Obtaining reliable microRNA expression profiles from high-throughput microarray data is an essential concern in many studies. The crucial step is the raw data normalization that allows unbiased between-array comparisons in the downstream analyses. However, microRNA data differs slightly from "standard" gene expression profiles. In this talk we present a novel microRNA-specific normalization method based on controlled assumptions. The method is evaluated based on quality control criteria and on its ability to detect the biology stimulated in the underlying experiment. These two metrics allow classifying other normalization methods and are applied to an Affymetrix/Exiqon platform comparison, complemented by QRT-PCR validations.