I am working on analyzing data from an Illumina beadchip microarray with samples from the following experimental set-up:
7 tissue donors x 3 treatments x 3 RNA collection time points
Treatments: Mock infection, virus A infection, virus B infection
RNA collected from tissue at time points 1,2 and 3
At least initially, I am planning to analyze the data separately for each time point, so I am not including time as a factor in the design matrices that I am making. I would like to do an analysis where virus A treatment and virus B treatment are paired with the Mock treatment of their common tissue donor and look at mock vs A and mock vs B and A vs B
Problem: Some of the samples failed on the array, including two of the mock treatments, so for those donors, I am left with just virusA and virusB infection.
So far I have been looking at pages 43 and 44 of the limma manual to figure out how to make appropriate design matrices, but was wondering how limma works when a comparison is indicated in the design or contrast matrix and a sample is missing one of the replicates necessary to make that comparison.
Thanks in advance!
University of Washington OBGYN
Fred Hutchinson Cancer Research Center VIDD