Digital PCR: Counting DNA molecules

Several decades ago, would scientists have laughed if you told them that you knew how many molecules of DNA were in a tube? Perhaps—but not today. It may have been nearly unimaginable to scientists of the past that we could count anything so tiny. But with a new technique called digital PCR (dPCR), today’s scientists can indeed count molecules of DNA.

Digital PCR derives its name from the method of counting. Researchers partition a sample of template DNA into many tiny partitions, such that each one essentially contains either 0 or 1 copy of template DNA. Counting the positive wells gives an absolute number of molecules in the original sample. In general, the more partitions a sample is divvied up into, the better the precision and the wider the dynamic range of measurement.

Using dPCR rather than its predecessor, qPCR, researchers are better able to measure small changes in amounts of DNA in a sample. For example, dPCR can distinguish between say, five and six copies of a gene, a feat of detection that is out of qPCR’s league. Scientists use dPCR to detect and quantify low-level pathogens, rare genetic sequences or mutations and copy-number variations, among other applications.

If dPCR is something you’d like to try, you’re in luck: Several such systems are now commercially available. Here’s a look at a few of them.

Droplet-based dPCR systems

Generally, dPCR systems are divided into two camps, those that use droplets for isolating DNA molecules, and those that don’t. There currently are two manufacturers of droplet-based dPCR, Bio-Rad Technologies and RainDance Technologies. Bio-Rad’s QX100 Droplet Digital PCR system  has been available the longest and currently is used by hundreds of labs. It also recently received the 2012 Frost & Sullivan North America Laboratory Researchers' Choice Award for Future Market Leader.

The QX100 system uses droplets to partition a 20-microliter sample into 20,000 tiny reactions, each about a nanoliter in size. “That number of partitions is sufficient to enable you to get nearly five logs of dynamic range,” says George Karlin-Neuman, scientific applications director of Bio-Rad Technologies’ Digital Biology Center. “And it’s able to give you precision within a single well down to several percent.” This precision is important when trying to detect copy-number difference in genes. “You need to have sufficient precision to be able to not blur one measurement to the other and [still] be able to distinguish [between] them,” says Karlin-Neuman.

Another key QX100 application, says Karlin-Neuman, is detection of rare mutations involved in cancer. Typically, because the QX100 creates 20,000 partitions, researchers can detect sequences that occur at a frequency of about 1/20,000. But they also can detect rarer sequences if they use multiple wells. For example, researchers can increase their precision five-fold by running five dPCR reactions in five wells. “With 20,000 [partitions] within each of those wells, you can merge them together in silico, essentially, and get the benefit of having 100,000 partitions,” explains Karlin-Neuman, who notes that cancer researchers have used that strategy to detect rare activating mutations such as BRAF and KRAS.

The other droplet-based dPCR system, from RainDance Technologies, is also well suited for the detection of rare mutations. The RainDrop™ System partitions a sample into 5 to 10 million droplets, each about 5 picoliters in volume. According to Rena McClory, RainDance’s marketing director for digital PCR, this extensive partitioning makes the system incredibly sensitive. “RainDance’s RainDrop System can detect one mutant amongst 250,000 wild-type molecules, with a lower limit of detection of one in more than 1,000,000,” says McClory.

The system creates droplets in eight channels simultaneously (in parallel), which yields about 80,000 PCR reactions per second, according to RainDance’s website.  During droplet formation, the size of the droplets is monitored by a CCD imaging system, which feeds that information back to the system to adjust droplet volume if necessary, ensuring uniformity.

RainDance also offers higher-level multiplexing than most dPCR systems, using up to 10 markers. “The RainDrop System shifts the PCR paradigm from a single color per marker to a more scalable and precise multicolor and intensity-per-marker method,” says McClory. RainDance accomplishes this not by the traditional method of using a different color for each marker (usually limited to four colors in PCR because of spectral overlap of the fluorophores), but by varying the concentrations of just two fluorescent-probe colors. Different probes are distinguishable according to their different concentrations, which are directly related to their measured fluorescence intensities [1].

Bio-Rad’s QX100 system also is capable of multiplexing; using two markers is routine for determining frequencies of single-nucleotide polymorphisms (SNPs) or copy-number variations. Karlin-Neuman says four-plexes can be used, for example, for phasing of two heterozygous SNPs, and that Bio-Rad plans to increase multiplexing opportunities in the future.

Chip-based dPCR

Fluidigm and Life Technologies offer nondroplet systems. Life Technologies’ new QuantStudio™ 3D Digital PCR System, launched in June of this year, is based on nanofluidic chips that enable collection of up to 20,000 data points per run. The chips are like dPCR plates with tiny wells, into which the sample is partitioned.

Paco Cifuentes, Life Technologies’ product applications director, says one strength of the QuantStudio is that unlike droplet dPCR, it is a closed system. After the researcher loads a sample into the chip, the PCR runs on a standard thermal cycler, and the QuantStudio reads the results. “Once the sample is deposited on the chip, the chip is sealed and disposed [of] at the end of the experiment,” says Cifuentes. “This approach minimizes the chances for contamination.”

Life Technologies also plans to launch an automatic chip loader in 2013 to reduce opportunities for human error. Life Technologies recently collaborated with academic researchers to show that its dPCR technology can detect BRAF mutations on thyroid-tumor samples occurring at a rate of less than 1%. 

Another nondroplet dPCR system is Fluidigm’s qdPCR 37K™ IFC. Because the system also performs qPCR, Fluidigm refers to it as a quantitative real-time digital PCR system. The instrument partitions a sample using tiny reaction chambers in a microfluidics chip, and according to Fluidigm director of molecular biology and assay development Ramesh Ramakrishnan, the system has five-color multiplexing capabilities.

Ramakrishnan says that what distinguishes Fluidigm’s dPCR system is its ability to collect data from each dPCR reaction in real time—in contrast to other systems, which are essentially endpoint assays. “This allows you to monitor reaction efficiency in real time, if it is necessary for your experiments,” he says. “You don’t have to, but in some cases it might be of great importance, like in a clinical setting where a false positive for a cancer diagnosis is at stake.” In a research environment, the ability to check the real-time data can help to discriminate real signals from the artifacts resulting from amplification errors. “In other words, it’s a built-in way to verify a positive,” says Ramakrishnan.

Digital PCR is still in its infancy, or perhaps toddlerhood. Cifuentes says the novelty of the techniques represents both a challenge and an opportunity. Researchers are discovering new dPCR applications and new ways to adapt the technique to their needs. “Higher throughput, increased multiplexing abilities, increased sensitivity and other key system performance features will evolve following the needs created by specific dPCR applications,” he says.

Although the basic principle of the technology works well, Karlin-Neuman points out many improvements that could open new doors for researchers—automation, sorting capabilities or even expanding dPCR technology beyond nucleic acids. Bio-Rad’s Digital Biology Center is exploring multiple applications of digital-counting technology. “You can certainly envision going off into proteins, potentially metabolites, and other macromolecules,” he says. And the company’s customers are following up with their own ideas. “People are interested … in putting organisms into droplets.”

Reference

[1] Zhong, Q, et al., “Multiplex digital PCR: breaking the one target per color barrier of quantitative PCR,” Lab Chip, 11:2167-74, 2011.