Computer predicts anti-cancer molecules

ATLANTA, June 17 (UPI) -- U.S. scientists have created a computerized method of analyzing cellular activity that correctly predicts the anti-tumor activity of several molecules.

Researchers Jeffrey Skolnick and John McDonald led a Georgia Institute of Technology team in developing the tool, called CoMet, that studies the integrated machinery of the cell, predicting which components can have an effect on cancer.

"This opens up the possibility of novel therapeutics for cancer and develops our understanding of why such metabolites work," said Skolnick.

He said metabolites are small molecules that are naturally produced in cells. Enzymes, the biological catalysts that produce and consume the metabolites, are created according to a cell's genetic blueprints. Importantly, however, he said metabolites can also affect the expression of genes.

"By comparing the gene expression levels of cancer cells relative to normal cells and converting that information into the enzymes that produce metabolites, CoMet predicts metabolites that have lower concentrations in cancer relative to normal cells," the scientists said, adding their findings prove that by adding such putatively depleted metabolites to cancer cells, they exhibit anti-cancer properties.

The research that included Adrian Arakaki, Roman Mezencev, Nathan Bowen and Ying Huang appears in the open access journal Molecular Cancer.