The effectiveness of the normalization was checked via the Shapiro-Wilk test for normality and the folded form F statistic to test for equality of group variances.
During normalization, we checked the single, array, and probe outliers such that these less reliable data were excluded.
Before including additional normalization we checked whether the presence of batch effects in AML data sets existed.
Before performing the data normalization, we checked the PCR efficiencies of all miRNA assays on the array cards by performing a 10-fold dilution series and subtracting the Ct values of the undiluted sample from the Ct values of the diluted sample.
pRL-TK (Promega) was used as a normalization control to check transfection efficiency.
The performance of the normalization step was checked by quantitative PCR with FastStart Universal SYBR Green Master ROXX) (Roche).
The first stage was a standard 12 degrees of freedom registration to the MNI152 template after which the normalization has been checked manually.
Variation between PCR replicates and deviation of normalization factors were checked and outliers with values outside 2 SDs from the mean were removed.
Expert users carefully evaluated each step to check for normalization or segmentation errors.
DNA quantification, quality check and normalization to 5 ng/μl were done on agarose gel (0.8%) using lambda DNA standard (MBI Fermentas, USA).
