NEW YORK (GenomeWeb News) – A study published in the November issue of Nature Biotechnology has found that RainDance Technologies' microdroplet-based PCR method is well suited for large-scale targeted resequencing studies, and more scalable than the "gold standard" of traditional PCR for that application.
The study was led by researchers at the Scripps Translational Science Institute; the University of California, San Diego School of Medicine; and RainDance. It found that the sequence-enrichment technology developed for RainDance's RDT-1000 instrument — which sets up millions of PCR reactions in picoliter-sized droplets and combines them in a single vial for amplification — offers a number of advantages over competing sequence-capture methods, such as standard PCR, molecular inversion probes, or hybridization-based methods.
In particular, the study determined that when compared with other enrichment methods, "microdroplet PCR generates substantially greater uniform coverage of targeted sequences." This uniformity results in a higher variant detection rate when coupled with second-generation sequencing, the study authors noted, citing a variant detection rate of 94.5 percent for microdroplet PCR, compared to a range of 64 percent to 89 percent for solution-based hybridization and 75 percent for molecular inversion probes.
The authors note, however, that the accuracy of called variant bases is "currently high across all enrichment strategies," and the proportion of reads mapping to targeted sequences is also "similar" across the different methods.
A key question that the study set out to determine was that of scalability, according to Steve Becker, vice president of commercial operations at RainDance. "PCR in general is the gold standard for sequence enrichment, [but] the reason that traditional PCR has not been used for large population studies is that it didn't scale well," he told GenomeWeb Daily News.
"So while the sensitivity and specificity were strong attributes, there were some detractors related to PCR [and the study authors] wanted to see whether those detractors went away or were mitigated" with microdroplet PCR, Becker added.
Kelly Frazer, chief of the division of genome information sciences at UCSD's department of pediatrics and senior author on the study, said in a statement that "the problems we typically encounter with standard PCR, such as difficult primer design and high allelic bias, were not observed with the microdroplet-based workflow."
Jeremy Lambert, product manager for the RDT-1000, told GWDN that the primary scalability issue with standard PCR "comes from the practical standpoint of how many 384-well plates can be pipetted and thermal-cycled in a given day." In order to get the same throughput of the RainDance platform, "you would need the capability to do a minimum of 30,000 PCR reactions per day. So that’s the equivalent of around 100 384-well plates per day, and that's outside the accessibility of most researchers."
In addition, he noted that the RainDance system requires much less DNA than standard PCR. "Even in a very highly optimized, low-volume PCR reaction, you typically need about 5 nanograms of DNA, so to do the equivalent of 4,000 PCR reactions per sample, that would be 20 micrograms of input DNA required with standard PCR."
The RainDance platform can do the same amount of amplification "with just 2 micrograms — an order of magnitude less input DNA," Lambert said.
There are also cost advantages. For the 4,000-target library described in the study, "it would cost $4,000 per sample with standard PCR to do the amplification," Lambert said. "With our technology, we can do that same process for about one quarter of the cost. So with the same budget you can enrich four times as many samples or do a study that has four times as many cohorts, which improves the ability to detect rare variants."
Becker said that RainDance views the paper as validation that the approach is well suited for population sequencing studies.
"When we speak to customers, they indicate to us that they really want to be processing, in a typical population study, between several hundred to thousands of patient samples. So having a very robust, scalable approach using the gold standard technology of PCR, we see is going to have important medical implications."
RainDance first unveiled its microdroplet PCR-based sequence enrichment system at the Advances in Genome Biology and Technology conference in Marco Island, Fla., in February.
The system can currently amplify up to 4,000 targeted sequences, but Becker said the company has "a path forward" to scale up to 20,000 targets — or about a tenth of the human exome — in the first half of 2010 using an expanded content format with five sets of primers in each droplet.
By comparison, Frazer and colleagues note in the study that the molecular inversion probe-based method can enrich for 13,000 targeted sequences, the solution-based hybridization approach can target 22,000 sequences, and array-based capture has been shown to target 26.6 megabases of coding exons "with similar mapping efficiency as that of solution-based hybridization." They note, however, that "although the approach has excellent concordance and uniformity, array-based capture is difficult to scale for population studies in individual laboratories."
The current throughput of the instrument is eight patients a day or 2,000 patients per year, which "fits nicely with the workflow for next-gen sequencers, where there's a typical processing time of 10 to 12 days per sequence run," Lambert said.
In addition to the scale-up in enrichment that is expected for the first half of 2010, RainDance is working on several additional applications for the microdroplet PCR technology for use with second-generation sequencing. The first, an application for methylation sequencing, "includes treatment of genomic DNA with sodium bisulfate reagent, which will modify the bases that are differentially methylated, and then we can amplify that converted DNA using PCR in droplets for subsequent analysis of the differential methylation pattern by next-gen sequencing," Lambert said.
The second is the capability to do ultra-deep resequencing to identify low-concentration mutations in a background of wildtype sequence — an application of particular interest for cancer studies.
"Because of the number of PCR reactions that we're able to generate, we can get very high levels of replicate reactions to allow us to amplify not only the background DNA, but also specifically amplify the tumor-specific DNA that might be at a very low percentage of the total DNA," Lambert said. "We're compartmentalizing the reaction, so there will be no competition between the tumor DNA and the wildtype DNA because they'd all be individual PCR reactions."
RainDance plans to launch both applications in the first half of 2010.