Most DNA ancestry tools run your genome against the same worldwide panel regardless of who you are: a Han Chinese customer, a Nigerian customer and a Norwegian customer all get sorted through the same generic buckets, calibrated for nobody in particular. A G25 Focus report does the opposite. It starts from the specific ancient and modern populations that plausibly shaped your heritage, and only those, then asks a sharper question: not "what is humanity made of," but "what are you made of, layer by layer, back to the Neolithic." This article explains what that actually means, what the model can and cannot tell you, and why the same honesty checks that catch a genuine outside signal in a customer who was sure they had none are the reason these reports are worth trusting.
One report per heritage, not one report for everyone
There are currently fourteen Focus reports, each built around a specific regional or diasporic heritage rather than a generic worldwide split. Grouped by region:
Europe: Thalassa (Mediterranean islands: Sicilian, Greek, Maltese, Corsican, Sardinian), France Profonde (France's historic regions, Brittany to Alsace), Nordsee (North Sea Germanic: English, Dutch, Frisian, Danish, Norwegian), Balkan Frontier (Serbian, Croatian, Bosnian, Macedonian, Bulgarian, Albanian), Drom (Roma and Romani, traced to the Swat Valley founder homeland).
Asia: Ararat (Levant and Caucasus: Maronite, Druze, Armenian, Georgian, Circassian, Assyrian), Simorgh (Iranian plateau), Sindhu (South Asia), Fenghuang (Sinosphere: Han North/South/Fujian, Korean, Japanese, Vietnamese, Manchu, Zhuang).
The Americas: Raíces (Latin American bloodline, Indigenous/European/African by region) and El Gringo (North American, Colonial/Anglo/Acadian and more).
Africa and the wider diaspora: Sankofa (African ancestry and diaspora), Am Yisrael (Jewish diaspora: Ashkenazi, Sephardi, Mizrahi and more).
Oceania and the Pacific: Lapita (Austronesian and Pacific: Filipino, Indonesian, Malay, Taiwanese Aborigine, Malagasy, Polynesian, Micronesian).
Every one of them is built the same way, and it is worth being specific about what "built" means here, because it is the whole point of the collection.
What a Focus report actually measures
Under the hood, every report runs the same core model: your Global25 coordinates, scored against a reference panel using non-negative least squares, the same method behind every article on this site. What changes from report to report is the panel itself. Ararat's panel is built from Levantine and Caucasus populations, ancient and modern. Fenghuang's is built from Yellow and Yangtze River Neolithic farmers through the Han, Tang and Manchu/Qing periods. Nobody gets modelled against a panel built for somebody else's history.
That specificity is what lets a Focus report do something a generic worldwide calculator cannot: break your ancestry down period by period, inside your own regional story, rather than as one flat percentage. A typical report walks through:
- Deep ancestral sources: the Neolithic-era streams that fed into your region in the first place, often eight to ten thousand years back.
- Bronze Age, Iron Age, Classical and Medieval layers, each modelled independently, against the specific populations that occupied your region in that era. A Fenghuang customer sees Erlitou and Shang Dynasty Bronze Age ancestry, then Zhou Dynasty Iron Age, then Han Dynasty and the Yayoi migration into Japan, then the Tang Dynasty and the Manchu/Qing conquest era, as four separate, independently fitted models rather than one blended average.
- A regional core versus outside admixture split: how much of you traces to the ancient peoples of your region at all, before anything else is broken down.
- Your closest of the report's modern communities: which specific community (Han South Cantonese, Croatian, Sardinian, whichever the report covers) your DNA resembles most, modelled directly against those modern reference populations rather than inferred from the deep layers.
- Y-DNA and mtDNA haplogroups, when you upload a raw file rather than pasting G25 coordinates.
Every layer reports its own fit quality, not just a percentage. A period that fits your DNA well is flagged GREAT or GOOD; a period that does not is flagged FAIR or DISTANT, with an explanation rather than a silently unreliable number. That is a deliberate design choice, and it is one very few consumer tools bother to make: it would be easy to always report a clean, confident-looking percentage for every era. It would also be dishonest, because some eras are genuinely better resolved by ancient DNA sampling than others, and a report that hides that is a report you cannot trust when it tells you something is real.
The honesty check: what happens when you are not who the model expects
Here is the part that actually differentiates a Focus report from a report built around a single "your ethnicity is X per cent this" number.
Every Focus report's reference panel includes, alongside the regional ancient and modern sources, a set of deliberately unrelated outside proxies: European, South Asian, and specifically Amerindian reference populations, chosen because they sit far outside the region the report is built for. Their job is not to explain your ancestry. Their job is to catch it when your ancestry genuinely does not belong to the report's region, and to say so plainly instead of quietly forcing your coordinates onto the nearest regional population regardless of fit.
This matters because the alternative failure mode is common and almost invisible to the customer. A calculator with no outside proxies has no way to say "this doesn't fit." It will always return its best regional guess, confidently, whether or not that guess means anything. A customer with a real, unexpected great-great-grandparent from outside the region, one detectable Amerindian, East Asian or African contribution, for instance, would simply have that signal absorbed into the nearest regional bucket and reported as if it were local. The result looks clean. It is also wrong.
Our Focus reports are built the other way round. If your DNA carries a real trace of ancestry the AmProxy or other outside sources can explain better than any population in the regional panel, the model assigns it there instead, and it shows up in your results as exactly what it is: outside admixture, not a distorted regional percentage. We have seen this happen in practice, a customer whose report for one Mediterranean island community surfaced a small but real steppe-and-Central-Asian signal nobody expected from the family's own history, confirmed independently across two separately modelled eras rather than appearing once and vanishing. That is the model working as designed: not manufacturing a surprise, but refusing to paper over a real one.
The same honesty extends to when a report is simply the wrong fit altogether. If your overall coordinates do not resemble the region a report is built for, at least not closely enough to trust, we say so before you pay rather than after, and point you to the Advanced Ancestry Report for the broader worldwide picture instead. A regional report that always says yes is not being generous. It is being unreliable.
Going back through time inside your own history
The other thing a Focus report gives you that a single ethnicity percentage cannot is a sense of sequence. Ancestry is not a snapshot; it is a series of layers laid down over thousands of years, and most of those layers have names, dates and archaeological contexts, even if a generic worldwide test has no reason to show them to you.
A Thalassa customer does not just learn they carry Sicilian ancestry. They see how much of that traces to the region's Bronze Age populations, how much to Classical-era Greek and Roman-period movement, how much to Medieval-era arrivals, each period modelled against the ancient samples that actually represent it, each with its own fit quality attached. A Simorgh customer walks the same path through the Iranian plateau's own Bronze Age, Achaemenid and Sassanid-era, and Medieval layers. The report reads less like a pie chart and more like a timeline, because that is what ancestry actually is.
What you get, concretely
- 15€ per report, instant delivery after payment.
- Upload your raw DNA file (23andMe, AncestryDNA, MyHeritage, FTDNA and others) or paste your own G25 coordinates, whichever you already have.
- 100% real, official Global25 coordinates from Davidski's published datasheets. None of our reference data is simulated or synthetic.
- An AI-written narrative tying the deep, era-by-era and community results together into a readable story, not just a table of numbers.
- Several reports available in multiple languages matching their community: Fenghuang in seven, Lapita in five, and so on.
If you already know roughly where your family comes from, there is a good chance one of these fourteen reports was built specifically for that heritage, ancient regional core through to the present, with the same honesty checks running underneath every one of them. If you are not sure yet, the Store's search-by-origin tool will point you at the right one in a few seconds.