The Huntsman Cancer Institute has applied for a patent to help researchers more accurately identify what role genes play in the development of cancer.
Contrary to existing practice, researchers would be able to take into account how genes interact, instead of simply studying individual genes.
Dr. Andrei Yakovlev, director of the biostatistics program at the institute, believes the normal process is too simplistic, ignoring the complexity of biological processes involved in cancer formation.
"The issue is how variation in the expression of groups of genes impacts their function in the body," he said.
"What researchers usually do is look at each gene individually. Take Gene A and compare how it is expressed in a cancer tumor and in normal tissue. If you see a huge difference, you include it as a candidate gene for further exploration. The distinctive feature of our approach is we look for subsets of genes from the very beginning, taking into account complex interactions between genes.
"It's a multidimensional search. We came up with a method that allows us to find such subsets, which then can be targeted as candidate genes for different interventions to gain different understanding. It has many applications, not just cancer."
HCI's biostatistics team approach lets researchers focus on small subsets of genes that are most likely to malfunction in cancer. Then they concentrate on how to analyze the data. The patent covers the method of analysis.
Researchers can monitor thousands of genes at the same time, Yakovlev said. They can obtain the samples from different organs or from normal tissue, and "we are interested in which genes are involved in some pathological or developmental process that would lead to cancer or another disease."
A problem, said Yakovlev, is that "the data is quite noisy. That's why it needs statistical approaches to identify sets of genes expressed differently."
So far, their methodology has passed every test they've thrown at it.
He describes it as a "tool for finding genes that change their expression during tumor genesis or progression. It can be used for diagnostics, and we hope also to link this to data on the efficacy of cancer treatments so we can use it for prognosis as well. This is a long journey."
Another HCI patent will let researchers determine the size of the sought-after subset of "differentially expressed" genes.
Besides Yakovlev, others involved in research leading to the patent are Aniko Szabo, Kenneth Boucher and Alexander Tsodikov, all of the Division of Biostatistics at Huntsman Cancer Institute; and David Jones and Dr. William Caroll of the University of Utah Department of Oncological Services.
E-mail: lois@desnews.com