Revisiting the sigmoidal curve fitting applied to quantitative real-time PCR data

Amplification of a cDNA product by quantitative polymerase chain reaction (qPCR) gives rise to fluorescence sigmoidal curves from which absolute or relative target gene content of the sample is inferred. Besides comparative C(t) methods that require the construction of a reference standard curve, other methods that focus on the analysis of the sole amplification curve have been proposed more recently. Among them, the so-called sigmoidal curve fitting (SCF) method rests on the fitting of an empirical sigmoidal model to the experimental amplification data points, leading to the prediction of the amplification efficiency and to the calculation of the initial copy number in the sample. The implicit assumption of this method is that the sigmoidal model may describe an amplification curve quantitatively even in the portion of the curve where the fluorescence signal is hidden in the noise band. The theoretical basis of the SCF method was revisited here for defining the class of experimental amplification curves for which the method might be relevant. Applying the SCF method to six well-characterized different PCR assays illustrated possible pitfalls leading to biased estimates of the amplification efficiency and, thus, of the target gene content of a sample.