A high-content screening platform is only as strong as its image-analysis software, an axiom that German software firm Definiens is banking on as it makes its first official foray into the high-content screening arena.
Last week, Definiens announced it had signed a non-exclusive agreement with Danish biotech BioImage in which the two companies will co-market their respective products for HCS to existing and potential customers.
Specifically, Definiens will provide Cellenger, its automated image-analysis software, while BioImage will contribute its Redistribution cell-based fluorescent protein-translocation assay technology.
In addition, the companies said in a statement that they will further develop specific applications for HCS based on these technologies to enable customers to “perform robust data extraction from high-content cell-based assays.”
The deal is certainly welcome to BioImage, which has forged several agreements over the past year in relation to its assay technology — including ones with other tool providers, pharmaceutical companies, and biotech firms — in an attempt to generate revenue in continued support of its own internal drug-discovery program.
But the agreement may be even more important for Definiens, which, according to CEO Thomas Heydler, looks to grab a share of an estimated $50 million to $100 million market for high-content image analysis software for cell-based assays.
“We see many areas of interest [for image analysis], but we see a lot of market momentum developing in specific categories, and one of those is high-content screening,” Heydler told Inside Bioassays last week. “We have done some very rough estimates to make a decision as to whether we want to engage in an active investment in that area, and our conclusion is that it is worthwhile.”
Cellenger is based on Definiens’ core IP, which protects a method for analyzing images of any kind using proprietary object-recognition algorithms. Heydler said that the company expects this method to give Definiens a unique advantage in HCS analysis.
“Anything a human can see, we can actually find with our software,” he explained. “It’s based on a cognitive, or object-oriented, approach to image analysis, and is different from other [products] traditionally found on the market, which are pixel-based methods.”
Heydler said that in those approaches, “you’re trying to match images to something you have already described previously. But as soon as the images begin to get more complex, or you want to use more colors, or you want to extract multiple parameters out of those images ... traditional methods are no longer working.”
The BioImage deal will mark Definiens’ second foray into the HCS marketplace, the first being a 2003 deal that saw Evotec and Definiens create an interface between Evotec’s Opera HCS instrument and Cellenger. Definiens has also utilized this core object-recognition technology with some success for a number of other biological applications such as tissue analysis and proteomics.
Early this year, it signed an agreement with Palm Microlaser to co-market and co-develop automated image analysis for laser microdissection of tissue sections. And just last month, it inked a deal with Beecher Instruments to use Cellenger to analyze tissue microarrays (see Inside Bioassays, 11/23/2004). Definiens’ object-recognition technology also underlies Proteomweaver, a software platform for 2D protein gel analysis.
“We basically have a platform technology, which is a very general image analysis platform that can be applied to all sorts of images in the overall pharmaceutical sciences values chain — whether we talk about cell-based images, tissue-based images, organs, and all the way to diagnostic and MRI images,” Heydler said.
“So we have a pre-developed package of image-analysis capabilities — specifically root sets — allowing us to address those high-content screening opportunities very quickly,” he added, “and to first and foremost allow a biologist to co-develop the biological assay together with the image analysis without requiring a very deep IT specialist.”
Although exact numbers are not known, it is likely safe to say that the biggest players in the image-analysis portion of the high-content screening market are Cellomics and Molecular Devices, each of which offers soup-to-nuts HCS platforms, including image analysis software. Both of these companies sell their image analysis software separately, but the products are optimized for their respective screening instruments.
In many ways, the Definiens-BioImage agreement is akin to a co-marketing agreement inked last year between Cellomics and Norak Biosciences, a US-based biotech firm that, like BioImage, sells a fluorescent protein translocation assay to support its fledgling drug-discovery business.
But Heydler said he doesn’t view such companies as competitors. In order to drive its product into the high-content screening market, Heydler said, Definiens plans to partner with “leading high-content screening companies both on the image acquisition side and assay side,” and thus would not rule out companies such as Cellomics and Molecular Devices.
“I wouldn’t call them competitors actually,” he said. “You can list a whole bunch of people offering these products, and we see all of them as possible partnership opportunities to broaden our market reach, and to actually complement their offerings with a much more advanced image-analysis method.
“If you look at some[thing] like Cellomics’ portfolio of software, I think there is a distinction to make between the pure-breed image analysis on one side, and what they have, the [Cellomic Store], which is really around the data management, and some of the decision-support stuff,” he added. “We do not consider that a competitive offering, but rather a complementary offering.”
Although Heydler declined to disclose specific potential partners, he did say that the company is currently speaking with several instrumentation and assay providers about possible agreements, and that Definiens would likely release news on such partnerships in the next few months.
One might speculate that PerkinElmer would be a future participant in such a deal. In September, BioImage and PerkinElmer forged a partnership to develop cell-based assays based on Redistribution technology on the PerkinElmer EnVision HTS plate reader. Although the EnVision is not a high-content screening instrument per se, the company has indicated that it may be developing a platform in this area (see Inside Bioassays, 9/7/2004).
On the flip side of the coin, the Definiens agreement may help serve as a bridge between BioImage and other high-content screening instrumentation providerts. In fact, Len Pagliaro, BioImage’s vice president of business development, yesterday told Inside Bioassays this week that it has validated five of its assays on the Evotec Opera in what essentially constitutes a co-marketing agreement, although it hasn’t yet been officially announced.
BioImage has also been using GE Healthcare’s IN Cell Analyzer 3000 platform, and believes the Definiens software may come in useful as an alternative to the default image analysis software packaged with that instrument. Of course, Cellomics also has an agreement with GE Healthcare in which its image analysis and data mining software is interfaced with the IN Cell.
“There are clearly customers who have preferences for screening instruments,” Pagliaro said. “So it is beneficial to us to be able to say, ‘Yes, our assays work on that platform.’”
Pagliaro said that the image algorithms that accompany the IN Cell are “OK,” but that BioImage had been working with Amerhsam, and now GE Healthcare, to improve them. “But we haven’t been able to improve them as fast as we would like,” he said.
He also said that Cellomics’ ArrayScan algorithms provide good analytical capability, but “are not the best in terms of throughput.”
“We felt it would be very advantageous for us to partner with a company, in Definiens, offering what we feel is a state-of-the-art software platform,” Pagliaro said. “We also thought we could provide a lot of value to them by providing HCS data that we’ve obtained using other platforms, so that it can be compared with their results.”
As the HCS market quickly expands, it will be interesting to see whether companies that traditionally provide entire HCS solutions will scrap certain portions to focus on single aspects, such as software, instrumentation, or assay technology — or whether demand will remain high for an integrated solution. Either way, companies continue to scramble to get a piece of the pie.
“It’s very hard to quantify the HCS market space, because there is so much going on right now, the data is changing every minute,” Heydler said. “The good news is that it’s changing in the right direction because it is getting bigger. More and more people are really starting to think about this.
“To us it really … looks like cellomics — using it in the context of genomics and proteomics — is the next new wave,” he added. “We see a major investment effort by some of the large [pharma] clients, but also in biotech companies, where they’re starting to see a huge opportunity for using these cell-based assays.”