Calculate mean and variance of pSTAT5

[aarmey]

This can be on Fig. C2. We can get rid of the inverse Gini coefficient (but keep the Gini coefficient). In similar plots, we should calculate the mean and variance of the pSTAT5 signal. We'll need this for the model fitting later. This would work really well on a two-y-axis plot, so that we can directly compare how mean pSTAT5 varies with its variance.

[aarmey]

These values should also be output to a csv, because the fitting will occur in Julia within a separate repo.