[BioC] A general data analysis question

[Khan, Sohail]

Dear Member,

I have a control and an experimental chip from Affymetrix.  I am thinking of analyzing the data as following:
Rank the genes:  R= E - C / E + C, where E is the probset intensity from MAS5 analysis for the experimental and C for the control.
Choose a cutoff of R > 0.6 as significant.

Could you point out the flaws and drawbacks of this procedure?.....(I am sure this is probably not the best approach).  Could you suggest another method?
Thanks in advanced.

Sohail Khan
Scientific Programmer
COLD SPRING HARBOR LABORATORY