A complex case involving an abandoned child and a serial killer inspired a new way of solving crimes through cousins’ DNA and family tree data. A similar technique was later used to identify a suspect in the Golden State Killer case and has led to arrests in more than 10 other murder and sexual assault cases in the past five months. Lisa Jensen was abandoned by her abusive father in 1986 when she was five years old. He left behind a fake name and a fingerprint.
Sixteen years later, the man who abandoned Lisa was arrested on charges of killing his girlfriend. When investigators looked into his history, they began to wonder if he was Lisa’s father after all. A DNA test showed he was not related, and it seemed likely he’d kidnapped her as a baby. Lisa’s kidnapper died in prison in 2010, without revealing any clues to Lisa’s past. By then she was in her early 30s, married with kids of her own. She still often wondered: who was she really? One night, Lisa saw a TV program about how DNA testing was helping people learn about their roots. She contacted Peter Headley, a detective who had tried unsuccessfully to find her real identity years before. Maybe this approach would help them? The detective emailed DNAadoption.com, a nonprofit that helps teach adoptees how to find their biological parents. For years, adoptees and the “search angels” who volunteer to help them have used genealogy websites and public records to try to find unknown parents. An adopted child will typically have at least one piece of biographical data, such as a birth date. But in this case, all Lisa had was her DNA. (Lisa declined to be interviewed but consented to have Dr Rae-Venter and Mr Headley recount the story of the investigation. To maximize her chance of finding relatives, Lisa joined four genealogical databases: 23andMe, which required her to spit in a tube; Ancestry.com, which required her to swab her cheek; and FamilyTreeDNA and GEDmatch, which accepted the genetic profiles created by those other sites. At that time, those sites together held roughly three million potential matches. Family relationships can be estimated by the amount of DNA shared by two people:
Great-Great-Grandparents
Great-Great-Aunt or Uncle 3–13%
Great-Grandparents 7–22%
Great-Aunt or Uncle 4–31%
First Cousin twice removed 1–8%
Grandmother or Grandfather 17–34%
Aunt or Uncle 20–32%
First Cousin once removed 2–13%
Second Cousin once removed 0–5%
Mother or Father Amount of shared DNA: 49–55%
SELF (100%)
Sibling 33–50%
First Cousin 8–18%
Second Cousin 1–8%
Third Cousin 0–3%
Niece or Nephew 20–32%
First Cousin once removed 2–13%
Second Cousin once removed 0–5%
Third Cousin once removed 0–3%
Child 49–55%
Great-Niece or Nephew 4–31%
First Cousin twice removed 1–8%
Second Cousin twice removed 0–4%
Third Cousin twice removed 0–2%
Grandchild 17–34%
Had Lisa been very lucky and matched with someone who shared 25 percent of her DNA, that person could have been a grandparent, aunt, uncle, niece or nephew.Instead, Ancestry.com identified a possible second cousin: Paul, an 81-ear-old man.
And 23andMe found two other relatively close cousins, including Adam, 40. To understand how Lisa was related to her three newfound cousins, Dr Rae-Venter had to convince them to upload their DNA profiles to GEDmatch, which has advanced tools for comparing genetic and family tree data. But the cousins were suspicious of the claim that their DNA could help a kidnapped woman find her real identity. Eventually, Dr Rae-Venter and detectives convinced Paul and Adam that it was not a scam, and they agreed to help. (With the exception of Adam, who asked to be identified, family members’ names are pseudonyms used by researchers to protect the family’s privacy.)When Dr Rae-Venter compared Lisa’s DNA sequence with Paul’s DNA, she found they had a surprisingly large match on the X chromosome.
Paul’s
Grandfather
Paul’s
Grandmother
Paul’s
Father
Paul’s
Mother
Paul’s
Maternal Aunt
Paul’s
First Cousin
Paul
Lisa’s
Parent?
Lisa?