German software firm Definiens, which in the past two years has found a niche in the life sciences market for its broadly applicable image-analysis software, last week further solidified its position by inking a licensing and co-marketing agreement with RNAi-based research firm Cenix BioScience.
Under the terms of the agreement, announced last week, Cenix will use Definiens' Cellenger image-analysis software as part of the high-throughput RNAi and high-content cellular analysis service Cenix offers to its customers, Cenix CEO and CSO Christopher Echeverri told CBA News.
In return, Definiens will receive an undisclosed licensing fee from Cenix. Further financial details were not disclosed.
According to Echeverri, the formal agreement was just the culmination of an unofficial partnership the companies have enjoyed for two years. "We've just been getting to know each other and have been working together on various things formally and informally over the past two years," Echeverri said. "I've been advertising [Definiens' software] already for some time. In many ways, this agreement is just the formalization of many things we were already doing."
Cenix's main business is using high-throughput RNAi knockdown combined with microscopy to provide target discovery and in vitro target validation in cells to "everyone from large pharma companies to academic labs," Echeverri said.
"In the late 1990's people spent a lot of money and put a lot of effort into what I would consider the first wave of functional genomics technologies — things like expression profiling and large-scale two-hybrid screening — to try to identify targets," Echeverri said. "Unfortunately, most of those technologies couldn't really give a direct link between genes and function.
So in the vast majority of those efforts, people end up with large collections of poorly validated target candidates, which still require a lot of validation before they can be selected and prioritized to start drug screening programs," he added. "This is really the problem that RNAi has answered, because it does tie the function directly to the gene, and you get a causal relationship."
Cenix has publicly disclosed service relationships with Bayer and Schering. In addition, Echeverri said, the company partners with several other large pharmas that have not yet been disclosed.
Echeverri said that Cenix decided on Definiens' software after exploring a number of other image-analysis packages looking for features that would best fit Cenix's needs, such as "how accurately you get your cell segmented, how versatile and how many different types of images and analysis routines you can do, and also how well it fits into our high-throughput data flow infrastructure.
"It's not so much that we hated everything else that was out there," he added. "We felt that in some ways, in our experience, we found that their package was perhaps less limiting than some of the other ones we saw."
As a specific example, Echeverri noted Cellenger's ability to initially identify where the cells are in a given image. He said that this task is relatively easy when working with a brightly labeled, non-confluent monolayer of cells.
However, cell populations with heterogeneous morphologies or that heavily overlap one another is another story. "Murphy's Law dictates that this is exactly what you'll end up with in these screens," Echeverri said.
"We were very impressed with the accuracy of [Cellenger's] segmentation of a number of different samples that we threw at [it]," he added.
Echeverri believes that most image-analysis packages — including those that are included with high-content screening instrumentation platforms — can perform crucial tasks such as "scoring cells that show good, bright fluorescence intensity above a given threshold, translocation events within cells, co-localization or loss of co-localization of multiple markers, cell morphology attributes, or nuclear morphology attributes.
"All of these things, I think, are offered by most of the packages," he added. "But the devil's in the details — the question is how well they do it."
In fact, Echeverri stressed that he sees Cellenger as complementary to platform software, and thinks that it could be used successfully in conjunction with such packages.
"I would love to see something like that happen," he said. "You may worry that you're getting a lot of redundancy, but in our experience, that kind of redundancy is not a bad thing. If one tool, for whatever reason, is not giving you accurate results, you try another one, and it may help."
Truer competitors to Definiens might be start-up biotech Vala Sciences, which provides reagents and software for conducting image-based cell assays for drug screening and cell biology research (see CBA News, 2/22/2005); and researchers from the Whitehead Institute who are in the process of developing an open-source cell-based image-analysis software package called CellProfiler (see 1/24/2005 issue of BioInform, a CBA News sister publication, and CBA News, 6/8/2004).
Although Cenix will be using Cellenger primarily for well- and plate-based cellular image analysis, the software finds application in a number of life sciences areas.
Definiens has already inked agreements in 2003 with German screening instrument provider Evotec Technologies for analysis of tissue and cell-based assays; in November with Wisconsin-based biotech Beecher Instruments to use Cellenger for tissue microarray image analysis (see CBA News, 11/23/2004); and in December with Danish biotech BioImage for co-marketing with that company's Redistribution fluorescent translocation assays (see CBA News, 12/14/2004).
Next up for the company may be pathology applications, at least according to Echeverri.
"[Definiens'] software originated completely outside the life sciences sector," Echeverri said. "They've been porting it to the life science sector, but among other things, it was developed for analyzing satellite imagery.
"They already are showing beautiful results with histology applications," he added, "and of course that gets into the huge field of pathology image analysis, which is a big, big need, I think. My impression is that they're just scratching the surface here."
Calls to Definiens were not returned in time for this publication.
— Ben Butkus ([email protected])