05 Feb 2004

Two papers in tomorrow's Science describe genetic maps that may prove useful for studying the genetic basis for human disease. A team of Canadian scientists led by Charles Boone at the University of Toronto report that they have constructed a genetic interaction map of yeast, while researchers in Germany led by Norbert Hubner at the Max-Delbruck-Center for Molecular Medicine, Berlin, report a single nucleotide polymorphism (SNP) map of the rat genome.

Genomes have a certain amount of redundancy built in. A mutation in a single gene may often exhibit no phenotype by itself, but in combination with a similarly innocuous second mutation, it may prove lethal. In an effort to understand the genetic basis for such redundancy, first authors Amy Tong, Guillaume Lasage, and 50 other colleagues in Canada, constructed the yeast map by making a synthetic genetic array. These arrays are essentially crosses between one batch of single mutations and another, to give an array of double mutants. In this case, the authors crossed a collection of yeast with single mutations (missing or partially functional alleles) in one of 132 genes, with another collection comprising over 4,500 strains, each completely devoid of a single gene. Though cells in each collection were completely viable despite their mutations, the array revealed combinations of mutations that are deleterious or even lethal to the offspring. Whenever this happened, Tong and colleagues scored the mutant pair as an interaction.

In total, the technique revealed about 4,000 interactions. Assuming genes not included in this map behave in a similar fashion, the authors calculate that there are about 100,000 interactions in the complete yeast genetic interaction network, with each gene having an average of 34 interactions.

So what does this mean? For one, these maps have the potential to reveal functionally related genes, because the authors found that genes with similar functional attributes, also called gene ontologies (such as cell wall maintenance, DNA synthesis, transcription, etc.), are more likely to interact. Birds of a feather, in other words…. Similar relationships have been found in protein-protein contact maps of roundworms (see ARF related news story) and fruit flies (see ARF related news story). For another, the maps may also be used to predict function of uncharacterized genes. As an example, the authors use the yeast gene csm3, which has an interaction pattern similar to DNA replication checkpoint genes, suggesting that csm3 may also be involved in the regulation of replication.

The map contains genes that have few interactions and genes with many. The latter "hub" genes may be more important for fitness, the authors suggest, and could be good targets for anti-cancer drugs because cancerous cells often have accumulated many mutations and might, therefore, lack genetic redundancy. Yeast maps may also serve as a starting point for building maps of human modifying mutations, like those that enhance the severity of diseases such as cystic fibrosis, suggest the authors. In Alzheimer’s, even afflicted members of families with an APP or presenilin mutation develop AD in different decades of their lives. This large difference in disease onset is chalked up to the action of modifying genes, but identifying these genes has been an elusive goal to date. Finally, the map may help reveal the etiology of diseases that are thought to result from mutations in multiple genes, such as late-onset Alzheimer's or schizophrenia.

An accompanying essay by Lee Hartwell, University of Washington, Seattle, puts the possibility of developing human genetic interaction maps in perspective. That Tong and colleagues found, on average, 34 interactions per mutant gene is a “daunting” result for those who aim to uncover the genetic basis of disease susceptibility in humans, Hartwell writes. This result predicts that current genome-wide screens will be largely unsuccessful, as indeed they have been in identifying genes for complex diseases. Instead, Hartwell proposes, one would need to focus on the right set of candidate genes. If the yeast genetic interaction map is completed, then it may well guide the selection of candidate genes, pathways, or “ontologies” on which to focus.

Meanwhile, Heike Zimdahl and colleagues in Berlin provide a separate tool with their map of over 12,000 SNPs in the rat genome. Zimdahl made the map by comparing cDNAs from four different rat strains. About 10 percent of the SNPs are in coding regions of genes, almost one-third are in the untranslated regions, and the remaining majority were found in the regions between genes of the genome. The map, which the authors hope will be a useful tool for comparing mammalian genes, is available at the ENSEMBL database.

The yeast interaction map is also available online at several sites, one being the biomolecular interaction network (BIND) database.—Tom Fagan

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

News Citations

1. Mapping Protein Networks in the Roundworm 3 Jan 2004
2. Mapping Interactions in the Drosophila Proteome 6 Nov 2003

External Citations

1. ENSEMBL
2. biomolecular interaction network

Further Reading

No Available Further Reading