Golden Helix and Agilent Technologies this week announced that Golden Helix will provide researchers with software tools and services for Agilent’s expanding menu of copy number variation microarrays.
Agilent markets its own microarray analysis software under the GeneSpring and DNA Analytics brands, and the company said it plans to eventually develop the informatics tools necessary to run and analyze data from copy number association studies. However, a company official said this week that the arrangement with Golden Helix will immediately provide customers interested in running CNAS with the software and services required to begin their studies.
As part of the deal, Golden Helix said in a statement that its SNP & Variation Suite has been "modified to directly import, analyze, and visualize Agilent’s genome-wide copy number data, including association studies, data prep, quality assurance, interaction analyses, study review, and predictive modeling for research purposes." Bozeman, Mont.-based Golden Helix has also expanded its analytic service offering to incorporate studies based on Agilent’s CGH/CNV microarray data.
The companies did not discuss any financial components of the deal, but pledged to engage in co-marketing activities that will become more frequent as customers progress with their copy number association studies.
Agilent has launched two CNV-themed arrays over the past two months to serve customers interested in looking at copy number variation as a component of genome-wide association studies. Last month, the Santa Clara, Calif.-based firm debuted its Human CNV Association 2x105K microarray, designed for studying associations between genomic CNVs and disease. The 105K array is based on a design by the Wellcome Trust Case Control Consortium, which is currently conducting a study of CNVs and their relationships to a number of widespread disorders, including Crohn’s disease, tuberculosis, diabetes, and bipolar disorder (see BAN 4/7/2009).
In March, Agilent also launched its 2x400K CNV array, which was designed to cover known CNV regions from the Database of Genomic Variants. Agilent is also planning to launch a million-feature chip in a few months that will include probes evenly tiled across the genome. All CNV-themed arrays are made possible by the firm's next-generation SurePrint G3 microarray platform, which contains up to 1 million probes on a standard 1x3 inch glass slide — up from 244,000 probes on the last version of its chips (see BAN 3/3/2009).
As part of the CNV array launch, Agilent has been upgrading its bioinformatics support for dealing with the arrays, specifically within its DNA Analytics software. At the same time, the company has been looking to work with third parties to enable use of its chips in genome-wide association studies, according to Dione Bailey, Agilent’s CGH and CNV array product manager. The firm's new relationship with Golden Helix, with whom it has not directly worked before, is an example of such a partnership.
"With the new CNV arrays that we have been launching, one of the things that has been lacking has been a [software] tool that can support large-scale CNV association studies," Bailey told BioArray News this week. "This is something that Agilent does not have right now within its existing analysis software," she said. Golden Helix has "a lot of experience doing association studies that involved the Illumina and Affymetrix genotyping platforms," she added.
Bailey said that a number of Affy and Illumina SNP array customers now want to expand their association studies to survey CNVs using Agilent's chips. Since "a lot of those customers" are familiar with Golden Helix's SVS software, it "put Golden Helix at the top of the list" when Agilent sought a partner to provide informatics support for its new chips.
"When you look at CNV associations, you have to handle very large data sets and you need to use analysis tools that are set up for doing these types of studies," Bailey said. She said that Agilent is "definitely looking at developing its own analysis tools," but in terms of a "short-term solution, it wanted to partner with a leader in the field."
Agilent is new to the association studies market. Since launching its microarray platform, it has regularly increased the density and multiplex formats of its arrays while expanding its applications menu into areas such as gene expression, DNA methylation, and miRNA expression profiling as well as chromatin immunoprecipitation and comparative genomic hybridization. But, until now, it has left the GWAS market to Affy and Illumina. Because of this, Agilent's current informatics toolbox contains functionalities for those areas with which it has the most experience, but not for copy number association studies.
Bailey added that most of its customers that are interested in conducting such studies have not yet even begun analyzing data, but that the deal with Golden Helix should allay any worries about not having the necessary tools in place once they reach that stage.
"Many customers are at the beginning of studies in the collection phase," Bailey said. "There are a few customers that are farther along, such as at the Sanger, but they tend to write their own algorithms anyway," she said. "We have many customers, though, [who] are scoping out CNV association studies, and we hope that this will encourage them to proceed."
Bailey declined to provide names of customers but said that the two firms will turn to researchers using Agilent CNV chips and Golden Helix's analysis tools to demonstrate the integrated setup in the future.
[ pagebreak ]
Representatives of Golden Helix agreed that it may take some time before Agilent users actually begin adopting SVS to analyze data from copy number association studies. They hope the deal with Agilent will turn those customers their way.
"We are a bit ahead of the curve, but anyone who is considering doing these kinds of studies might be put off by the concern of availability of appropriate data analysis tools," said Andy Ferrin, director of sales and marketing at Golden Helix. "I suspect that now, with the availability of these analysis tools, there is more of an incentive for people to begin their studies."
In February, Golden Helix launched the seventh version of its SNP & Variation Suite for SNP and copy number analysis. The new version was designed to overcome data-capacity constraints, limited visualization capabilities, and data-workflow challenges associated with high-density SNP-genotyping arrays and second-generation sequencers (see BAN 2/10/2009).
Among the new features included in SVS 7 were interactive visualization and an integrated genome browser; real-time data editing and enrichment; advanced predictive modeling; quality assurance and stratification correction; and enhanced genome-wide association analyses including the ability to use genetic covariates in regression.
Though the upgrade to handling Agilent data for CNV association studies comes just a few months after Golden Helix rolled out the new system, Ferrin said that making adjustments within SVS was "not difficult" and the company's software and services are now compatible with Agilent data.
"This was largely a matter of importation," Ferrin said. "We had the infrastructure already and it was primarily a matter of making the importation of Agilent data simple and fast," Ferrin said. The format of Agilent input files is "structurally different" from Illumina or Affymetrix data files and so Golden Helix had to create an importation wizard that would be able to handle Agilent-based projects, he added.
Though Golden Helix has experience working with Affy and Illumina-driven GWAS projects, Ferrin said that informatics challenges for Agilent-based CNV association studies may be different.
"Copy number analysis is gaining in importance and popularity," Ferrin said. "But the challenge is the quality of the data. It has long been assumed that SNP data is of decent quality, but that is not the case for copy number data," he said. According to Ferrin, SVS 7 includes special quality control infrastructure to mitigate the batch effects and noisy data that degrade CNV data on different vendors' platforms.
However, he said that Agilent's chips so far do not exhibit "significant quality problems" and that the QC functionality designed to handle CNV content on other arrays will make Agilent data better, though it is not as necessary as it would be for other platforms.