Paradigm Genetics, which launched a computational systems biology research group last year [BioInform 04-28-03], views its recent acquisition of pathology software firm TissueInformatics as a key step toward filling out its systems biology capabilities, according to Thomas Colatsky, VP of healthcare research at Paradigm.
TissueInformatics uses machine vision to extract high-resolution, quantifiable information about specific tissue features from normal histopathology slides. The ability to assign discrete numbers to features that were previously described observationally is extremely valuable in bridging the gap between genomic-level data and phenotypic-level data, Colatsky said. “Phenotypic data right now is probably the biggest challenge for people doing systems biology research, simply because it tends not to be as quantitative as you would like it to be,” he said. “Having the ability to dive down to the histological level, and do that in a highly quantitative and reproducible way, and, more importantly, to take all of the objective bias out of that, really extends the systems biology platform in a way that is going to be uniquely suited to our goals.”
Systems biology research hangs on the integration of large-scale data sets, and most groups start with some combination of genomics, proteomics, and metabolomics data. From there, some groups are adding data from the scientific literature, high-throughput screening, cellular assays, or other resources, but few have thrown pathology data into the mix. But Colatsky said that pathology data is the cornerstone of one of Paradigm’s primary research activities: toxicology. “It is probably unique right now because pathology has been something very hard to take out of the observational realm and put into the analytical realm,” he said, but added that access to TissueInformatics’ technology may enable the company to “move systems biology from where it’s currently sitting, which is more on the discovery side, into an area where very important decisions are being made in terms of the viability of a drug or it’s ability to harm or benefit patients.”
Paradigm’s acquisition of TissueInformatics grew out of a collaboration the companies entered last March to study drug-induced liver toxicity. “At Paradigm, doing phenotypic analysis was something that was outside of our current core competencies, and understanding what was going on at the tissue level we began to believe more and more was something that we needed to have in house, and be able to work with and manage,” Colatsky said. The company took steps to acquire TissueInformatics before they got “snapped up by another suitor at some point downstream.”
The all-stock acquisition, which could be worth as much as $8.5 million based on Paradigm’s stock price on Feb. 5, is scheduled to close in March. Following the acquisition, Paradigm will employ 225 people and it will continue to be based in Research Triangle Park. Colatsky said that the 24-person TissueInformatics group will continue to develop and market its software products from its Pittsburgh headquarters, “but the supporting structure — both in terms of pathology support, as well as the extensions of that data stream into our other data streams in the overall systems biology approach that we have — that will become something that will be company wide.”
Based on their existing collaboration, the two companies have already made some headway in integrating pathology data with gene expression and proteomics data, and TissueInformatics is also engaged in a similar project with biotech firm DiAthegen that involves “pulling correlations out between gene expression and tissue feature changes,” Colatsky said.
Paradigm’s healthcare research group is currently engaged in several pilot studies with biotech and pharmaceutical firms, but Colatsky said he was unable to disclose the names of any potential collaborators. While the company’s previous association with TissueInformatics helped it secure some of those pilots, he said, bringing that technology in house “gives us the capability of presenting a more complete package in the systems biology area for life sciences than we had before, and it actually makes the integration of that data stream a more realistic part of the overall systems approach.
“Rather than have it be kind of Chinese menu add-on, it’s actually part of the offering at this point, and should be very beneficial in getting some additional interactions with other companies,” he said.