System-Specific Periodicity in qPCR Data and its Impact on Quantitation
Statistical noise is a feature of every quantitative PCR (qPCR) curve. In principle, two different forms of noise can be encountered: (i) the dispersion of fluorescence signals about their true values cycle to cycle (within-sample noise) and (ii) the dispersion between different replicate qPCR curves at the same cycle (between-sample noise). The most obvious between-sample noise effect is an overall shifting of the qPCR curve on the y-axis, sometimes with scaling of the signal above baseline, such that both the baseline and the plateau-baseline difference may vary reaction to reaction. These effects have long been known, but not fully understood.
In recent work (Tellinghuisen & Spiess, 2014), we showed preliminary results on a published large-scale technical replicate dataset (Ruijter et al., 2013) that revealed unexpected between-sample periodicity for fluorescence values at all cycle numbers. The origin of these periodic patterns in qPCR data remains elusive.
To examine this phenomenon in more detail, we have employed autocorrelation analysis on a larger cohort of published and self-generated high-replicate qPCR data acquired from different platforms and have analysed Cq values with respect to intrinsic periodic patterns when obtained under several different quantitation regimes.
In recent work (Tellinghuisen & Spiess, 2014), we showed preliminary results on a published large-scale technical replicate dataset (Ruijter et al., 2013) that revealed unexpected between-sample periodicity for fluorescence values at all cycle numbers. The origin of these periodic patterns in qPCR data remains elusive.
To examine this phenomenon in more detail, we have employed autocorrelation analysis on a larger cohort of published and self-generated high-replicate qPCR data acquired from different platforms and have analysed Cq values with respect to intrinsic periodic patterns when obtained under several different quantitation regimes.
