Could someone kindly explain to me how to select the Third Quartile (Q3) gene expression values? I'm processing RNAseq data from TCGA studies. For the purpose of further gene cauterization I first performed log2 normalization, than median centering, after which I obviously got a lot of negative values. I can't understand how to perform Q3 by columns, should I calculate Q3 on all data or separately for each column? If the second is true-for one column gene might be above Q3 and for others no. Aren't we loosing the important difference between samples like that? And finally aren't we making data skewed by choosing only values that are >Q3? Sorry, I am a beginner in RNAseq data analysis. Thank you very much in advance for your help!