The software company has built a Personalized Medicine working group to tackle workflow bottlenecks.
By Allison Proffitt
September 25, 2009 | ‘My view of the world is that if you have a tough question to answer, it takes a village of diverse expertise to really answer it,” says Bruce McManus, head of the Prevention of Organ Failure Centre of Excellence (PROOF Centre) in Vancouver, Canada.
In this case, the village is the Informatics for Personalized Medicine working group, organized by IO Informatics.
“We’re interested in one focus, which is how do informatics, analytics, and technologies like ours contribute to personalized medicine,” says Robert Stanley, chair of the working group and IO Informatics’ president and CEO (see, “Building a Google for Bioinformatics,” Bio•IT World Jul 2007). The mandate of the working group is “developing practical applications of informatics technologies for personalized medicine.” Stanley concedes that there is certainly a strong bias, “but we think it’s a bias of general value.”
IO Informatics started the working group in January, their second such group, as a combination of customers, prospects, and other colleagues. The 11-member group comprises “individuals representing the spectrum of translational medicine,” said Stanley. There are representatives from early drug discovery, research hospitals, technology experts, systems biology experts, and three representatives from IO Informatics.
“To me the way that the group is set up is beneficial in that, as opposed to other working groups I’ve been a part of in the past, IO Informatics really listens to real world feedback. They’re a small company and I don’t see as many big egos in place.” says Kathy Gibson, who after 13 years at Pfizer started her own consulting firm, Helio Consulting, and earlier this year joined the IO Informatics’ advisory board. “The biggest part is not as much how they’re organized as their operating philosophy,”
It’s the philosophy that’s at the core of the working group, says Bill Hayden, director, international business development at IO Informatics. “Knowledge integration is fine within a group, it’s fine within a company, but if you really believe in that—which we do, obviously, and all the members do—integration of data throughout the whole spectrum should be better, should be the best, where you’re pulling from bench to bedside and back again. When you integrate all that data, you really do get the full picture.”
The topics up for discussion are brought by the working group members themselves. “Basically the working group offers an interface with clinician scientists who are creating large, complex, uneven datasets that need to be analyzed in a very rapid fire way, in a continuously refined way,” explains McManus. “IO Informatics benefits from this partnership—an intellectual partnership—by seeing the various settings and all of the challenges we’re faced with, and trying to extract information from the data.”
The working group conversations have resulted in talks given at scientific conferences and could, Stanley hints, lead to publications in the future.
An area focus for the current group is bringing biomarkers to point of care and early toxicity detection. “If you’re looking at different compounds, we can help the customer create profiles for different types of toxicity, and do that same kind of screening,” explains Stanley. “Go out and get different data sources and pull data though and get automated alerts—you know, this experimental compound is starting to show some toxicity in those assays—and do that much earlier and cast a broader net for toxicity detection.”
In the working group, McManus finds the discussions helpful. “IO Informatics comes at this in a very scholarly way, and they have some terrific people within that organization that are sort of intellectual relatives of our computational scientists within our own teams,” he says. “Having access to the network of people who are either on the working group or who are connected to the working group and having the free flow of information from a variety of media… has been very beneficial to our team here.” McManus has found the conversations to be a sort of “arms length validation of some of our own computational conclusions.”
Gibson agrees. “Most of us are coming at this having been in health care or biomedical research for a number of years, and know very well the opportunities and frustrations of having tons of data out there… and not being able to make effective use of it.” She says the problem is one she saw firsthand at Pfizer. “Lots of people have talked about this over this over the years, and to be frank not much has happened to improve mining large data stores and having it be intuitive to people. I think this working group is the best opportunity to whittle away at the problem.”
ASKed and Answered
From IO Informatics’ perspective, the working group has provided access to user questions and data. “It’s a two way street,” says Stanley. “Our products should be talking about their problems… [group members] talk methods, goals, and problems. We talk about solutions.”
Those relationships were instrumental in shaping IO Informatics’ new product, the Applied Semantic Knowledgebase, or ASK. “It’s a different type of knowledgebase,” says Stanley. “It’s not the kind of knowledgebase that you normally think of that has all of the information in a therapeutic area, all the proteins, genes, and pathways. It’s a knowledgebase containing the patterns that ‘make a difference’, the clusters of biomarkers that can be used for stratifying drugs according to their activity, or compounds according to their activity… It’s a practical hypothesis base.”
ASK is an enterprise product that works with other IO Informatics products to streamline toxicity profiling, target validation, patient stratification, and other tasks relying on a semantic database and arrays of SPARQL queries. It can automate screening and queries. Stanley said that the company had some ideas for a product like ASK, but time spent with the working group learning the users’ challenges and problems refined the concept. “It’s a learning relationship,” he says, “we share a high level vision for predictive biology and systems biology to change the face of drug discovery and health care.”
McManus echoes Stanley’s analogy. “[Our] only goal eventually is to find out ways to segment patients to care for them better... IO Informatics and their interest in trying to solve these complex problems fits perfectly with the [needs of] the independent working group members.” Besides, he continues, “It’s a lot of fun. Very few people get a chance to think about the world as a system and the biology of health as a system and I’m getting the privilege of doing that, and this working group is one of the vehicles for that privilege. In the end knowing that it’s very likely we’re going to be able to help.”
This article also appeared in the September-October 2009 issue of Bio-IT World Magazine.
Subscriptions are free for qualifying individuals. Apply today.