By Molika Ashford
Researchers at Emory University's Department of Human Genetics have harnessed RainDance Techologies' capture technology to sequence the human X chromosome exome, reporting in a study this month that the technique offers greater uniformity relative to other capture methods for this application.
With the intention of researching the role of X-linked genes in autism spectrum disorders, the group, led by Michael Zwick, used a RainDance expanded content library designed to enrich the human X chromosome exome in a single tube with 11,845 unique amplicons. Illumina sequencing of 24 samples showed coverage for 97 percent of the targeted sequences, they reported in the journal Genomics this month.
Beyond the group's research interest in autism, the efficiency of the method suggests a wide variety of potential clinical and research applications, the authors wrote. Zwick told Clinical Sequencing News that, having used many different capture approaches in his previous research, he considers reproducibility and efficiency to be the main strengths of the RainDance-based X chromosome exome assay.
"Most of the fragments we attempted to amplify worked, and only a very small percentage failed," he said. "The even better news in terms of the clinical world is that when they failed, they failed consistently across all 24 samples."
"In clinical applications you're under deadline to get a sample done rapidly and have it work the first time and have good quality data. [With this method] there's a pretty low variance between experiments, which makes it a pretty ideal application for clinical diagnostics," he said.
Zwick's team's main interest in developing the method was to use it for research into X-linked autism genes. He said his group felt the best way to approach this goal would be to focus on the coding regions of the X chromosome.
"We think that the variants we have the best chance of interpreting are variants in the exome — not that non-coding is not important, but this is really the first place you would look because you have the best probability of interpreting the variants you find," he said.
In their study, the group identified target regions from the RefSeq Genes track using the University of California, Santa Cruz, Genome Browser.
"We gave this to RainDance," said Zwick, and the company then designed a series of PCR reactions that could amplify all the targeted regions the team had identified. "The total size was around 2.5 megabases," Zwick said. The RainDance library consisted of 11,845 amplicons, which were reduced to a final synthesized multiplex primer library of 11,576 amplicons and associated primer pairs: covering 97.7 percent of the initially designed amplicons and 98.05 percent of the targeted region, the authors reported.
While other groups have already adapted Rain Dance's microdroplet amplification for targeted sequencing, Zwick and his coauthors reported in their paper that expanding the amount of targeted material to comprise a whole chromosome exome had yet to be well tested.
At the same time, other X chromosome sequencing studies have been limited by their use of traditional PCR amplification, they wrote.
"PCR is a really well-established technique," Zwick said. "It works very well at amplifying regions of complex genomes, but the problem is that it's really tough to find a way to multiplex."
He added that 11,000 reactions in a single tube "is a real technical challenge that really required a RainDance-like solution to solve."
With the libraries purchased from RainDance, he said, the researchers sequenced a group of 24 male samples from the SFARI Simplex collection, a resource of the Simons Foundation Autism Research Initiative, using an Illumina Genome Analyzer II. The group also sequenced two HapMap samples to compare genotype calls with the data from their sequencing.
"I was really trying to describe the method, and then say how many samples we could multiplex in a single lane of [the] Illumina," Zwick said.
Using four of the 24 SCC samples, the researchers tested a variety of multiplex configurations in order to determine the optimal level with minimal sequencing costs and adequate coverage. "We demonstrated one, two, three and four-plex configurations, and it looked like three was probably about the right number," Zwick said.
To assess repeatability, they sequenced 18 additional samples in a three-plex configuration and two more in a two-plex configuration. One sample failed because of insufficient genomic DNA, while the remaining 19 performed "well, or better, than our original set of four test samples," the authors reported.
The group also compared the performance of their RainDance-based enrichment method with microarray-based genomic selection. The accuracy of the two methods was similar, the authors reported, but array-based capture required between three and seven more reads to achieve a similar performance level.
"The big advantage of RainDance was you could actually multiplex more samples," Zwick said. "And the RainDance assay is more uniform than your typical capture approach."
In the paper, the authors wrote that array-based capture's main advantage is that it allows for rapid design for a small number of samples. In contrast, they said, the RainDance approach is more efficient for a large number of samples.
While in-solution target-capture methods are generally higher throughput than array-based approaches, the researchers did not evaluate in-solution capture for the study.
Zwick said he is now moving forward to use the assay in studying X chromosome gene mutations in autism, starting with the sample reported in the study. "We're evaluating that data right now," he said, and "we already think we have some interesting findings we want to publish."
He suggested that for research purposes, the method could be useful in investigating other diseases where there may be potential sex biases, to test the hypothesis that there are variants on the X chromosome that contribute to the disorder.
Beyond that, he said he thinks the RainDance technology may be "very competitive" for doing any experiments that require capture of a subset of genes, not necessarily all on the X chromosome.
Another team at Emory, led by Madhuri Hegde, senior director of Emory Genetics Laboratory, has been using RainDance capture technology with Applied Biosystems' SOLiD to offer sequencing-based genetic testing for X-linked intellectual disability, congenital muscular dystrophy, and congenital disorders of glycosylation, (CSN 5/25/2010) as well as two newer research-only panels, one for mutations associated with autism spectrum disorder as well as disease-causing mutations that occur on the X chromosome. (CSN 3/22/2011)
Zwick said that the ability to do every gene on the X chromosome might be overkill for applications like Hegde's.
But, he said, "the fact that we can do all the genes on the X [chromosome] quite repeatedly suggests they could build custom libraries that would really focus on specific applications for specific testing."
"I think there is a really nice niche for this type of technology in the diagnostic framework," he said.
Zwick added that the RainDance amplification method "probably doesn't scale to whole exomes."
However, he said, "I imagine as we hone in more and more on pathway analysis and a systems biology framework, where you have a collection of, say, 1,000 genes you want to sequence, but you don't want the whole exome, this could be a really valuable research tool."
Joris Veltman, a researcher at Radboud University in the Netherlands who has worked primarily with whole-exome studies for investigating autism and intellectual disability mutations, told Clinical Sequencing News in an e-mail that Zwick's method may be useful for X-linked disorders, but that it is often difficult to predict whether the abnormality is X-linked or not.
"In general, I am in favor of using more unbiased whole-genome or whole-exome sequencing approaches," he said.
While these approaches require more sequencing and are therefore more expensive, "the big advantage is that they give an unbiased view of all/most genomic variants that can be linked to disease in a single experiment, do not require updating when new disease genes are revealed, and because they can be used for any genetic disease, they will have favorable turnaround times, and more control data can be obtained to improve variant interpretation, which is very important," he said.
Zwick agreed that the application for more targeted studies is limited, but said that particularly for diagnostics, using a more focused approach can be beneficial. While most studies will eventually benefit from sequencing the whole genome, "there is a huge bioinformatic burden when you sequence [more] DNA," he said.
For diagnostic sequencing, "it's actually kind of a huge advantage to be able to look at a lot of the genome, but only the portion you really want to look at carefully."
"Doing more, in the clinical world, could almost be a distraction in some sense," he said.
Have topics you'd like to see covered in Clinical Sequencing News? Contact the editor at mashford [at] genomeweb [.] com.