Predicting functional similarities based on gene expression time series plots?

In order for predicting similarities from time series curves to give meaningful results, do plots for the same GO term (i.e. MF, BP or CC) or pathway need to be more correlated to one another than to the remaining genome?  What if I see two major but mutually exclusive themes for the same GO term / pathway?  Which correlation measure should  I use for this purpose in R?

More Thomas Hahn's questions See All
Similar questions and discussions