Dear R users, I am have a question on SAM analysis - two class unpaired. I am not sure how exactly SAM calculates the fold change for logged2 transformed data. The output produced by SAM - rbind(siggenes.table$genes.up,siggenes.table$genes.lo) - has numerator, denominator, d-statistic, fold change and q value. Now if the data is log2 transformed then numerator = "difference of means" and the fold change should be = 2^(numerator). But this is not always true. For some genes it looks OK but for others the fold change calcuated by SAM is far different from that of 2^(numerator). So I think SAM considers a slightly different approach to estimate fold change. Do you know excatly how SAM estimates the fold change? Would really appreciate if you can help me with this. Thanks, ~Nikhil. Nikhil Garge Research Programmer Center for Biotech and Genomic Medicine Medical College of Georgia CA 4100, Augusta GA 30912 [[alternative HTML version deleted]]