Update mc2 agg methods to subset to only t wells

[madison-feshuk]

Aggregation method bug related to #99. Median and mean methods should only include t wells to allow user to maintain baseline variability

Treatment technical replicate rvals have been median'd and there is just one cval per bio rep. For the controls, it seems they have been median'd but there are duplicate cvals one for each technical rep with same value. After level 3, all normalized control response values are 0 since the mc3 bval method also takes a median.