Agilent Technologies and Roche NimbleGen are both offering users of second-generation sequencing systems new tools to select portions of the genome for targeted resequencing.
While both companies offer hybridization-based methods and are serving a growing demand for targeted resequencing, they each market their solutions for use with different sequencing platforms and for distinct project sizes.
Later this week, Agilent is scheduled to launch kits for an in-solution method, called SureSelect Target Enrichment System, that allow researchers to select an average 3.3 megabases of genomic DNA using up to 55,000 customer-specified RNA probes. Later this year, the company also wants to offer a microarray-based product for targeted selection aimed at smaller-scale experiments.
Agilent's SureSelect kits are based on a method the company licensed from the Broad Institute last year (see In Sequence 4/8/2008) that uses biotinylated RNA "baits" to capture DNA targets in solution.
In a paper published in Nature Biotechnology earlier this month, Broad and Agilent researchers described the method in detail, showing that they could use it to target approximately 15,000 exons, or 2.5 megabases of DNA sequence, and four genomic regions, totaling 1.7 megabases of sequence.
Since bringing the technology in house about a year ago, Agilent has "put it into a commercial kit, with all the quality control steps along the way," according to Fred Ernani, the company's emerging genomics applications product manager.
One of the main differences between the Broad's original method and Agilent's commercial product is the length of the bait, which is now 120 mers instead of the original 170 mers.
According to Ernani, the kits are geared at customers wanting to sequence 3 megabases or less DNA in 10 or more samples, for example to follow up on genomic regions of interest from genome-wide association studies.
An advantage of solution-based hybridization, he said, is that it is easy to scale and requires less DNA than microarray-based products. In addition, the hybridization reaction is completed in only 24 hours instead of several days.
Using Agilent's eArray online design tool that already exists for its microarray users, customers can specify their genomic regions of interest — discontiguous regions like exons, or contiguous regions — and design up to 55,000 baits per library to capture them. Right now, they can target DNA from the human genome as well as over 20 model organisms, including mouse, rat, and Arabidopsis.
Agilent's genomics division then synthesizes the oligos, and its Stratagene division generates the biotinylated RNA, which is shipped to the customer with the appropriate buffers.
Customers can purchase the kits in different sizes, depending on the number of samples in their experiment, starting from a 10-reaction kits up to a 5,000-reaction kit.
At the moment, the kits are optimized for shotgun sequencing with the Illumina Genome Analyzer, but Agilent is in the process of optimizing the technology for the Applied Biosystems SOLiD platform, as well as for paired-end sequencing with either platform.
[ pagebreak ]
The company is not disclosing the price of the kits, but the price per sample decreases with larger kits sizes, and "as a researcher's project scales, we will be quite competitively priced" compared to competing projects, such as Roche NimbleGen microarrays, Ernani said.
So far, Agilent has made the kits available to early-access customers, including the Wellcome Trust Sanger Institute and Scripps Genomic Medicine.
According to Dan Turner, head of sequencing development at the Sanger Institute, results from several experiments with the Agilent SureSelect probes have been "very good," using the Genome Analyzer as the sequencing platform.
Compared to NimbleGen array-based sequence capture, which his group has also used, the Agilent approach "is technically less demanding, uses less DNA, should be simple to scale up and, in our hands, gives very specific enrichment," Turner told In Sequence by e-mail.
Kelly Frazer, director of genomic biology at Scripps Genomic Medicine, agreed that Agilent's protocol is "easy to use" and to scale up to larger numbers of samples. Furthermore, the method provides "good fold enrichment" and "uniform coverage of targeted bases."
Later this year, Agilent also plans to launch an array-based capture method. It is currently testing the product with early-access customers, including the Broad Institute and Cold Spring Harbor Laboratory.
According to Ernani, the array product will be more suitable for projects with fewer than 10 samples, and "is going to be very powerful for studies where you need iterative designs."
"As soon as the experiment is completed, you can go back on the array and refine the design for use on different regions, or remove regions that are not interesting anymore," he said.
Like SureSelect, the arrays will initially be optimized for the Genome Analyzer, though "we also see a way forward to make our application work for SOLiD," according to Ernani. He mentioned, though, that Applied Biosystems has already come out with a technical note describing a capture method for SOLiD on the Agilent arrays. "So there is a solution out for SOLiD using Agilent arrays, [but] it wasn't developed by Agilent," he said.
Jay Shendure, an assistant professor in the department of genome sciences at the University of Washington, told In Sequence that his group has used both Agilent and NimbleGen arrays for hybridization capture, “and both work very well.” He added that he plans to use “either” now or in the future.
Researchers at Cold Spring Harbor Laboratory switched from NimbleGen arrays to Agilent arrays, mainly because they already had "a lot of experience" and equipment to handle microarrays from Agilent, and because they wanted to optimize their capture methods for the Illumina sequencing platform, "and it seemed to us that Nimblegen was going to be optimizing for 454," according to Dick McCombie, a professor at CSHL.
He said that is group is also planning to try the solution-based SureSelect method for large sample sets where they want to study the same group of targets.
NimbleGen's Sequence Capture 2.1 Human Exon Arrays
Last month, Roche NimbleGen launched Sequence Capture 2.1 Human Exome microarrays, which allow researchers to capture nearly all human exons, covering a total of more than 30 megabases — approximately 180,000 protein-coding exons and about 700 micro RNAs exons — on a single array. The design is built on the 2008 build of the Consensus CDS database project, which defines a core set of high-quality human and mouse protein-coding regions.
The US list price for the 2.1 Human Exome array is $2,000, which includes ancillary equipment such as a disposable mixer, elution chamber, and gasket.
The company also plans to start selling 2.1-million feature custom arrays in the first half of this year that will allow users to choose a capture region of up to 30 megabases, either contiguous regions or exons.
[ pagebreak ]
Both types of arrays are built on NimbleGen's new high density HD2 platform, which uses 2.1 million oligonucleotide probe, each more than 60 nucleotides in length, on a single microarray.
The launch comes less than a year after NimbleGen introduced a sequence capture service, under which scientists can request up to 5 megabases of target DNA from the human genome to be captured on 385,000-feature arrays (see In Sequence 4/8/2008).
These custom-designed arrays have also been available for purchase since last October for users to perform their own capture reactions. In addition, customers can order the arrays for organisms besides humans.
No service is currently available for the 2.1M Human Exome arrays, but the company plans to start offering such a service "in early 2009,” according to the NimbleGen website.
NimbleGen has optimized its arrays for sequencing with the 454 Genome Sequencer FLX, using the Titanium long-read chemistry. The arrays yield enriched DNA “that can be directly and easily integrated into the workflow of the 454 Genome Sequencer FLX instrument,” according to NimbleGen’s website, which is “the most appropriate sequencing technology for the NimbleGen Sequence Capture solution.”
The company is also working on 2.1M and 385K Sequence Capture products that use a protocol optimized for the 454 Titanium series kits and provide “a seamless transition from capture into sequencing.” Captured fragments will be "ready to go into the emPCR step” with no need for further library preparation, according to the website.
Though some customers are using the arrays with other sequencing platforms, such as the Illumina Genome Analyzer, NimbleGen does not provide validated protocols for use with these instruments. Customers have to develop their own protocols for preparing the captured DNA for sequencing with platforms other than the 454 GS FLX, according to Xinmin Zhang, senior marketing manager for sequencing products at Roche NimbleGen.
The Sanger Institute’s Turner agreed that the NimbleGen protocols are “very well established for the 454” but that his group has found “results with our Illuminas [sequencers] to be a little variable.”
“The read length and the fragments captured using our sequence capture platform are a good fit for each other,” explained Dan Burgess, a senior scientist at Roche NimbleGen, during a company workshop at the Advances in Genome Biology and Technology meeting in Marco Island earlier this month.
According to Burgess, the company started using a new design algorithm last fall that has allowed it to increase the capture uniformity, achieving “a much more balanced distribution of sequence coverage over the target.”
He cited as an advantage of array-based capture over solution-based target enrichment that it is easier to optimize the target design on microarrays.
According to Zhang, NimbleGen's Sequence Capture arrays will "definitely compete" directly with Agilent's SureSelect product for projects with small target regions up to 3 megabases, for example follow-on studies to genome-wide association studies.
At the moment, Agilent's product is not competing for whole-exome projects, he said, because it requires several probe libraries to cover all exons, while NimbleGen delivers these on a single array. That might change in the future if Agilent develops its product further, he said.
Turner said that the main advantage of the NimbleGen arrays is that “a whole exome array is available.” He added that “our results with this approach have improved dramatically over the last few months.”
[ pagebreak ]
Zhang told In Sequence this week that Roche NimbleGen recently conducted an internal comparison between data from the Broad Institute's recent Nature Biotechnology publication with Agilent, using an early version of its SureSelect technology, and data NimbleGen generated internally last year using its own Sequence Capture arrays. Both data sets shared a 468-kilobase contiguous genomic region from the ENCODE project and approximately 11,300 exons.
According to NimbleGen's analysis, its own arrays achieved better coverage of the contiguous region, while both platforms covered exons equally well. Also, the arrays appeared to generate more uniform coverage, Zhang said. However, the two groups used different sequencing platforms to analyze the captured DNA, which could account for some of the differences in the results: NimbleGen the 454 GS FLX Titanium, and the Broad Institute the Illumina GA II.
Zhang said NimbleGen plans to make the raw data for its own experiment available to customers, so they can perform their own comparison with the Broad data, which is available from the Broad Institute's website.
The company says it has approximately 80 customers for the capture arrays now, several of whom talked about projects involving the arrays at the recent AGBT conference. They included researchers from Baylor College of Medicine, Washington University School of Medicine, and the Jackson Laboratory.
For example, Jon Armstrong, a researcher at the Genome Center at Washington University, presented an internal performance comparison between NimbleGen arrays, coupled with 454 sequencing, and an in-house solution-based capture method called WUCap, combined with Illumina sequencing. He found the results from both methods to be comparable, though the cost of WUCap per sample was lower “when amortized over larger sample sets.” He also said WUCap would be easier to automate and requires less input DNA.
Agilent and NimbleGen are not the only companies working on selection or capture methods for targeted resequencing. RainDance Technologies, for example, launched a technology this month that allows users to amplify up to 4,000 targets by PCR in microdroplets (see other article in this issue). Also, Febit recently said it is going to launch its HybSelect method next month (see In Sequence's sister publication, BioArray News, 2/17/2009), and Olink Genomic said last year that it plans to start offering a multiplexed amplification service in 2009 (see In Sequence 9/9/2008).
Further, a researcher from the molecular informatics group of Rosetta Inpharmatics, a subsidiary of Merck, said at the AGBT conference that the company has been developing a solution-based capture protocol and plans to “introduce” whole-genome exome sequencing in the near future. A spokesperson for Merck provided no further information on the technology or how the company intends to commercialize it.