Computer model helps predict drug adverse effects

A set of computer models holds promise to predict negative side effects in drugs, helping researchers develop safer medications and potentially save billions of dollars spent on developing drugs that fail.

The work, appearing online this week in the journal Nature, features research from the University of California, San Francisco, Institutes for BioMedical Research and SeaChange Pharmaceuticals--a UCSF spin-off company launched by two of the paper's authors. They set out to create a model to predict which drug prospects were most likely to have adverse side effects, according to a UCSF article.

First, the researchers created a database of 73 proteins associated with adverse side effects. Then they used a computer program to analyze whether a library of 656 approved drugs were likely to interact with any of them.

The researchers made more than 1,000 predictions, and found about half of them to be correct, verified through new tests in the laboratory. They also found 151 previously unknown interactions.

"The biggest surprise was just how promiscuous the drugs were, with each drug hitting more than 10 percent of the targets, and how often the side-effect targets were unrelated to the previously known targets of the drugs," said Brian Shoichet, a UCSF professor of pharmaceutical chemistry and joint adviser on the project. "That would have been hard to predict using standard scientific approaches."

A commonly cited cost for bringing a drug to market is $1.2 billion across 15 years, although that can swell to $4 billion to $12 billion per drug, depending upon how many failures are included in the estimate. Drug reactions cause about 20 percent of drugs in clinical development to fail.

Kenneth Kaitin, director of the Tufts Center for the Study of Drug Development, who was not involved in the study, told the Boston Globe that one of the big challenges for the industry is that compounds enter clinical testing without a solid idea of how the drugs work and what the side effects ultimately will be. "You're essentially entering blind," he said.

The model not only helped drug makers find potential interactions earlier in the process, which reduced costs, but it also helped in repurposing medications to treat other conditions.

Earlier this year, researchers at Stanford University also developed an algorithm that allows doctors to differentiate between adverse events related to drugs from those related to another illness.

What's more, two companies--AdverseEvents and Clarimed--have developed websites backed by algorithms that sort FDA adverse event reports, making them easier for consumers to understand.

To learn more:
- here's the UCSF article
- read the study abstract
- check out the Boston Globe piece


The Real Payback of Healthcare Analytics

With the unpredictability of healthcare today, organizations are sharpening their focus on analytics to make more data-informed decisions. Join us for a roundtable session in which thought leaders will discuss how they are maximizing their analytics investments.