I have count data from an experiment where we used 8 chemicals tested at 4 concentrations each, both in the presence of estrogen and without estrogen. There were 3 technical replicates and 3 biological replicates. The DESeqDataSetFromMatrix command looks like this (without interactions):
dds <- DESeqDataSetFromMatrix(countData = countData_matrix,
colData = colData, design= ~ Conc_uM + chemical + estrogen)
Problematically, I just learned that some of the chemicals share controls. In other words, chemical #1 and chemical #2 share positive and negative estrogen controls; chemical #3 and chemical #4 share positive and negative estrogen controls, and so on.
So I need to figure out how to represent this in my colData and countData. Does it make sense to copy and rename the countData that I need to be used twice, and create new colData conditions that match the new colnames in the countData? Will this mess up the statistics at all, since there will be times that I will want to pool all the data and just compare all chemicals without estrogen to all chemicals with estrogen?
Thanks in advance,