SALT LAKE CITY — At first, it looks like a mass of colored dots spiraling out from the center like a galaxy.
But then, Lewis Frey tells the computer to rotate it and what emerges is a 3-D model of what looks more like a swirling tornado. This is one of the first graphic representations of the complex interrelations of cancer mutations.
"It's like Facebook for cancers," Frey said.
From breast and colorectal cancers, to brain cancer and other more rare types, each type of cancer is interrelated with other kinds based on shared genetic mutations. If one follows the map of shared mutations, they can be traced backward to core mutations often inherited by a person from their parents.
As genetic technology has advanced, the amount of genetic information to sort through has began to flood researchers.
The human genome contains several terabytes of genetic information, and cancer is caused by more than one mutation. The average type of cancer can have between 65 and 75 separate mutations and can reach up to more than 100. This makes identifying cancer by its genetic properties immensely complex.
Frey, an assistant professor of bioinformatics at the University of Utah and a researcher at the Huntsman Cancer Institute, has created a software than can quickly sift through genetic data and map out the relations between certain types of cancer based on mutation. The new technology could revolutionize the way cancers are identified and treated.
"This looks across multiple cancer types and sees commonality between the genes that are mutated," Frey said. "So the mutations that occur in colorectal cancer can also occur in breast cancer, or in brain cancer."
Armed with this information, doctors will be able to apply drugs and treatments used in one cancer, to a related cancer. This can also be used to identify what treatments would be good to treat rare, or "orphaned" cancers in which no treatment has been developed.
Frey doesn't call his creation "Facebook for cancers" by whim. He actually used computer algorithms in his software that are commonly found on social networking sites.
"I think it's very interesting that he would want to apply algorithms from social networking to genes. They do relate in a social way," said Dr. Mary Edgerton, a surgical pathologist at the MD Anderson Center out of the University of Texas.
This month, Frey, Edgerton and Stephen Piccolo of the University of Utah co-authored a paper on this new system, which was published in the journal BMC Medical Genomics.
"I see this as something that will really contribute to cancer patients across the field," Edgerton said.
The cost of genetic sequencing has gone down dramatically. What used to cost millions of dollars to do can now be done for around $5,000 to $3,000. Frey said he foresees that in about five to 10 years people will go in to have a routine genetic analysis done by their doctor to see what types of cancer they are prone to.
"You could get a mutation profile of a particular patient's tumor and then compare it to the nearest neighbor on the profile," he said. Eventually he hopes his program will be used as a 3-D encyclopedia to identify cancers by their genetic code.
Frey said he used data from the Utah Population Database to start his computer model. The database is the largest and most complete population database in the U.S. used for biomedical research. It is co-managed by the Huntsman Institute and University of Utah and includes information on 6.4 million people, linked through 9 million records.
It will be a few more years before Frey can build up the cancer database to where it can be used on a clinical level.
Email: gfattah@desnews.com