May 18, 2012

Resolving The Ortholog Conjecture

Evidence that comparing genes across species is biologically useful

Researchers at the SIB Swiss Institute of Bioinformatics and the EMBL-European Bioinformatics Institute have confirmed the long-held conjecture that studying the genes we share with other animals is a viable means of extrapolating information about human biology. The study, published in the open access journal PLoS Computational Biology, shows how bioinformatics makes it possible to test the conjecture.

Scientists have long looked to model species — mice, for example — to understand human biology. This is at the root of what is called the 'ortholog conjecture': the idea that studying genes which are separated by a speciation event, but retain the same evolutionary ancestor, is useful.

In genetics, scientists must address these issues in order to pinpoint the best genes to study. For example, it may be better to compare genes in mice and humans that directly descend from a common ancestor (orthologs), than to compare copies of genes with a different function (paralogs).

Consider hemoglobins, for example, which are protein complexes used to carry oxygen across diverse animals. Normal hemoglobin in humans is made up of two subunits — alpha globin and beta globin. These proteins are related to each other by a duplication event that happened long ago in animal evolution. If one were interested in the human beta globin, would the best model be to study beta globin in mouse (this is the ortholog of human beta globin) or would it be to study alpha globin in humans (alpha and beta globins are paralogs)?

For the past 40 years, scientists have used orthologs by studying genes in model species, and this has provided invaluable insights in all areas of biology. Until now, there hasn't been enough data to use orthologs with empirical authority. However, with advances in biotechnology producing vast quantities of data, there is finally enough data to settle the debate.

Using advanced computational techniques on data derived from tens of thousands of scientific articles, the researchers analyzed 400,000 pairs of genes (orthologs and paralogs) from 13 different species. The team compared the two approaches and observed only a weak decrease in functional similarity between orthologs.

"We have the data to prove that the study of orthologs is indeed useful, but we are only at the beginning," says Prof. Marc Robinson-Rechavi of SIB and the University of Lausanne. "This is at the heart of all of comparative genomics, in which we try to extrapolate knowledge from a handful of organisms and apply it to all of life."

"We found that current experimental annotations do support the standard model," explains Christophe Dessimoz of EMBL-EBI. "Our work corroborates the assumption that studying the genes of other species — whether mice, yeast, or even bacteria — can elucidate aspects of human biology."

The same question has recently been addressed by Nehrt and colleagues, whose different conclusion sparked some debate. The new research suggests that these controversial results were due to overlooked biases in the collective knowledge of gene function. Controlling for these, the new study strongly supports the ortholog conjecture and the fact that studying species we are only distantly related to — even worms, flies, yeasts or bacteria — is relevant and useful.


On the Net: