What Investigative Genetic Genealogy Is and How It Differs
To understand what happened in 2018, it is necessary to understand something about what investigative genetic genealogy is and how it differs from the criminal DNA database work that investigators had been doing for years.
CODIS and similar law enforcement DNA databases work through direct matching. A crime-scene DNA profile is compared against a database of profiles from known individuals. If the crime-scene profile matches a profile in the database exactly — or closely enough to meet the statistical thresholds that govern match declarations — the result identifies a specific person. The match is direct: crime scene to person.
The limitation of this system, for cases like the Golden State Killer investigation, is that it can only match against people who are in the database. If the person responsible has never had their DNA entered into CODIS — through a prior conviction, arrest, or sample collection — the profile sits as an unidentified entry, waiting for a match that may never come.
Consumer genetic genealogy databases work differently. Services like 23andMe, AncestryDNA, MyHeritage, and the user-uploaded platform GEDmatch contain DNA profiles from millions of people who submitted their own saliva samples or uploaded raw data files from commercial tests, primarily to learn about their ancestry and find relatives. These profiles represent people who have not committed crimes and are not under investigation. They are in the database because they chose to participate in consumer genealogy.
The relationship between a crime-scene DNA profile and profiles in a consumer genealogy database is probabilistic and relational rather than direct. A crime-scene profile compared against a genealogy database will not typically find the exact person responsible. It will find people who share portions of their DNA with that person — relatives, by various degrees of relationship. The closer the relationship, the larger the shared segment of DNA. A first-degree relative (parent, child, sibling) shares approximately fifty percent of their DNA. A first cousin shares approximately twelve to thirteen percent. A second cousin shares approximately three percent. Distant cousins share smaller and smaller segments.