Celiac Disease Genes in Raw DNA: HLA-DQ2 and DQ8

Two tag SNPs, rs2187668 and rs7454108, track the HLA types behind almost all celiac disease. Why a negative is more useful than a positive, and where tags fail.

Celiac disease is an autoimmune reaction to gluten that damages the small intestine, and it has one of the strongest genetic gatekeepers of any common condition: almost everyone who develops it carries one of two specific immune-system types, HLA-DQ2 or HLA-DQ8. Those types can be estimated from a raw DNA file using two tag SNPs. The catch is in the word “estimated”, and in the fact that a positive result is far weaker evidence than a negative one.

HLA types and why they matter here

The HLA genes on chromosome 6 encode the proteins that present fragments of foreign material to the immune system. The DQ2.5 and DQ8 versions are unusually good at presenting fragments of gluten, which is the first step in the celiac reaction. Around 90 percent of people with celiac disease carry DQ2.5, most of the rest carry DQ8, and a small minority carry a related type called DQ2.2.

The reverse is not true. Roughly 30 to 40 percent of people of European ancestry carry DQ2 or DQ8, and only about one in a hundred develops celiac disease. Having the type is necessary, more or less, but nowhere near sufficient.

The two tag SNPs

HLA genes are so variable that a SNP chip cannot read them directly. Instead, researchers found nearby single-letter variants that travel with particular HLA types often enough to stand in for them. Monsuur and colleagues showed in 2008 that a handful of such tags could detect the celiac-risk HLA types with high accuracy in European populations.

The two most useful, and the ones most raw files include, are:

  • rs2187668 - tags the DQ2.5 type (specifically the DQA105:01-DQB102:01 haplotype). On the forward strand the common allele is C; the tag allele is T. One T means one copy of DQ2.5; TT means two.
  • rs7454108 - tags DQ8 (DQB1*03:02). Common allele T; the tag allele is C.

So a file showing CC at rs2187668 and TT at rs7454108 suggests neither DQ2.5 nor DQ8. A CT at the first or a TC at the second suggests one copy of the corresponding type.

Reading the result the right way round

Because the HLA types are so common and the disease so much rarer, the two possible results have very different weight.

A negative result is the informative one. If you carry neither DQ2 nor DQ8, celiac disease is very unlikely - clinicians use HLA typing for exactly this purpose, to rule celiac out in ambiguous cases or in relatives of patients. The tag-SNP version is a little less reliable than a proper HLA test, but a clean negative still moves the probability a long way.

A positive result mostly tells you that you are like a third of the population. Two copies of DQ2.5 does carry more risk than one, and DQ2.5 more than DQ8, but even the highest-risk genotype leaves most carriers without the disease. A positive tag is not a reason to start a gluten-free diet. It is, at most, a reason to be quicker to mention persistent gut symptoms, anaemia or a family history to a doctor.

Where the tags fail

Tag SNPs are proxies, and proxies have holes:

  • DQ2.2 is not covered. The minority of celiac patients with DQ2.2 rather than DQ2.5 will look negative on rs2187668. A tag for DQ2.2 exists but is not on every chip.
  • Ancestry changes the linkage. The tags were validated mostly in Europeans. In other populations the SNP and the HLA type can come apart, in either direction.
  • No-calls happen in the HLA region. It is one of the most variable parts of the genome, and probes there fail more often than average. A -- at either marker means no information, not a negative.

A clinical HLA-DQ test types the genes directly and does not have these problems. If a result would change a decision, that is the test to ask for.

The test that actually diagnoses celiac disease

Genetics can rule celiac out; it cannot rule it in. Diagnosis rests on blood antibodies (tissue transglutaminase IgA is the usual first test) and, often, a biopsy - and both only work if you are still eating gluten. This is the most important practical point in the whole topic: people who go gluten-free because of a genetic result and then seek a diagnosis often find the tests are no longer interpretable.

If you are exploring what else your file can and cannot say about autoimmune and inflammatory conditions, the health and wellness analysis reads a range of markers locally in your browser, and our guide on single-SNP claims versus polygenic reality is useful context for the many weaker associations you will find online.

References

  • Monsuur AJ, et al. Effective detection of human leukocyte antigen risk alleles in celiac disease using tag single nucleotide polymorphisms. PLoS ONE. 2008. PubMed 18509540
  • National Institute of Diabetes and Digestive and Kidney Diseases. Celiac Disease.
  • Celiac Disease Foundation. Genetic testing.
  • dbSNP entries for rs2187668 and rs7454108, NCBI.

This article is educational only and is not medical advice. Do not start or stop a gluten-free diet on the basis of a genetic result; talk to a doctor while you are still eating gluten.

Further reading