todo

use absolute values in the correlation matrix

ask question about whether clustering makes sense

plot(as.dendrogram(..,horiz=TRUE))

WGCNA to colour the dendrograms?

What is difference going from clustering with k = 2 to k=3?

Discuss whether to remove the outlier or not? Is this mentioned in the paper

Get the order of the genes (rows) in the heatmap?

heatmap of a list of genes, rather than cell-cycle pathway

other packages for heatmap (heatmap.2, heatmap.plus....) that give a density plot, or multiple sample colours

dendextend - comparing dendrograms

inSilicoDB - combining multiple datasets

summary(pca)

boxplot of pca versus clinical variables

head(pca$rotation)

no ggplot2 example for pca