Forty minutes before midnight on November 18, 1997, a dark-haired woman in her mid-twenties with a blanket wrapped around her waist walked into a Chevron gas station store on Lake City Way in Seattle, Washington.
The woman went straight to the restroom in the back of the store and locked the door behind her. While she was there, a customer in line for the cashier heard the sound of a baby’s cries coming from the restroom. Fourteen minutes later, the woman emerged from the room and left the store.
Early the next morning, a store clerk entered the restroom and found streaks of blood on the floor, toilet, and sink. The clerk was not alarmed by the blood; this was a public restroom, and someone coming in with a cut or bloody nose was not unusual. Another clerk mopped up the blood at 7:00 a.m., and the day proceeded as usual.
It was not until the following day, November 20, around noon, that a third store clerk, Lakota Murray, emptied the trash in the restroom and made a horrifying discovery.
At the bottom of a clear plastic trash liner, covered by debris, Murray found a newborn baby. The baby’s placenta and umbilical cord were still attached.
Murray called police at 12:40 p.m. Seattle detectives arrived, examined the infant, and determined that the child was dead. An autopsy later revealed that the child weighed seven pounds, seven ounces, and had been born alive.
At first, the death was reported as the result of natural causes. But, quickly, a medical examiner changed that to homicide. The exact phrasing used on the death certificate was “Full term birth, perinatal death, etiology undetermined.” Perinatal meant the death occurred around the time of birth. Etiology referred to cause of death.
The unanswered questions: Who had left the newborn child there to die, and why?
The detectives assigned to the case sifted through evidence: a surveillance tape showing the woman entering and leaving the store; the testimony of the customer who had heard a baby crying; and, most important, a placental blood clot smear found with the infant.
A medical examiner preserved the smear and a laboratory produced an STR DNA profile of the newborn’s mother. But back in 1998, DNA analysis was still a developing discipline, and the STR profile did not lead to any suspects.
Detectives kept at it, but the case went cold.
It stayed cold for the next twenty years.
In that time, the victim’s remains were buried in Seattle’s Calvary Catholic Cemetery, in a plot marked by a plaque etched with the figure of Winnie-the-Pooh and the words “We Care.” The plaque also featured the only name the victim had: Baby Boy Doe.
One of the people who contacted me in the months after Joseph DeAngelo’s arrest was Seattle detective Rolf Norton, who had heard about me from Paul Holes. Rolf filled me in on a homicide case he was handling: Case No. 1997-504314.
The case of Baby Boy Doe.
After two decades with no new clues or leads, detectives had reopened the investigation in 2018, and the Washington State Patrol Crime Laboratory had reextracted DNA from the placental blood clot smear using more advanced technology than was available in 1998. The extract was used to create a new, more complete DNA profile to run through CODIS, the FBI’s criminal database, but there was no match.
Next, the Seattle Police Department sent some of the DNA extract to an Oklahoma City lab called DNA Solutions to develop a SNP profile. Finally, the lab sent the SNP profile to Detective Norton to upload to GEDmatch. After a few days, a list of people who shared DNA with Baby Boy Doe’s mother was posted on the page for Baby Boy Doe’s kit. It was at this point that Detective Norton contacted me and asked for my help. Could these matches somehow lead us to the mother of Baby Boy Doe?
Once again, I said I believed that they could.
The child in this case would be the youngest unidentified homicide victim I had encountered: an infant killed when he was likely just minutes old. An infant who essentially did not exist in any records or even memories—a child never given the chance to live. If I could somehow succeed in identifying his mother, we might be able to finally add this child to his rightful family tree.
I got to work in July 2018. I pulled together a small team of genealogists to help me with the case. The DNA had been isolated from a placental blood clot, which meant that we had the mother’s DNA but not the father’s.
Among our initial jumble of distant relatives to the mother was a woman named Molly. She was a close match at 513 cM, and she was most likely a first cousin once removed to the mother of Baby Boy Doe.
I started a speculative tree for the mother by building a tree for Molly. I identified a close relative in Molly’s tree, Teresa, whom we could test as a reference sample to give us another data point. Because Teresa was such a close relative to Molly, I asked Detective Norton to surreptitiously DNA-test her—a trash pull—so as not to alert her to our investigation, as it was possible that she could be the mother.
Detective Norton’s officers surveilled the woman and picked up some chewing gum she tossed away. DNA was taken from the gum, and the resulting extract was used to produce a SNP profile for upload to GEDmatch. The match results revealed that Teresa was not herself the mother, but she was a match with the unidentified mother at 706 cM, suggesting she was likely a first cousin to the mother as well. It was another solid match, and I now had an MRCA, allowing me to build down all the descendant lines in search of the missing mother.
My next step: entering the mother’s speculative tree into What Are The Odds? (WATO), along with the match information for Molly and Teresa.
The WATO tool gave me the odds of precisely where on the speculative tree the mother likely fit in: in this case, ten-to-one odds that she belonged in one of six possible slots in the tree—slots occupied by six female first cousins.
One of these six cousins was Baby Boy Doe’s mother.
But how to determine which one? All I knew was that three of the women were sisters and the other three cousins were each from different families. To get any further, I had to employ a technique I called the Jonny Perl method.
Jonny Perl, the genealogist who created the wonderful mapping tool DNA Painter, figured out a way to distinguish among cousins when trying to identify an unknown parent. He did this by building back from the spouse of the parent who is a descendant of the MRCA.
I did just that with our matches and came up with the names of the four sets of parents for all six cousins. One parent in each set was a descendant of the MRCA. Now I needed to build back up the ancestral lines of the spouse of the parent who was the MRCA descendant. What I was looking for was a connection to a match with the unidentified mother’s DNA matches. Once I found such a connection, I would know I had the right cousin. I did not know how far back I would have to build, but I knew that it would likely be several generations.
It turned out that the father of one of the cousins, whose mother was a descendant of the MRCA, had the surname Warren. Luckily, one of my team members, Lorrie Burns, remembered coming across the name Warren in the huge jumble of our original matches to the mother on GEDmatch. Now I needed to build back in time to see if these two Warrens were related.
I built back two generations and found no connection.
I went back one more, and still nothing.
But four generations back, in each of the two Warrens’ trees, I found it: the most recent common ancestor between Warren the original DNA match and Warren the spouse whose ancestral line we had built back.
This was it! Finding the MRCA for the two Warrens told me that the daughter of Warren the spouse had to be Baby Boy Doe’s mother.
That daughter’s name was Christine M. Warren.
Several bits of data pointed to her. Physically, she matched the descriptions provided by witnesses. She resembled the woman in the surveillance video. She also would have been twenty-seven years old when Baby Boy Doe was born, matching estimates of her age in 1997. She had lived in Seattle most of her life and lived there still. And she was a first cousin to Teresa. I presented these findings to Detective Norton, who was now prepared to take the next step: surreptitiously retrieving Christine’s DNA and comparing it to their crime-scene DNA.
Detective Norton had to be careful. He did not want to risk confronting Christine Warren before he knew for sure that she was the mother. Instead, he devised a sting.
In November 2020, detectives mailed Christine Warren an invitation to take part in a survey for a fictional flavored-water drink called Sparkling Icy. The invitation included a complimentary gift card. Christine Warren took the bait, filled out a survey, and mailed it to a P.O. box set up by the Seattle PD. First, detectives compared the signature on the survey to the signature in Christine Warren’s driver’s license records, to make sure they had the correct Christine Warren.
The signatures matched.
Next, in December 2020, they sent the envelope in which the survey was mailed to the Washington State Patrol crime lab. The lab extracted DNA from the envelope seal and the stamp, both of which Warren had licked. An autosomal STR DNA profile, or a short tandem repeat DNA profile (see Glossary, page 247), was created from the extract. This new STR DNA profile was compared to the STR DNA profile created from the placental blood clot smear gathered in 1997. The results came in on December 31, 2020—New Year’s Eve.
The two STR DNA profiles were a match.
Beyond any doubt, Christine Warren was Baby Boy Doe’s biological mother.
Detective Norton tied up all the incriminating details before, in March 2021, interviewing the suspect. In the interview, Christine Warren admitted to everything. By her telling, the baby’s father was agitated by the news that she was unexpectedly pregnant, and for that reason she could not bear the thought of having the child. Instead, she simply “blocked the pregnancy out of her mind,” investigators said in a 2021 charging document. “She ignored the fact that she was pregnant.”
For the next several months, she received no medical attention, nor did she mention the pregnancy to a single other soul.
Finally, late on November 18, 1997, Warren was riding with a friend in a car when she felt a wave of contractions. She asked the friend to pull over at the Chevron store, covered her stomach with a blanket and walked into the restroom. Calmly, she walked out again fourteen minutes later and went on with her night. She told detectives “the baby dropped into the toilet” and remained there for several minutes. Then she panicked and put the baby in the trashcan and covered him with paper and other debris. “In her mind,” detectives explained, “she did not believe the baby was alive.” But under questioning she admitted that she never checked and did not know for sure if her child was living.
Police arrested Warren, then fifty years old, set her bail at ten thousand dollars, and later charged her with murder in the second degree—a crime that could send her to prison for as long as eighteen years (as far as I can tell, she has not yet been convicted as of this writing).
My team and I had cracked the mystery of Baby Boy Doe. Still, this was not a resolution anyone wished to celebrate. The details of the case were just too sad. Detective Norton shifted attention to all the dedicated investigators who had worked on the case over the last twenty years, men and women who “created a foundation of evidence that we were able to parlay into a conclusion,” as he explained to Seattle’s KIRO 7 News. “This was an incredibly sad case in 1997, and it’s an incredibly sad case in 2021.”
He was right. The death of Baby Boy Doe was senseless and tragic. Even so, identifying his mother and filling out the details of the crime were meaningful. Twenty-three years after his death, the boy was no longer an unidentified phantom, no longer erased from his family tree. He existed, and his short life mattered, and he had not been forgotten.
And when it was finally possible, good people gathered to ensure that justice was done in his honor.
