Cogenics, the genomic services arm of Clinical Data, is working with Rosetta Biosoftware in hopes of gaining an informatics edge in the competitive market for microRNA expression-profiling services.
This week, the company said that it is offering miRNA-analysis services based on an updated version of the Rosetta Resolver microarray-analysis software with new features for handling miRNA data — and that it is the only service provider in North America with access to this capability prior to the official launch of the next version of Resolver later this year.
Cogenics officials said that this arrangement, which grew naturally out of a collaborative relationship between the firms dating back to 2000, should give it an important head start in a quickly growing market.
“This gives us a period of time where really we’re the only company who can provide these capabilities,” said Clive Higgins, general manager, North America. “It’s a way of us continuing to differentiate ourselves as a fee-for-service organization from any of the other providers out there that may wish to get into this space or compete with us.”
Cogenics began offering miRNA profiling services on the Agilent Technologies miRNA array platform last March. Higgins said that the company’s first customers were primarily in academia, but that it is now seeing more interest from large pharmaceutical firms.
Ed Lobenhofer, director of scientific affairs at Cogenics, said that the company is seeing “very significant growth” in this part of its business, but declined to provide specific details.
Certainly, microRNA research is on the rise. A PubMed search on “microRNA OR miRNA” shows that only four papers on the subject were published in 2004, but that rose very quickly to 367 in 2005, 636 in 2006, and 942 in 2007.
Cogenics isn’t the only service provider looking to meet this demand, however. Other miRNA profiling service providers include Invitrogen, which uses its own NCode miRNA arrays; Asuragen, which uses both TaqMan assays and Agilent arrays; Miltenyi Biotec, which uses its own miRXplore microarrays; Febit, which uses its own arrays; Exiqon, which uses its miRCURY array platform; and Toronto’s University Health Network Microarray Center, which uses the Agilent and Exiqon array platforms.
Higgins said that Cogenics views improved analytical abilities as a key competitive advantage in the market. “I think it will help us to establish our reputation as really a top-notch provider, and to differentiate ourselves from the other service providers,” he said.
‘A Learning Process’
When Cogenics began offering miRNA profiling services last year, “we found that early customers wanted us to not only provide the actual sample analysis, but the analysis of their data as well,” Higgins said. However, he added, “options at the time were pretty limited given the newness of the technology and the science. They were either very limited or they were very labor-intensive.”
At first, Lobenhofer said, Cogenics relied on “brute force methods” for analyzing the data, but soon began discussing with Rosetta about adapting Resolver to handle miRNA data in addition to gene-expression data.
Now, he said, “we’re seeing dramatic improvements in the efficiency and speed at which we can analyze data by having an integrated software package as opposed to a brute force method.”
Higgins estimated that the speedup has been as much as 10-fold. One of the key advantages, he said, is that the company can use the pipeline it had already established for feeding gene-expression data into Resolver. It also allows Cogenics researchers and customers to integrate miRNA data with gene-expression data in the same analytical environment, he said.
“It enables us to do all of the standard analyses that we’ve been deploying for quite some time using mRNA data,” such as hierarchical clustering, biomarker discovery, principal component analysis, and the like, Lobenhofer said.
“We’re getting a lot of the same questions that we do with mRNA gene expression profiling, but we’re now looking at a smaller number of targets with microRNAs.”
However, he said, “by having the data integrated into a common data warehouse with built-in analysis tools, it really facilitates the speed and efficiency at which we can perform the exact same analyses that we have been doing on mRNA, but this time on miRNA.”
Lobenhofer warned that the miRNA field is still young, so there are a number of unanswered questions regarding data analysis. “A lot of this is still a learning process for us,” he said. “We’re getting a lot of the same questions that we do with mRNA gene-expression profiling, but we’re now looking at a smaller number of targets with microRNAs — in the high-hundred level” as opposed to tens of thousands for gene-expression profiling.
One issue Cogenics and Rosetta are currently looking at is whether the normalization schemes that have been developed for mRNA analysis can also be applied to miRNA analysis. In mRNA profiling, which involves tens of thousands of transcripts, “the null hypothesis is that a given transcript is not changed when you compare, and that is what allows you to normalize your data and identify things that are different,” he said. “What we’re trying to determine now is whether that null hypothesis can still hold true in the analysis when you’re looking at a smaller number of biomolecules, specifically the known microRNAs.”
Another challenge is data integration. “We have seen some clients interested in just profiling miRNA, but I think the trend is toward profiling both microRNA and mRNA and seeing how the data integrate with each other,” Lobenhofer said. “The standard theory right now is that each miRNA regulates between 200 and 300 transcripts, so obviously the integration of those data becomes a bit of a challenge and that’s why we’re very pleased to have both of those data streams in the same analysis system.”
Other companies are looking at these same issues. Software firm Partek is currently in the “advanced stages” of developing miRNA expression-analysis capabilities for its Genomic Suite platform, Mike Lelivelt, vice president of genomics, told BioInform this week. He declined to provide a timeframe for when the company plans to release the new software.
Like Rosetta and Cogenics, Partek believes that the key to a successful miRNA-analysis platform in a seamless analytical workflow and an ability to integrate with other types of data.
“We do see growing interest in miRNA profiling, and I think … it’s interesting in and of itself, but the power of microRNA profiling really comes in when it’s intersected with other profiling aspects of the genome,” such as copy number data, mRNA expression data, and ChIP-chip and ChIP-seq data, he said.
Lelivelt said that the current version of the Partek Genomics Suite can already analyze miRNA expression data, but with “some limitations.”
The new version of the software will offer “streamlined workflows” that will serve as “out-of-the-box solutions that can be approached by the biologist instead of the bioinformatician,” and “workflows that pull together expression data with miRNA data in a way that you can derive real biological inference from them in context.”
Not every microarray analysis firm plans to modify its software for miRNA analysis, however. Kurt Zingler, head of US business at Genedata, told BioInform via e-mail that the company’s Expressionist software can already “easily analyze miRNA arrays.”
In addition, he said that the company has not yet seen much demand for this capability, though he noted that this may be due to Genedata’s focus on pharma and biotech companies, which “have been a bit slower to adopt newer technologies as compared to academics.”
Another firm selling microarray analysis software, SAS subsidiary JMP, doesn’t have any immediate plans to update its JMP Genomics platform. The software “doesn't currently have any specific applications for analyzing miRNA data, but if the data is contained in a text file, JMP Genomics should be able to handle it,” a company spokesperson told BioInform via e-mail. “We have not had customer requests for miRNA-specific capabilities, but they may be incorporated in future versions of our software,” she said.