Milton Taylor (rear) and Takuma Tsukahara (in front)

Doctors have known for a long time that drug regimens and other medical treatments don't always improve their patients' fates. Even when the treatments work, some patients respond more positively than others.

Patients' differential responses to a treatment have been presumed to be the result of environmental (diet or air quality, perhaps) and genetic factors. So far, that theory is holding true.

But it is only recently that doctors have been able to begin identifying the differences that make certain people susceptible or resistant to treatment.

By identifying genes that respond to interferon -- a drug commonly used to treat hepatitis C viral infections, and certain types of cancers -- scientists from Indiana University Bloomington and three collaborating institutions have devised a novel way of predicting patient response to treatment.

The researchers used a blind, statistical approach to identify 36 genes that are not only actively expressed in the presence of interferon, but also are turned on in patients whose virus counts are dramatically diminished. The researchers describe the technique and report the results of the first test in Public Library of Science: ONE in July.

"This method gives us the opportunity of identifying genes that are important in the response to any drug," says IU Bloomington biologist Milton Taylor, who led the study. "This method is not necessarily confined to hepatitis C. In this case we were just using interferon and hepatitis C to see if the method works."

How can scientists know which genes -- out of the 50,000 that make up the human genome -- are actually involved in mitigating diseases such as the flu, or herpes or Hodgkin's Lymphoma?

It is not enough to simply look at what genes are turned on in the midst of an infection, or even to look at which genes are most active in patients who are faring well with a prescribed treatment, Taylor says. So, with IUB biologist Takuma Tsukahara, University of Haifa mathematician Leonid Brodsky, and others, Taylor decided to delineate the most important genes by combining viral counts in the blood with gene expression across time for each individual patient.

The scientists examined expression patterns of 22,000 genes in 69 patients at six different time points during treatment. Their analysis turned up 36 candidate genes that are closely associated with virus removal in patients. A quarter of these 36 have no known function. Lending credibility to their methodology, however, a sweep of the literature shows that nearly all 36 genes have been previously identified as playing a role -- known or unknown -- in the human response to interferon treatment. Using other methods, Taylor says, a researcher would have to examine perhaps 1,000 genes altered by the treatment and from these decide by other means which were most important.