So I've come across a bit of a conundrum. My PI wants me to represent the delta-CT on a linear scale (says she thinks linear, not log), and so wants me to "simply" convert the delta-CT to "2-to-the-negative-power" of the delta-CT. Sounds easy, but I've hit a wall in terms of how to proceed.
The problem arises from the fact that I have multiple biological replicates for two groups, and she wants the mean of each group and it's standard deviation (SD) represented as fold-change of the control RNA.
Here's the catch: Do I convert to the linear value before I calculate the mean and SD, or do I calculate the mean and SD of the delta-CT values and then convert those to the fold-change?
I thought that this would be equivalent, but the two methods give you quite different results (SEE ATTACHED). So which one is the right way to represent this? OR, as may be the case, is the very prospect of representing the delta-CT value this way flawed (if so, please try to explain it to me in a way a non-statistician could comprehend)? All and any advice would be greatly appreciated.