By Sabrina Rubakovic

The Duke Chronicle, Duke University

Researchers at Duke and other universities have discovered a way to diagnose viral infections before symptoms even appear.

And their findings could have practical applications for preventing the spread of swine flu at Duke.

The research concluded that as the body responds to infection, the expression of specific genes is slightly altered. These baseline changes in human health can then be detected, allowing the diagnosis of infections days before the onset of symptoms. This study, which focused on upper respiratory viral infections, was funded by the Predicting Health and Disease program of the Defense Advanced Research Projects Agency, a branch of the U.S. Department of Defense.

"We are using genomic technology to predict who will get sick and who won't," said Dr. Geoffrey Ginsburg, professor of medicine and pathology and director of the Center for Genomic Medicine in the Institute for Genome Sciences and Policy, who worked on the study.

The project consisted of three studies, in which researchers inoculated healthy volunteers with live Rhinoviruses, and Influenza A and Respiratory Syncytial viruses in a closed environment. Blood and other biological materials were collected from the subjects every three to four hours. The complex data obtained from these samples were heavily analyzed using genomic technology to build models that predict whether an individual will become symptomatic.

The underlying aim of the research program is military in nature - the early detection and subsequent prevention of disease in U.S. soldiers. Infections among troops can be debilitating, especially to those stationed in closed environments, such as submarines, said Lawrence Carin, professor of electrical and computer engineering at Duke, who also worked on the study. Being able to determine which soldiers will get sick days before the appearance of symptoms would allow a battalion commander to separate an infected soldier from other troops, and appropriately apportion tasks.

The project, however, has evolved to encompass the realm of medical treatment in general. The current aim is to develop a portable health assessment system able to conduct thorough medical analysis in short time spans and at low costs. The goal is to develop "pre-diagnostic tools for upper-respiratory viral infections, and refine those tests so that they can be applied to clinical medicine over the next few years," Ginsburg said. SRI International, a non-profit research organization participating in the program, has a similar vision for the distant future. SRI envisions an easy-to-use, non-invasive device that predicts the course of disease, severity of symptoms and infection level of the patient in 10 minutes at a cost of $10. The researchers believe, however, that achieving this goal will require research beyond the scope of current DARPA PHD funding, according to a statement on the SRI International Web site.

Despite its biomedical appearance, the study took an interdisciplinary approach at Duke. Carin was able to apply his expertise in statistical analysis to the project.

"The gene expressions we examined included over 20,000 genes. Our job is to find the 10 to 20 genes that are responsible for a viral infection. This is a definite statistical challenge," Carin said. "That was my aspect of the project. We also have electrical engineers that are trying to build electronic chips so that the task of pre-diagnosis could be very cost-efficiently implemented."

These chips aim to monitor specific genes which are responsible for battling viral infections. When a virus enters the body, more copies of these genes are produced, making it possible to detect the presence of an infection prior to the onset of symptoms.

Duke faculty are preparing to embark on the second phase of the research study that will focus on the validation of the findings thus far, and the building of technology to apply these findings to real-world diagnostic medicine, Carin noted.

With these goals in mind, the project has begun to focus on the early detection of the H1N1 virus at Duke. Researchers have been collecting blood samples from students to determine who will contract the virus. If it is determined that a student has this flu, blood samples will be collected from the people they recently came in contact with. This will allow the university to track and prevent proliferation of the virus.

"If you know someone will come down with H1N1, you can get them out of the classrooms and offices before they develop symptoms," Carin said. "If you can get people who are sick out of the population, you can mitigate the spread of the disease."

Researchers aim to produce a prototype able to carry out pre-symptomatic diagnosis by the end of next summer. They envision that the development of a marketable device, however, is still several years away, Ginsburg said, adding that he sees this as a very accelerated timeline, cutting the normal time span for production of a working device in half.

Click here to read the original story.