGE table. I've worked with Drosophila a while ago and I am not up to the most recent databases, but I am sure you will find several interesting data sets if you look around. Then you should / could have stronger candidates to work on.