11.10.2015
This is experiment #232.
The heatmap for each time point did not quite reflect the eigenvector.
Greg says that this is because each group of genes was scaled to the max value in the group.
That is, for the 40 minute genes, the FPKMs at 0 minutes were scaled to the max FPKM at 0 min.
Greg wrote the recolor.py script to add the max value that I want to each track file. 
It is extremely important to use the same max for all time points and all gene groups.
Greg also recommends that I remove genes that are >130 in mitosis in each of my filters.
-would eliminate the need to use an addend (which I didn't do anyway in #229)
-would ensure that only genes that are off in mitosis are included in the analysis
I will do this in experiment #233.
For now, just recolor and repeat heatmaps.