G-PF3148 is a branch of the FamilyTreeDNA Y tree with 328 kits that gave a country, counting all sub-branches. Stands out most: Armenia (1.4%), Morocco (1.4%), Georgia (1.1%). The branch's common ancestor lived about 13k years ago: the FTDNA Discover estimate runs from 11k to 15k years ago. YFull has 128 samples in this branch and dates its common ancestor to 13k years ago.
Countries and regions given by the branch's carriers. “Share of branch” is the part of the branch's kits that belongs to the country. “Share of country” is the part of that country's kits that falls into the branch; it is computed only for countries with at least 50 kits.
−10,000−6000−4000−2000CE10001500
PalaeolithicMesolithicNeolithicChalcolithicBronze AgeIron AgeMiddle AgesModern eraMinus means BCE; epoch boundaries are conventional.
Colour is the branch's share of the country's kits. Hover over a country or tap it.
Where the ancestors of FamilyTreeDNA geographic project members on this branch lived, from the project maps as of 2026-10-20. Colour is the branch's share of the members at a place, or the number of its carriers; places with fewer than three are left blank. The centre and the 68% circle are corrected for some countries testing more than others. This is a picture of the sample: members chose whether to join and which place to give.
97
People on the maps
at 2 places of 1°; 91 more where there are fewer than three carriers
1,920 km
Radius of 68% of carriers
around the branch centre 42.3° N, 23.6° E
Ancient samples of the branch and its subclades with a date and a find site, from FamilyTreeDNA Discover, 184 in all. The slider picks the years; the older the era, the wider the window.
How far from the branch's present-day centre its ancient samples were found. The centre and the 68% circle are the same as on the project map above. Find sites are taken from the Mitotree in FTDNA Discover.
Ancient samples with a known site: 184. 168 lie inside the 68% circle; half of all are within 860 km of the centre. The centre of the ancient finds (Croatia) lies 670 km from today's.
Horizontal: years ago; vertical: km from today's branch centre. The dashed line is the edge of the 68% circle.
Farthest from the centre
Aligrama, Swat Valley, Pakistan, 760 BCE ? G-FGC58137 — 4,260 km
Cabeço da Arruda I, Torres Vedras, Lisboa, Portugal, 3,100 BCE ? G-BY201413 — 2,730 km
Cueva de los Murciélag, Córdoba, Andalusia, Spain, 5,177 BCE G-BY200976 — 2,430 km
Today's carriers are geographic project members: they chose whether to join and which place to give. Ancient samples are the ones studies have covered, and some regions are better excavated than others. The chart puts samples of different eras side by side.
How the branch's ancestral line moved, as modelled by FamilyTreeDNA Discover: ancestor locations are estimated from present-day carriers, the path between them is the least-cost route over terrain. Dates are node TMRCAs, good to centuries. The path ends at a fork of child branches: click one and the path continues into it.
Left to right runs from the distant past to today, in years ago; the past is packed tighter. Upward are the kilometres the line covered along Discover's route. A dot is an ancestor of the branch, the shaded band is the fastest stretch, the dashed line is G-M201, the root of the letter. Steep means the line was moving, flat means it stood still.
The line moved fastest 27,900–26,200 years ago: 640 km, about 380 km per millennium. From the root of the letter on, the line moved slowest 26,200–22,200 years ago: 200 km in 4,000 years. In all, the line covered 14,000 km along Discover's route. Joining the ancestors with straight lines gives 10,800 km; a straight line from the first point to the branch gives 3,900 km.
Show the mapHide the mapThe map unfolds by itself as you scroll down to it.
From YFull samples that gave a mother tongue: the part of a people's samples that falls into this branch. The list has peoples with at least 30 samples, 3 or more of them in this branch.
Sub-branches that have their own page (50 kits or more), as a tree: every direct one, nested ones from 1% of the branch's kits. The bar is the kits in the sub-branch with its own sub-branches; the age is when its common ancestor lived.
Each of these sub-branches has fewer than 50 kits with a country, so none has its own page. Below are all with 5 or more kits, with their age and top countries.
Each source's estimate on one scale, in years ago. A dot is the estimate, a bar its spread (a 95% interval for YFull); no bar means the source publishes no spread. The “common ancestor” rows show when the last common ancestor of all carriers lived, the “split” rows when the branch broke off from its parent. Discover does not estimate the split, so its row shows the age of the parent branch itself.
FTDNA Discover common ancestor12.5k years (10.8–14.4)
YFull common ancestor12.2k years (11.2–13.2) · 128 samples
TheYtree common ancestor12.0k years · 82 samples
FTDNA Discover split from the parent13.7k years (11.9–15.8)
YFull split from the parent13.0k years (12.1–14.0) · 128 samples
12k14k16k
Thickness against the parent: by ancient samples ×1.1, by living carriers ×0.90: seen in the past, thinned out. All such splits on a map
How many sub-branches already existed at each moment: a sub-branch counts from the time its common ancestor lived (FamilyTreeDNA Discover estimates). Sub-branches with a known age: 197.
explosive growth: the number of sub-branches doubles in under 400 yearsgrowth: the number doubles in 400–1,500 yearsstagnation: no doubling within 1,500 yearsgrey: fewer than 5 sub-branches at the start of a stretch — too little to tell the rate
Across: years ago; up: number of sub-branches, log scale. The shaded band marks the period when the count multiplied most; the last thousand years are left out.
The branch grew fastest 9,000–7,000 years ago: its sub-branches went from 25 to 51.
Sub-branches of the last centuries show up only where many relatives have tested, so the right end of the curve says more about who took a test than about population growth.
FTDNA: how much the kit count grew between the tree copies of 2026-10-09 and 2026-11-13. A kit that moved to a new sub-branch after refinement is not counted as new. YFull: samples YFull itself marks as new, in the copy of 2026-11-02; YFull drops the mark after a few weeks.
Surnames that testers entered at FTDNA, over the whole branch with its sub-branches: 16 distinct among 37 kits. FTDNA gives surnames per branch only, with no country attached, so they cannot be split by country.
186 samples from excavations belong to this branch or its subbranches. The data come from FamilyTreeDNA Discover and the Allen Ancient DNA Resource, the database of the David Reich lab. AADR gives the links to the papers, some of the places and the samples Discover lacks. “?” means Discover is unsure of the sample's place on the tree. 93 of 186 samples are marked.
A dot is a sample, a line is the range it is dated to, the colour is its culture. A hollow dot: its placement on the tree is uncertain. The dashed line is the branch's common ancestor as FTDNA Discover estimates it.
Years ago. The scale tightens towards the past: the further back, the closer the centuries.
Linear Pottery cultureHorgen cultureNeolithic AnatolianNeolithic BalkanChalcolithic IberianNeolithic Iberianotherplacement on the tree is uncertain
People of the past whom FamilyTreeDNA Discover places on this branch or its sub-branches, by their remains or by their male-line descendants. Only people who lived before 1850.
Ötzi the Iceman, ≈ 3250 BCE G-L166
Data on this page: FamilyTreeDNA Y tree, 13 November 2026: 328 kits; YFull Y tree, 2 November 2026: 128 samples.
How to cite: ftdnastat, "Haplogroup G-PF3148: age, countries, sub-branches", https://ftdnastat.github.io/en/clade/g-pf3148/, data of 2026-11-17, build 20261117-b6de9f1.
The data is licensed under CC BY 4.0: take it, rework it, just credit ftdnastat.