Biologist crunches 'big data' to personalize cancer treatment
April 10, 2013 | by Darrin Joy
Everything from the human genome project to social media is producing “big data.” And to help medicine become more personalized, some adventurous biomedical researchers are tapping into these vast amounts of information to pinpoint the key factors affecting health.
Andrei Rodin, Ph.D., associate professor in the Department of Diabetes and Metabolic Diseases Research at City of Hope, is one of those researchers.
Recently appointed the Dr. Susumu Ohno Chair in Theoretical Biology, a position held previously by his late father, Rodin is using advanced methods derived from computer science to gather all possible data surrounding a patient and isolate the bits that could one day guide prevention or treatment.
We all leave a trail of data in our daily lives — about where we live, the food we eat, even the time we spend in the gym (if any). For patients battling cancer, diabetes and other life-threatening illnesses, this mass of data also includes biomedical information — not just about mutations in their genes or their medical history but how many hours they spend in freeway traffic each day. And much, much more.
Researchers like Rodin aim to figure out which pieces of data are important to health or which have a connection to how well a treatment might work for an individual patient.
“So we look at the genome, metabolome, epigenome and other variables about people at different levels of biological abstraction … up to and including, I don’t know, annual peanut butter consumption, and based on those variables and their interactions we can try to predict a single variable — perhaps a treatment outcome, or response to a particular drug, or whether or not a person will get a certain disease in the first place,” Rodin says.
Still in its early stages, Rodin’s work could result in high-powered computational methods that would be much more specific than current traditional statistical techniques at enabling clinicians to tailor treatment to a specific patient.
“Right now when we talk about [personalized therapy], it’s being done, but it is not exactly easy,” says Rodin. “It’s not yet mainstream.” Current methods just aren’t selective, and scalable, enough to do the job well, he explains.
Rodin proposes building real-time biological network models, and looking at these networks to zero in on the factors that are important to how a patient’s disease might progress or respond to treatment.
The work could one day empower physicians to scan a patient’s “personal big data” relatively quickly and use the information to guide treatment decisions. It also could help clinicians understand a person’s risk of developing a given disease, so they could work to lower that risk.
You may also be interested in