Self-correction of cycle threshold values by a normal distribution–based process to improve accuracy of quantification in real-time digital PCR

The digital polymerase chain reaction (dPCR) is a new and developing nucleic acid detection technology with high sensitivity that can realize the absolute quantitative analysis of samples. In order to improve the accuracy of quantitative results, real-time digital PCR emphasizes the kinetic information during amplification to identify prominent abnormal data. However, it is challenging to use a unified standard to accurately classify the amplification curve of each well as negative and positive, due to the interference caused by various factors in the experiment. In this work, a normal distribution–based cycle threshold value self-correcting model (NCSM) was established, which focused on the feature of the cycle threshold values in amplification curves and conducted continuous detection and correction on the whole. The cycle threshold value distribution was closer to the ideal normal distribution to avoid the influence of interference. Thus, the model achieves a more accurate classification between positive and negative results. The corrective process was applied to plasmid samples and resulted in an accuracy improvement from 92 to 99%. The coefficient of variation was below 5% when considering the quantitation of a range between 100 and 10,000 copies. At the same time, by utilizing this model, the distribution of cycle threshold values at the endpoint can be predicted with fewer thermal cycles, which can reduce the cycling time by around 25% while maintaining a consistency of more than 98%. Therefore, using the NCSM can effectively enhance the quantitative accuracy and increase the detection efficiency based on the real-time dPCR platform. Graphical abstract