Although qPCR sounds quantitative when you say it (it’s in the name!) and looks quantitative when you analyze the results, its potential for accuracy, sensitivity, and dynamic range lags well behind that of digital PCR (dPCR). There are several limitations to any traditional PCR platform, either in endpoint or real-time assays. One example is that exponential amplification always produces copy numbers in multiples of two. Another is that target gene expression must be derived by comparison to a standard curve generated by amplifying a reference gene. Two-fold measurements offer a disconnect from reality, and normalizations to housekeeping reference points can assume incorrect kinetics of amplification. Therefore, qPCR is an inadequate technique to achieve some high-sensitivity objectives, such as measuring copy number variants of disease biomarkers, or identifying rare point mutations against an overwhelming background of wild type allelic expression.
Alternatively, dPCR provides a substantial upgrade by detecting single molecules, quantifying their abundance, and obviating the need for generation of standard curves, which can consume excess time and reagents and introduce amplification biases. With increased sensitivity and accuracy, dPCR can additionally be applied to cataloguing water-borne microbial pathogens refractory to traditional methods, and therefore accelerate and validate crucial use and recycling decisions.
Digital PCR: Enhanced sensitivity and accuracy through partitioning
Work by investigators, including Kary Mullis at Cetus Corporation, in the early days of PCR enabled detection and amplification of a single copy of the λ-globin gene. Their approach was enhanced by Pamela Sykes and colleagues to allow copy number quantification, in a procedure first referred to as digital PCR by Bert Vogelstein and Kenneth Kinzler. The critical step Vogelstein and Kinzler contributed was using oil emulsion to maximize separation of individual PCRs into minimized volumes. By partitioning identical reactions into thousands of tiny individual microreactors, one can treat them simultaneously as a population, with a demographic distribution of on/off (or digital) signals corresponding to starting material with one or zero target molecules. Signals in this case consist of emitted fluorescence via accumulation of a TaqMan-style probe. Although each individual readout is analogous to what emerges from a larger single-well qPCR, dPCR has an endpoint readout, instead of a contemporaneous threshold cycle that varies between samples.
To generate quantitative data, dPCR employs several assumptions and statistical conditions: 1) each partition begins with an equivalent, random probability of containing target molecules; 2) all partitions have the same volume; and 3) one can therefore use binomial probability and Poisson distribution to extrapolate absolute numbers of independent “events” occurring at a constant rate during a fixed period. Because of these relationships, there is an optimal partition occupancy rate, λ, that drives considerations, including partition number and volume that can impact the dynamic range and accuracy of dPCR. For example, the Wilson method of direct calculation incorporates the probability that a partition is empty, the total number of partitions, and a confidence interval of 95 percent. In this algorithm, a λ of 1.6 is optimal for an assay with 10,000 partitions and corresponds to about 20 percent vacancy. Optimized thusly, dPCR can detect variations within a linear range of less than 30 percent, an obvious improvement over the classic two-fold limitation. However, deviations toward much lower or higher occupancy can skew accuracy, and bracketing an assay’s median intrinsic dynamic range promotes the highest fidelity. One way to ensure this is to develop dPCR partitioning strategies that allow for different volumetric ranges across subsets of partitions, with larger volumes promoting sensitivity, smaller ones enabling optimal detection limits, and medium volumes for precision. Finally, the incorporation of microfluidic devices into dPCR workflows adds an aspect of massively parallel throughput that can diversify analytic potential and improve accuracy by reducing pressures on expense and reagent use to allow unfettered reiteration of technical replicates.
“While bacterial pathogens grab many of the headlines, waterborne viruses can be silent killers because they often occupy wastewater at levels below facilitative thresholds of detection.”
Leveraging dPCR to characterize and solve wastewater problems
In developing nations, there is often economic pressure to mitigate water waste through reuse. There are analogous pressures in historically privileged areas newly plagued by drought or population influx to conserve household, commercial, and municipal waters downstream of their initial use. In both cases, there is a quandary over whether these waters can be recycled safely, particularly for agricultural purposes. However, policy makers have already implemented many such programs, with active procedures and regulations taking place well ahead of a detailed understanding of what’s in the water before reusing it. While bacterial pathogens grab many of the headlines, waterborne viruses can be silent killers because they often occupy wastewater at levels below facilitative thresholds of detection. Moreover, small populations are often infected in a localized manner, with houses, neighborhoods, and cruise ships all serving as highly variable foci. However, the ubiquity of enteric viral pathogens is almost prosaic in nature. Norovirus is the most common source of viral acute gastroenteritis. Human adenoviruses are omnipresent, non-seasonal, UV-resistant, and can cause fatal infections in immunocompromised people, but also respiratory, mucosal, and gastric issues in otherwise healthy people.
Because qPCR is an insufficient platform to assess enteric pathogens in wastewater, investigators have developed robust epidemiological models to derive the statistical likelihood of and quantity of their presence from overall infection rates, and to predict whether ultrafiltration, membrane bioreactors, and other treatments are sufficient before downstream reuse (the developing consensus: they are not). Recently, dPCR has begun to serve as empirical validation for these methods, and subsequently to extend its own legitimacy in ever-improving waves of sensitivity and accuracy. The incorporation of array- and microfluidic-based droplets (ddPCR) has allowed researchers to assess log removal values in ground water downstream of agricultural runoff, graywater, blackwater, and mixed wastewater for genogroup I and II noroviruses, and for various adenoviruses including HAdV41, a common diarrheal agent and bellwether for water treatment safety. Commercial manufacture of plate-based dPCR instruments continues to improve throughput, which will facilitate making decisive and accurately informed policy changes that can broadly impact human health.