By Ben Butkus
In an effort to retain its early lead in the rapidly growing digital PCR arena, Fluidigm is developing versions of its Digital Array PCR chips that will enable "hundreds of thousands" of PCR reactions to be performed in parallel, a company executive said this week.
In addition, the company is working with undisclosed academic collaborators to push its technology into the realm of "millions of dPCR reactions per chip," according to the official — a throughput that other companies are just beginning to envision by using nanodroplet or other nanoscale partitioning technologies.
In an interview with PCR Insider this week, Fluidigm President, CEO, and co-founder Gajus Worthington outlined the company's plans to "remain the market leader in dPCR" in light of recent assertions by up-and-coming competitors in the space that the physical footprint of Fluidigm's chip-based platform limits its ability to perform PCR reactions in a massively parallel fashion.
Specifically, as reported last month by PCR Insider, companies such as Life Technologies, RainDance Technologies, QuantaLife, and Inostics have in recent months developed or acquired technology designed to gain a foothold in the dPCR space (PCR Insider, 7/8/10 and 7/15/10).
Most of these companies are working on solutions to partition a sample into the millions of individual nanoliter-scale-or-less PCR reaction volumes thought necessary to truly unleash the power of dPCR as a tool for copy number variation, rare allele detection, validation of next-generation sequencing, and diagnostics.
Meantime, Fluidigm's Digital Array technology, which it first introduced to the market in 2006 well ahead of its would-be competitors, uses microfluidics and patented integrated fluidic circuit technology to partition each of a fixed number of samples into a fixed number of PCR reactions on a chip format with physical dimensions that are similar to a 384-well screening plate.
The company's first iteration of this technology divided 12 samples into 765 PCR reactions for a total of 9,180 individual reactions. Then, in 2008, the company introduced a Digital Array chip that could partition each of 48 samples into 770 PCR reactions for 36,960 individual reactions in volumes as small as 4 µL.
However, according to Worthington, Fluidigm is "nowhere near the limit of our technology, not by many logs, with respect to what we can do with dPCR. We don't have any limitation with the physical footprint. We chose to implement our initial products in a 384-well footprint for convenience only. People are used to pipetting into a 384-well plate, and they're used to handling them.
"People are pointing to that and assuming that just because we did that, it's an inherent part of the technology," Worthington added. "It's not."
Worthington conceded that the number of reactions that can be performed on a Digital Array chip is a function of the amount of area needed for each reaction. However, "everything just keeps getting smaller, and the rate at which we've been able to reduce the feature sizes on our chip has been very rapid. That's why we've been able to go from 10,000 to 40,000, and will go from 40,000 to hundreds of thousands."
On that point, Worthington said that Fluidigm "has chips now that we're testing internally that can do hundreds of thousands of dPCR reactions per run." However, he declined to provide a development timeline for such technology.
"We don't want to confuse our customers right now with premature offerings," Worthington said. "But if the market needs that, it's something we will be able to do handily. You can look at our historic trajectory of how we've increased density, and we're taking a big step now with devices that do hundreds of thousands of reactions in parallel."
In addition, Fluidigm is working with undisclosed academic collaborators "that unfortunately we can't be very specific about because they're getting ready to publish, that are using our technology to do upwards of millions of dPCR reactions per chip," Worthington said.
But is the ability to do millions of dPCR reactions in parallel a useful tool or just overkill? Some representatives from would-be Fluidigm competitors told PCR Insider that it might be the latter, and Worthington seemed to echo this sentiment.
"There are certain applications where having millions of dPCR reactions would be helpful," Worthington said. "We aren't necessarily convinced that those are the large applications."
For instance, Worthington pointed to copy number variation analysis as one of the most promising applications for massively parallel dPCR.
"Conventional technologies such as PCR and microarrays and such are great for certain applications — gene expression and genotyping in particular," he said. "CNV is different, and the amount of discrimination capability that is needed can vary from two-fold to discriminating the difference between 18 or 19 copies. That can be very important clinically and from a research perspective. So we see CNV in a broad sense as being a big application for dPCR, and really for all kinds of areas: cancer, developmental biology, cellular research, and as follow up to [genome-wide association studies.]"
However, Fluidigm's customers are using the current iterations of the Digital Array technology to do CNV analyses right now, said Worthington, estimating that the bulk of the 30 or so papers published using Fluidigm's dPCR technology were for CNV applications.
Otherwise, to date "the dPCR applications have not been a high-sample-throughput market," Worthington said. "Our current device has 48 sample inputs, and that's actually quite a lot for some of the applications we see today. The push, I think, is for more pixels per sample, so you can have greater dynamic range of discrimination, or more specificity."
Fluidigm has been developing its thermal cycling and detection technologies in concert with its chips, and Worthington said that the company "potentially" would need to rejigger its instrument platforms to accommodate higher-density chips. However, the chips being tested in house at Fluidigm can be read on the company's current BioMark real-time PCR system.
"To go up to millions [of reactions], there will probably need to be some upgrades, not necessarily in the optics, but in the detection," he said. That push is dependent primarily on detection technology such as CCD cameras, which Worthington said are "fortunately advancing very rapidly. As the number of reactions per unit area increases dramatically, your camera needs to have more pixels to be able to discriminate those."
The company believes it has maximized its thermal cycling capabilities for the time being with the introduction in May of its FC1 Cycler, which is designed to reduce thermal cycling time three-fold over conventional devices (PCR Insider, 5/20/10).
This platform, which can be combined with Fluidigm's BioMark reader or lower-end EP1 reader for non-real-time PCR applications, can perform a full PCR reaction in about a half hour. "That means you can run these chips extremely quickly, so you're done in an hour, with literally hundreds of thousands of reactions," Worthington said.
Lastly, on the IP front, Fluidigm believes it has a "substantial" dPCR IP portfolio, Worthington said.
While many of the company's would-be competitors — particularly Life Technologies — have acquired or licensed IP surrounding the compartmentalization of PCR reactions that is at the core of dPCR, "Fluidigm has been in the market for several years now, and that has given us the opportunity to establish a very strong IP position," Worthington said.
"A lot of the earlier stuff wasn't commercial, and when you commercialize things, you make all kinds of important inventions," he added. "And companies that do that first … benefit from that greatly."
Overall, Fluidigm sees the coming competition as a good thing. "Companies announcing their intention to get into this space is certainly a sign that others are recognizing what we've seen for a while — that there is a large potential opportunity here," Worthington said. "That's all good. That draws attention from potential users and customers and other stakeholders."
But, he added, "a lot of the plans companies have for entering the market in the near future are things we are already doing today. We view ourselves as having quite a head start to the potential entrants to the market."