Most quantitative assays are analog in nature, with an average signal measured from the total sample volume. The current gold standard for measuring specific DNA amounts is quantitative PCR (qPCR). qPCR utilizes specific complementary oligonucleotide primers that amplify the target of interest in combination with a fluorescent readout, based either on DNA intercalating dyes or hydrolysis-based probes (with hydrolysis probes providing additional specificity for highly stringent assays) .
However, qPCR has a number of technical limitations including the need for assay calibration with standards that are similar in quality to the samples being evaluated. This can lead to an iterative workflow process and challenges to provide qualified standards for comparison. Multiplexing assays may not be straightforward due to potential competition between assays. Also, the theoretical limit of quantification between two samples is 2-fold (corresponding to a single PCR cycle threshold), which is often not sufficient for copy number variation applications with heterogeneous material [2,3]. Typical assay sensitivities range from 1–10% for the detection of mutant gDNA in wild-type gDNA (Minor Allele Frequency or MAF), with some highly developed qPCR assays providing a limit of detection of ~0.1% MAF .
Digital PCR (dPCR) is a novel method for precise quantification of nucleic acids. It uses similar assay reagents as used in standard analog measurements, but counts the total number of individual target molecules in a digital format, enabling many applications that require high sensitivity and have restricted sample availability.
How dPCR works
Digital PCR measurements are performed by dividing the sample and qPCR assay mixture into a very large number of separate small volume reactions, such that there is either zero or one target molecule present in any individual reaction [5,6,7] (Figure 1). This is the fundamental concept for making ”digital” measurements. Thermal cycling is performed to endpoint. Any target-containing compartments will become brightly fluorescent while compartments without targets will have only background fluorescence. A reaction with no target molecule is counted as a 0 (PCR-negative), and a reaction that has one target molecule is counted as a 1 (PCR-positive). When the entire set of divided reactions is counted, the total number of ‘positive’ reactions is equal to the number of original target molecules in the entire volume, and the total number of reactions multiplied by the individual reaction volume equals the total volume assayed. Thus, the absolute concentration of the target is easily calculated as being equal to the total number of target molecules divided by the total measured volume. Uncertainty in this “absolute” measurement comes only from error in the measured volume or the presence of more than a single target molecule in a compartment, so dPCR methods which control for these two factors provide the highest accuracy.
When target molecules are divided into separate reaction compartments, the chances for more than one target molecule to be co-located in the same compartment can be calculated using Poisson statistics [8,9]. When the number of target molecules is significantly smaller than the number of compartments (low occupancy), the chance of co-compartmentalization is small. Poisson statistics can be used either as a small correction factor (at low occupancy) or it can be used to calculate an estimated concentration (at high occupancy). dPCR platforms which divide the sample into a larger number of compartments will have the highest accuracy, by directly counting single molecules (low occupancy). Similarly, dPCR performed using higher numbers of compartments provides the highest sensitivity—with limits of detection approaching 1 in a million, and the widest dynamic range of inputs—over 6 logs.
Multiplexing with dPCR
When in the digital range (where all compartments contain either 0 or 1 target molecule) it is possible to multiplex qPCR assays without concern for competition or cross reactivity, as each target-containing reaction will proceed with the target binding to its primers/probe specifically, whereas no reaction will occur in compartments without targets. Having each molecule in a separate reaction compartment allows both high and low abundance targets to be counted in the same experiment without concern for “swamping out” the low abundance target (since each compartment has at most 1 target, independent of its concentration in the average sample volume). When more than one target is counted (e.g., in a duplex assay format), ratios of the counts for one target relative to another (e.g., mutant allele vs. wild type allele) enable “absolute ratios” to be quantified, using one of the targets as an internal normalizing reference (e.g., how many amplifiable genome equivalents were loaded) that has gone through the identical experiment as the other targets assayed.
In addition, since dPCR is performed as an endpoint reaction (PCR is run to completion before measuring fluorescence), having true single target molecules in isolation allows multiplexing based on probe intensity . By adding the target-specific fluorescent assay at a limiting concentration, a compartment with that target molecule will be PCR-positive, but with a limited brightness at PCR endpoint. To count a second target type, a different target-specific probe with the same “color” is added at a higher concentration. A compartment with the second target will have a brighter signal at PCR endpoint than a compartment with the first target, enabling separate counts for each target. Combinations of both different color probes and different concentration probes can be used to multiplex at higher levels (Figure 2).
Applications for dPCR
Applications where dPCR has been demonstrated or may be well suited include: rare allele detection in heterogeneous tumors or other genetic-based diseases ; liquid biopsies of solid tumor burden using peripheral body fluids ; non-invasive prenatal diagnostics ; viral load detection ; gene expression; copy number variation in heterogeneous samples ; assays with limited sample material, such as single cell gene expression and FFPE samples; DNA quality control tests before sequencing [15,16]; and validation of low frequency mutations identified by sequencing.