July 29, 2010 | It is unlikely that too many software-as-a-service companies feature quotes from T. S. Eliot on their web site, but Parthys Reverse Informatics, based in Chennai, India, is the exception:
Where is the Life we have lost in living?
Where is the wisdom we have lost in knowledge?
Where is the knowledge we have lost in information?
Parthys Reverse Informatics is the brainchild of Parthiban Srinivasan, a computational chemist by training. After various stints in academia and industry, Parthiban defined what he saw as a problem with the current approach to data analysis in life sciences:
“We’re not working on bioinformatics, we’re creating chaos! Everybody’s worried about having too much data, but we’re delivering data from existing knowledge. So informatics goes from data to information to knowledge. We [Parthys] go from knowledge to information to data. It’s reverse informatics.”
Parthiban worked for NASA in the late 1990s, applying his computational chemistry training to analyzing molecules in outer space. He later moved to the Weizmann Institute in Israel, working with Jan Martin, a recipient of the Dirac Prize. From Israel, Parthiban joined AstraZeneca back in India, working on pharmacophore and virtual screening for tuberculosis (TB). “I became more of a bio-IT person,” he said.
In 2002, Parthiban joined Jubilant Biosys, which like other Indian bioinformatics companies was “facing the tempest of the industry to succeed,” he recalls. “The Indian companies built products, but the products in the international market were superior and available for free. Nothing was moving.” Parthiban had the opportunity to travel alongside some consultants from McKinsey. “They knew strategy, not science. I was living in a plane for a year.” Jubilant turned around, and within a few months, Parthiban was tasked with recruiting some 400 scientists. “The key to success: Instead of focusing on pure bioinformatics, include chemistry,” he said.
However, Parthiban questioned scientists sitting in front of expensive computers designing compounds. “Why should a pharma company give you a biological target and ask you to design a compound?” he asks rhetorically. “It’s such a sensitive matter. There’s no business there. We were investing in too many Silicon Graphics machines!”
Instead, he chose to focus on small molecules with big opportunities. “We gathered all related information, built thematic databases and moved on the value chain. It was a phenomenal success.” Jubilant was competing with the likes of Ingenuity, but with a greater focus on chemistry. “From the beginning, for 25-30 years, chemistry-based databases were all privately held, whereas life sciences databases are government supported. We couldn’t make money out of it. We have to make money—sustainability was the question.”
After three years, Parthiban joined another Indian informatics company, GVK Bio. He got an opportunity to make a new turn from informatics to patent analytics and he developed market intelligence reports for pharma research. An example was looking for ways to synthesize compounds without infringing intellectual property (IP). “It worked surprisingly well! What a success,” he said.
Finally, Parthiban decided to strike out on his own. After tasting success in informatics and patent analytics, he moved to social networks, specifically building networks of key opinion leaders. His team worked with an American marketing firm to build a large database of some 20,000 key leaders in breast cancer for a major pharma company. Part of that project involved disambiguation of scientist’s names. “We developed a web-spider that fetches relevant information from various sources, then did a ‘de-duplication.’”
Parthiban coined a pet phrase to describe his company’s business model: “KIDding is our business: Knowledge to Information to Data.” The goal is to move from unstructured knowledge to structured data, he says. Computational science compliments experimental and theoretical approaches to science. “What is common in all the three modes of science is that you start with a question, you collect data, find an answer. Today, it’s the reverse. You have answers everywhere. The secret of doing science is asking smart questions.”
For example, say you want to know all the compounds that interact with Jun protein kinases. “You already have the answer, it is already there,” says Parthiban. “But you are not able to access the information… If you hear a dog barking and it is raining, you might conclude that whenever a dog barks, it rains—that is a scientist’s bias. We eliminate this bias and pull out the data from different studies—the naked facts.”
Another buzz word in pharma, says Parthiban, is “interoperability” or “actionable data.” Finding data from one study that will interact with data from another is where the most value lies. “This is what we’re facilitating. Knowledge-Information-Data.” The key elements of reverse informatics, he says, are extraction, normalization, de-duplication, and integration.
The name of the company relates to India’s strict regulations over company names at the time of registration, which either had to include the name of the owner or describe what the company was about. The complete name of the company Parthiban was obliged to register is “Parthys Reverse Informatics Analytics Solutions Private Limited.” (Parthiban thinks he might be a contender for the Guinness Book of World Records.) Eventually, the Indian government relaxed its rules, but Parthiban was already promoting the name.
Networking at conferences, including the Bio-IT World Expo, Parthiban started to win contracts from some small companies. He weathered the economic crisis in 2009 and has built the company up to 100 staff.
Parthys focuses on informatics, social networking, and IP. For example, patent search services are now done by Parthys. In another project with Collaborative Drug Discovery, led by Barry Bunin, Parthys built a thematic database capturing biological assay details, activity data and chemical information with a focus on disease. Part of that work involves 25 chemists “sitting in India looking at life sciences and chemistry information from patents across the globe in languages we don’t read.”
Another new project is construction of a next-generation pharmacodynamics database, which will be Parthys’ first product. “A leading PK/PD software company is inquiring about this. Fingers crossed”
Parthiban also uses the term “in litero” (in line with in vivo) to describe his belief in the value of literature searching. He recounts a story of a big pharma who filed a patent application on a new technology with the U.S. Patent and Trademark Office. “They said, ‘Bad news, for innovation, there is already a patent. The good news: you’re the owner.’ They didn’t know they had the patent! It’s not just what you’re competitors are doing. Tracking back is the biggest problem. You don’t know what your company has already patented.” Big pharmas don’t know how many compounds they’ve patented, and nor do the scientists. “I would build a beautiful database,” for one of them, he says.
This article also appeared in the July-August 2010 issue of Bio-IT World Magazine. Subscriptions are free for qualifying individuals. Apply today.