The humble yeast organism, the same one that makes beer and bread, has made science headlines for the second time in a week.
Last week scientists showed yeast could help test whether drugs will work. Now bioinformaticists -- researchers who combine biology, information technology, and other disciplines -- have used the yeast genome to demonstrate a breakthrough technique for learning what genes do.
Using yeast Saccharomyces cerevisiae as a model, researchers at the Dana-Farber Cancer Institute, in Boston, Massachusetts, for the first time combined two sets of data to create a more fine-tuned map than just a genome map.
The researchers examined gene expression –- the instructions a gene sends out -- and the interactions between the proteins the genes create, of a large number of genes at once. They then created profiles for the various genes.
By creating profiles of genes according to the data sets, the researchers assign them to groups -- in this case by noting the similarities that overlap between the gene expression data and the protein data. They believe genes in the same group will have related functions.
The same technique could eventually be used to determine human gene function.
"Maybe this overlap we saw in yeast will also be observed in mammalian cells," said Mark Vidal, an assistant professor of genetics at Harvard Medical School and researcher at the Dana-Farber Cancer Institute, who led the study.
Although researchers completed most of the human genome map in February, figuring out the function of the estimated 30,000 human genes is a distant goal.
Vidal and his colleagues hope to make it a little less distant with their new technique, which is described in the November 5 online issue of Nature Genetics.
"They have taken two data sets and said that when transcription patterns correlate, then protein-protein interactions correlate too," said Richard Gibbs, director of the human genome sequencing center at Baylor College of Medicine. "Together these sets of data provide a more powerful indicator that you've mapped a biological control system."
For the past 20 years or so, researchers have examined genes one at a time to figure out what they do. But it's becoming more and more clear that this is an inefficient method.
"It's pretty obvious genes don’t act alone, they act as networks," Vidal said.
Vidal and his colleagues chose yeast to test their method because it's one of the most thoroughly studied organisms. Its entire genome of 6,116 genes has been available for several years, and scientists have tested it in a plethora of situations.
Computer informaticists have created databases full of gene expression and protein interaction information for yeast, which the researchers needed to test their method.
"These bioinformaticians are using existing data sets, which once again shows that a thoughtful analysis of the data by the computer people can be a very good thing to add to the biological work," Gibbs said.
