I am conducting a differential expression analysis of RNA-seq data in edgeR. Our experiment has four factors (GLM = the population biological replicates come from X treatment conditions in the lab X interaction). We have between three and four biological replicates for each combination of factors described above, and a common dispersion of 0.6778. When we contrasted eight biological replicates from one population, four subject to each treatment condition, we found that only 10 of the 25,163 tags were differentially expressed. For a similar contrast, in a similar study on the same organism (eastern oyster Crassostrea virginica) in the same system, with six biological replicates per-factor, thousands of DE tags were discovered. I am wondering if the low number of DE tags and high dispersion is a result of low sample size, or inherent qualities of the organisms used.
