Rated 4.98-stars across 4.8K+ reviews
Rated 4.98-stars across 3.9K+ reviews Rated 4.98-stars across 3.9K+ reviews Rated 4.98-stars across 3.9K+ reviews Rated 4.98-stars across 3.9K+ reviews Rated 4.98-stars across 3.9K+ reviews

Do DNA Results Differ by Ancestry? Inside Geneus DNA's Ethnicity-Specific Imputation

GeneusDNA profile image By
GeneusDNA
|
Nov 22, 2026
|
52
Did you know
Genetics
DNA results by ancestry
Summary
DNA results by ancestry

Do DNA results differ by ancestry? Most genes like ALDH2, LCT, CYP1A2 and COMT point the same direction worldwide, but allele frequencies and imputation accuracy differ — so Geneus DNA uses a customized Illumina GSA v4 chip and a multi-ethnic reference database for results that are accurate for Thais.

Key Points in 1 Minute

  • Ancestry genuinely affects how DNA results are read — but mostly through allele frequency and imputation accuracy, not because genes work differently in different ethnicities.
  • Geneus DNA tests on a customized Illumina GSA v4 chip (our own patent) and runs imputation on a multi-ethnic reference database (Thai, Asian, Chinese, Japanese, European, African), making results more accurate for Thai and Asian people.
  • Most of the best-studied genes point the same direction worldwide — ALDH2, LCT, CYP1A2, COMT — the same genotype means the same thing whatever your ancestry.
  • In some cases results do differ by ancestry, especially polygenic risk scores (PRS) and variants found mainly in certain populations.
  • Already tested with 23andMe or AncestryDNA? You can upload that raw data and have Geneus DNA analyse it further.

A question we hear often: "If a Thai person and a Westerner test the same DNA, will the results be the same or different?" The answer has two sides. On one hand, the biology of most genes is universal — the same genotype usually points the same way worldwide. On the other hand, ancestry matters more than many expect for the accuracy of how results are read. This article explains why, and how Geneus DNA handles it.

Why Ancestry Changes How DNA Is Read

A MicroArray test reads SNPs at millions of predefined positions, then uses a process called imputation (statistical gap-filling) to estimate the positions it did not read directly, relying on a "reference panel" — a genetic database of many people.

Here is the crux: allele frequencies and the co-inheritance patterns of SNPs (linkage disequilibrium) differ across ancestries. If you use a European-only reference panel to impute data for a Thai person, accuracy drops, because the genetic patterns do not match. This is why having an ancestry-matched reference database matters so much for result quality. Read more about the testing technologies in how MicroArray, WES and WGS differ.

How Geneus DNA Tests and Imputes by Ancestry

It starts at the testing step: Geneus DNA uses an Illumina GSA v4 chip that we customized under our own patent, so it covers genetic positions that matter more for Thai and Asian people. The data then goes through imputation against a multi-ethnic reference database — Thai, Asian, Chinese, Japanese, as well as European and African — making results for people in the Asian region far more accurate. Accuracy (R²) is mostly around 0.95 on average, and imputation can fill in up to roughly 30 million positions to simulate a resolution close to whole-genome sequencing (WGS).

Just as important is the size of the database. Geneus DNA has now collected genetic data from more than 200,000 Asian and Southeast Asian individuals — larger than a national programme such as Genomics Thailand, which set a target of sequencing around 50,000 Thai genomes. The bigger the regional database, the more accurate and fair the results become for Thai people.

Most Well-Studied Genes Point the Same Direction Worldwide: ALDH2, LCT, CYP1A2, COMT

Even though allele frequencies differ by ethnicity, the biology of most genes is universal: if you carry a given genotype, the result usually goes the same way whatever your ancestry. Some heavily studied, globally used examples:

  • ALDH2 (key SNP rs671) — makes an enzyme that breaks down alcohol. People carrying the ALDH2*2 allele have reduced enzyme activity, so acetaldehyde builds up and causes the "alcohol flush" (facial flushing when drinking). The allele is common in East Asians, but anyone who carries it gets the same effect.
  • LCT (rs4988235, in the nearby MCM6 gene) — relates to digesting the milk sugar lactose. The allele for "digesting milk well into adulthood" (lactase persistence) is common in Europeans; most Thai and Asian people lack it and so tend to be lactose intolerant.
  • CYP1A2 (rs762551) — the main enzyme that clears caffeine. "Fast metabolizers" clear coffee quickly, while "slow metabolizers" are more sensitive to caffeine and clear it more slowly.
  • COMT (rs4680, Val158Met) — an enzyme that breaks down dopamine in the brain. The Val form clears it fast (nicknamed "warrior"); the Met form clears it slowly ("worrier"), linked to stress handling, pain and focus.

For this group of genes, interpretation does not depend on your ethnicity — the same genotype means the same thing. What differs is only the "frequency" at which we find that allele in each population.

When Results Really Do Differ by Ancestry

There are cases where ancestry creates a meaningful difference — and this is exactly where an ancestry-matched database helps most:

  • Large differences in allele frequency — some variants are common in Asians but rare in Europeans (e.g. ALDH2*2); referencing a mismatched population can lead to errors.
  • Polygenic risk scores (PRS) — mostly built on European data, so accuracy drops in Asians unless the reference is adjusted. See more in what a polygenic risk score (PRS) is.
  • Population-specific variants — variants important only in certain ancestries can be missed if the dataset does not cover that group well.

Already Have Results From 23andMe or AncestryDNA? Upload and Build on Them

If you have tested your DNA with an overseas service like 23andMe or AncestryDNA, you do not have to start over. You can upload your raw DNA data and let Geneus DNA analyse it further with a multi-ethnic reference database and ancestry-matched imputation, giving you fresh insights from data you already have. (The upload-page link will be added shortly.)

Limitations and What to Keep in Mind

  • Imputation is most accurate for common variants, so it works well for PRS and general traits, but does not replace WES or WGS for rare disease-causing variants.
  • DNA results are supporting information, not a diagnosis, and a statistical association is not the same as a direct cause.
  • Always interpret results with a doctor or genetic specialist to plan your health appropriately.

1. If a Thai person and a foreigner test the same DNA gene, will the results differ?

For most genes the result points the same direction worldwide — the same genotype means the same thing whatever your ancestry. What differs is the allele frequency and imputation accuracy, which is why Geneus DNA uses a multi-ethnic reference database to make results more accurate for Thais.

2. What is imputation, and why does it need to be ancestry-specific?

Imputation statistically fills in genetic data at positions not read directly, using a reference panel from many people. Because allele frequencies and SNP co-inheritance patterns differ across ancestries, an ancestry-matched reference panel gives more accurate results.

3. I already have results from 23andMe or AncestryDNA — can I reuse them?

Yes. You can upload your raw DNA data and have Geneus DNA analyse it further with a multi-ethnic reference database and ancestry-matched imputation, giving you fresh insights from data you already have.

4. How large is Geneus DNA's Thai/Asian database?

Geneus DNA has now collected genetic data from more than 200,000 Asian and Southeast Asian individuals — larger than a national programme like Genomics Thailand, which targeted around 50,000 Thai genomes. The bigger the database, the more accurate the results for Thai people.

References

  1. MedlinePlus (NIH). ALDH2 gene. medlineplus.gov
  2. MedlinePlus (NIH). LCT gene. medlineplus.gov
  3. Cornelis MC, et al. Coffee, CYP1A2 Genotype, and Risk of Myocardial Infarction. JAMA. 2006. pubmed.ncbi.nlm.nih.gov
  4. Tunbridge EM, et al. Catechol-O-Methyltransferase, Cognition, and Psychosis: Val158Met and Beyond. Biological Psychiatry. 2006. pubmed.ncbi.nlm.nih.gov
  5. Martin AR, et al. Clinical use of current polygenic risk scores may exacerbate health disparities. Nature Genetics. 2019. nature.com
  6. Sirugo G, Williams SM, Tishkoff SA. The Missing Diversity in Human Genetic Studies. Cell. 2019. cell.com
  7. National Science and Technology Development Agency (NSTDA). Genomic Medicine in Thailand (Genomics Thailand). nstda.or.th
Written by GeneusDNA
chat whatsapp chat line chat facebook