Hi

Is it compulsory that differentially expressed mirna should be found only from matched samples (sequencing data from both normal and cancer tissue in the same patients) of TCGA data ?? Can i combine both matched as well as unmatched samples to find DEmirna?