I am interested in selecting genes associated with a linear outcome (drug dose) and including adjustments for multiple testing/false discovery. As an example, the Excel version of SAM, for example, has an implementation where the delta statistic is based on the linear regression coefficient. Am I correct, this response type is not supported in either the development or current release of siggenes? I re-reviewed siggenes and reviewed the limma package (with which I am less familiar), as well as some of the on-line Bioconductor tutorials, and it was not clear to me that there was an application that I could select as a "standard" method (or at least a method with published references) of selecting genes associated with a linear response. This strikes me as a question similar to one that has been asked many times before, but on review of the mail archives, I couldn't find an answer. I am sure I am missing something simple. Any suggestions? Neil Hayes UNC Chapel Hill