by Caitlin Smith
Digital PCR (dPCR) is a dream come true for anyone who has struggled with references and standards in making quantitative PCR (qPCR or real-time PCR) measurements. Though the techniques are related—they both aim to estimate the amount of nucleic acid in a starting sample—there is one important difference. Quantitative PCR allows you to measure the amount of nucleic acid relative to an internal control or reference standard, whereas dPCR lets you count individual molecules of DNA for an absolute quantification of starting material. Digital PCR is valuable for applications including rare allele detection, identifying mutant genes against a wild-type background, discriminating between low-fold copy numbers, validating next-generation sequencing results and more recently, gene expression analysis.
Digital PCR works by dividing or partitioning a sample into many individual qPCR reactions. Measurements with say, a fluorescently labeled probe, will reveal that some of these reactions contain the target molecule and others don’t. This generates an absolute number of target molecules in the sample, without reference to standards or endogenous controls. That doesn’t mean dPCR is always the best tool, however. “Digital PCR can have a significantly easier workflow than conventional qPCR but can be lower throughput,” says Ramesh Ramakrishnan, director of molecular biology at Fluidigm. “Conventional qPCR methods can be of higher throughput than current digital PCR technology, whereas digital PCR is potentially more specific, particularly when searching for rare mutations.” Copy number variation (CNV) analysis is a good example of the value of dPCR. “If you’re running a CNV assay, where you’re trying to distinguish between, say, one or two copies of a target in the genome—you could do that in a well-optimized qPCR experiment,” says Richard Kurtz, division marketing manager for the Gene Expression Division at Bio-Rad. “But when you get to higher-order levels of variation when you’re trying to distinguish between, say, four and five copies, or five and six copies, it’s pretty much impossible with qPCR unless you are running a tremendous amount of replicates.” As dPCR evolves to achieve greater throughput, though, it might begin to supplant qPCR in more areas of bioresearch.
Partitioning the sample for greater accuracy
One of the most important determinants of accuracy in digital PCR is the partitioning of the sample. A single sample is divided into many smaller portions—so small, in fact, that each tiny partition usually contains either one copy of the target nucleic acid or none. Using fluorescently labeled probes to measure the presence or absence of target in each partition yields an estimate of the absolute number of molecules in the original sample. “The beauty of digital PCR is the sheer simplicity of the approach—a positive signal indicates the presence of a target, while the absence of signal indicates the lack of a target,” says Ramakrishnan.
Life Technologies offers TaqMan® OpenArray® Digital PCR Kits that run on an OpenArray® Plate, a microscope slide-sized stainless steel nanofluidic plate. The plate contains 3,072 “through-holes” etched into the surface, which act as independent PCR reaction wells. The through-holes are arranged as 48 subarrays of groups of 64 wells. The outer surfaces of the plate have a hydrophobic coating, and the inside of the through-holes has a hydrophilic coating. This separates all the tiny partitions so they cannot touch one another and mix.
The plates can be run on the existing OpenArray® Real-Time PCR System and on the new QuantStudio™ 12K Flex instrument for dPCR, both from Life Technologies. The QuantStudio 12K Flex supports 96-/384-well plate formats, TaqMan® Array Cards and mid-density OpenArray Plates. “With the ability to run up to four OpenArray Plates simultaneously, the QuantStudio 12K Flex instrument can generate 12,228 independent PCR reactions in a single run, with PCR amplification data being collected in real time,” says Iain Russell, senior product manager for qPCR at Life Technologies. “Due to the open nature of OpenArray Digital PCR Plates, the number of digital reaction replicates required to generate a single answer can be easily tailored to meet application needs, balancing the need for higher system performance with achieved throughput and experiment cost. It is feasible to run as little as a single subarray per sample, enabling 192 digital PCR answers to be generated in a single instrument run or, conversely, a single sample can be run across all four plates (192 subarrays), maximizing replicates and overall system performance.”
Unlike Fluidigm and Life Technologies, which use microfluidic chips and channels, Bio-Rad (and soon, RainDance Technologies) uses tiny droplets for partitioning. “Our unique benefit is that we are able to generate these very uniform, very reproducible one-nanoliter droplets. One benefit of this is that we have 20,000 droplets per sample, so we have many more partitions,” says Kurtz. Other systems range from about 760 to 3,000 partitions. “We have a lot more partitions, and having more partitions gives us a lot more precision in our analysis.” Bio-Rad’s system can accommodate up to five targets per droplet, because they have so many partitions; this provides a larger dynamic range of detection. “Our system can handle anywhere from 1 to 100,000 target copies. So we don't have to do a serial dilution of the sample before running it,” Kurtz adds.
RainDance will release its new system later this year, taking advantage of its RainStorm microdroplet technology. “The RainStorm microdroplet-based method utilizes up to 10 million reactions in five-picoliter droplets (or 200 more reactions per one-nanoliter droplet) per sample to enable detection of rare mutations in a background of wild-type at levels better than one in 200,000. We are also able to perform much higher multiplexing due to use of varying concentrations of fluorescent probes,” says Rena McClory, marketing director for digital PCR at RainDance Technologies. “This system [is] a powerful genomic analysis platform for new research in cancer, including rare variant detection, absolute quantification of biomarkers, tumor profiling and the ability to monitor residual disease. Researchers should consider the level of sensitivity needed and the degree of multiplexing needed, not only now but in the future. For example, with the RainDance digital PCR platform, we expect sensitivity that would allow a user to see one mutation in 100,000 to 200,000. Most current systems struggle to see one in 5,000 mutations.”
Flexibility built into systems
Digital PCR systems are coming out with additional features that can lighten your load. For example, although most digital PCR workflows begin with either a genomic DNA or cDNA sample, some dPCR systems can use RNA as a starting sample. For workflows which begin with an RNA sample, such as gene expression or viral RNA detection research, the processing and reverse transcriptase steps upstream of cDNA generation usually entail separate, preparatory steps. “This increases the number of hands-on steps, negatively impacts the workflow and raises the specter of contamination,” notes Ramakrishnan. “In order to deal with these issues, we have developed protocols where we can directly detect specific RNA molecules in our existing commercial digital chips by integrating the reverse transcriptase reaction with the detection chemistry. This has particular relevance in clinical settings but also has broader implications in general gene expression studies.” In addition, Bio-Rad is launching a one-step kit for digital PCR in gene expression called One-Step RT-ddPCR Kit for Probes. “Instead of partitioning your cDNA after you do the reverse transcription, now you can actually partition RNA and then do the reverse transcription and the PCR, all within the droplets,” says Kurtz. “We’re excited about this approach, because it’s a really great way of going from RNA to counts.”
For labs that use qPCR and dPCR, Life Technologies offers a system that lets you switch between them. “In digital mode, because many applications do not require large numbers of replicates to derive a suitable answer, the flexibility achieved through the use of the OpenArray Plate enables [users] to actively manage how they utilize the replicates available to them: increase sample throughput for low-precision applications or maximize platform performance for high-precision applications,” says Russell. “This flexibility in the use of the technology can be achieved without any compromise to the digital capabilities.”
Greater throughput? Yes. Higher multiplexing? Not quite yet . . .
The throughput of dPCR is increasing as researchers demand more from the technique. However, many say they are looking forward to greater multiplexing capabilities in the future, in addition to more automation. McClory notes that “conventional qPCR has limited multiplexing capability due to spectral overlap of fluorescent probes” and expects that improvements in dPCR multiplexing will continue. “Multiplexing enables development of assays for biomarker panels [with] reduced cost and sample consumption, increased throughput and the potential for built-in assay controls,” McClory says.
Russell also expects robust development of dPCR in the near future. “Despite the fact that digital PCR as a concept has been around for over two decades, I would argue that today it is really still in its infancy,” says Russell. “It is the ability now to miniaturize and run hundreds to thousands of PCR reactions in parallel that has made digital PCR a viable alternative to standard qPCR both from practical and economic standpoints. I fully expect the scaling of this technology to continue over the next three [to] five years, changing digital PCR from a niche technology to something that is used as a standard tool in the lab.”