Scholars develop algorithm to investigate cancer genetics

Researchers from Princeton's Lewis-Sigler Institute for Integrative Genomics and Computer Science Department have developed a new tool to systematically identify the chromosomal alterations that cause cancer.

Olga Troyanskaya, assistant professor of computer science at the Lewis-Sigler Institute for Integrative Genomics, Chad Myers GS and fellow researchers created a computer algorithm that works to measure gene expression and pinpoint where chromosomal amplifications and deletions that cause cancer occur.

"Most types of cancer involve cells growing too fast according to some capacity and not dying when they are supposed to," Myers said. Tools that can identify the origins of these problems provide useful information for developing drugs for treatment.

Describing the complications of data collection, Troyanskaya said, "Microchips measure the amount of protein made for each gene, and there are approximately 50,000 genes in the entire genome." She likened the problem to trying to edit a book that was written in a different language and filled with typos.

Essentially, the algorithm organizes this data and enables researchers to assemble an accurate profile of where the amplifications and deletions of chromosomes have occurred, compiling a genomic footprint. From there, scientists try to discern a pattern from "the hundreds of genes involved in the common alterations and identify which few lead to cancer," Myers said.

Both Troyanskaya and Myers attested to the extremely complicated nature of the data and the difficulty in differentiating between which patterns actually cause cancer and which do not. While some patterns are easily recognizable, others require much closer analysis.

Troyanskaya and Myers worked in close collaboration with Lewis-Sigler fellow Maiytreya Dunham, who helped analyze the patterns of amplifications and deletions.

Thus far, the researchers have applied their technique to yeast and breast cancer cells. Troyanskaya said the particular diseases studied at Princeton are largely determined by the available data.

However, Myers said, "The algorithm is supposed to be robust enough to be applied to any data. We developed an interface so that people all over the country can download it and use it on their own information."

The wide applicability of the technique will enable people to study chromosomal changes in a multitude of cells and organisms.

"The hope is that this research will allow us to understand cancer progression, how cancer cells become malignant . . . a breakthrough with the potential to help many people," Troyanskaya said.