Using yeast to hunt down bad mutations
Yeast can be used to identify which genetic mutations cause disease, a new study has found. The work will help researchers develop a reference for clinicians to interpret their patients’ genomes and predict their susceptibility to certain diseases.
The study, led by CIFAR Senior Fellow Frederick Roth (University of Toronto and Mt Sinai Hospital), used yeast to test mutations, or ‘genetic variants,’ in 22 human genes that are known to cause disease. The researchers identified which variants in these genes were harmful, and which weren’t, with more precision than current computational methods.
The researchers examined a subset of human disease genes that have yeast ‘cousin’ genes. In many cases, the human gene is still able to substitute for its cousin yeast gene, despite the one billion years since the common ancestor of yeast and humans. To test whether the variants of the human gene affect function, they used strains of yeast from a collection built by CIFAR fellows Charlie Boone and Brenda Andrews (both at the University of Toronto). These strains carry specific mutations that allow the yeast cousin gene to function at low temperature, but cause it to fail at high temperature. For each yeast cousin gene, they introduced the human cousin to the yeast. If the normal version of the human gene saved the mutated yeast cell from dying at high temperature, but a variant of the human gene did not, then the researchers knew the variant was affecting the human gene’s function.
Pinpointing variants that cause disease is a huge challenge for genetics. Most research has studied common variants in the human genome, but the rarest variants are often the most damaging. Each person has between 100 and 400 rare variants that change how genes produce proteins. But even if such a rare variant is found in a patient, it’s difficult to know what that means.
“Very often, you are seeing a patient in the clinic with many variants that have never been seen before by anyone in the world,” Roth says. “How do you decide which of this person’s variants are damaging?”
There are computational methods to estimate how damaging a given variant might be. These work to accurately detect up to a fifth of disease-causing mutations, but the study shows that experiments using a model organism such as yeast can work just as accurately for up to as many as 80 per cent of disease variants.
“If you can manage to do an experiment it’s going to be better than using the computer programs that are out there, by a lot,” Roth says.
For researchers who are familiar with the mighty power of yeast as a genetic model, the findings are not surprising. But Roth says it might be for others. “You’ve got a single-celled organism that diverged a billion years ago from humans. Am I really going to believe that the way this variant is working in yeast is telling me anything about what it’s going to do in humans? The doubters may now believe that the yeast that brought us wine, beer and bread can now help us learn about our own personal genomes.”
The researchers’ goal is to create a look-up table of problematic mutations for clinicians to use when they are interpreting patients’ genomes. “For human genes with yeast counterparts, you could actually test all the possible mutations before you ever have seen them in a human,” Roth says. A look up table could help patients know if they are susceptible to certain diseases, so they can seek therapy or take preventative measures.
Only about five per cent of human disease genes can be tested this way with yeast. But Roth says other research has shown that yeast can provide other proxies for variants. For example, yeast can provide a test of whether the protein product of the breast-cancer-associated gene BRCA1 can bind tightly to another specific protein. If a variant version does not bind, this can accurately predict that it is a a damaging variant. There are many other model organisms, including flies, worms, and zebrafish that can each be useful for testing specific human disease genes.
The study was published in Genome Research.
On December 4, 2016, Fellows from CIFAR’s program in Genetic Networks held an introductory workshop with other clinical, academic and...