Genetics by the numbers
New statistical technique helps target relevant gene variations in mental illness
When it comes to our DNA, the numbers are staggering. Humans have 6 billion DNA molecules organized into pairs, which is collectively known as the human genome. A fraction of this DNA makes up the 20,000 or so genes that code for proteins, while the rest is called “non-coding” DNA. The task facing geneticists is to find which genes and other pieces of DNA may be relevant to illness.
A new technique developed at CAMH, published in the journal PLOS One, may help. Dr. Jo Knight, Senior Scientist and an expert in statistical genetics, answers a few questions about this approach, which she developed with her student Sarah Gagliano and collaborators from the United Kingdom.
How do scientists find genes related to mental illness?
In the earlier days of genetics, people typically searched for changes or mutations to genes they thought were relevant to an illness. Any such changes to a gene’s code could affect normal biological processes. In the most straightforward case, one genetic mutation could cause a specific illness.
But for complex conditions like mental illness, we now know the challenge in linking genes to illness is much greater than we thought. Hundreds of genes and importantly, other parts of the genome may be involved in such conditions.
We’ve learned much more about our DNA. Most of it isn’t organized into genes, and doesn’t code for any proteins. Yet this non-coding DNA is important. Some of it regulates or controls genes, so it still influences whether or not proteins are produced. Therefore, much of the genome, not just the genes, is now characterized as “functional elements.”
We need to consider the role of all these functional elements of DNA, when we’re trying to understand the genetic basis for complex illnesses.
How do we approach this challenge?
One type of study that can provide detailed information on our DNA is called a genome-wide association study (GWAS). Rather than focusing on a particular gene of interest, researchers scan the entire genome to investigate DNA variations. The goal is to find differences in DNA between people with a psychiatric illness, and those without the illness. If a particular variation shows up more frequently in people who have the illness compared to those who don’t, we think it could be associated with the illness.
The problem is that sometimes there are hundreds of thousands of these “signals,” and we need to do more work to find out which ones are truly linked to the illness. This is where our technique comes into play.
What does your technique do?
Usually when people analyze results from a GWAS, they look at the variations between the “cases” (DNA of individuals with illness) and the “controls” (DNA of individuals without illness). But if there are thousands of signals that appear to be linked to an illness, some of these signals may actually show up by chance. How do you decide which ones are really relevant and worth further study?
There is a growing amount of information available on functional elements of DNA, those large stretches that don’t code for genes. We thought this information could help.
Our technique takes data about the functional role of DNA, and combines it with the case-control evidence from a GWAS. The goal is to help scientists identify which variations should be prioritized for further study after they’ve scanned the whole human genome.
Where do you get information on functional elements of DNA?
Much of the functional information has been made available for all researchers to use. One major source is from the ENCODE project (Encyclopedia of DNA Elements) based at the University of California at Santa Cruz, which is funded by the U.S. National Human Genome Research Institute. Another is the Road Map Epigenomics Project of the U.S. National Institutes of Health. Within CAMH, we also access functional DNA evidence from samples from our colleague, Senior Scientist Dr. Art Petronis.
By using multiple lines of evidence, we hope to identify those variations that are truly causing the illness rather than those that are just associated by chance.
What’s happening next?
Sarah Gagliano, my undergraduate student, was the person who undertook much of the work for this project. She was recently awarded the Peterborough K.M. Hunter Graduate Studentship to refine the method further.