Digital Cell Imaging Labs, a 15-person cellular-imaging software firm based in Edegem, Belgium, is taking a slightly different tack from its HCS image-analysis rivals.
The company, founded in 2001 as a spin-off from an image-processing research group at the University of Antwerp, has carved out a comfortable niche in the market by specializing in custom-written software for complex or novel biological assays.
Though the high-content screening market in which it competes is becoming increasingly crowded, DCI Labs says its bespoke offerings — in which each algorithm is made “from scratch” — give it a competitive advantage.
DCI Labs offers a portfolio of algorithms for common assays, but around 70 percent of the company’s time is spent on custom work, according to Barbara Weyn, managing director of DCI Labs.
“Our slogan is, ‘If it can be simple, keep it simple,’” Weyn said. “We’re not here to do the simple things. If people come to us and say, ‘We want to do a cell count,’ I usually say, ‘Go and download an NIH software package — you can do this yourself.’
“But images are becoming more and more complex, and also there are more imaging modes — more and more brightfield images — and to analyze those very difficult images, you need to have core expertise,” she said. “You need to have background in image analysis to extract the information.”
Weyn said that the company also sees demand for customization of standard assays that need to be tweaked to work with different formats or specific cell types, as well as novel algorithm development for proprietary assays.
Weyn declined to name specific customers, but she said that the company’s client base is around 80 percent pharmaceutical firms and 20 percent academic groups. All of DCI Labs’ customers to date are based in Europe, but Weyn said that the company is hoping to add some US customers soon.
DCI Labs operates in a segment of the high-content screening market that is growing increasingly competitive. A number of software companies, such as Definiens, Bioimagene, SVision, and Vala Sciences offer products targeted toward HCS image analysis.
These firms join instrumentation vendors like Cellomics, Molecular Devices, and BD Biosciences, which offer their own software, as well as several open source packages. Recently, scientific software firm Accelrys threw its hat in the ring by launching a suite of image-analysis applications for its SciTegic Pipeline Pilot workflow platform [CBA News 01-26-07].
But DCI Labs sees a competitive advantage in its focus on customization.
Weyn noted that most software packages in the market are black-box offerings that work fine for standard assays, but don’t offer any flexibility for researchers who decide to take a different approach that might require modification of the algorithms.
“We’re different in the sense that we focus on more difficult algorithms — brightfield [and] confocal assays, where customization is a priority — so we really make sure that it works on their cells under every biological condition,” she said.
|
“We think that algorithm development is an area for specialists and not for cell biologists, so we make each algorithm from scratch and we have very highly trained specialists who do that.” |
One company that does bear a slight resemblance to DCI Labs is Definiens, whose Cellenger platform allows customers to build their own analysis workflows using a suite of components [CBA News 11-10-06].
However, Weyn said, “we think that algorithm development is an area for specialists and not for cell biologists, so we make each algorithm from scratch and we have very highly trained specialists who do that.”
The advantage of this approach, she said, is that the algorithms are written from the ground up by specialists so that they are not only accurate, but fast.
As an example, Weyn noted that in a comparison of a micronucleus assay algorithm, “ours is a little bit better than Definiens’, but probably not enough to make a difference, to buy an entirely new package if you already have it.” However, she noted, “it’s 20 times faster, so when speed is an issue, we have an advantage over the others.”
Development time for a typical algorithm can take several months, she said.
Weyn said the company takes a platform-independent approach to algorithm development. It offers importers for the Evotec and GE Healthcare platforms, and has a software-bundling agreement with Maia Scientific for several of its applications, but none of these are exclusive agreements, she said.