HLA-DQ2 and HLA-DQ8 explained: the allele names on a celiac gene report

Short answer

Celiac disease almost always occurs in people who carry one of two groups of gene variants, DQ2 and DQ81. Coeliac UK reports that 99.6% of patients carry DQ2.5, DQ8 or DQ2.22, yet the NIDDK says about 30 percent of people have DQ2 or DQ8 and only about 3 percent of them develop celiac disease3. A positive result therefore does not confirm the diagnosis4.

Published by Gluten Verdict · updated · 11 sources, each linked below

Reviewed by Mark Peter, Registered dietitian (RD), on

Desk with a stapled report seen edge-on, a pen and a mug

Key points

  • DQ2.5 is the type most tied to celiac disease: Coeliac UK lists DQ2.5 as highest risk, DQ8 as moderate and DQ2.2 as low, but not no, risk5.
  • A report that says only “DQ2” is ambiguous, because two different molecules have been called DQ2: DQ2.5 and DQ2.26, 7.
  • In one US study of 10,191 people at risk, 28.28% of those with two copies of DQ2.5, or DQ2.2 plus DQ2.5, were positive for anti-endomysial antibodies, against 9.09% of those with one copy8, 9.
  • Carrying DQ2 or DQ8 is common: the NIDDK puts it at about 30 percent of people3, and Coeliac UK at up to 40% of the general population10.
  • Coeliac UK calls patients who carry none of DQ2.5, DQ8 or DQ2.2 “very rare”42, and ESPGHAN says earlier reports of up to 5% DQ2/DQ8-negative patients had several causes, one being the testing method11.

What the names on the report mean

HLA is a gene system on the short arm of chromosome 6, where Coeliac UK says the so-called disease association genes, or permissibility genes, are situated12. A review from a Sardinian research group says the HLA genes involved in celiac disease are located at HLA-DQA1 and HLA-DQB113. Each codes for one chain of a two-chain protein: the review describes DQA1 05:01 and DQB1 02:01, the two alleles most often linked to celiac disease, as encoding the alpha and beta chains of the protein called DQ214, which is DQ2.5 in the finer naming15.

The report then names the finished protein. The table shows which alleles stand behind each name, with what the sources say about its link to celiac disease.

Protein names on a report and the alleles behind them
Name on the reportAlleles that encode itWhat the sources say about celiac risk
DQ2.5DQA1 05:01 with DQB1 02:0115Highest risk of the three types, per Coeliac UK5. The Sardinian review calls it the type most permissive for celiac disease16.
DQ2.2DQA1 02:01 with DQB1 02:0215Low risk, but not no risk, per Coeliac UK5. The review says its alpha chain makes it less suited than DQ2.5 to present gluten-derived peptides to T cells17.
DQ8DQA1 03:01 with DQB1 03:0218Moderate risk, per Coeliac UK5. The review says DQ8 without DQ2.5 confers a low risk19 and DQ8 together with DQ2.5 a high risk20.
DQ7.3 or DQ9DQA1 03 with DQB1 03:01 (DQ7.3), or with DQB1 03:03 (DQ9)21Not DQ8: the review says only DQ8 is permissive for celiac disease21.

Two cautions when matching a report to this table. First, “DQ2” alone does not say which molecule: the DQ2.2 molecule “has been named” DQ2 as well, although its alpha chain is different6, and the two differ in how they relate to celiac disease7. Second, allele numbers are written at different levels of detail. A 2009 US study wrote DQ2.2 as DQA1 0201-DQB1 028; the 2025 review writes DQA1 02:01 and DQB1 02:0215. Coeliac UK says of UK laboratories that the format of the reporting is not standardised22.

How often each type appears in people with celiac disease

The figures depend on who was studied and how they were typed, so the table keeps each one next to its source instead of merging them.

Share of people with celiac disease carrying each type, by source
Source and groupDQ2 or DQ2.5DQ8Other, or neither
Coeliac UK: patients, adults and childrenNot given separatelyNot given separately99.6% carry DQ2.5, DQ8 or DQ2.22
Review from a Sardinian group (2025)Approximately 90%-95% positive for alleles that encode DQ2, more specifically DQ2.5 in cis or trans23About 5%-7%18The remaining 3%-5%: DQ2.2, or very rarely half heterodimers such as DQ2.3 or DQ7.524
World Gastroenterology Organisation guideline (2016): patients of northern European descentApproximately 95%25The rest25Not given
437 Italian children with celiac disease (2009)91% carried DQ2 and/or DQ8, reported together26Counted with DQ2266% had only the beta2 chain, 2% were alpha5 positive, and four children were negative for DQ2, DQ8, beta2 and alpha526

The rows come from different groups, methods and years and are not meant to be added together. “Cis” and “trans” in the review’s wording describe where the alleles sit: cis means alleles at different loci on the same chromosome, trans means they sit on opposite chromosomes27.

How the combinations compare in one US study

A 2009 US study typed the DNA of 10,191 people who were at risk for celiac disease and counted who tested positive for anti-endomysial antibodies (EMA), the marker the authors used to identify celiac disease28. The people were selected for being at risk, so the rates below are not rates for people in general.

EMA-positive rate by DQ genotype group (Pietzak and colleagues, 2009)
Genotype groupShare of the 10,191 samplesEMA-positive within the group
Two copies of DQ2.5, or DQ2.2 with DQ2.55.38%828.28% (95% CI 24.55-32.26)8
DQ2.5 heterozygousNot in the abstract9.09% (95% CI 7.82-10.51)9
DQ8 homozygousNot in the abstract8.42% (95% CI 3.71-15.92)29
DQ8 heterozygousNot in the abstract2.11% (95% CI 1.43-3.00)29
DQ2.2 with DQ8, or DQ2.5 with DQ85.08%3011.78% (95% CI 9.13-14.87)30
Neither DQ2 nor DQ842.03%, 4283 samples310.16% (95% CI 0.07-0.34)31

“Homozygous” means the same allele on both chromosomes and “heterozygous” means different alleles, in the wording of the 2025 review’s glossary27.

One copy or two copies of DQ2.5

In the US study, the group with two copies of DQ2.5 (or DQ2.2 plus DQ2.5) had a higher EMA-positive rate than the group with one copy: 28.28% against 9.09%8, 9. The authors wrote that the data suggest a quantitative relationship between the type and proportion of DQ heterodimers and the risk of celiac disease32.

A laboratory study from 2003 found a matching pattern: cells with two copies of DQ2 presented gluten to T cells better than cells with one DQ2 copy and a non-DQ2 allele33. The same group reported that DQ2.5 presents a large repertoire of gluten peptides, whereas DQ2.2 can present only a subset34, and said the findings explain the strongly increased risk for people homozygous for DQ2.5 and for DQ2.2/2.5 heterozygous people35.

That is why DQ2.2 reads differently depending on what sits beside it. The 2025 review lists DQ2.5/DQ2.2, DQ2.5/DQ2.5 and DQ2.5/DQ8 among the high-risk genotypes36. It explains that DQ2.2 can behave as a donor of beta chains for building the DQ2.5 protein37, and that when DQ2.2 combines with DQA1 05:05 and DQB1 03:01 to form the DQ2.5 protein in trans, it confers a similar risk to DQ2.5 formed in cis38.

How common DQ2 and DQ8 are in people without celiac disease

Both types are common. The NIDDK says about 30 percent of people have DQ2 or DQ8, and only about 3 percent of people with DQ2 or DQ8 develop celiac disease3. Coeliac UK says the DQ2 and DQ8 haplotypes are found in up to 40% of the general population10, and the World Gastroenterology Organisation guideline gives 30–35% for the general population, against approximately 70% of first-degree relatives39.

Coeliac UK reasons that, with celiac disease estimated at 1% in the UK, “there must be other factors” that work with these haplotypes10.

What the DQ protein does with gluten

A 2015 review by du Pré and Sollid says CD4+ T cells that recognize gluten peptides bound to predisposing HLA-DQ molecules play a key role in celiac disease. These T cells recognize deamidated gluten peptides better, and the enzyme transglutaminase 2 (TG2) does the deamidating40.

A 2017 review by Sollid puts it as certain HLA-DQ allotypes predisposing to the disease by presenting these modified gluten peptides to CD4+ T cells41.

Celiac disease without DQ2 or DQ8

Coeliac UK calls patients who carry none of DQ2.5, DQ8 or DQ2.2 “very rare”42. The sources give these figures and explanations:

  • An Australasian group of 356 people with celiac disease: 99.6% were DQ2/DQ8-positive, as ESPGHAN’s 2020 guideline reports11.
  • ESPGHAN says the higher percentage of DQ2/DQ8-negative patients (up to 5%) in earlier publications had several causes, one being the HLA testing method11. SNP-based tests recognize the common variants (DQ2.5, DQ8, DQ2.2, DQ7.5), while in-depth allele typing is needed to identify rare ones, so the accuracy of excluding celiac disease depends on the method used43. Other causes it lists are a more heterogeneous group of patients with compatible symptoms and, in some cases, false-positive histopathology44.
  • In the US study, 0.16% of the 4283 samples with neither DQ2 nor DQ8 were EMA-positive31.
  • Among 437 Italian children with celiac disease, four were negative for DQ2, DQ8, beta2 and alpha526.
  • The 2025 Sardinian review says a multicenter European study detected patients with genotypes other than DQ2 and DQ8 in 2% of cases45. Its own series had 10 such patients, all diagnosed in the late 1980s before the tTG blood tests were available, and the authors could not re-verify those diagnoses46.

These figures come from different groups, methods and decades, so they are not one estimate.

What a positive or negative result does and does not show

The sources treat a negative result as more informative than a positive one. ESPGHAN’s guideline says that if a patient tests negative for DQ2 and DQ8 the risk of celiac disease is very low, while a positive result does not confirm the diagnosis4. The 2023 American College of Gastroenterology guideline says HLA testing may be helpful in selected situations47, and that if it is negative, celiac disease is ruled out48. Coeliac UK says a positive result does not diagnose celiac disease and, in its example of a patient with suspected celiac disease who tests positive, is not very helpful, because so many of the general population are DQ positive49. A 2025 review adds that the test has value only when negative50.

ESPGHAN also says HLA typing is not required in patients with positive TGA-IgA who qualify for diagnosis with biopsies, or who have high serum TGA-IgA (10 times the upper limit of normal or more) and positive EMA-IgA4.

The World Gastroenterology Organisation guideline says research suggests HLA haplotypes alone confer approximately 35–40% of the genetic risk51.

This guide explains terms. What a particular result means for a particular person is a question for that person’s clinician.

Questions people ask

Do DQ2 and DQ8 cause celiac disease on their own?

No. The NIDDK says research suggests celiac disease only occurs in people who have certain genes and eat food that contains gluten52, and not everyone with DQ2 or DQ8 who eats gluten develops it53.

Does a report without DQ2 or DQ8 mean celiac disease is impossible?

The sources word it differently. The NIDDK says people without these gene variants are very unlikely to develop celiac disease1; ESPGHAN says the risk is very low4; the 2023 American College of Gastroenterology guideline says that if HLA testing is negative, celiac disease is ruled out48. Coeliac UK says patients with none of DQ2.5, DQ8 or DQ2.2 are very rare and that being negative, though “never a guarantee”, would be very reassuring42.

What do homozygous, heterozygous, cis and trans mean on a report?

The 2025 review’s glossary defines homozygous as the same allele sequence at the same locus on both chromosomes, heterozygous as different sequences, cis as alleles at different loci on the same chromosome, and trans as alleles at different loci on opposite chromosomes27.

More in Celiac genetics and family risk

Sources