I did relative comparative qPCR (ddCt) and during data analysis I have selected the most stable gene for endogenous control (from several I have used), also tried multiple endogenous controls. However when I did Global Normalization (Mestdagh et al 2009, Genome Biology 10:R64) for some genes the expression was quite opposite than it was after normalization with endogenous control (I think about genes that their expression was changed with statistical significance p