Improved strategies and optimization of calibration models for real-time PCR absolute quantification

Abstract

Real-time PCR absolute quantification applications are becoming more common in the recreational and drinking water quality industries. Many methods rely on the use of standard curves to make estimates of DNA target concentrations in unknown samples. Traditional absolute quantification approaches dictate that a standard curve must accompany each experimental run. However, the generation of a standard curve for each qPCR experiment set-up can be expensive and time consuming, especially for studies with large numbers of unknown samples. As a result, many researchers have adopted a master calibration strategy where a single curve is derived from DNA standard measurements generated from multiple instrument runs. However, a master curve can inflate uncertainty associated with intercept and slope parameters and decrease the accuracy of unknown sample DNA target concentration estimates. Here we report two alternative strategies termed ‘pooled’ and ‘mixed’ for the generation of calibration equations from absolute standard curves which can help reduce the cost and time of laboratory testing, as well as the uncertainty in calibration model parameter estimates. In this study, four different strategies for generating calibration models were compared based on a series of repeated experiments for two different qPCR assays using a Monte Carlo Markov Chain method. The hierarchical Bayesian approach allowed for the comparison of uncertainty in intercept and slope model parameters and the optimization of experiment design. Data suggests that the ‘pooled’ model can reduce uncertainty in both slope and intercept parameter estimates compared to the traditional single curve approach. In addition, the ‘mixed’ model achieved uncertainty estimates similar to the ‘single’ model while increasing the number of available reaction wells per instrument run.