Case Study 2 — The Grim Sleeper: When the Database Found a Father Through His Son

A real, publicly documented case (Los Angeles, California), used here to teach familial searching — a technique distinct from the investigative genetic genealogy of Case Study 1, and the complementary half of §29.5. The case is widely cited as the first in the United States in which a deliberate familial DNA database search led to an arrest in a major investigation. The crimes were grave; they are described only as clinically as the analysis requires. The defendant was convicted. The technique, not the verdict, is the subject; this study uses it to illustrate the method's power, its specific limits, and the equity questions it raises — not to relitigate the case.

1. Background: a series, a profile, and an empty database

Over a span of years in and around South Los Angeles, a series of killings shared a common signature, and — crucially for this study — biological evidence at the scenes yielded a single crime-scene DNA profile (Chapter 7) that linked the series to one offender. The press eventually attached the name "Grim Sleeper" to the case, a reference to an apparent gap in the series.

The investigators had what the Golden State Killer investigators had: an excellent crime-scene profile and an offender who was not in the criminal database. Repeated CODIS searches for an exact match returned nothing — the same structural wall Case Study 1 described. The offender had no qualifying profile on file, so a database of known offenders was silent on him. For years, that silence held the case.

The eventual break did not come from a consumer genealogy database (that is IGG, Case Study 1). It came from searching the criminal database differently.

2. The method: familial searching, step by step

Familial searching is the deliberate search of a criminal DNA database (CODIS) for a partial match to a crime-scene profile — a profile that is similar to a known offender's but not identical — on the theory that a close biological relative of that known offender (a parent, child, or full sibling) may be the true source of the crime-scene DNA. It uses the same STR profiles and the same criminal database as ordinary CODIS searching; it simply asks a different question. Ordinary searching asks, "Is this exact profile in the database?" Familial searching asks, "Is a profile in the database similar enough to have come from a close relative of our unknown source?"

The logic rests on inheritance. A parent and child share half their DNA; full siblings share, on average, half. So a close relative of the true offender, if present in CODIS, will produce not an exact match but a partial one — sharing many more alleles with the crime-scene profile than an unrelated person would, but not all. A familial search uses specialized software and kinship statistics to flag those near-misses and rank them by how likely each is to represent a true close relative rather than a chance partial overlap.

In the Grim Sleeper case, the steps ran (mapped to the §29.5 framework):

1. A near-match in the criminal database. A familial search of California's offender database flagged a profile that partially matched the crime-scene profile — consistent with a close relative of the unknown offender. The flagged profile belonged to a younger man recently added to the database.

2. From relative to candidate. The partial match pointed investigators not at the flagged individual as the offender (his own profile did not match the crime-scene evidence), but at his close family. The age and relationship indicated by the kinship statistics suggested the true offender was an older male relative — a father. Investigators identified Lonnie David Franklin Jr. as the candidate.

3. Confirm with conventional, direct DNA. As with IGG, the familial search produced a lead, not an identification. Investigators obtained an abandoned DNA sample directly from Franklin (famously, from a discarded item collected during covert surveillance) and ran a conventional STR comparison against the crime-scene profile. That direct comparison — the gold-standard method of Chapter 7 — is what identified him. He was arrested in 2010 and later convicted.

FIGURE (case study) — Familial searching vs. the IGG of Case Study 1     [schematic]

   CRIME-SCENE PROFILE (offender not in CODIS by exact match)
                    │
        search CODIS for a PARTIAL match (kinship software)
                    │
        partial match → a younger relative ALREADY in CODIS
                    │
        kinship stats → true offender is a close relative (a father)
                    │
        identify candidate → CONFIRM with direct abandoned-DNA STR
                    │
                    ▼
        gold-standard STR match makes the courtroom identification

   CONTRAST (Case Study 1, IGG): SNP markers, CONSUMER genealogy database,
   DISTANT cousins, family-tree reconstruction. Same ending: a LEAD that
   conventional STR then confirms.

3. Familial searching vs. investigative genetic genealogy — the distinction that matters

Students and even practitioners routinely conflate these two techniques. The Grim Sleeper and the Golden State Killer are the textbook contrast, and holding them apart is the analytic payoff of pairing them.

Both end at the same honest place: an investigative lead that the validated, gold-standard STR comparison confirms before anything reaches a jury. Neither is offered as the identification itself. But they reach the relative by entirely different routes — one through the criminal database and close kin, the other through consumer genetics and distant kin — and, as the next section shows, they raise different equity questions because they leverage different populations' data.

4. What the technique did — and didn't — establish; and its specific costs

State the strengths and limits precisely.

• Familial searching did convert an exact-match CODIS failure into a productive lead by asking a kinship question of the same database — reaching the offender through his son's presence in the system.
• It did not identify Franklin by itself. The identification was the direct STR comparison between his abandoned-DNA sample and the crime-scene profile. The partial match was the path; the direct match was the proof.
• A partial match is probabilistic and noisy. Kinship software flags candidates by likelihood, and a flagged partial match can be a coincidental overlap rather than a true relative; familial searching produces candidates to investigate, some of which are false leads, not confirmed family relationships. The technique narrows; it does not conclude.

The equity question familial searching raises is specific, and different from IGG's. Familial searching leverages the criminal database — and because of long-documented unevenness in who ends up in that database, its burden falls unequally. The relatives of everyone already in CODIS become, in effect, partially searchable; communities that are over-represented in the offender database by decades of uneven enforcement therefore have more of their families subject to this genetic reach than communities that are under-represented. In the Grim Sleeper case specifically, this dimension drew comment because of the demographics of the affected community. The point for a forensic scientist is not to adjudicate the policy but to see the structure: a technique can be valid and effective and still distribute its intrusions unevenly across groups, for reasons that have nothing to do with the technique's accuracy and everything to do with whose data populates the database it searches.

⚖️ In the Courtroom As with IGG, the genetic-genealogy or familial-search work itself usually does not reach the jury; the jury hears the direct STR match. So courtroom challenges have centered less on the reliability of the confirming science (which is old and validated) than on the propriety of the search — whether a familial search was authorized under the jurisdiction's policy, whether the covert collection of abandoned DNA was proper, and broader privacy questions. Several U.S. states permit familial searching under defined protocols; others restrict or prohibit it; the policies vary precisely because the propriety question (not the validity question) is genuinely contested. A forensic scientist must keep the two apart: the confirming STR match is sound; whether the search that produced the lead was lawful and proportionate is a separate, jurisdiction-dependent matter.

5. The lesson

Three lessons, each central to the chapter.

1. Two routes to the relative, one honest ending. Familial searching (close kin, criminal database, STR markers) and IGG (distant kin, consumer database, SNP markers) are different techniques, and conflating them muddles the science and the ethics. Yet both share the feature that makes them defensible: they generate a lead that the validated, gold-standard STR comparison confirms. The new, contested step points; the old, validated step identifies. That hand-off is the signature of honest forensic progress (compare Case Study 1, and contrast the self-confirming bite mark of Chapter 16).

2. Valid and effective is not the same as costless. Familial searching works, and it distributes its intrusions unevenly across communities for reasons rooted in who is already in the database, not in the technique's accuracy. Seeing that structure — and refusing to let a method's effectiveness silence the question of who bears its costs — is the mature forensic posture this book asks for (§29.6: "who bears the cost, and did they consent?").

3. A partial match is a lead, not a relationship. Kinship software flags candidates probabilistically; some flags are coincidence. Familial searching narrows the field to people worth investigating, and the actual family relationship — like the actual identification — must be confirmed by direct evidence. This is the chapter's, and the book's, first theme once more: the technique excludes the "no lead at all" dead end and narrows, but it does not, by itself, prove either the relationship or the identity. Direct, validated DNA does that.

The Grim Sleeper case shows familial searching at its most powerful — turning a son's database entry into a father's arrest — and, paired with the Golden State Killer, lets you hold the two relative-finding techniques apart with precision. Both are genuine advances; both hand off to the gold standard; both impose costs unevenly. That triple awareness — power, honesty, and cost — is exactly what §29.6 asks you to bring to every method the future will offer.

Discussion questions

1. Build the contrast table from memory: on markers, database, and degree of relatedness, how do familial searching and IGG differ? What is the one thing they have in common at the end, and why does that shared feature make both defensible?

2. The partial CODIS match pointed at Franklin's son, not at Franklin. Explain the inheritance logic that makes a close relative show up as a partial (not exact, not absent) match, and why kinship software is needed to flag it.

3. Familial searching produced a lead; a direct abandoned-DNA STR comparison produced the identification. Why is this division of labor essential to the technique's place on the validity spectrum, and how does it parallel IGG (Case Study 1)?

4. A partial match can be a coincidental overlap rather than a true relative. Using Chapter 9's likelihood-ratio thinking, explain why a flagged familial-search candidate is a hypothesis to investigate, not a confirmed family relationship — and what confirms it.

5. The equity concern with familial searching is rooted in who is already in CODIS. Contrast this with IGG's equity concern (rooted in who uploads to consumer databases, §29.4 and Case Study 1). Why do the two techniques distribute their intrusions across different populations, and why is that a fact a careful court should weigh even when the technique is accurate?

6. Several states permit familial searching under protocols; others prohibit it. Using the distinction between the validity question and the propriety question (§29.5, In the Courtroom), explain why reasonable jurisdictions can reach opposite policies without disagreeing about whether the confirming DNA science is sound.