Amplification Enumeration: dPCR or qPCR?

The polymerase chain reaction (PCR), as imagined by Kary Mullis and practiced for more than three decades, is not inherently quantitative. But variants such as quantitative real-time PCR (qPCR) and digital PCR (dPCR) can make it so. Both qPCR and dPCR have their place, with a large overlap between them. Here we look at some reasons, besides history and existing infrastructure, a researcher may choose to do one rather than the other.

qPCR and dPCR explained

We’re all familiar with PCR, in which a DNA target is exponentially amplified between flanking primer sequences. In theory, only a single template is necessary to generate millions of copies of that intervening sequence in a small number of cycles. Traditionally, that process is evaluated by gel electrophoresis once the reaction is over—a process called “endpoint PCR.”

Endpoint PCR is the parent of both qPCR and dPCR; therefore, both technologies share its DNA (pun intended). In the case of qPCR, though, rather than running the contents of the reaction out on a gel, researchers observe its progress in real time. The doubling of amplicons between subsequent cycles is detected as a concomitant increase in fluorescent signal, using a variety of detection technologies involving quenched or cleaved sequence-specific probes, for example, or dyes that fluoresce when bound to double-stranded DNA.

As the qPCR reaction progresses, the fluorescent signal is plotted as an amplification curve. The point at which the curve intersects a predetermined level, indicating that it has risen above background noise, is called the “threshold cycle,” or Ct, which can be used to back-calculate the initial target concentration. But because many other factors also can influence Ct, the value provides only relative quantitation; a standard curve must be run as well to achieve absolute quantitation.

On the other hand, dPCR partitions a single reaction into hundreds, thousands or even millions of individual endpoint PCR reactions. This can be performed in microtiter plates (used in home-brew methods), microfluidic chambers (in the case of Fluidigm’s and Life Technologies/Thermo Fisher Scientific’s systems) or oil droplets (Bio-Rad Laboratories and RainDance Technologies). Fluorescence is used to detect the presence of amplicons, with the ratio of amplification-positive to amplification-negative partitions used to back-calculate the target concentration in the initial sample. In this case, absolute quantitation is achieved without reference to a standard curve. But in practice, dPCR has a more limited dynamic range than qPCR, so it’s necessary to begin with at least a rough idea of the starting concentration.

The gold standard

Like many researchers, Thermo Fisher Scientific senior product manager Iain Russell considers qPCR the gold standard for a wide variety of genetic analysis studies, citing its well-proven track record, simple workflows, low per-sample cost and high throughput. Russell points out that qPCR’s vast catalog of publications over the years speaks to its “ability to deliver sufficient performance to answer many of the questions of interest to the research community.” He adds that “this technology maturity reflects the decades of investment that have been made in the qPCR platforms.”

Among qPCR’s broad range of applications are gene-expression analysis, genotyping, pathogen detection, viral quantitation, DNA methylation analysis and high-resolution melt analysis, says Jennifer Dennis, senior product manager for gene-expression marketing for Bio-Rad Laboratories. “qPCR is particularly well suited to gene-expression analysis where expression level is normalized to a reference gene,” and in which samples from different experimental conditions are compared.

To ‘d’ or not to ‘d’?

According to Ramesh Ramakrishnan, director of microbiology at Fluidigm, dPCR serves more as a complement to qPCR than as a replacement. “There are only a small set of applications that tend to favor the particular characteristics of dPCR,” he says.

The ability to provide absolute quantification, Ramakrishnan says, gives dPCR the ability to identify rare events (the proverbial needle in a haystack) as well as to precisely define copy-number variation.

Using dPCR, researchers can “find that rare variant that you couldn’t really get with real-time [PCR] no matter what, because you’re not partitioning the sample in the same way,” explains Lawrence Jennings, director of molecular diagnostics laboratories at Lurie Children’s Hospital of Chicago and associate professor in pathology at Northwestern University.

In qPCR, a rare variant will be in such a low concentration compared with related species that its signal will likely not rise above background. dPCR helps to spread out the contaminating DNA so it’s less concentrated, says Stephanie Fraley, assistant professor of bioengineering at the University of California, San Diego. Similarly, inhibitors that may plague qPCR reactions can be diluted out in dPCR. So a rare variant present in at least one dPCR partition, (relatively) free from closely related species, will be amplified and detected.

“The very limit of sample differentiation of qPCR is a doubling of sample input, or one Ct value,” notes Fritz Eibel, senior vice president of strategic marketing for RainDance Technologies. So although qPCR can distinguish two copies of a gene from one, a digital approach can distinguish seven copies from eight.

Not the be-all, end-all

Powerful as it is, dPCR isn’t necessarily the be-all and end-all of PCR quantitation. “In a lot of ways, qPCR is much better still for high throughput, especially compared to chip-based dPCR platforms. It’s much more time- and materials-intensive to load multiple chips,” says Jocelyn Henline, research technologist at the Johns Hopkins School of Medicine’s Genetic Resources Core Facility, who uses the Thermo Scientific QuantStudio® 3D dPCR system. Admittedly, the recently introduced autoloader has made chip loading much more user-friendly, Henline says, yet, “you can probably set up a 96-well [qPCR] plate in the same amount of time it would take you to set up 24 digital samples.” It’s also necessary to get a rough estimate of the target concentration prior to running dPCR, which is oftentimes determined by running qPCR.

Jennings, who uses Bio-Rad’s original QX100™ ddPCR™ system, notes also that “there is hands-on time in the beginning and end [of the dPCR setup process] that you don’t have with qPCR, namely making the droplets and acquiring the reads,” which may be partially compensated for by not having to produce a standard curve.

PCR sometimes generates random false positives in no-template controls, which can have serious consequences in the clinical realm [1]. These can be teased out by observing the amplification behavior of the reactions—the Ct curves. But, Ramakrishnan argues, qPCR does not offer the requisite statistical power to detect rare events. Fluidigm’s solution in the Biomark HD is a hybrid in which “we actually do dPCR but also collect the real-time curves for each amplification. So it’s still dPCR, which means that it’s positive or negative. We have the ability to go back to our images and say which one of those so-called positive signals is actually based on a true specific amplification.”

The RainDance’s RainDrop Digital PCR system generates up to 10 million picoliter-sized droplets and interrogates them in a relatively slow, expensive protocol. But the company is not trying to compete with qPCR; instead, it is focusing on the liquid biopsy market, says Eibel. “The RainDance dPCR platform enables scientists to interrogate liquid biopsy samples at a much greater depth”—one mutation in a million, vs. qPCR’s 1%.

Of course, when it comes right down to it, researchers likely don’t care how they quantify target molecules, just that they do it—and accurately. In that sense, the qPCR/dPCR debate is somewhat academic. Says Henline, “If they’re looking at something they’re not going to be able to get a good quantification [of] using the qPCR, then they’ll say to do dPCR.”

Reference

[1] Kiselinova, M, et al., “Comparison of droplet digital PCR and seminested realtime PCR for quantification of cell-associated HIV-1 RNA,” PLOS ONE, 9:e85999, 2014. [PubMed ID: 24465831]

Image: RainDance Technologies