Courtesy: Flickr/epSos.de

At the Semantic Technology and Business Conference in New York City last week, attendees got to hear a lot about how semantic technology is influencing various sectors, such as government (see our stories here and here) and media (see this article and this one). Another prominent one on display: pharmaceuticals.

Pharma, for example, was the driving use case for the update to Callimachus that focuses on helping users deal with data that’s external to the framework for data-driven applications, David Wood, CTO of Callimachus project sponsor 3 Round Stones, told The Semantic Web Blog at the event. (To learn more about the update, see our story here.)

A session on Tuesday last week saw Lee Feigenbaum, vp of marketing at Cambridge Semantics, which makes the Anzo express and Anzo Enterprise solutions,  put forth a case for semantic tech as being key to data integration and interoperability in the sector, as well. “Can semantic web technologies break down enterprise data silos just as they break down document silos on the web?” he said. “The answer to the question is, “Of course.” Compared to the web, the data silo challenges of even the largest pharmaceuticals organization is relatively minor.”

Semantic technologies, he says, are used less for linking scientific data together to answer scientific questions, and more for R&D, drug development, clinical, sales, marketing and finance. One example that perhaps isn’t as often cited: supply chain management. One pharma company Cambridge works with deals with a few dozen contract manufacturing organizations, collecting data feeds in random formats and having to combine them with internal database information to produce metrics.

“In the past they could track 7 KPIs (key performance indicators) to tell them how well the manufacturing part of the supply chain is doing,” he said. “Leveraging the ultimate flexibility of these technologies, within a couple of months they went up from 7 KPIs to over 30 and added the flexibility of letting end users define new metrics as they saw fit.”

In a conversation with The Semantic Web Blog following his presentation, Feigenbaum said that in pharma – as well as in other industries, like financials – semantic technology “lets experts drive the direction of things instead of being constrained by IT systems.” Take the area of competitive intelligence: Instead of being held back by fixed and limited reports, business people at pharma firms can consider as wide a set of information as possible and as quickly as possible, so that it doesn’t take six months to a year after making a strategic shift in direction to figure out what new companies they should be tracking.

Among Cambridge’s business-specific apps is Competitive Intelligence for Pharma. Cambridge shortly will release V.3.1 of its Anzo technology which will let users auto-generate relational tables from ontologies.” You take the semantic data and create an optimized, normalized database from it,” he says. “That’s one way to deal with the quality of big data and also to work with existing tools.”

Last Wednesday continued the conversation about pharmaceuticals, with a panel discussion on using linked semantic data in biomedical research and pharmaceuticals. One point made there was how semantic linked knowledge can advance more targeted medicine. “The information out there is staggering,” said Eric Neumann, co-founder and CTO of PanGenX. The mission is to “collect and bring in information from all different sources and help stakeholders make more informed decisions.”

PanGenX envisions what Neumann described as a knowledge base, a large named graph that different views can be applied to. “The future is about application lenses,” he said. In that scenario, semantic technologies can be applied to pull together information from different sources in the right context, depending on whether the query comes from a clinician, for instance, or an insurance company. “It’s about providing powerful and insightful lenses to look at this data or any data a customer has that you want to link to, and do the merging dynamically,” he said.

AstraZeneca principal informatics scientist Tom Plasterer, who helped co-found PanGenX, discussed how that pharma company has begun to build a Linked Data architecture. “Building linked data apps is where it comes together,” he said, and opening the door to letting scientists, clinicians and informatics teams freely interoperate. The focus clearly is data interoperability, not integration. “A data warehouse,” he said, “is just another silo.”

The effort there is about one year old and so far focused on clinical trial assets. About 50 users are actively engaged in the prototype, generating queries and exploring data. “But I will say there are an awful lot of places outside of R&D starting to look at this and say, ‘This [approach] is really cool,’” Plasterer said, from corporate IT to marketing.

Stay tuned for more on the pharma sector and the semantic web, coming soon.