An interview with genetic epidemiologist Dr. Ashley-Koch about which genes are correlated with suicidal behaviours and why learning about them is crucial for early intervention.
Allison Elizabeth Ashley-Koch is a genetic epidemiologist at Duke University School of Medicine and she’s dedicated her life’s work to figuring out the genetic basis of several diseases, especially those pertaining to mental health.
She has recently published a paper pinpointing four specific genes which seem to be correlated with suicidal thoughts or actions, by poring over medical data from over 600,000 veterans. Her team’s research, which was published in the peer-reviewed scientific journal JAMA Networks, can help learn more about the biology of suicidal behaviors and help target early intervention and therapeutics for those affected. Suicide is the cause of over 700,000 deaths annually.
I’m a science journalist, and my work — which has appeared in the New York Times, Guardian, BBC, and National Geographic — is all about telling stories of how the mind works. I had the pleasure of chatting with Ashley-Koch, where I unpack the behind-the-scenes of science that’s changing the way we think about the world.
A collage of images made by Sofia Quaglia
Why is it that you’ve decided to look into the mental health of veterans, specifically?
I've probably been working with veteran mental health for about 10 years or so. I'm particularly interested in that for several reasons: I have family members, and close friends who have been in the military, and I think military veterans do an enormous service to our country, and globally to other countries as well. This research is an opportunity to help, in some small way, and understand some of the conditions that plagued them. Hopefully, this will enable them to get better treatments and be able to identify people at risk a bit faster.
Where does one start to look for “red flags” to identify people at risk of mental health issues?
We did our analysis on the Million Veteran Program and as the name implies, the goal is to collect health information, medical records, as well as blood samples on ultimately a million veterans. We had the opportunity to work with this large biorepository. Suicide is a little less frequent than that. So, you can imagine, statistically, the larger your data set is, the more statistical power you will have to detect associations.
And with this data, you started getting a little bit closer to finding early identifiers for people at risk of mental health issues, right?
Our study focused on over 600,000 veterans who have given a DNA sample and have agreed to make their medical data accessible. From that, we were able to identify a little over 120,000 of them that had some sort of “suicidal behavior”, such as suicidal ideation or a suicide attempt. This was according to both their self-reports of how they’re feeling and the reports filed by their doctors and therapists over the years — so it was probably the broadest definition of a “suicide phenotype.” Once those patients were identified, we conducted what's called a ‘genome-wide association study’, which is where you test a lot of genetic markers, in this case, snips, or “single nucleotide polymorphisms” all across the person’s genome, and you look to see: are those specific snips associated with risk for the suicide behaviors in any way? And this became one of the largest and certainly the most diverse genome-wide association studies for suicidal thoughts and behaviors.
What correlations between genes and suicidal behaviors did you find?
We were able to identify four genes that had some evidence for being associated with risk for suicidal behaviors — and then we went ahead and replicated the findings in an independent study too.
A few things are particularly important to note here before we go ahead though.
First of all, these four genes were across all the ancestral groups and ethnicities, so they weren't limited to a particular ancestral group. Therefore, they are important for every one of all ancestries. I think is particularly important because historically a lot of the work has focused almost exclusively on European ancestry groups.
Secondly, the data used to replicate the experiment, and see whether certain genes made a difference in suicidal behaviors in other people external to the experiment, was almost exclusively civilian population. So, we identified these risk factors in a veteran population, but they've replicated in a primarily civilian population. This suggests that these genes are not only imported across ancestries but also across civilians and veterans.
So what genes are we talking about specifically then, and why are they so important?
These are the four genes with the strongest links to suicidal behavior, according to our data analysis. ESR1, which is an estrogen receptor and DRD2, which is a dopamine receptor. Then, DCC gives the body the instructions for making a protein called the netrin-1 receptor, and the gene TRAF3, which has been associated with antisocial behavior, substance use, and ADHD.
What do we do with this new information now, and why do these genes even matter?
We need to start trying to figure out the mechanism. Our work, essentially, shone a light on these particular genes. But the next step is to answer the question of how do these genes increase risk?
That said, those four genes were really important because they have been implicated in other mental health disorders previously. For example, the estrogen receptor has been previously associated with post-traumatic stress disorder, depression, and anxiety and is often also associated with an increased risk for suicidal behaviors. Therefore, there's a lot of additional statistical support in other mental health disorders suggesting that this gene is important.
But it’s also crucial to remember that these genes don’t act in a vacuum, correct?
Correct! That’s why we’ve started a new study always using the One Million Veterans data and looking at a lot of the environmental factors in these behaviors too. We want to better understand what are the additional socio-demographic factors that may be interacting with genes to lead to suicidal behaviors.
Therefore, having the genetic markers for suicidal behaviors doesn’t guarantee that you have suicidal behaviors?
Even if we were, for example, to screen people to see whether they have that dopamine receptor gene, just because they have that genetic risk factor, does not guarantee that they're going to develop suicidal behaviors. Sure, it may put them at greater risk, but it's not 100% guaranteed that they would develop it. And that's why we are working on these other environmental risk factors too. Ultimately, we would like to identify people before these behaviors develop, based on a series of risk factors. Then we can perhaps target special interventions, to help those individuals, and possibly even stop the behaviors from developing in the first place. That would be our ultimate goal.
What would early intervention for something like this even look like though?
In the immediate future, the more modifiable factors are going to be some of the environmental factors, right? If you know someone is particularly at risk, genetically, to suicidal behaviors perhaps you want to also identify what the environmental risk factors are, as well as reflect on how we can potentially modify or figure out how to intervene.
From the genetics perspective, we hope that understanding the underlying biology and genetics may lead us to better therapeutics. With a more complete genetic profile, you can have more information to work with, and use better and more targeted pharmacological treatment or holistic therapies.
You need to really act on this in multiple directions, so I think you have both pieces developing in parallel.
This interview, like all interviews of Behind the Science, was edited for clarity.