Two recent studies have illustrated the utility of RainDance Technologies' MethylSeq Solution, a sequencing-based cytosine methylation analysis method that relies on the company's microdroplet PCR technology to amplify targeted sites in the genome.
"These two papers really showed quite well the application of our technology for doing targeted methylation analysis," Andy Watson, RainDance's vice president and chief marketing officer, told In Sequence. The studies demonstrate that the RainDance approach captures methylated regions of the genome more effectively than other targeted approaches and is more cost-effective than genome-wide methylated sequencing.
Both studies focused on cytosine methylation in CpG islands, sites in the genome where cytosine frequently neighbors guanine. The CpG motif shows up in the promoter region of about two-thirds of genes in the human genome and, typically, methylation of CpGs in a promoter corresponds to muted expression of the associated gene.
In one of these studies, published online in Genome Research in July, researchers from Scripps Research Institute and RainDance reported that they successfully used the microdroplet PCR method, combined with Illumina sequencing, to measure CpG methylation patterns for 50 genes before scaling up the approach to target more than 2,100 genes. In the future, they hope to use a similar strategy to test many more samples, first from healthy individuals and eventually from those with autoimmune conditions or variable responses to organ and cell transplants.
"A good method would be one that we could do relatively large numbers of patients," senior author Daniel Salomon, a molecular and experimental medicine researcher at the Scripps Research Institute, told IS. "We wanted a method that, in the end, we would be able to create a reasonably high enough throughput to do hundreds of patients."
In another study, published in PLoS One last month, a team led by investigators at Christian-Albrechts University in Germany, which also included investigators from RainDance, described how they employed a similar sequencing approach to track methylation on both DNA strands at targeted sites in lymphoma and colorectal cancer samples.
By combining data on microdroplet-amplified regions of interest from bisulfite-treated and untreated DNA, the team obtained both methylation and SNP patterns, allowing them to find so-called hepitypes, DNA haplotypes with characteristic cytosine patterns, which they believe may offer clues to cancer evolution.
"We established a pipeline allowing a simultaneous, targeted assessment of DNA sequence variation and methylation," senior author Jochen Hampe, a researcher at Christian-Albrechts University, and co-authors wrote.
"The pipeline has a considerable scalability potential," they added, "because the degree of redundancy may be reduced, the proportion of longer amplicons might be increased, and larger numbers of primers … may be incorporated in the libraries."
The RainDance PCR system, Rainstorm, is a microdroplet-based PCR method in which reactions take place within miniscule droplets that are produced, combined, and processed on a disposable chip. Droplets containing a library of primer pairs targeting specific DNA sequences are merged with another droplet containing template DNA before being thermocycled, leading to the amplification of a targeted stretch of DNA that can then be sequenced.
Using this approach, researchers can run up to 20,000 amplification and primer pairs for a single sample, Watson explained, generating between one and two million PCR reactions. For the MethylSeq application, the company can run up to 4,000 amplification and primer pairs per sample.
In the past, RainDance has touted its microdroplet-PCR technology for medical genetics, cancer sequencing, and other applications, such as multiplex enrichment of sequences of interest and ultra-deep resequencing to track down rare mutations in cancer and other tissues (IS 6/22/2010), including formalin-fixed paraffin-embedded and fresh frozen samples (IS 2/8/2011).
To make the approach amenable to methylation analyses, researchers use the same microdroplet PCR technology to amplify the promoter regions for genes of interest from genomic DNA that's been treated with sodium bisulfite, which converts unmethylated cytosines to uracil. RainDance researchers have also tweaked their primer design pipeline with CpG island targeting in mind, to be sure that the sequences they target actually do get amplified and are compatible with both the RainDance instrument and any sequencing platforms that might be paired with it.
"We redesigned our pipeline to ensure that the bisulfite-converted regions would be targeted with primer pairs that would bind uniquely in the genome," Watson said. "That turned out to be not trivial," he added, "but once we've done it we've seen tremendous success in both the specificity and yield of those primers."
The approach provides high-quality and high-coverage sequence targeting for the gene regions of interest, Watson said, allowing quantitative methylation analysis. And because it relies on sequencing, the method offers a detailed picture of the entire promoter region for each gene of interest.
"With the quantitative ability that our technology enables, along with the sequencing readout, that allows, on a base-by-base basis, the methylation status to be quantified across the entire promoter region rather than just targeting individual residues," he noted.
Though a wide range of strategies — from enzymatic digestion or bisulfite conversion to affinity-based methods — can be used to gauge cytosine methylation in CpG islands, the price, process, and precision of these approaches vary.
For instance, array-based methods for detecting cytosine methylation with thousands of probes are simple, inexpensive, and very useful for some studies, Scripps' Salomon explained, but are not designed to look at individual CpGs and typically can't be customized to target any gene.
"You get what they give you … If your gene or CpG is not there, you don't get it," he said. "It certainly gives you CpG status, but it doesn't tell you about any individual CpGs, so if there was substructure you would never see it."
Such details can be important, Salomon added, particularly for researchers interested in looking at details within a CpG region or profiling the methylation patterns for genes that lack full-fledged CpG islands in their predicted promoter/enhancer regions but which may still contain functionally important CpGs.
[ pagebreak ]
"Rather than average data across a promoter region, you get very specific data across each individual base across that promoter region," Watson agreed, "which gives you a much easier way to analyze exactly what's going on."
Other targeted methylation sequencing methods do exist, including a method involving multiplexed bisulfite patch PCR in conjunction with deep sequencing of barcoded samples (IS 7/20/2010). But Salomon and his co-authors argued that this approach and other strategies involving multiple PCR amplification rounds could lead to amplification bias.
On the other hand, they argued that strategies such as methylated DNA immunoprecipitation coupled with sequencing (MeDIP-seq), methylated CpG island recovery assays (MIRA), and methylation-sensitive restriction enzyme sequencing (MRE-seq), are difficult to scale up for use in large sample sets.
At the extreme end of that spectrum, whole-genome bisulfite sequencing provides a snapshot of cytosine methylation across the entire genome but carries a hefty price tag and bioinformatics burden.
Whole-genome bisulfite sequencing is "the ideal way to do it" and might be tempting to apply to numerous samples at a sub $1,000 per genome price point, Salomon said. But at $5,000 or more per genome, not including the analytical investment, he said the whole-genome approach remains out of reach for many labs.
Immunity and Cancer
Salomon's interest in the molecular genomics of immunology and immune function and their role in organ and cell transplant rejection risk led him to epigenetic studies as a complement to some of the genetic strategies that he and his team are using to study transplant outcomes.
So far, the team is focusing on cytosine methylation in the CpG islands of immune-related genes in healthy individuals to find baseline epigenetics involved in immunity. They ultimately hope to track down epigenetic markers for predicting an individual's response to immune challenges, along with clues to understanding autoimmune conditions such as type 1 diabetes and rejection risk for those receiving organ and cell transplants.
"There's an incredible amount of diversity in the reaction of different individuals to different immune challenges," Salomon said, explaining that past research hints that at least some of these differences are a consequence of epigenetics.
He was drawn to the RainDance approach while on the hunt for an unbiased, scientifically rigorous method that could not only detect methylation down to the individual CpG island level, but that was also targeted and scalable.
He and his colleagues reported in Genome Research that they got 96 percent coverage of CpG sequences in their initial experiments targeting about 400 amplicons representing around 1,000 bases up- and downstream of the transcription start sites for 50 genes using genomic DNA from Jurkat cells.
When they scaled the method up to target roughly 3,500 amplicons containing 77,674 CpGs in the promoter regions of more than 2,100 genes using bisulfite converted DNA from human CD4 T cells, the overall amplicon coverage was 99 percent and CpG coverage was 97 percent.
For both the 50 gene and 2,100 gene libraries, the method generated 100 reads per CpG targeted starting from two micrograms of bisulfite-converted DNA and using a single lane of sequencing on an Illumina GAIIx flow cell.
For their part, the German team is using the PCR-based bisulfite sequencing method to target CpG sites for genes that are predicted to be important in cancer, while simultaneously looking at DNA sequence patterns associated with these epigenetic marks.
"Virtually all cancers are in some way associated with aberrant DNA methylation," Hampe and colleagues wrote. "Ideally, a parallel, deep analysis of DNA should be conducted in order to understand the interaction of germline risk factors, somatic genetic, and epigenetic evolution within the tumor, not the least because altered methylation may affect cytosine residues outside a CpG context."
The team designed primers targeting 34,083 base pairs of sequence. After filtering out redundant primer pairs and primers that amplified off-target sequences, they were left with about 1,000 primer pairs, which they used to amplify targeted sequences from matched-normal colon cancer and follicular lymphoma DNA samples that had been bisulfite treated.
Meanwhile, to amplify nearly 59,000 bases representing target regions in DNA that had not been bisulfite converted, the researchers used 249 primer pairs that were designed using the RainDance primer design pipeline.
The team then sequenced both the bisulfite-converted and non-converted libraries using the Roche 454 GS FLX platform to generate CpG methylation and SNP data, respectively.
They reported that the microdroplet PCR-based approach yielded at least 20-fold coverage for 95 percent of the 34,083 bases targeted in the bisulfite-converted DNA.
In a statement, Hampe said the microdroplet PCR-based methylation analysis method "allows us to leverage the inherent power of next-generation sequencing in a more efficient and targeted manner."
"We are now able to interrogate any region of the methylome to better understand the
progression of different types of cancer," he added.
The two studies demonstrate the utility of the MethylSeq approach in immune- and cancer-focused studies, but RainDance's Watson said the approach is suited to any research that involves characterizing cytosine methylation, including studies of development and disease in humans as well as research focused on epigenetic patterns in plants and other organisms.
The method is compatible with virtually any sequencing platform, he said, and RainDance has developed workflows on the back end of its platform that allow customers to convert amplified samples into libraries that can be sequenced with a variety of second- and third-generation platforms.
While Salomon was hesitant to put a firm price on the experiments done in his study, he said the cost per sample was comparable to commercial array-based methods. That estimate does not include the RainDance RDT 1000 instrument purchase price.
Watson said the instrument is available for as low as $100,000 under the company's new pricing programs. He noted that chip prices have dipped in the wake of a recent deal with Sony, which will now manufacture and supply RainDance chips, and estimated that chip and primer prices are around $150 per sample. Sequencing costs are heading south as well, he added, and depending on the platform used are likely in the few $10s to $50 range per sample, Watson estimated.
Researchers interested in using the RainDance system for methylation analysis can order custom-designed libraries and then run their samples on the RainDance instrument.
The complete methylation analysis workflow will also be available through service providers such as Expression Analysis, Ambry Genetics, and GeneDx that are already offering other services that rely on RainDance technology.
Have topics you'd like to see covered in In Sequence? Contact the editor at anderson [at] genomeweb [.] com.