Hi, I am analizing a set of microarrays with SAM strategy using SIGGENES package. The microarray data is the following: Name Description FLAG <- 7 arrays of an IP over RNA from a transgenic animal with a TAG in the protein of interest for the IP Wt <- 7 arrays of an IP over RNA from an Wild type animal WAFLAG <- 3 arrays of total RNA from the Whole Animal with a TAG in the protein of interest WAWt <- 3 arrays of total RNA from the Whole Animal I am interested in those transcripts that are significatively change in FLAG vs Wt samples, as those transcripts would be the hipothetical targets of my protein of interest. However, I would like to avoid noise from the FLAG epitope of my protein in the total RNA samples. To these end, I compared WAFLAG vs WAWt samples. So, here is my problem. I am not sure what to do in order to avoid the noise I described above. What I was thinking is to substract the fold changes generated in WaFLAG vs WAWT to the expression values to the FLAG and Wt samples before make the SAM analysis over these FLAG and Wt. Also I don´t know if there is a way to construct in some way the CL vector in siggenes. Any suggestions will be highly appreciated. Another thought I have, is that only a have to worry about those genes that significatevely changed in both contrasts, WaFLAG vs WAWT and FLAG and Wt. Mark [[alternative HTML version deleted]]