This video includes a discussion of serious crimes which may be disturbing for some viewers so I wanted to let you know that upfront, but I think it's necessary to talk about these crimes in some detail for reasons that will become apparent. In the small town of Visalia, California in the mid 1970s, one of the state's most prolific criminals got his start. He repeatedly broke into houses and stole small items like cash, coins, and jewelry. He was dubbed the Visalia ransacker but soon his crimes escalated.

- Six months later in 1976, he moves up to Sacramento and over the course of the next three years, he's moving all across Northern California committing numerous sexual assaults, over 50 sexual assaults. - Here, he was called the east area rapist. - [Male Reporter] The police are saying lock up tight. Sacramento's infamous east area rapist may still be in town.

- And then in 1979, he moves down to Southern California, starting in Santa Barbara and starts killing. And he was known as the original night stalker in Southern California. - [Reporter] Variously known as the east area rapist and the original night stalker, these were brutal meticulously planned crimes that spanned numerous California counties between 1976 and 1986. - Police believed that all of these crimes were committed by the same person now known as the golden state killer. How was it known that these were committed by the same person?

- Investigators were relying on MO. MO, modus operandi is basically how a criminal will commit a crime. - He was wearing some type of a mask or a hood. He was wearing shorts and a T-shirt. - The golden state killer had a very distinctive MO - What he would do is he would break into a house with a gun, a house with a man and a woman. He would tie them up tell the man I'm gonna put plates on your back. If I hear those plates move, I'm going to kill your wife and then kill you, take the wife into another room and sexually assault her. Then he would go through the house for however much time, eating, stealing stuff, stealing little things or whatever and then he would leave.

- How was he not caught based on fingerprint evidence? - Well, he didn't leave his fingerprints. He always wore a mask. He always had gloves on. This was an offender that did everything to prevent himself from being caught. - [Reporter] The golden state killer has been linked to at least 12 murders, 50 rapes and a string of burglaries throughout California. - All right, how's it going? - How you doing? - Good. - So we supply DNA sequencing solutions to the forensic laboratories that process crime scene DNA. At a crime scene, there's typically less than a nanogram of DNA left and it's usually degraded, right? So it breaks down. - Why is there so little? at a crime scene? A nanogram? - Yes. - That's insane.

- Yes. - There's no accident that he stops in 1986. 1986 is when the first DNA case of a murder case is actually solved with DNA and he's, I think he's following that and he's saying, you know what, what I'm doing here, I'm leaving my DNA. I can't do this anymore. - [Reporter] Investigator Paul Holes tracked the golden state killer for decades. - He couldn't it account for DNA technology and that really was his big mistake. That was the critical mistake. He left his DNA all over California. And it turns out that in three of the cases, I had three sexual assault kits and that's where I got the golden state killer's DNA from Northern California.

- But having the unknown killer's DNA wasn't enough. They needed to match it to a known sample. In 1990, the FBI started work on what would become a national genetic database. It mainly stores DNA profiles from convicted criminals and persons of interest. It's called the combined DNA index system or CODIS for short. In each of your cells, there are 23 pairs of chromosomes, one from your mom and one from your dad. And at particular places on some chromosomes, there are short sections of DNA that repeat like A A T G, A A T G, A A T G, and so on. These are called short tandem repeats or STRs and different people have different numbers of repeats. Forensic labs produce STR profiles by counting up the number of repeats at each location. Initially, there were only 13 places on the DNA where STRs were counted, but in 2017 that was expanded to 20. It feels to me like that's not that much information. - It isn't. - But the chances of two people having the exact same number of repeats at all locations is incredibly small. The CODIS database now contains over 18 million STR profiles.

- The golden state killer's DNA profile has been up at the national DNA database that's run by the FBI called CODIS since 2001 and has been searching ever since with no hits. We did Interpol searching, trying to search other country's DNA databases thinking well, maybe he came from a out of the country that's why after all this time we have yet to be able to identify him and obviously we didn't get any hits.

- But genetic sequencing technology was advancing rapidly. - [Female Reporter] The genome project was completed. Since 1986 scientists had been mapping and sequencing the three billion nucleotides contained in the human genome. It's anticipated that

- Soon private companies began offering genetic tests directly to consumers and they provide far more genetic data than what's in the CODIS database. I traveled to Houston, Texas to visit Family Tree DNA headquarters. - We'll go in the lab and we'll start talking about the life cycle of a sample. There you go. - Thank you. These samples, what are they? - So they're cheek cells. They just scrape the inside of their cheek, they put a cap on it and they send it back to us. - It's like a robot ballet. - Yes, so for DNA extraction, we need to remove all of the protections that are on the DNA and just separate it by itself so we can do testing on that.

- Instead of looking at 20 places on your DNA, these tests examined 700 to 800,000. - The Family Finder test is based on what's called a micro array and this is what a micro array looks like. Each one of those boxes is an individual. We can fit 24 individuals on here. And within that box is about 710,000 positions of your DNA, that we're measuring, we're getting data for. - 710,000? - Yes, and they're called snips. - So hang on, that's like 710,000 individual bases, like ATGC? - Yes. - For each person, you can measure 24 on that single chip? - Correct. - The human genome consists of three billion bases. All people share about 99.9% the exact same DNA, which means on average, only about three million basis differ from person to person. These individual letter changes are known as single nucleotide polymorphisms or SNPs.

So on there, are there a whole bunch of like little pieces of DNA then? - Yes, so what we have is a short piece of DNA that is specific to one SNP. So it's one position in the genome where we know there can be a difference. - Okay. Is there any example of a simple SNP that confers some sort of? - Actually blue eyes is a single base change. - Really? - Yeah. - So everyone with blue eyes has this one letter change? - mm hmm - You can see now it's starting to scan the first box.

The results of your 710,000 SNPs can tell you about your geographic ancestry or possible medical conditions, but they can also quantify how related you are to someone else. The genetics of having a kid are a bit like cutting two decks of cards together. The resulting deck will share about half its cards with each parent and there'll be long sequences of cards which match identically sequences in each parent. Now, if this deck has a kid, well then its offspring will share only about a quarter of its cards with each grandparent and the sections of identical cards will be shorter, and this is how it works with DNA.

The more closely you're related to someone, the more DNA you share and the longer the blocks of identical DNA are. But to compare the DNA of two people, you don't need to read every letter. The snips are sprinkled throughout the DNA roughly every 2000 bases. So if you find long sections where the snips match up, you can bet that the DNA in between is identical too by comparing hundreds of thousands of snips sprinkled throughout the DNA, you can see where these identical blocks start and stop and so how much total DNA is shared. - Everywhere that is highlighted in blue is where I share DNA. - With your father? - With my father which is exactly what you expect because I should share DNA with him everywhere because he gave me 50% of my DNA. So this is a second to third cousin. I have no idea who this person is. And you can see here, we are definitely not as closely related as me and my dad because where the blue is, those are the only regions of the chromosome where we share DNA.

- We all know that these genetic tests can help identify our relatives but the question is, can you do the same thing with law enforcement databases? Can you identify family members using the CODIS database? - You can and it's only limited to really first order and second order relatives. - That means if I'm looking for the golden state killer, I'm looking for either his father, his brother or his son. We did that over and over again with no hits. - Why is it so limited? - Because you've only got 23 loci. If you wanna do proper kinship analysis, you need to have more points assessed on that genome to see how related you really are.

- We know the answer is behind a locked door and the answer is behind the locked door of 23andMe and Ancestry.com. We know that they have millions of profiles. Somebody related to him, a second cousin I would guess, would be in there. - Sure. - And that was very hard to deal with. You know, that the answer is there behind this locked door. - Law enforcement is not allowed to search the databases of 23andMe and Ancestry.com but luckily for investigators, there was another way. An independent website called GEDmatch had been set up to allow people to connect with relatives. You can upload the raw data from any of the big genetic testing companies and search for matches.

- So when we initially search the GEDmatch database, our best hits come back at third cousins which means they shared you know, our top hit, I think shared about 1% of the DNA. It's very little DNA that's shared but that's a starting data point. And theoretically, by taking multiple of these individuals that share DNA and building their family trees using traditional genealogy and public source documents like census records and obituaries and newspaper articles and findagrave.com, you build this family tree back until you get two of these relatives of the golden state killer, you get them to intersect where now you see, oh, they've got, they share great-great grandparents together. Theoretically the golden state killer is a descendant from those same great-great grandparents. So once I identify that common ancestor, it's now identifying every single descendant from those great-great grandparents and this becomes a huge process. People in the 1840s would typically have 15 kids. You have to identify each of those kids and then all of their children and then all of their children. So now we have this exponentially growing family tree.

- How wide did that tree get? - I think we had over 1000 individuals in that family tree but we knew a lot about our offender. We were confident he was born between 1940 and 1960. We knew he was a white male. We knew he was 5'8" to 5'10". You know, he's in California. He's up in Sacramento in 1976. He's in Southern California in the '80s and we basically narrowed it down to about five males and then it's just investigations.

- With a suspect at their fingertips, this spring, they followed him to a Sacramento area Hobby Lobby store and took a DNA sample from his car door handle. Days later, investigators recovered another DNA sample from one discarded tissue which registered a match to DNA evidence left at one of the crime scenes. - Investigators in California say DNA evidence led them to one of the country's most notorious serial killers. Former police officer Joseph James DeAngelo was arrested yesterday and he is believed to be the so-called golden state killer.

- For 44 years, countless investigators have worked it and have failed to solve it. Within four and a half months, a team of six were able to figure out who the golden state killer was and that was Joseph DeAngelo. - After the golden state killer was identified, the flood gates opened and we were seeing every week not only single killers, but multiple killers, serial killers, every week were being identified from cases in the '70s and the '80s. - 'Cause, there's been over 70 cases now that they've solved this way and they continue to solve them at a fairly regular rate and I think - Presumably that rate's just gonna increase. - Yeah. - And there's gonna be more because it's working and we are going to help find these people.

- But the scary thing is in the US, you have 100,000 cold case murders with DNA. You have more cold cases but the ones with DNA is estimated about 100,000 There's roughly about 650,000 sexual assaults, cold cases with DNA. - With DNA? It seems like a real paradigm shift. - It is. Multiple law enforcement agencies has said that this is the most revolutionary tool they've had since the adoption of the fingerprint.

- With traditional DNA, it required getting a sample from the actual individual. It's a one-to-one type of process. And so these offenders who let's say they have committed crimes and haven't been caught and they're just living their life and they're thinking, well, my DNA has never been taken from me. I may have left DNA in the case that I committed back in 1975, but they've never gotten my DNA, they won't find me. But now we don't have to rely upon that. Somebody distantly related to that offender has put their DNA into a database and they have to start getting nervous because that's outside of their control. They can't account for a third cousin or what that third cousin is doing. And so they start getting nervous. And I've kind of wondered how many of these individuals recognize that law enforcement is eventually gonna be knocking on their door as a result of a third cousin putting their DNA in the database?

- Each person who uploads their information to a database illuminates the identities of hundreds of other people around them, parents, siblings, aunts, and uncles and cousins. A study by ancestry DNA found the average person in the UK has 175 third cousins and it doesn't stop there. The DNA in one person will be shared by hundreds of people yet to be born. Children, grandchildren, nieces, and nephews, and cousins once or twice removed. Realistically, each person shares identifiable DNA with nearly 1000 people past and future. So it doesn't take many people uploading their data to reveal the identities of everyone living in a country. Back in 2018, a study found that using an existing database of just 1.3 million people, they could identify 60% of all Americans of European descent. And they estimated with a database containing just 2% of the population, you could find a third cousin or closer match in 99% of cases.

- DNA storage is one of the things that we have that is really unlike anything else. If you look inside, this is at minus 20 degrees Celsius. - That keeps the DNA good for a long time? - We have DNA samples in here that are probably about 15 years old. - How many samples are stored in there? - Just over 2 million. - Wow. - Yeah. So the capacity is just over 2 million. - 2 million people's DNA stored in this one small room? - Correct. - Wow. By 2021, over 30 million people worldwide have taken a direct consumer genetic test, the majority of which are through Ancestry and 23andMe, which do not currently work with law enforcement but Family Tree DNA does.

- If you ask Americans, should we have a national database where everyone's logged in it? - As their DNA? - Yeah by their DNA and their name, most Americans will be somewhat uncomfortable or potentially very uncomfortable with that. But we're kind of sleepwalking into that scenario. - There are some people who are concerned with privacy. If anyone chooses or decides that they do not want their results to be used for law enforcement matching, they can remove themselves from that. - I do worry about genetic information being used to discriminate against someone. - My biggest thing is health insurance. I worry about if you have somebody's DNA profile and some laws enacted, it gets in the wrong hands and they say that this person has a proclivity, they're headed towards Parkinson's, they're headed towards this, headed towards that and then rates skyrocket. That's the biggest place that I see as a privacy concern.

- Once your genetic information is out there, it's out there. It's not like a credit card where you can get a new one. Yeah you're giving away something that you don't know what it can do. - Well, yeah our information isn't ours in the first place. - Right. - Our information is shared. - Yes with your relatives. - Yes. - And although you may be adverse to this sort of thing, if they've already done it, if it, you know, in a way you're out there.

- In society today, you have very clearly had two camps. There are those that are very comfortable having their genome available to be able to be searched for law enforcement purposes and there are others that are not comfortable at all and that's okay. - The problem is of course, those people who don't want theirs to be searchable could have their DNA uploaded inadvertently by their relatives. Do you know what I mean? So like, if I don't want my DNA to be searchable by law enforcement but my sister uploads hers, - Yes. - It's almost as good as me doing it myself.

- Right now, Family Tree DNA's policy is we only accept samples from law enforcement under certain circumstances. It has to be a sexual assault, identification of remains, DNA recovered from the scene of a homicide. And we also do child abductions. So if there's a child that's kidnapped and there's DNA evidence, we will process the DNA so that they can try to find out who the abductor is. Those are the only cases that we will accept. - Got it. - Unfortunately now GEDmatch has flipped off the switch and GEDmatch has actually said that you don't all have to have your information searched without opting in. - Right. - Was that decision by GEDmatch a direct result of this golden state killer case? - No, it wasn't. - No. - Wanna tell the story? - You know, it had to do with a Utah case.

- Where an elderly woman was playing the organ in a church. - And there was an individual who broke in and assaulted her. - But it wasn't a sexual assault, but it's a serious crime. - I think believed, tried to strangle her and left her for dead. I believe he thought that she was obviously dead and left. The police I believe approached GEDmatch because that they were concerned that he was going to recommit other crimes. - But GEDmatch's terms of service said they would only help law enforcement in cases of rape or homicide. - Her problem was that she didn't die. - 'Cause if she died then? - If she died it would have fit the conditions and then they would have uploaded the sample because it met the terms of service and everything. - So what happened? They didn't take that case? - They did. - They subsequently caught the individual from that search that they performed on GEDmatch.

- But people within the genealogy world kinda got their passions and flame saying that is stepping outside of what this tool should be used for and they ended up putting a lot of pressure on GEDmatch. - They flipped the switch because they got freaked out. - But at the end of the day, it was a violation of terms of service. And I think ultimately led to a good thing. Everybody got zeroed out and then everybody gets to sign back in. - Subsequently GEDmatch was sold to Verogen. - I can tell you a new user today when they sign up for GEDmatch, about 73% of them are signing up to allow law enforcement to search. We're pretty happy with that. - Yeah I mean, that's, it's a high number. - It's a high number. - And it suggests that people want that innocence.

- How many second cousins or third cousins do you now? You know what I mean? You don't know that many of them. Without even knowing it, you're helping justice. You're helping a victim's family get a little bit of peace, not closure, but a little bit of peace and why not? - They get an answer. - Partly, you could make an argument that if you wanna be an activist helping catch criminals, one of the best things you can do is encourage everyone you know to be tested and put in the database. - Correct and you know, there are millions of people who have tested at other companies that are not working with law enforcement. And if they want to help, all they have to do is download their raw data and they can upload it for free into our database.

- I absolutely believe in a person's right to privacy. But I also absolutely believe we have a right not to get murdered or sexually assaulted or it will be part having violent crimes committed against us. There has to be an equal balance. - That scale of my privacy has been taken because of a third cousin has been identified versus a mother who's talking about her daughter has been murdered, that scale is way the hell like this. - We're at a point now where in many ways technology has kind of outstripped where the laws are. I truly believe that ultimately this tool and the, all the concerns with it will probably end up in front of the nation Supreme Court and there will be a decision. My bottom line message, if you're gonna make a decision, if there's gonna be laws restricting law enforcement to use it, make sure that you're making an informed decision and not assuming what it is based on its DNA and genetics and law enforcement because it really isn't what you think it is.

- I gotta keep a couple of million people happy so that I have access to the identities, kind of in a sense. - In a sense, right? - Of millions, of hundreds of millions of people. - Right. - And that's a pretty selective sample. These are people who are willing to give up their information - Give up, yes. - To find their ancestors. - Yeah. - There may be many others who are militantly private but they don't get a say in the decision of whether they're gonna be searchable through that person or not. - True, that's the choice you make. - That's the choice you don't make. - It's a choice you don't make but it's also it's the choice we have to make as a society, is what's that balance. There is always gonna be winners and loses in my mind, right? And I don't wanna use losers in the negative sense there but in a decision at the end of the day, you have to balance in my mind the public safety with the public privacy.

- Is DuckDuckGo, is that good? - I use it. I'm a little paranoid. So it's supposed to not track you as much I guess, but yeah. - I have been working on this video for a couple of years so I really wanna thank Brilliant for sponsoring it. I appreciate their willingness to support me tackling complex and important stories. Brilliant is a website, an app that teaches you all kinds of stem concepts in an interactive and in-depth way, if you wanna learn more about DNA or solving mysteries with deductive reasoning, I would recommend their courses on computational biology and logic. Check out this interactive logic puzzle. You have to figure out if the light is on or off, knowing that some of the characters may be lying to you. And instead of just telling you the solution, Brilliant allows you to try different configurations and discover the answer for yourself.