What are at-home genetic testing companies doing with your DNA information? How does law enforcement use genetic databases to solve crimes? Can your genetic information be used against you? Bioethicist Dr. Amy McGuire explains.
Amy McGuire, J.D., Ph.D., is a professor of biomedical ethics and director of the Center for Medical Ethics and Health Policy at Baylor College of Medicine. Her professional interests include ethical and policy issues raised by emerging technologies, research ethics, informed consent and confidentiality.
See more information on Baylor College of Medicine’s clinical genetic testing program.
See below for some of the research studies mentioned in the episode:
Spit it Out (Part 2): Privacy and Prejudice
Erin: Welcome to Body of Work, an exploration of health topics in the news and important issues facing science with experts from Baylor College of Medicine. I’m Erin Blair, and this is part two of our interview with bioethicist Dr. Amy McGuire. If you haven’t listened to part one, we suggest you go back and start there.
It seems the popularity of at home genetic testing has skyrocketed in the last few years. Is that because the cost has come down? Will there be some unforeseen consequences of this trend?
Dr. McGuire: So the cost has certainly come down tremendously, and it's continuing to come down. So the cost now of some of the direct-to-consumer genetic tests are like $99, under $100. But it's important for customers to know that many of these companies - and I'll talk mostly about 23andme because I think they're the ones that have been really transparent about this - they're not making their money off of selling $99 tests. That's not how they're making money as a company. And 23andme recently did a huge three hundred-million-dollar deal with GlaxoSmithKline because they now have from their customer base one of the largest databases of DNA data. And when you sign up for their services and use their tests you are agreeing to allow your data to be used for research purposes. And it's not tied to your name, it's not tied to your account, but it's being aggregated and used for research purposes. So they're able to sell access to their database, their very large database, to companies like GlaxoSmithKline in order to do research using their customer’s data, and that's their business model really.
Erin: And what can GlaxoSmithKline discover? What benefit do they have from having this big population of data?
Dr. McGuire: Lots of things. I mean, what has been a challenge for us in this whole era of genomic medicine, is that the one of the biggest bottlenecks is not having enough access to patient information and patient data. We're in the world of big data now where everybody's trying to amass large amounts of data on individuals so we can better understand particular genetic variants and their association with disease. We can develop drugs that are more targeted to individuals based on their genetic makeup. So this is big business, and everybody's trying to get in on this game. And there's been some countries internationally that have tried to create population databases with their whole population in it, like Iceland and the UK bio bank. The United States is starting to try to do that through our “All of Us” research project, which is funded through the National Institutes of Health, and the goal of that is to get medical genomic and other types of wearable information on a million people. We also have an effort through the Veterans Association to collect information on a million veterans, and so you see these large-scale efforts to collect data for research purposes. 23andme still has one of the largest databases, just because they have such a such a huge number of customers.
Erin: There's also been a lot of talk in the news about using DNA from these online banks to solve some cold-case files. How much should a person consider that aspect in deciding whether or not to do an at-home genetic test?
Dr. McGuire: So the first big case that was solved using this technique was the Golden State Killer case. So the Golden State Killer was somebody who ravaged the coast of California for two decades committing over 50 rapes and over a dozen murders. And the police could never identify who the suspect was but knew it was the same person committing all of these crimes. That was in the 1970s and 1980s, and it's been a cold case ever since. They had DNA from the crime scenes, and the way that law enforcement typically tries to identify suspects based on DNA is that they use their national forensic database, which is the CODIS database. So this is a database that typically includes DNA from anybody in any state who has been convicted of a crime, and in some states it's also people who have been accused of crimes. But it certainly doesn't include everybody, and the technology that they use for the DNA analysis is fairly limited so that you're really looking for a direct match for the most part in that database.
Now what's happened over the last several years is that a tremendous number of people have gotten very interested in genetic genealogy. You submit a sample like a cheek swab, the company analyzes it, and it tells you what part of the world your ancestors might have come from. And then they want to make your data available to all of their other customers for the purpose of familial matching. So you'll go into your profile and it'll say here you are and you have in our database three second cousins, one first cousin, and an aunt, and it'll tell you who you're related to, many of them you may not know. And so some of these databases are larger than others and some companies have said, well regardless of where you do your genetic genealogy, you can upload all your information to our database and we'll have sort of a master database of all these companies. And then you can connect to all these other people who - I used ancestry you use 23andme, but we are related and I might want to connect with you, so I can use one of these other sort of super databases. So one of those companies is called GEDmatch. Those data are necessarily public, and they're connected to you because the whole purpose of it is to connect you to people.
So law enforcement picked up on this and said, hey wait a minute, maybe our suspect in some of these crimes are distantly related to people in these databases. So what they did in the Golden State Killer case, and they've done in subsequent cases since then, is they took the crime scene DNA, they sent it to a lab for processing, and they got all the DNA data from the crime scene sample, and they uploaded it to GEDmatch and said, “Here I am, John Doe, connect me to anybody I might be related to.” And they got connected to somebody who was a partial match, I think it was a second or third degree-cousin was what the match was. And then they went and they did an extensive family tree and a lot of investigation, and they said what's the family tree of this particular person who we matched to and is there anyone in that family tree who could have potentially been a suspect in these crimes? They identified Joseph D'Angelo as one of the second or third degree-cousins of the person who was in the database, and they followed him around. He seemed like he could have been a suspect, he was in the right place at the right time when those crimes were committed. They followed him around, he discarded a tissue one day into the garbage, they picked it up, they analyzed it for his DNA, and they matched it to the crime scene DNA and it was a perfect hit. Remarkable, right? So they were able to identify somebody who's committed a crime more than 20 years ago, has never been a suspect, they never would have thought of him as a suspect, and they've used this technique.
So this has sort of created a whirlwind within the direct-to-consumer genetic testing company space because the companies have all responded very differently about this. So I think there's a lot of different approaches to this now, and everybody's trying to figure out what the right thing to do is, and we're still in a space where we don't have consensus.
I'll just say one more thing that I think is interesting. So shortly after the Golden State Killer case hit the media, we did a survey of about 1,500 people in the United States and we said “What do you think of this? Do you think this is a good idea?” Now it might have been a little bit biased because the Golden State Killer case had just been solved and I think people were very enthusiastic about the idea of identifying that particular that suspect. But what was really interesting to me is that about in the 80s to 90 percent of the people that we surveyed felt that they were totally comfortable with law enforcement using these databases for the purposes of solving really violent crimes like murder and rape, for identifying missing persons and for crimes against children. And only about 40-something percent were comfortable across the board with them using this information for less violent crimes, more petty crimes so things like drug possession, immigration, things that you think might become a sensitive use of this information. So that gives a little bit more nuance into how we ought to be thinking about this from a policy perspective, I think.
Erin: Why do you think that users are putting affordability and curiosity above privacy and confidentiality with these tests?
Dr. McGuire: So I think everybody is different in sort of their appetite for privacy risks and in terms of their desire for information. So I think it's a very personal decision and it's like any other decision that we make in any other sphere of our lives where it's a trade-off. People make trade-offs whether consciously or subconsciously, they're trading off what they see as the risks and the benefits of what they're doing. So the example that I like to give is that I know that I am risking my privacy every time I shop online using my credit card. And if you were to ask me in the hypothetical, is that risk to your privacy really worth the benefit that you get from the convenience of using Amazon as opposed to driving five minutes to Target a couple times a week, I would say probably not. And yet I do it all the time. I use Amazon, I put my credit card in there, I give them all my information and I make that trade-off because in that moment the trade-off is worth it to me. I don't want to get out of bed and go to Target.
Those are the same sorts of trade-offs that people are making when they decide when they decide to get genetic testing or they decide to do direct-to-consumer genetic testing. Some people care a lot about their privacy, and they're not going to ever have these tests because they're worried that somebody might get access to their information. Other people don't care very much about their privacy, and they're really curious about who am I, and where did I come from, and what does my future look like, and can I get all of the information that I want about the circumstances of my life? So they're going to get this test. And other people just don't think that the privacy risks are that great.
So I think there is certainly a risk that if somebody gets your genetic information and they want to do something harmful with it they can, but we don't have a whole lot of documented cases of people actually using other people's genetic information in ways that are particularly discriminatory. Now that may just be because they haven't been documented because we certainly hear anecdotal cases of people feeling like they've been harmed based on their genetic information or family history information. But I think the jury's still out on what exactly are the risks in terms of people's privacy and what happens if somebody does actually get your information.
Erin: So I could assume that perhaps people would be afraid that they would not be employable if they had some high degree of risk for a cancer or something that would be medically expensive for an employer to take on. Is that the kind of concern that hypothetically is out there in the in the privacy tradeoff?
Dr. McGuire: Yes, I think most people when you talk to them and they they're worried about their privacy they're concerned about employment and insurance. Those are kind of the two big ticket things that come up. We actually do have a federal law, the Genetic Information
Nondiscrimination Act, that went into effect in 2008. And that protects people from being discriminated against in most employment situations, at least with employers who have over 50 employees and health insurance circumstances where they can't use predictive risk information based on your genetics to make discriminatory decisions about hiring, firing, coverage, reimbursement things like that. In health insurance, we also have some additional protections under the Affordable Care Act where even if you have symptoms of a disease they can't make discriminatory decisions because they can't base those decisions on pre-existing conditions any longer, so we have some protections there.
The areas where we don't have any protections are things like life insurance, long term care, disability - those types of industries still very much use risk assessment to make coverage decisions, and part of their risk assessment is going to be what is your family history? Do you have any genetic predispositions to things? That's kind of how they do business, so I think that's probably the area of most concern for people. And people who are getting these kinds of tests really need to understand that if they're subsequently going to apply for life insurance and they're asked whether they have had any genetic tests or they have any genetic predispositions to disease and they do have that information and don't disclose it, that that could interfere with their ability to collect on insurance should they need it in the future. And so having that information does put them a little bit at risk with regard to those types of things.
Erin: If I choose not to get involved in having my genetic testing done on myself, am I safe? Am I free from my privacy being violated, my personal information being accessed by law enforcement or someone else? If I haven't spit into the tube am I safe?
Dr. McGuire: So the whole premise behind this is familial matching, and that's what makes DNA so powerful in some ways, is that it doesn't just give you information about yourself, it gives you some information about all of your relatives. So the answer is no. Most of the people who have been identified as suspects in crimes did not individually participate in ancestry testing. They are not genetic genealogists, they did not do direct-to-consumer genetic testing, they don't even know anyone in their family maybe who has done direct-to-consumer genetic testing or genetic genealogy. It could be their third degree relative who they don't even know they're related to and they've never met, they you know have no relationship with who's in that database.
There was a really interesting study that came out in Science recently where the investigators looked at, just based on how many people currently engage in these genetic genealogy types of activities, how many people in the United States could be identified based on the information that are in those databases. One thing to note is that those databases tend to be biased more towards people of European descent, and this is a good and a bad thing. On the one hand, there's not a lot of racial diversity, there's not a lot of ethnic diversity. This is a good thing when you're talking about law enforcement access to these databases in some ways because one of the major criticisms of that forensic database that they typically use, the CODIS database, is that it reflects the racial biases that are inherent in our criminal justice system. So the argument has been that the majority of people that are in that database are African-American men, they're ethnic minorities, racial minorities, and so those people in those populations are more likely to be identified because they're the ones who you have access to. This sort of equalizes that because now we've got these huge databases of largely white men and women. And in the study that was done, my colleagues found that currently based on what information is in those databases, 60% of Americans of European descent could be identified by familial matching to those databases. And they suspect that with the industry growing at the rate that it's going that within a few years, everybody of European descent in the United States would be able to be identified through familial matching using those databases. So we're all sort of at risk of potential identification.
Erin: I have an ancestry account, and every major holiday they say, “Hey we've got a special on this at home test that you could you could give everybody in your family one, and then just think how much more data you'd have about your family tree! We could fill in all the little gaps!” How much of the marketing of these kits is purely for the entertainment value?
Dr. McGuire: I think a lot of people certainly find entertainment value in this, and I think the ancestry thing is a big sort of hook for that because a lot of people see that as low-risk, just kind of a matter of curiosity, and again this sort of innate drive to connect with people even if we never knew who they were. But even with the ancestry thing, I would say you have to be aware of what you're going get.
So I have a really interesting story, I'm actually very close to my second cousins. My grandmother on my father's side had two sisters, and the three sisters were very very close to each other their whole lives, so as a consequence all the second cousins are very close to each other. And not that long ago my one of my second cousins came down to Houston for a conference and said, do you want to have brunch. And we got together for brunch and he said did you hear about our family scandal? I said, what's our family scandal? We're a pretty boring family. So we sit down to lunch, and he had spent the previous three months doing this intensive investigative work because he took an Ancestry.com test and it connected him to several of our other known relatives who were second cousins. And then there was a second cousin who popped up who nobody knew who this person was. So he reached out to this person and he said how could we possibly be related? I know all of my second cousins. And this person said I have no idea, I am super confused. My dad is Hispanic and my mom is I think Irish Catholic from Ireland, and I did this test and I'm 50% Ashkenazi Jewish and related to all you people. She says I have no idea what's going on here.
Well she ended up being related to somebody else, they ended up contacting all these people, they had this huge investigation that the two of them did together. And it turns out that they figured out that there were two generations of infidelity that happened that led to her being born. And her biological father is not actually her biological father. And her parents are still alive and they're still married, they’ve been married for 50 years, and it created a lot of family drama. You know, she didn't want to tell them but there was this whole family secret. And needless to say also her biological father, who they figured out who that was, had no idea that he had a daughter and had no idea that his biological father was not the person who he thought his father was.
So it was a really complicated story and you hear that all the time. And so you can imagine that it can be upsetting, sometimes even more upsetting than getting health risk information, to find out family information like that. So I do think if you’re going to give this as a gift, people need to understand what they're giving and what it might reveal and make sure that the other person is on board with it. If you're giving something as a gift that includes health information and I think there's another layer of understanding in terms of what they're going to do, and you always should give the kit and not just surprise them by taking their saliva somehow because that would that would be problematic.
Erin: Grabbing the Kleenex out of the trash?
Dr. McGuire: Well you do have to actually submit quite a large sample of your saliva, so it requires extensive spitting into a tube so that might be difficult to collect without somebody's knowledge surreptitiously.
But the other angle on that is that a lot of people are doing this for their children who’re under 18 years of age. And the companies will say in their Terms of Service you have to be over 18 to do this, but they don't know who's submitting the sample. And I think that can be somewhat problematic because you know as we've talked about not everybody wants the same information. And so if you have a child and they haven't had the opportunity to make their own decision about whether they want this information, we probably want to give them the opportunity to reach the age of maturity so that they can make their own decision about this.
Erin: Thank you for tuning in to Body of Work by Baylor College of Medicine. If you enjoyed this episode, you’ll want to subscribe and be on the lookout for our next episode, where we’ll talk to addiction expert Dr. Thomas Kosten about American’s opioid crisis.
If you like the show, please give us a five-star review and tell your friends to listen. We’re available on Spotify, Apple Podcast and Stitcher as well as at bcm.edu/podcast. There you can also find the episode notes, including information about the experts featured on the show.
A quick note about the medical advice and opinions stated in this podcast: each individual’s health profile is unique, so please see a healthcare professional about any questions you may have.
Until next time, take care.