For more than 40 years, the Golden State Killer haunted California. A serial rapist and murderer active in the 1970s and ’80s, he eluded detectives for decades. By 2018, hope of identifying him was fading, until a woman — curious about her ancestry — spat into a plastic tube and mailed it to a genealogy company.

Her DNA became the key. Investigators traced a distant cousin match in the GEDmatch database, built a sprawling family tree and followed it to Joseph James DeAngelo Jr., a former police officer living in Sacramento. A discarded tissue confirmed the match, and in 2020, DeAngelo pleaded guilty to 13 murders and 13 kidnappings.

The case was the first high-profile crime solved using consumer genetic testing. Since then, millions have sent in their own samples — to companies such as 23andMe and AncestryDNA — to explore ancestry, uncover health risks or trace long-lost relatives. As people turn to DNA to unlock the stories of their past and future, the science often reveals something messier — and far less certain — than they imagined. In the process, our understanding of family and identity can be upended.

When DNA Reshapes Identity

In fact, the revelations can be shattering: a parent who isn’t biological, or siblings who, suddenly, aren’t your own. Yet, DNA also shows just how closely we’re all connected. The most recent common ancestor of all humans alive today lived only a few thousand years ago — a point USC Dornsife geneticist Michael “Doc” Edge emphasizes in his genetics class.

“I’ve got bad news for all of you,” Edge, assistant professor of quantitative and computational biology, tells his students on the first day of class. “You’re related to me. We’re all related.”

But if we’re all connected, why can uncovering ancestry sometimes be so destabilizing? Why can it unravel cherished family stories or shake a person’s sense of self?

Police Windfall, Privacy Minefield

Genetic data has proliferated at an unprecedented scale — and some of it is now searchable online. This development happened quickly, with almost no public debate.

Americans have long opposed the idea of a national DNA database, citing privacy concerns. But under the radar, consumer genetics has created something close to one. Only about 7% of Americans have taken a home DNA test, yet thanks to the mathematics of shared DNA, nearly everyone is now traceable.

“Only around 1% of the population needs to upload their DNA into a database searchable by law enforcement for virtually everyone to be findable,” says Edge. “So, a small group of participants has created a de facto national DNA database.”

For police, it’s a windfall. But it’s also a privacy minefield. When one person shares their DNA, they’re effectively making decisions on behalf of their entire extended family. Edge’s research has shown how bad actors could manipulate these databases, potentially exposing users — and their relatives — to phishing attacks or other malicious actions.

Meanwhile, forensic genealogists — the specialists who build family trees to identify suspects — have been caught dodging database rules. In 2023, a whistleblower revealed that some investigators routinely searched data from people who had explicitly opted out of law enforcement use.

And life insurance companies could, in theory, sift through DNA data to assess medical risks.

There’s no evidence this has happened — yet. But there is no law to prevent it. “I think society still doesn’t understand the risk landscape,” says Edge. “And we’re definitely underprepared.”

A small group of participants has created a de facto national DNA database.

Genes Are Not A Crystal Ball

Even as genetic databases expand and researchers find new uses for them, the information DNA can reveal remains limited. Many people imagine genes as a crystal ball, predicting everything from disease risk to intelligence. But the reality is far murkier.

“We’re learning that genes don’t matter as much as people thought,” says Patrick Turley, associate professor (research) of economics, who studies how genes influence health, behavior and social outcomes.

“We used to believe there might be individual genes that have major effects on diabetes, or even IQ. What we know now is very different: Traits are shaped by all our genes working in concert. Millions of genetic markers nudge us in subtle ways, shaping probabilities rather than certainties,” says Turley, who directs the Behavioral and Health Genomics Center at USC Dornsife.

Consider our health. Some conditions, such as Huntington’s disease, stem from a single faulty gene. But most hereditary illnesses involve vast networks of genes, each exerting only the faintest influence. In this chaotic landscape, predicting an individual’s risk becomes extraordinarily difficult.

Still, knowing you carry a higher risk than average can be helpful. Doctors already ask about family history; genetic testing can complement that by flagging patients who might benefit from earlier screenings or more frequent monitoring.

But having a genetic risk factor is not a prediction. DNA, Turley says firmly, is not “some mythical force.” Tell someone their diabetes risk is 25% instead of 20%, and that small bump can feel like destiny. “People tell themselves, ‘I’m a person who is at high-risk for diabetes,’” he says. But in fact, it’s a tiny difference that will have a negligible effect.

Even height — often cited as a success story of genetic prediction — reveals the limits of science. Researchers analyzed 5.5 million genomes tied to measured heights, yet their best model could explain less than half of the variation. “It’s a huge effort to obtain predictions that, in the end, are simply not that good,” says Turley.

The challenge goes deeper: Most genetic datasets come from people of European descent, making predictions less accurate when applied to other populations. The gap widens further for complex traits like intelligence or educational attainment. Using genetics to predict whether someone will graduate high school, Turley explains, “tends not to be useful for predicting a person’s individual outcome.”

Turley has also studied a new and often controversial frontier: embryo selection. Companies such as San Francisco-based Orchid Health now offer to test and rank embryos according to genetic risk factors. The potential benefit, Turley says, is modest. In the best-case scenario, choosing the “best” embryo from 10 might, for instance, reduce the risk of diabetes from 35% to 30%. Yet companies rarely communicate such caveats to their clients, leading to confusion and overconfidence.

When DNA Rewrites Family

DNA’s influence — real or imagined — extends beyond health, touching identity, ancestry and the stories we tell about family.

Turley points to a familiar pattern: People raised to believe they had Native American ancestry discover they have no trace of it in their genome. For someone whose supposed Cherokee heritage shaped their identity — or who suddenly learns the father they love isn’t biological — the fallout can be profound.

Genetic testing companies know this. They recognize that what genetic genealogists diplomatically call “nonpaternity events” can be emotionally impactful, and some provide referrals to support services when results upend families.

In his genetics class, Edge offers a different frame for family — one meant to soften the authority we sometimes grant to DNA. “When I talk about somebody’s mom or dad or their cousin in this class, I’m using it as a shorthand for sources of gametes: sperm and eggs,” he says. “In reality, there are many other ways to define a family, and those other definitions are sometimes more meaningful.”

On the bonds that truly make a family — the shared experiences, the chosen commitments, the lives lived together — DNA is silent, Edge notes. And while genetic testing promises to reveal who we truly are, it also shows how little of our ancestors we carry.

“You don’t share any more DNA with your ancestors 10 generations back than you do with a random stranger,” he adds.

Even as genetic databases grow, the ties that bind remain the ones we weave ourselves, strand by strand, across a lifetime.