ftdnastat Русский

Findings

What stands out in the data. Recomputed automatically with every update. Remember: these are FTDNA customers, not whole populations. How it is computed is explained under About the data.

Most distinctive branches #

Branches that occur many times more often in one region than worldwide. ×10 means ten times the world share. Branches of at least 20 kits are included.

Young but widespread branches #

The branch's common ancestor lived less than 2,000 years ago, yet the branch already makes up 1% or more in 5 or more countries (each with 100+ kits). This is what a recent expansion, a diaspora, or mistakes in the reported ancestral country look like.

Ancient branches that stayed home #

The branch's common ancestor lived more than 4,000 years ago, yet at least 50% of its kits still come from one country (branches with 200+ kits worldwide). These are lineages that never left home in thousands of years.

Countries whose male and female lines point different ways #

Countries whose five most similar neighbours by the paternal line and by the maternal line barely coincide. The male and female lines have different histories here.

Who tests the father and who tests the mother #

Maternal-line tests per one paternal-line test. 1 means equal numbers, 0.1 means almost only the male line is tested. Countries with 150+ kits in each tree.

Largest pile-ups on a single branch #

Cases where a single finest branch gathers a large share of a region's kits. Usually one surname or clan project.

YFull

Branch ages: YFull against FTDNA Discover #

Branches with the same name in both trees: 19,277. For each, the age of the common ancestor (TMRCA) estimated by YFull is compared with FTDNA's estimate. On average YFull's estimate is 7% younger; YFull gives a younger age for 67% of branches. For 98% of branches the two estimates stay within each other's margin of error. The date a branch split from its parent (YFull's formed) is not compared here.

Where the estimates disagree

Branches with 100+ YFull samples where the two estimates differ by more than their margins of error allow (the 95% ranges do not overlap).

YFull

Paternal and maternal lines: are they linked, and how much of it is just geography #

Both lines are known for one person, and they are not independent: knowing the paternal line lets you guess something about the maternal one. The link was measured over 10,473 YFull pairs with a known country (40 countries with 80+ pairs). Knowing only a person's country explains 27% of the whole link between their lines. The region explains even more, 51% (2,259 pairs in 24 regions with 50+ pairs). The finer the geography, the more of the link turns out to be simply “people from the same place”. What remains is structure inside regions, not kinship of the pairs themselves.

How the link vanishes inside a country

YFull

Fathers tell the people apart, mothers do not #

Do paternal and maternal lines paint the same picture of a people? For each people 40 members were drawn at random and the number of distinct major branches among them counted, for Y and for mt separately. Percentages are how much that is of the whole world's diversity in FTDNA data. Of 14 peoples the most uniform by mothers keeps 92% of the world's diversity, the most uniform by fathers only 35%: by mothers, peoples can hardly be told apart.

Most homogeneous by fathers

Most diverse by fathers

YFull

Countries: how far paternal and maternal lines stray from the world mix #

How far a country's mix of major branches is from the world mix, separately for the paternal and the maternal line. The number is the difference in percent: 0% is the world mix, 100% is nothing in common. FTDNA kits, 49 countries with 150+ kits on each line. In 44 of 49 countries the paternal line is the more distinctive one.

The rare exceptions: the maternal line is the more distinctive one

YFull

How detailed the tests are: how deep into the tree the samples reach #

YFull samples are placed on the FTDNA tree, so depth can be compared directly. In the typical one of 66 countries a YFull sample sits 37.1 branchings from the root, an FTDNA kit 31.5. 35% of YFull samples reached the finest branch, against 23% of FTDNA kits. That is why the same country looks different at the two companies: YFull's whole-genome tests go deeper. On the maternal line there is no gap: in the typical one of 27 countries the depth differs by only −0.1 branchings, against +4.6 for Y-DNA. Both companies read the whole mitochondrial DNA.

Largest gap

Nearly equal

YFull

Where the two companies show a different picture of a country #

For each country the major-branch shares at FTDNA and YFull were compared (starred branches left out). Part of the difference is just the chance of small samples: it was estimated by comparing two random samples of the same size 120 times, then subtracted. The number in a row is the difference beyond chance, in percent; hover to see the total difference and the chance level. Then the branch that contributes most (FTDNA share → YFull share). 29 countries with 150+ samples at both companies, typical difference beyond chance 2.4%.

Disagree most

Agree

The same for mtDNA

The same for the maternal line. Under-typed samples with a bare letter (H, U, J with no digit) are left out; branches under 0.5% at both companies are lumped into “other”. 16 countries.

YFull

Blind spots: a branch prominent at one company and nearly absent at the other #

The branch makes up at least 4% at one company, and the gap between the companies is too large to be chance: at least 4 times the expected random scatter. Usually the mark of one large project or of the different routes by which customers reach each company.

The same for mtDNA

YFull

Branch = people #

Branches of the YFull tree where almost every sample with a reported native language is from one people. Conditions: at least 80% of those who gave a language, with at least 10 such samples, and a language given for at least 25% of the branch's samples. The top-most such branches are shown. Only one customer in four reports a language, so the list reflects project activity as much as genetics.

YFull

Diversity of paternal lines by people #

How varied a people's paternal lines are. So that large peoples do not win by size alone, 40 samples were drawn at random from each and the number of distinct major branches among them counted. The number in a row is that count per 40 people: 1 means everyone is in one branch, higher is more varied.

Most uniform

Most diverse