The Daily - A New Way to Solve a Murder, Part 2: The Future of Genetic Privacy
Episode Date: June 7, 2019The police identified a suspect in a double murder after combing through DNA profiles on a website designed to connect family members. We look at what his trial will tell us about the future of geneti...c genealogy in solving crimes. Guests: Heather Murphy, a New York Times reporter, spoke with CeCe Moore, a genetic genealogist, and Curtis Rogers, a creator of the genealogy website GEDMatch. For more information on today’s episode, visit nytimes.com/thedaily. Background reading: The case of William Earl Talbott II, who is accused of killing a Canadian couple in Washington State 32 years ago, could result in legal precedents involving the use of genetic genealogy techniques by law enforcement.
Transcript
Discussion (0)
From The New York Times, I'm Michael Barbaro.
This is The Daily.
This is really too good for just my little Roger Cernan group.
Let's start another website and share this with other people.
We know the suspect kidnapped her, but we don't know who she is.
Could we, in fact, identify somebody who had no clue who she was
when all we really would have is her DNA?
My concerns were that there could be a violation of privacy.
Felice's case basically laid the groundwork for where everything's going now.
It's like when they first discovered that fingerprints could be used.
We've now started a whole new way of finding people.
Today.
today. Police identified a suspected double murderer after combing through DNA profiles on a website designed to connect family members. What his trial will tell us about the future of
genetic genealogy in solving crimes. It's Friday, June 7th.
It's Friday, June 7th.
So Heather Murphy, you've been covering this story for The Times.
We left off with Kurt Rogers, who'd just seen that the site he had built, GEDmatch, this database of DNA profiles that people are mostly using to connect with their relatives, was used to crack the case of the Golden State Killer.
Yeah. was used to crack the case of the Golden State Killer. Kurt then decided to change GetMatch's user agreement and officially allow for law enforcement to use his site
to find people suspected of murder and rape.
Mm-hmm.
So what happened next?
So what happened next is that law enforcement agencies across the country
are looking at this and thinking, hmm, what could this technique do for me?
So one of the people that they want to talk to...
Hello?
Hi, Cece, it's Heather.
Hi, Heather, how are you?
...is this genetic genealogist named Cece Moore.
The only other way that they can learn about their heritage
and their birth families is DNA.
Cece Moore was already very well known in the genetic genealogy community.
She's been on TV.
She's running this Facebook group that helps people use these techniques.
Genealogist Cece Moore says Janet's only hope is DNA and offers to help.
She's one of the leading genealogists using DNA testing and GEDmatch to help adoptees to find their biological parents.
So it was the day of the Golden State Killer announcement. I remember I got lucky and
randomly called you and you were very upset and you were very concerned. Why were you upset and
concerned in that moment? I had had many sleepless nights over the last few years because many law enforcement professionals had
approached me and asked me if I thought I could use my techniques to help them stop a serial
killer. And it was something I really wanted to do and something deep down I felt was the right
thing. But I didn't feel it was the right time. She knew what was possible with genetic genealogy
techniques. But she and others in this little world that they're in of sort of advanced genealogical puzzle masters, you could say, they had some concerns.
I was pretty anticipatory. You know, how was this going to go? Would people see this as a good thing or a bad thing? Would they think it was an invasion of privacy?
or a bad thing? Would they think it was an invasion of privacy? She's wondering what Kurt Rogers of GEDmatch is wondering, which is, are people going to be upset? People have joined
these sites to find relatives. It feels very personal. How are people going to feel when
they learn that law enforcement is also using that site, combing through all that information
and using it to arrest relatives they may or may not have ever met for crimes that they may or may not have ever done.
I felt a responsibility, an ethical responsibility to my community,
in part because I had helped grow these databases.
I think I was certainly one of the biggest promoters,
if not the biggest promoter, encouraging people to upload to GEDmatch specifically.
If this went badly, using it for a murder case, then what if there were some sort of legal
objection and the site got shut down? And if people were upset about it and started to leave
the site, then the site would cease being useful for all the things that different genealogical
researchers needed it for. Many bioethicists that I expected to be concerned about it,
many of them had been quite negative in the past.
So I was really quite surprised to see the really enthusiastic reaction.
And what she sees is that people are pretty positive.
And so when Kurt decides to open up the site explicitly to law enforcement,
she decides to get on board.
That was really important to me,
and that was one of the big reasons I was able to move forward.
So what does she end up doing?
So right after the arrest in the case of the Golden State Killer,
Cece gets a call from this company called Parabon.
Parabon Nanolabs of Reston, Virginia
has developed sophisticated software
capable of unlocking information contained in DNA
left at a crime scene.
Up until this point, their primary service
involves taking somebody's DNA
and making predictions about their eye color,
hair, race, and turning that into this sort of mugshot.
Because of Parabon's snapshot, investigators will have their first clue of what a suspect looks like.
Not all scientists are totally convinced that these facial predictions are really accurate,
but they have made a successful business of doing this for law enforcement agencies across the country.
When a lot of police departments are desperate, they come to Parabon to have something.
And so when this company comes to her and says, we'd like you to work with us.
We decided to join forces to take that same crime scene DNA that Parabon had been using to predict what someone looked like, but now try to actually identify that person.
She said, yeah, sure. And because they already have DNA from dozens of law enforcement agencies
that they were creating these mugshots for.
I kind of had my pick of cases right away.
Immediately, she's able to dive right into cases.
And at the time of the Golden State Killer suspect arrest, a lot of people were opining that
this was going to be a one-off, that it was
too difficult, too much time investment, etc. And I wanted to prove that that wasn't true,
that the power of genetic genealogy could absolutely be applied to cold cases and even
active cases. There was no reason that that couldn't happen. And what does Cece do when
she starts this? What's her first
step? So one of the first cases that lands in her lap involves this couple who was killed in 1987.
The very first case I worked on was one up in Snohomish County, Washington. It was a young
couple from Victoria, British Columbia, Canada. Their names are Tanya Van Kulenberg and Jay Cook.
They were traveling from Victoria, British Columbia to Seattle, Washington,
took a ferry, and then they just disappeared.
I believe I was already aware of Jay Cook and Tanya Van Kulenberg's case.
And she's very intrigued by this case.
Because my family's from Washington State.
A, she's from Washington State herself, and intrigued by this case. Because my family's from Washington State.
A, she's from Washington State herself.
And B, this woman.
Tanya was born the same year I was.
Had she survived, would have been about the same age that Cece was.
They were barely adults.
Jay was 20, Tanya was 17.
Just starting out their life.
And I felt a real connection to this couple, even though, of course, I never knew them. They were going to pick up something in Seattle and then come home.
And the more she looks into this case, the stranger it gets.
The woman, Tanya, was very reliable.
She would always call her parents if she wasn't coming back.
But one day passes, two days pass, and she hasn't returned home when they expected her. And they start to get really worried. Finally, police find Tanya's body in a ditch in a
wooded area. She's partly clothed. She's been shot and she's been raped. And they don't find Jay
until a few days later. And he's 90 miles away in a different location. And he's been killed
through totally different means. He appeared to have been strangled,
and he has this pack of Camel Lights stuffed in his mouth.
Police follow many different leads over the years,
but none of them go anywhere.
And so for 30 years, they're left without a prime suspect.
And it was a terribly sad and horrible crime.
So as Cece's looking over this file,
what she sees is that Parabon has already made
this sort of predictive DNA mugshot
for the sheriff's department.
The crime scene DNA that Parabon had analyzed
was from semen.
I believe it was from a rape kit from Tanya.
So she's ready to move forward.
And what does she do?
The first thing she does,
sort of like what Barbara Ray Venter did in the Lisa case, is she takes that profile and she
uploads it to GEDmatch, Kurt's site. In the few years that have passed since Barbara was first
trying to solve the Lisa case, GEDmatch has grown. We're now at about a million profiles just in GEDmatch alone. So Parabon uploaded that raw data file on a Friday evening.
And so there's a lot of comparison going on, heavy algorithms.
And it takes between about eight hours and two days.
And so I was, you know, waiting with bated breath on pins and needles.
I stayed up really late waiting for it.
I finally fell asleep.
And so I woke up, logged in,
rolled out of bed right away,
opened my computer, logged in,
and saw that we had two quite significant matches.
And she gets lucky.
Both of the matches shared enough DNA
to be approximately second cousins.
This person who for 30 years,
nobody knows who they are. Immediately, they find
two of their second cousins on GEDmatch. And second cousin matches are really good on GEDmatch.
Two people that were approximately second cousins with our unknown suspect who didn't share any DNA
with each other. These cousins are on different sides of the family. One's on the father's side,
one's on the mother's side. So I knew that somehow the family trees of those two people had to converge in order to point me toward the suspect.
And so from there, she proceeds to build her family tree back to the common ancestor, and then she starts building it forward in time.
And so this triangulation was pointing at just one family, and in that family there was only one male
who could have that correct mix of ancestral DNA.
And then I found that he lived very close to one of the crime scenes.
In fact, really just down the road is where one of the bodies was found.
So once Cece has identified a potential suspect, what does she do?
So she's working with the local sheriff's department.
And what they do is they do some investigation to make sure this is plausible, that this really could be the guy.
And then they go out in front of his house and they watch him.
And he eventually discards a cup.
They go in, grab the cup, take it back to the lab, extract a genetic profile.
And they see, does it match our crime scene DNA left from 30 years ago?
And?
It does.
Yesterday, we took into custody a 55-year-old SeaTac man who was suspected of the 1987 murders of Jay Cook and Tanya Van Kylenberth.
And so just three weeks after the Golden State Killer announcement, there's a new
arrest. William Earl Talbot
II. William Earl Talbot
II. He stands charged
with two counts of murder from
1987. They have their suspect.
Six on your side at five.
Starts now. And
for all the law enforcement agencies that have been
watching and thinking, was that just a fluke
what they did in the Golden State Killer case?
Could this ever be done again?
It's Fort Wayne's most notorious cold case.
Oh, actually, yes, this could be done again.
Police have made an arrest in the 1988 homicide.
And so suddenly, this sets off...
DNA from condoms found in 2014.
Arrest after arrest after arrest.
Police say DNA evidence helped them find their guy. More than
40 people have been identified for murders and sexual assault. That's when Miller admitted to
abducting Tinsley. Some of them are starting to plead guilty, which is also reinforcing, okay,
this is working. When a loved one is the victim of a violent crime, families want answers.
Pretty soon, a site that's quite similar to GEDmatch called FamilyTreeDNA decides that they are also ready to start openly cooperating with law enforcement.
You don't have to buy anything, just participate.
They even advertise that their customers can help find a killer.
Go to join.familytreeDNA.com.
And all the meanwhile,
none of these cases
have gone to trial yet.
None?
No.
Until now.
The case of William Talbot II,
the person,
the alleged killer
of Jay and Tanya,
will be the first
to go to trial.
This will be the first time
that genetic genealogy
is tested in court.
We'll be right back.
So in just a few years, this website called GetBatch, created by Kurt Rogers, this retiree interested in genealogy, is used to help a kidnapped girl find her family
and figure out who she herself is.
It's then used to solve the Golden State Killer case.
And now it's being used by CeCe Moore
to find a suspect in a double murder case,
William Talbot II.
And he's now about to go on trial.
Yeah, it's pretty incredible, right?
And what do we know about William Talbot II, the suspect? and he's now about to go on trial. Yeah, it's pretty incredible, right?
And what do we know about William Talbot II, the suspect?
So what we know is that he tends to go by Bill among friends. He's spent a lot of his life working as a truck driver
and he's never gotten in trouble with the law.
Not once.
No, in any serious way.
And what we know is that what connects him to these two murders are two pieces of genetic
evidence. One is the semen found on Tanya's body, and another is semen found on Tanya's pants that
were found in the van that Tanya and Jade took on their trip. That's it. That's the only DNA evidence,
at least in the court documents that I've seen, connecting him to the crime. So
there's actually is no DNA evidence connecting him to the second murder. To Jay's murder.
That's right. And when you read through these court documents, what you see is he's a man who
has a lot of friends and he's not known to be an angry guy. Here, let me just read you really
quickly something that stood out to me from one of his friends. This was in the court documents.
I've never seen Bill get upset with or even raise his voice with anyone. He never would.
I would trust Bill with my life. With everything I hold near and dear to me,
I would trust Bill with my daughter. He is a caring, loving, sweet, compassionate person
and a good friend who wouldn't do harm to anyone ever. He's always confident and a happy, sweet soul.
This is from a friend who goes on camping trips
and rides on a Harley with him sometimes.
She's known him since 2006.
So there's a lot of notes like that.
And so you can imagine that all of this has come
as a big shock to the people who know Bill.
And so is there any possibility
that this DNA doesn't connect him to the crime, that there's another plausible explanation for why it was found there?
So all genetic genealogy can really claim to do is to identify a bit of DNA, to put a name to a bit of DNA that's found at a crime scene.
But just because DNA is present somewhere doesn't necessarily mean that somebody killed somebody.
So, yes, it is possible that we will see an argument made by the defense that Bill had sex with this woman but didn't kill her.
Or was involved in something that happened with Tanya but was not involved with something involving Jay.
Because what they may point out is there is no physical evidence,
at least as of a month ago in the court documents,
tying him to that scene.
So how much emphasis and weight are the investigators
putting on the DNA evidence
in the absence of any other corroborating evidence?
Yeah, so this is an interesting case
because in a case like the Golden State Killer,
you have multiple crime scenes with DNA evidence.
And so the suspect is being connected
through multiple crime scenes.
In this case, you just really see
that our suspect is connected to one of the murders.
And yet he's charged with both murders.
And what it reminds us of
is how much Kurt's site and genetic genealogy
have shifted the nature of these investigations.
Often before, DNA was something that comes in later in an investigation
once you have a list of suspects to either confirm or deny
that somebody should continue to be investigated.
In this case, DNA is actually a starting point.
Bill was not on the radar of investigators ever.
It was only once they put a name to this DNA that they began to look into him.
And so it shows this sort of flip in the order of operations
of how an entire investigation is going to work.
And it seems so compelling on the surface, DNA.
It's very present somewhere. is going to work. And it seems so compelling on the surface, DNA.
It's very present somewhere.
But in the absence of additional evidence,
the question becomes, what does it really mean?
Exactly.
For example, there's no time element to DNA.
We actually don't know when it was left behind in the timeline of this period of days
in which they both died.
And it just raises a lot of questions.
I think that's one of the things we'll see with this case is, is DNA enough?
I wonder why weren't police able to do anything with this suspect using traditional investigative techniques?
Yeah, they tried. And they did have DNA. And police do have their own DNA database that they've been using long before they realized that they could use Kurt's site, GEDmatch. It's called CODIS. It's the criminal DNA database. And there are a few
reasons why CODIS didn't do it for them. One is that in order to find a person in CODIS,
in some states, you need to find an exact match. Not just a relative.
Exactly. So if they have this crime scene DNA from William Talbot II, you'd need William Talbot
II to already be in their database in order to find him.
And in order for that to happen in Washington state, he would have had to be convicted for
a felony or some other major crime.
And he wasn't.
He wasn't.
Because Bill is not a convicted felon.
And so this actually highlights another piece of this, which is that CODIS traditionally has skewed towards certain populations, including its disproportionately African-American.
white retirees. And it highlights the fact that a lot of the people that we're seeing being arrested for these crimes using genetic genealogy are white and middle class, people with no criminal record.
So in a sense, and they probably did this unwittingly, they supplied the world with a much
bigger and more racially diverse group of potential suspects.
Yes. And I have talked to people who are very concerned about genetic genealogy,
but they say like, hmm, that one point is sort of interesting in sort of how it
inadvertently managed to sort of balance things out a little bit there.
So just to be clear, it's very possible that William Talbot II never uploaded any DNA
to any database, and yet he has still been arrested and he's about to be tried based on DNA that somebody else uploaded.
calculations by several different researchers could potentially be used to identify most white people in America, actually a good percentage of all people in America, even if those people have
chosen not to take 23andMe tests, not to take an ancestry test, just because somewhere out there,
two people in the third cousin range decided to upload to GetMatch. So basically, you couldn't
even escape this if you wanted to. No, there's no more genetic privacy. It's done.
And that's because of this site.
Yes.
That's because of what Kurt Rogers built.
It's because of what Kurt Rogers built.
Privacy is a big question that people have around this,
specifically if this is a constitutional search and seizure issue, if this is a Fourth Amendment issue.
So one lawyer told me that because most people on these sites aren't expecting that their profiles are going to be used by law enforcement, they're just looking to find some relatives or do research or find their biological father.
do research or find their biological father.
And so because police are building these cases and identifying this DNA off people who have no idea
their DNA is going to be used that way,
that the evidence wouldn't hold
and that she would challenge it in that way.
It'd be like an illegal, warrantless search of a house.
Anything found in it couldn't be used in court, theoretically.
Yes.
But I will say she was in the minority.
You know, another lawyer I spoke to pointed out
that really the Fourth Amendment
protects what people have decided not to expose.
But his point was that this is a case
where people have voluntarily exposed their information.
So good luck to the defense trying to challenge
that law enforcement shouldn't use this site,
is what he said. And I've heard that from a lot of people. But I will say that even these lawyers who are
saying that they think that this technique will hold up in court agree that this is uncharted
territory. And they're all very curious to see how it plays out. Because usually the law is a bit
behind science and technology. So even if it's not a constitutional
issue right now, they said it could be a constitutional issue in the future.
Even some of the lawyers who told me they see this working and holding up legally,
they have ethical concerns about it and said, well, just because it holds up in court
doesn't mean to them that there shouldn't be some sort of regulation guiding how this works.
However it shakes out, it's going to set some kind of precedent that is very important symbolically for law enforcement agencies across the country that are trying to decide whether they feel comfortable with this.
And in the end, all of this, of course, comes back to these sites, particularly GEDmatch, that allow the investigation of people's genetic information.
Yeah, it is incredible to think that it all began with Kurt.
And what does Kurt think of all this now?
You know, I've asked him directly. You essentially became the new default
government database. That feels like a lot of weight on one person. Was that a lot of weight on you?
of weight on you? I have not given that much thought and I don't feel that. Maybe that's a perception outside. I just have not thought about it. I think it's hard for him to feel how big it
is. I think he gets it at moments. I think when he gets all the
letters from people saying, thank you. And I think also when he gets the backlash. Recently, he did
something that made some people angry, which is that he allowed a detective and a genealogist
to use his site for something that wasn't a murder. It was an attempted murder. He was very
convinced and moved
and wanted to help with the case.
So that was outside of his privacy rules.
Yes, he violated his own privacy rules
and people got very angry
and he went through a bit of a reckoning
and decided, okay,
I'm actually going to listen to people
who are saying there's some issues
with the consent agreement
and he changed the user agreement yet again.
there's some issues with the consent agreement, and he changed the user agreement yet again.
We decided that we needed to change our terms of service and required that people opt in.
So users now have to choose to opt in to allow their profiles to be used by law enforcement.
That takes the pool from one million to how many that law enforcement can use?
Zero.
Can you explain?
To zero.
Immediately, we opted everyone out. Everyone had to come in and specifically say to us that they wanted to have their information used by law enforcement.
So you could say, OK, isn't this story over?
Actually, no, for two reasons.
One, they are trying to slowly grow their database back up by encouraging people to
opt in.
They're very quickly at 30,000.
But the bigger reason is because this has already moved way past GEDmatch, as Kurt fully
recognizes himself.
You know, Family Tree DNA has said they can allow their database to be used,
and their database is as big as ours.
Family Tree DNA also has a pool of one million,
all available to law enforcement.
And so this journey that Kurt has on
has already become something much bigger.
In other words, what he has created,
no matter his intentions, no matter their rules,
he puts in place the default settings, it can't be undone.
It can't be rolled back.
No, there's no going back.
I mean, I talked to a detective the other day who said to me, the genie's out of the bottle.
There is no going back.
Heather, thank you very much.
Thank you. Thank you.
William Talbot II is scheduled to stand trial
for the murders of Tanya Van Kulenberg and Jay Cook on Monday.
We'll be right back.
All next week on The Daily.
Within a few hours,
things go totally nuts.
My colleague Katrin Van Holt.
Got these people breaking shop windows, chanting angry slogans, shouting foreigners out.
On the state of Europe and its union.
Us, as European citizens, we are living in a bubble.
This populist wave would be able to destroy it from the inside.
We've had enough. We don't want to be governed by you. We want to govern ourselves.
The Daily is made by Our theme music is by Jim Brunberg and Ben Lansford of Wonderly.
Special thanks to Sam Dolnick, Michaela Bouchard, Stella Tan, and Julia Simon.
That's it for The Daily. I'm Michael Barbaro. See you on Monday.