Figure 2

Schematic of analytic approaches with associated figure and table numbers. We took a two-pronged approach in analyzing the vast array of data generated from RNA-seq. First, for VMN vs. H comparisons, we used a cyclic loess method, fastlo, to eliminate a non-linear tissue-dependent skew between datasets (left, red). We then treated time point samples as replicates for each tissue for differential expression with edgeR. Second, we followed the Trinity-supported downstream analyses of clustering differentially expressed transcripts and focused these analyses on clusters with biologically relevant expression patterns across the day and season (right, blue). Additionally, we analyzed expression and correlation patterns of transcripts belonging to six broad gene functional categories. *Genes were grouped in broad categories regardless of whether they were differentially expressed.