170104 limma and the design matrix - npslindstrom/DE-analysis GitHub Wiki

Today I have been spending a few hours trying to troubleshoot and refine the limma script I published here before. The script feels very rough and dirty but it seems to work when doing some testing on known data. I added some lines to write the results as a text file that I can download and examine in excel. At first I was certain that there was some error in the script since the results table looked really strange so I spent way to long trying to find the error and trying different design matrices etc. After Linnea explained to me what I was actually comparing the results seem in line with expectations so overall it seems to do the trick.