161207 Post Seminar 2 - npslindstrom/DE-analysis GitHub Wiki

So today we had the second project seminar. We presented our thoughts regarding the project and how far we have come. Since our data was in such good shape when we recieved it a lot of work had already been done for us. Since the project is supposed to be a few weeks long we suggested to our supervisor Olof that we needed some help to expand the project scope.

Linnea managed to get DESeq2 up and running and told us that it found more significant hits as compared to DESeq1 when used on the same data, which is curious.

During the meeting with Olof we decided on three approaches we could take to expand the project:

1. Redo the RNA-data pipeline with different a different aligner software, i.e. STAR instead of Tophat2 for instance.

2. Try to do some sort of multiple comparison of several conditions at once, this is supposedly possible with DESeq and LIMMA but I am not sure I understand what this means really so that is something for me to look into.

3. Expand on the analysis of the DESeq results, do some sort of functional analysis regarding the differentially expressed genes we find based on protein domains etc.

I feel like I am most interested in point 1 just to get a sort of "bottom up" understanding of how the analysis is performed and the intricacies of it. I think this is what I will work on tomorrow. We need to determine which aligner we want to use and how to run it as well as determining which parts of our data we want to redo.