Why Does My DNA Test Show Persian or Iranian Ancestry When My Family Is Arab From Iraq or the Gulf
A Persian or Iranian share on an Arab DNA test is common. It comes from shared deep ancestry, recent contact and how a test labels its regions, not from a hidden Persian grandparent.
A Persian, Iranian or Anatolia/Iran share on an Arab person's DNA test is normal and expected. Iraq, the Gulf and Iran sit on one continuous stretch of West Asian population history, and a test can only use the region names its reference panel has, so shared ancestry gets the nearest name on the menu. It is not proof of a Persian grandparent, and it says nothing about your nationality or how Arab you are. A Persian share is also not evidence of a Shia, Ajam or non-Arab origin: Sunni and Shia Iraqis both carry this ancestry.
A test matches you to reference groups, not nations
A DNA test does not detect nationality or ethnicity. It reads several hundred thousand positions in your genome for a chip file, or millions for a whole genome, compares them with reference people whose ancestry was sampled in particular places, and finds the mix of those references that fits you best. The label on each slice is the name of the reference group, not a fact about your family.
Neighbouring populations are not separated by clean genetic borders. They form a gradient, and people on either side of a line on the map share many of the same variants at similar frequencies. A person from Iraq or the Gulf can therefore fit partly to an Iranian reference group, and the software then prints the name of that group. Arab identity is built from language, history, family and culture. A percentage from a reference panel measures something else.
| Cause | What it looks like | How to tell |
|---|---|---|
| Deep shared ancestry | A steady Iran-related share, often larger in Iraqi results | Similar share in both parents and across companies; no close Iranian matches in relative-matching tools. |
| Recent contact (Huwala, Ajam, Iranian-coast families) | A larger share, sometimes with a known family story | Family history, plus close Iranian matches in relative-matching tools (IBD) at the big companies or GEDmatch |
| Thin or coarse reference panel | A Persian label or a merged Middle East bucket that changes between tests | Compare panels, look at the regions each can name and the closest reference populations |
Scroll the table sideways for the rest →
Each cause can produce the same label, so the label alone does not say which one applies.
Zagros ancestry is older than any border
Ancient DNA work on early farmers has shown that people from the Zagros mountains, in what is now western Iran, carried a distinct ancestry. Lazaridis et al. 2016 (Nature) found Zagros farmer ancestry, and Iran-related ancestry more broadly, reaching the Bronze Age Levant. Ancient DNA from Mesopotamia itself is thin, so much of what is said about Iraq is inferred from Zagros, Anatolian and Levantine samples plus a few later Mesopotamian genomes. Even so, it makes this ancestry part of the deep background for many people across the region, Arab and non-Arab alike.
How much of it a person carries differs by region and by family. As a general observation from published PCA and admixture studies of Middle Eastern populations, the size of the Iran/Caucasus-related component varies between Iraq, the Arabian interior and the coastal Gulf, and between individual families. Sampling and methods differ between studies, so this is a broad pattern, not a rule for your family. A large share can still be ancestral rather than recent. Ancient groups are also used as stand-ins, modelled proxies for something that was itself a mixture, not named ancestors. Exactly how much of the modern Arab signal traces to which ancient movement is still argued over.
Iraq and the Gulf have mixed with Iran for millennia
Iraq shares a long border with Iran, and Persian empires ruled Mesopotamia for centuries before Islam. Pilgrimage, trade, marriage and migration have gone both ways ever since. Across the Gulf, people have crossed the water in both directions for centuries.
Some recent movements are well known. The Huwala are Arab families who lived on the Iranian coast and later returned to the Arabian side. Ajam communities, families with roots in Iran, live in Bahrain, Kuwait and the UAE. If your family belongs to one of these histories, it may explain a larger Iranian-like share than the deep background alone would give.
This is a population-level effect, and the percentage alone cannot tell you whether your own family has a recent Iranian line. Admixture timing comes from the length of ancestry tracts in your own genome (local ancestry): old mixing has been cut into short tracts by many generations of recombination, while recent mixing leaves longer ones. That is different from matching segments with relatives. Relative matching (IBD) at the big testing companies, or on GEDmatch, can show a recent Iranian line through matches with Iranian relatives. HealthOSX /dna has neither relative matching nor admixture dating, so it cannot tell recent from old mixing.
The label depends on which regions the test can name
Two people with identical DNA can get different words depending on the test. What changes is the reference panel and the list of regions it is allowed to output.
- A company with few Middle Eastern references may have one Middle East bucket, or none that fits Iraqis and Gulf Arabs, and your DNA falls on the nearest thing it has.
- A panel that has Iranian or Persian samples but few Iraqi or Gulf Arab ones may give that name to anything that resembles them.
- Some merge Anatolia and Iran into one region, and others split Arabia, the Levant and the Caucasus apart.
- Companies revise their panels, so the same file can give a different answer next year.
| Panel design | Regions it can name | What an Iraqi or Gulf Arab result may say |
|---|---|---|
| Coarse, few Middle Eastern references | One Middle East or West Asia bucket | Middle East, or Persian if only Iranian samples exist |
| Anatolia and Iran merged | Levant, Arabia, Anatolia/Iran | A share under the merged Anatolia/Iran name |
| Dense Middle East panel | Levant, Arabia, Anatolia/Iran, Caucasus, North Africa separate | Arabia, Levant and Anatolia/Iran shown separately; the Anatolia/Iran share can be small or large depending on where the family is from |
Scroll the table sideways for the rest →
Illustrative of how panel design changes the wording, not the output of any named company.
A percentage is not a count of ancestors. It is a statistical fit against references, and it comes with a range. A share near 25% fits a grandparent and a share near 12% fits a great-grandparent, but deep shared ancestry can produce the same numbers, so the percentage alone cannot decide.
What the result does not mean
- It does not prove a Persian parent or grandparent.
- It is not a statement about nationality, citizenship or sect.
- It is not evidence against your Arab identity.
What to do next
- Test, or upload, a parent's file and compare their shares with yours. A similar share in both parents points to older, shared ancestry.
- Ask older relatives about Huwala or Ajam roots, or family links to the Iranian coast.
- If you have a close-relative match list at a testing company, check whether it contains Iranian surnames.
- Read the ± range before acting on any small share. A small share with a wide range may be noise.
If two tests give different answers
They will, and that is expected, because the panels differ. The useful comparison is not the headline number but a few specific things:
- Which regions each test can name, and whether Arabia, the Levant and Anatolia/Iran are separate or merged.
- Whether the closest reference populations are listed, since those say more than a regional percentage.
- Whether results from different companies agree on the broad picture, even where the labels differ.

Choosing a file. The report is read on your own device.

Ancestry across 15 regions and 157 reference populations.
If you want a second reading with a panel built for this part of the world, the free reader at healthosx.com/dna, also available in Arabic at healthosx.com/ar/dna, takes a raw 23andMe, AncestryDNA, MyHeritage or whole-genome VCF file. Its ancestry runs only with your consent and uses the AADR Human Origins panel, where Anatolia/Iran, Arabia, the Levant, the Caucasus and North Africa are separate regions, each share shown with its likely range. It also lists your closest reference populations, meaning groups ranked by genetic distance among 298 present-day populations, which include Kurdish, Lebanese and Iraqi Jewish groups. The list shows whether you sit nearest Kurdish, Iranian-Jewish, Iraqi-Jewish, Lebanese or other groups, without naming any single historical population. The report also shows your closest ancient groups, such as Neolithic Anatolian farmers, worded as similarity rather than descent.
New to raw DNA files? Start with the full guide: 23andMe Raw Data Interpretation Free in Your Browser
Written by
HealthOSX Research