RNA-seq: biological replicates - (May/26/2013 )
I have planned transcript profiling by RNA-seq of rat tissues before and after a treatment. However, because of economical limitations, I can have only two biological replicates of each tissue for RNA-seq and I have the tissue available from four animals (biological replicates) in each case. In such case for RNA-seq I can (1) randomly select two animals out of four or (2) pool RNAs from each of the two animals together to have two pooled samples for minimizing the biological variation and obtain only treatment-induced effects. Which of these two strategies (1 or 2) do you vote for and why? I have already read few papers on microarray analysis with the similar approach, but I wanted to know general opinion of the Bioforumists.
(Dear administrator: I don't understand if this topic is more suitable for microarray section)
Given the current budgetary situation, it is understandable that you can only sequence two samples as biological repeats. Out of the two strategies, I prefer the first one: sequence randomly selected two of the four mice as separate RNA samples. Pooling two into one reduces your ability to detect inter-individual variation and makes interpretation of result more complicated unless you can do multiplex sequencing as has been done in the literature.
So i have read many of the statisitcal papers and i ended with the following question in RNA-Seq experimental design - if i take samples from 3 mice pool them together (all 3 of the same biological group but from different litters) and repeat this with 3 different mice pool them together as well - at the end i have 2 different barcoded libraries constructed independatly but repearsenting the same biological group. i would of course do the same for the compared biological group (different developmental stage)
Do i have 2 biological repeats (replicates) at each group, or just 2?