[BioC] Statistical design help before proceed with biological experiments

[Marcelo Luiz de Laia]

Dear Bioconductors,

I known that this mailing list is about bioconductor software, but, I 
read a few vignetes and, after this, I have a doubt about my future 
experimental design. I am sorry for this.

We biological experiments is:

A bacterium parasite of plants in its wild type state will be compared 
with a mutant in three different times (0, 1, and 3 days after the 
inoculation). In the time zero, bacterial cells, wild type and mutant, 
they will be inoculated in its host. In the all times above, the RNA of 
each one of them will be collected. Soon, for each time we will have RNA 
of both, of the wild type and of the mutant.

MT1 x WT1
MT2 x WT2
MT3 x WT3

Our doubt is: which experimental design to use?

For example:

This one?
RNA mutant in T1 compared to RNA mutant in T2 compared to RNA mutant in 
T3 compared to RNA mutant in T1 (loop)

        MT1
MT3            MT2

Or this one?
One pool of wild type RNA would be used as reference (control)? In this 
case, RNA mutant in MT1 will be compared to WT RNA pool, RNA mutant in 
MT2 will be compared to WT RNA pool, and RNA mutant in MT3 will be 
compared to WT RNA pool.
MT1
MT2       WT (pool)
MT3

Or we compare, in each one of the times, the mutant RNA versus the wild 
type RNA and later we proceed the
identification the difference of expression between the times using, for 
example, limma?

MT1         MT2        MT3
  x               x               x
WT1        WT2        WT3

We are interested in verifying the differential expression to the long 
one of the time.

I ask a statistician here, but he dont have expertise for this situation.

Any suggestion is very apreciate.

Thansk a lot

Marcelo