378. 23andMe (and You, and Everyone Else)

Episode Summary

Executive Summary: The episode examines 23andMe’s rise from consumer DNA testing to a major genetics and research platform, using the Golden State Killer case as a gateway into debates over privacy, regulation, health risk prediction, behavior change, diversity in datasets, and the ethics of monetizing genetic data. Anne Wojcicki argues direct-to-consumer genomics empowers individuals and advances medicine, while the transcript raises concerns about accuracy, equity, and unintended consequences for families.

Main Topics: Golden State Killer and consumer DNA policing (Priority: 5/5): The episode opens with the Golden State Killer arrest, showing how genealogical databases and relative-matching helped police identify a suspect through uploaded DNA and public records. 23andMe’s mission and origin story (Priority: 5/5): Anne Wojcicki explains that 23andMe was founded to shift healthcare toward prevention, consumer empowerment, and better use of genetic information rather than monetizing illness. FDA regulation and clinical legitimacy (Priority: 4/5): The company’s health reports were initially blocked by the FDA, forcing 23andMe to prove its product safety and validity before regaining authorization for selected risk reports. Privacy, consent, and data-sharing ethics (Priority: 5/5): The discussion focuses on customer control over genetic data, limits on sharing, and concerns about law enforcement, insurers, employers, and foreign governments using DNA data improperly. Predictive health reports and behavior change (Priority: 4/5): The episode explores polygenic risk scores, diabetes risk reports, and whether genetic information actually changes behavior; evidence cited is mixed and the host challenges the strength of the claims. Diversity gaps in genetic databases (Priority: 4/5): Criticism of Eurocentric datasets and concerns about accuracy for non-European populations lead to discussion of recruiting more diverse participants through global testing initiatives. Research partnerships, drug discovery, and monetization (Priority: 4/5): 23andMe’s data partnerships with pharma companies are framed as a way to accelerate discoveries, but also raise questions about whether customers should be compensated for supplying valuable data. Family discovery and identity effects (Priority: 3/5): The transcript highlights how DNA testing can uncover adoptions, unknown relatives, and paternity surprises, raising questions about identity, family secrets, and what genetic ‘roots’ mean.

Key Arguments: Consumer genomics can empower people by giving them direct access to health and ancestry information without needing a physician or genetic counselor. The Golden State Killer case shows that publicly available genealogical DNA data can be used for major public-safety outcomes, though this also heightens privacy concerns. 23andMe argues its privacy model is based on choice: customers should decide whether and how their data are used. The FDA’s intervention forced 23andMe to validate its reports, which the company presents as evidence of responsibility and product legitimacy. Genetic risk information may help people make better decisions, but evidence that it reliably changes diet or exercise behavior is weak and contested. Polygenic risk scores are presented as the future of medicine because they combine many small genetic effects rather than relying on a single gene. 23andMe acknowledges its database is mostly European and tries to improve representation through targeted global recruitment and free testing. The company sees pharma partnerships as a way to transform customer data into medical research and eventually therapies, even if customers are not directly paid. DNA testing can produce both joyful reunions and painful revelations, so users should be aware that results may affect families in unexpected ways.

Data Points: Golden State Killer crimes: More than a dozen murders and 50 rapes - Described in the opening example of forensic genealogy JEDMatch relatives found: At least 24 relatives - Police cross-referenced the suspect’s DNA with public genealogy data 23andMe health report price: $99 ancestry; another $100 for health information - The transcript describes the company’s consumer pricing 23andMe total customers: About 10 million - Company scale mentioned during the discussion of valuation and data 23andMe health-report customers: 5 million - Used to describe the size of its medical research database FDA warning letter: November 2013 - When 23andMe was told to stop returning health interpretations until approved FDA-authorized conditions: 10 ailments or conditions - In 2017 the FDA cleared selected direct-to-consumer risk reports CDC lifetime diabetes risk: Roughly 40% of U.S. adults - Used to frame the significance of the diabetes report Type 2 diabetes report dataset: More than 2.5 million customers - The report was developed using customer data with consent Non-European customers: About 20% - Wojcicki cites diversity efforts and existing representation 23andMe publications: Over 150 publications - Referenced as evidence of research output Parkinson’s participants: Over 19,000 people - A scale example of research cohorts Alzheimer’s high-risk individuals: Over 1 million - Used to illustrate dataset size Heart disease participants: 800,000 people - Another cohort size mentioned Colorectal cancer participants: Over 10,000 people - Research cohort scale Depression participants: 750,000 people - Research cohort scale GSK investment: $300 million - In exchange for rights to use data for drug discovery Compound pipeline: 13–14 compounds - Wojcicki says they are in research/development stages Study result on behavior change: Over 40% at six months - Wojcicki cites a Harvard-linked study of customers making diet changes Customer participation in study: About 1,000 users - Referenced in the critique of self-reported behavior change research

Pivotal Quotes: "How remarkable is that? That a bunch of civilians just looking to fill out their family trees had inadvertently crowdsourced the capture of a murderer." — Narrator: Opening framing of forensic genealogy and the Golden State Killer case "In our consent form, we specifically thought about like, how do we make sure that we're not set up for crime and you know for like the FBI to come and use this." — Anne Wojcicki: On privacy and anticipating law-enforcement use of consumer DNA data "If I stay healthy, like no one really cares except for me." — Anne Wojcicki: Explaining why she left Wall Street to build a prevention-oriented health company

Implications: Direct-to-consumer DNA is becoming a powerful tool for health, research, and policing, but it raises major questions about consent, equity, identity, and who benefits from genetic data. The future likely brings more discovery, more regulation, and more personal consequences for users and their families.