Robert, The point is that the standard way of calculating p-values for significance of expression does not work especially for the hetergenous dataset. So I need to show why the standard analysis does not work while showing co-expression or patterns of expression does give meaningful results. Lana