Real-time PCR has been used for gene expression analysis for over a decade (Heid et al. 1996, Higuchi et al. 1992). Most gene expression assays are based on comparing two or more samples, and require uniform sampling conditions for valid comparisons. Many factors can contribute to variability in sample analysis, making experimental results difficult to reproduce. Variability is most often related to events upstream of the qPCR assay—namely, the quantity and quality of the extracted sample and reverse-transcription efficiency (Fleige and Pfaffl 2006). Two methods of sample normalization for accurate comparison between genes of interest are normalizing to input RNA and normalizing to a reference gene.
Normalization to Input RNA
Normalization to input RNA implies starting with the same amount and quality of material in each sample. We monitored the expression of four genes in HeLa cells subjected to different treatments. Standard curves were used to determine individual amplification efficiencies. The relative expression of the four genes across treatments was normalized to the amount of input RNA (Figure 1).
Normalization to a Single Reference Gene or Multiple Reference Genes
Although normalizing to input RNA ensures that equivalent amounts of RNA are compared, it cannot compensate for variations in reverse transcription efficiency. Therefore, researchers often normalize target gene expression levels to that of a reference gene. The ideal reference gene does not vary as a function of treatment or condition; however, it is often difficult to identify a gene meeting this criterion (Thellin et al. 1999). This is illustrated in Figure 2, which shows an analysis from the same input RNA, of three common reference genes. Although equal starting amounts of RNA were used, the expression levels of the three genes varied considerably, depending on treatment. These data indicate that genes considered to be housekeeping genes may be expressed at variable levels within an experimental system.
A more accurate normalization strategy has been proposed by Vandesompele, et al. (2002). They proposed selecting a set of genes that display minimal variation across the treatment, determining the geometric mean, and normalizing the target gene(s) to this geometric mean. Figure 3 shows the expression levels of four genes in HeLa cells across five treatments, as calculated by multiple reference gene normalization.
Proper normalization is essential for accurate gene expression studies. To simplify data analysis, iQTM5 and MyiQTM real-time PCR detection systems have analysis software that permits normalization to a standardized input amount, a single reference gene, or the geometric mean of multiple reference genes. Additionally, the software can consider individual assay efficiencies, and combine multiple datasets for a complete gene study.
Click here for more information