As the rise of next-generation sequencing has driven demand for new analytical tools, many software companies that have long served array users have shifted their resources to this rapidly growing segment of the market.
But microarray users haven't been left behind as a result of this trend, according to sources familiar with the industry who told BioArray News that these new NGS analysis tools will have a "trickle-down effect" or "fringe benefit" for array clients.
Furthermore, some software vendors expressed skepticism that next-gen sequencing would eventually replace array technology, and stressed that they market their tools for data analysis in general, forgoing any platform-specific specialization.
"Our service business is increasingly bridging NGS and microarrays, and combining array data sets with targeted sequencing and large-scale sequencing," said Christophe Lambert, CEO of Bozeman, Mont.-based Golden Helix, which sells an analysis tool called SNP and Variation Suite.
Lambert argued that NGS analysis is a must-have for bioinformatics firms amid a federal funding environment that "favors" the technology.
"Because the funding environment is pushing toward the bleeding edge, we have to serve the research community and their push toward that edge," he told BioArray News. "You have to allocate your resources with the idea of the future, more than sitting and capitalizing on a cash cow," he said.
At the same time, Gabe Rudy, vice president of business development of the 13-year-old firm, said that the company's recent focus on building tools for NGS does not mean that the same tools could not be applied to array-based studies.
"We still see microarrays as being a very useful tool, and there's a price point there where they will continue to make sense, especially … the exome arrays that are going to include rare variants," Rudy told
Richard Resnick, CEO of Westborough, Mass.-based GenomeQuest, described a situation where the advancement of sequencing data-analysis tools benefits users of both technologies "There is both a trickle-down effect from NGS to arrays and a trickle-up effect in the opposite direction," Resnick told BioArray News.
According to Resnick, as both DNA resequencing and genome-wide arrays generate information about sample-specific genetic variation, the firm's customers are interested in viewing that data side by side. "We have many pharma customers who actually start by uploading all of their array data so that they can then use it in the context of new genomic data," he said.
He also noted that whole-genome sequencing is feeding the array market, as some research groups choose to build custom genotyping arrays that are "better designed for specific applications such as population-specific genotyping … or diagnostic approaches where very particular genotypes are being measured."
Soheil Shams, CEO of El Segundo, Calif.-based BioDiscovery, said that his firm is indeed developing a new product that will have NGS analysis capabilities. At the same time, he said that the 14-year-old firm does not intend to offer platform-specific tools.
"In the end, it doesn't matter what the tool is, it's just data," Shams told BioArray News. The company currently offers two main analysis tools, Nexus Copy Number for comparative genomic hybridization and SNP array data, and Nexus Expression, for microarray gene expression analysis.
In terms of adding sequencing-specific capabilities, he said that there are "issues about how you manage the large volumes of data, but to me, those are technical issues that are not of concern to the biologist."
[ pagebreak ]
Shams claimed that most NGS analysis tools are targeted to bioinformaticians, rather than end users. "Everybody out there that has a product for next-gen sequencing, that product is not for the biologist," said Shams. Biologists "are asking questions about how CNVs might relate to a certain disease, not off of which technology did I generate the data," he said.
Shams did not offer a release date for the NGS-compatible software BioDiscovery is currently developing, but said that the company must make "architectural changes in how the data is handled," though it will retain the same interface as Nexus.
He also cautioned users against believing that new tools or more bioinformaticians would expedite the answers to their research questions.
"Just like you cannot get nine women to give birth in one month, because you can't spread [the pregnancy] across, you cannot get 100 informaticians to solve your problem in a day," he said.
He was also wary of what he referred to as "hype" surrounding next-gen sequencing. "If I labeled a product for next-gen, I would probably get some customers," said Shams, "but it will be obsolete within a year."
Defying Categorization
Bioinformatics firms that BioArray News spoke to were split on how best to tackle the challenge of developing data-analysis tools for next-generation sequencing users. While some prefer to make tools for analyzing particular types of data, such as SNP, CNV, or expression data, others market their products for use with specific technology platforms. Ultimately, though, software companies interviewed for this story said they seek to serve users of both arrays and sequencing — at least to some extent.
A spokesperson for DNAnexus, a Mountain View, Calif.-based sequence analysis company founded last year, told BioArray News that while the shop has no plans to add array analysis capabilities to its software, it does hope to add functions to allow customers to integrate legacy SNP array data into the analysis.
"We do want to be able to have customized files, so that users can upload different tracks into their genome browser [and] can look at sequencing data and SNP data," the spokesperson said. "In terms of the primary analysis of SNP chip data, I am not aware of any plans to have that," he said. "We are really focused on the sequencing data."
Others are more platform-agnostic. "Aside from data import, we tend not to categorize data into NGS data or array data, but rather numeric versus variant or SNP data," said Doug Robinson, a life sciences specialist at SAS's JMP Genomics business.
He said that by approaching data analysis based on the type of data rather than the platform, "data processing is consistent across various sources and integration of data from different sources is quite simple."
Robinson told BioArray News that the firm has been developing methods for analyzing combined data sets in its flagship software tool, JMP Genomics, "for quite some time.
"We can examine correlations between two numeric data sets from the same samples across all possible combinations," said Robinson. "This can be used for copy number versus gene expression, miRNA versus gene expression, and so on, regardless of the data source," he said. "Additionally, our existing genetic analysis applications can use variant call files from NGS or SNP data from arrays or combinations of the two."
Robinson noted, however, that the Cary, NC-based firm is currently focusing its efforts on developing sequencing-related tools. "We are actively adding modules that address new issues found with NGS data," said Robinson, who cited normalization of RNA-seq data as a particularly "hot topic."
Additions to the next version of JMP Genomics will include a host of sequencing-focused tools including new methods for data normalization and variant calling, he added.
"Our future software development will continue to concentrate on providing analytic tools for genomic data analysis — whether it is NGS data or data from other sources," said Robinson. "Clearly, NGS data provides somewhat different information than array data, and so we will be building tools to take advantage of this, but many of these applications will also be able to take advantage of array studies or allow for combining array information with NGS data."
[ pagebreak ]
Rise and Fall
Jens Hoefkens, head of Genedata's Expressionist business unit, said that the Basel, Switzerland-based company is unfazed by the need to serve next-gen sequencing customers.
"This is the second time this has happened, because we have been doing a lot of work in the areas of proteomics and metabolomics based on mass spectrometry," said Hoefkens. "For us, the whole process of expanding our product is one that we have already done before."
According to Hoefkens, Genedata does not intend to develop a sequencing-specific software tool. Instead, the firm plans to add functionality to its current software for sequencing users. And he noted that those additions often benefit array users too.
"This is something that has kept our existing microarray clientele over the past few years, in that we continue to provide new methods and tools for their data, despite the fact that our main effort is no longer in array analysis," said Hoefkens. "But it is a fringe benefit of the way we structure our software that we continue to provide new methods to them."
Hoefkens said that Genedata is keen to serve array users, as he expects the market for array analysis tools will continue to exist for some time.
"I think it will be awhile until someone publishes an NGS study with 2,000 samples," said Hoefkens. "I think microarray data are going to continue to play an important role in many applications for a long time to come," he said. "It is a no brainer for us to continue to support them, though that is not to say that NGS isn't gaining importance."
One software firm that has chosen a technology-specific path is Madrid, Spain-based Integromics. The company currently offers three tools as part of its OmicsOffice analysis tool: Integrated Biomarker Discovery, or IBD, for array expression analysis; real-time StatMiner for quantitative RT-PCR analysis; and SeqSolve for analysis of next-generation sequencing data.
Michael McManus, CEO of Integromics, told BioArray News that the company wants to enable users to "compare between these three types of analyses because we don't think that transition to sequencing is going to happen as fast as everyone is saying."
He also noted the relevance of array data in submissions to the US Food and Drug Administration. "Part of that process has to do with how to validate and certify that a small molecule that you are suggesting is a drug actually is affecting the protein you address. Right now the standards for that are microarrays," he said.
According to McManus, proponents of sequencing have not yet addressed how the newer technology will be able to replace arrays in the regulatory environment and other settings.
"I have seen this many times — that a new technology comes out and the guys who are using the technology instantly assume it will replace everything and don't take the practical point of view of how this future method will supplant those standardized methods," he said.
"Of course it is disruptive," McManus said of sequencing. "What is not being talked about is how that disruption is converted to revenue by displacing these historical systems."
Golden Helix's Lambert predicted, however, that in markets where array technologies are just now becoming the standard, next-gen sequencing tools will ultimately become the technology of choice.
"There is likely to be a migration of those people over to sequencing," he said. "Eventually sequencing technologies will replace it all, except for very specific clinical applications."
Have topics you'd like to see covered in BioArray News? Contact the editor at jpetrone [at] genomeweb [.] com