Noncompetitive RT-PCR relies on the observation that prior to the onset of the plateau effect there is a linear relationship between the quantity of input RNA and final product during PCR amplification. To determine the number of cycles at which this linear relationship occurs, the initial sample of RNA should express high levels of target mRNA. Exogenous target should not be added to the RNA sample; instead, the sample should have endogenous target gene expression that is high, but no more than two to three times the highest levels that would be expected in an actual experiment.
Using this procedure, one can obtain three levels of specificity: (1) amplification of the product with specific primers; (2) correspondence of the actual product size to the original estimated product size; and (3) hybridization of the product with an internal probe not corresponding to either primer. In addition, the low cycle number reduces artifacts such as nonspecific amplifications that could pose major obstacles at higher cycle numbers. Another technique for detecting the product involves the incorporation of labeled nucleotides into PCR products that are then resolved by gel electrophoresis. However, this approach is often associated with trace amounts of unincorporated label that can produce a "trail" of label throughout the lane of an electrophoretic gel. One can also use labeled primers at the beginning of the assay or as a final PCR step with a new internal labeled primer annealed to one strand of the PCR product and extended using Taq DNA polymerase, but both approaches often result in considerable background radioactivity and rarely give signals as well-defined and quantifiable as does the Southern blot.