In 1348, during the panic of the Black Death, a mob attacked the Jewish quarter of Tàrrega in Catalonia. The dead were buried in six mass graves in the Jewish cemetery of Les Roquetes. In 2026 a team led by Santos published the DNA of the people in those graves.
Four of the men have Y-chromosome results. Now apply a rule you will find all over genealogy forums: if your paternal branch has no Jewish matches, your line wasn't Jewish. Run the four Tàrrega men through it, and the rule excludes two of them — men buried in a Jewish cemetery, killed for being Jews.
The rule is wrong. Here is why, and what to use instead.
Two men the rule gets wrong
- ROQ1 carries
J-Y31564. On the public trees, his branch has no modern Jewish matches at all. Its nearest relatives are in Turkey. - ROQ7 carries
E-Y231455. The entire branch holds three testers worldwide — one Ethiopian (Amhara), one Saudi from Makkah, one Belarusian Jew.
And the community they belonged to is about as unambiguously Jewish as ancient DNA gets: modelled at roughly 69% Bronze Age Levantine and 31% medieval Iberian, buried in a Jewish cemetery, dead in a documented anti-Jewish massacre.
So a Jewish man in 1348 can carry a Y-line that, today, shows no Jewish relatives at all. How?
Four reasons the databases under-count Jewish lines
1. Ordinary extinction. Most male lines alive in 1400 have no living male-line descendants today. A man has daughters only, or sons who die young, and the Y-line ends. This is plain arithmetic, and it happens to every population before anyone is persecuted.
2. Targeted destruction. Massacre, expulsion and the Inquisition did not remove families at random. They removed Jewish families, and they removed them disproportionately.
3. Converso descendants don't test as Jews. Five centuries of Catholic life later, the male-line descendants of Iberian conversos are recorded — and identify — as Spanish, Portuguese, Mexican, Colombian, Filipino. A Sephardic Y-line can sit in a database under an entirely non-Jewish label.
4. Who buys tests. The Jewish labels in commercial databases come overwhelmingly from self-identified Ashkenazi testers, because Ashkenazi genealogy testing is popular and well organised. "No Jewish matches" partly measures who has paid for a test.
Proof it isn't theoretical
In the San Luis Valley of Colorado and in northern New Mexico, Hispanic Catholic families with no Jewish identity carry BRCA1 185delAG — the classic Jewish founder mutation for breast and ovarian cancer risk. When researchers looked at the DNA around the mutation, it sat on the Jewish founder haplotype: it had been inherited from a Jewish ancestor, not arisen independently.
No Jewish practice, no Jewish label, nothing in any database. The ancestry was invisible to every method that relied on identity, and visible to a test that looked at the DNA itself.
The honest answer: undetermined
Keep the honesty on both sides. An empty match list is not a negative result — but it is not a positive one either. A branch with three testers tells you almost nothing about who else belongs to it, in either direction.
The correct report is a third state: undetermined — insufficient sampling. Most people with rare branches belong there. A good tool tells you so, and shows you the denominator — "3 testers in this branch worldwide" — instead of letting silence read as "no."
There's a second trap hiding here. An empty match list can also be an empty reading of the tree. Some trees show ethnic and language labels only as hover text, and a quick look at the page misses them. We learned this ourselves: a branch we first reported as having "no Jewish testers" turned out, on a proper check, to contain a Libyan Jewish sample. Before believing any "no," check that you've actually looked.
Where to go instead
If your branch is sparse, the useful question is not "who matches me?" It is: how recently does my branch meet the nearest Jewish branch? If that meeting point is within the last two or three thousand years, the question is live. If it is five thousand years back, the Y chromosome is describing the Bronze Age, not your family. We explain how to read that number in our post on Jewish branches.
And the only thing that turns undetermined into yes is a document: a named person, in a dated record.
Sources
- Santos et al., Genes 17(3):358 (2026) — Tàrrega (Les Roquetes) genomes; Supp. Table S10 (Y results). Data CC-BY.
- YFull YTree — J-Y31564, E-Y231455
- Studies of BRCA1 185delAG in the San Luis Valley and New Mexico Hispanic population (founder-haplotype analysis)
- Velez et al., "The impact of Converso Jews on the genomes of modern Latin Americans"