[BioC] rma on affy experiment with different cell lines

[Dipl.-Ing. Johannes Rainer]

hi,

has anyone experience how good rma performs on affymetrix experiments where the
data from different cell lines are compared against each other?
i have 4 chips, one with cell line 1 with treatment, one with cell line 1
without treatment, one with cell line 2 with treatment and one with cell line 2
without treatment. the question is, should i normalize all them together using
rma, or should i normalize cell line 1 (+/- treatment) and cell line 2 (+/-
treatment) separatly.
As i have no replicates, i am hoping, as the cell lines are not that different,
that the model parameters are better fitted into the data.

thanks, jo