I have a question regarding how to do statistics on real-time PCR data.
Let's say I am treating cells with inhibitor X and measuring expression of gene Y, and I want to see if induction of gene Y by some stimulus is reduced by inhibitor X. Let's say I repeat this experiment 5 times, independently, and get the following results, expressed in fold-induction by the stimulus:
untreated: 10, 20, 15, 12, 8
+ inhibitor X: 5, 6, 9, 4, 7
Do I do a t-test based on the fold-change values? i.e. [10, 20, 15, 12, 8] vs. [5, 6, 9, 4, 7]
Or, do I do express the inhibitor data set as a % of the untreated and do a t-test based on [100%, 100%, 100%, 100%, 100%] vs. [50%, 30%, 60%, 33%, 87.5%]? Assuming the data set is paired.
I have also heard that it is more accurate to do statistics based on the delta-Ct values from real-time PCR and not the fold-change data. Any thoughts on this?