Selection of superior reference genes' combination for quantitative real-time PCR in B-cell lymphomas

Normalization of real-time quantitative PCR data to appropriate reference genes is crucial to accurately interpret results. Many genes commonly used as reference standards do not perform as expected, depending on cell type and experimental design. In our previous work, we addressed the issue of suitable reference genes for lymphoid tissue and successfully applied the normalization factor-based approach to discriminate lymphoid malignancies according to their cyclin D1 mRNA level. Here, we addressed the problem of reference gene selection and sufficient number on an enlarged sample set with seven candidate genes. The experimental set included 165 samples of spleens, lymph nodes, and peripheral blood mononuclear cells from patients with different types of non-Hodgkin lymphomas along with non-neoplastic lymphoid specimens. For the first time, we compared all major stability ranking algorithms of Visual Basic for Applications (VBA) applets geNorm, BestKeeper, and NormFinder and tested candidate reference genes on a large and heterogeneous set of fresh clinical lymphoid samples. We concluded that a normalization-based approach using three reference genes (YWHAZ, UBC and ACTB) allows for robust reduction of the variance in real-time PCR results and that the further addition of reference genes does not improve data normalization. This creates a clinically applicable tool for PCR-based lymphoma diagnostics.