Rare GLP1R Variants Blunt Semaglutide, Not Tirzepatide: ASHG Data

Complete guide: GLP-1 drugs: a complete guide to semaglutide and beyond

What Helix presented at ASHG 2026

On Oct. 22, 2026, at the American Society of Human Genetics annual meeting, Kelly Schiabor Barrett of Helix gave a platform talk titled “Rare GLP1R variants selectively attenuate weight loss with semaglutide but not tirzepatide, implicating GIP receptor agonism as a compensatory mechanism.” The analysis drew on 13,725 people taking either drug, from Helix exome-linked health records². A Helix press release describes the same findings¹.

Carriers of rare, damaging variants in GLP1R, the gene for the GLP-1 receptor, lost 2.4% of body weight on semaglutide (n = 17) compared with 8.8–8.9% in matched controls. On tirzepatide, carriers lost 13.3% versus 15.2% in matched non-carriers (n = 19). The test of whether the two drugs behaved differently gave an interaction P = 0.009². Helix's release states 8.8% for the semaglutide controls and the talk summary states 8.9%; we show both¹².

Variants in GIPR, the gene for the GIP receptor, “were not associated with differential response to either medicine,” according to the release¹.

The proposed explanation: GIP agonism as a backstop

Semaglutide acts on the GLP-1 receptor. Tirzepatide acts on both the GIP and GLP-1 receptors¹⁰¹¹. In a 2020 pharmacology study from Eli Lilly and academic collaborators, tirzepatide engaged the GIP receptor more than the GLP-1 receptor at clinically relevant exposures, and showed biased signaling at the GLP-1 receptor that favored cyclic AMP generation over β-arrestin recruitment¹⁰. That is the logic behind the title of the talk: if the GLP-1 receptor works poorly, a second receptor may partly carry the effect².

This is a hypothesis. The abstract-level data show an association between carrier status and weight change, not a measured contribution of GIP signaling. The summary of the talk notes that the most damaging variants still lost less weight on tirzepatide, which fits partial, not complete, compensation². The authors suggest that carriers who respond poorly to semaglutide might switch to tirzepatide; that is a proposal and was not tested².

What 17 and 19 people can and cannot support

The weak point is size. Seventeen and nineteen carriers produce wide uncertainty around any average, and a few outliers can move a mean of that kind. The release lists no confidence intervals or formal limitations¹, and the summary of the talk says plainly that the carrier numbers are small and that conference abstracts are not peer reviewed and may change before publication². We could not find the abstract on a public ASHG page, so the numbers here come from the company release and one independent summary.

Also note the baseline. The matched controls lost 8.8–8.9% on semaglutide and 15.2% on tirzepatide in real-world records¹². In randomized trials, participants lost 14.9% on semaglutide 2.4 mg at 68 weeks (STEP 1) and 15.0–20.9% on tirzepatide at 72 weeks (SURMOUNT-1)¹²¹³. In the head-to-head SURMOUNT-5, the figures were 20.2% with tirzepatide and 13.7% with semaglutide at 72 weeks¹⁴. Records-based cohorts differ from trials in dose, adherence and follow-up, so only the within-cohort comparison between carriers and controls is informative.

How this differs from the 2026 Nature paper

This result should not be confused with the 23andMe genome-wide association study in Nature. That paper analyzed self-reported weight loss and side effects in 27,885 people on GLP-1 receptor agonists and found a common missense variant in GLP1R (rs10305420) associated with an additional −0.76 kg per effect allele (P = 2.9 × 10⁻¹⁰). It also linked GLP1R and GIPR variation to nausea or vomiting, with the GIPR association restricted to tirzepatide users³. Scientific American adds that about 40% of people of European and Middle Eastern ancestry carry the GLP1R variant, and that people with two copies of both it and a GIPR variant had an estimated 15-fold higher likelihood of vomiting on tirzepatide than people with neither⁴.

The two datasets answer different questions: a common variant with a small effect across thousands of people versus rare damaging variants with a possibly larger effect in a handful. Scientific American noted that rarer variants not captured in the Nature analysis could matter more⁴. We have not seen a Nature paper reporting the rare-variant result, and the Helix talk should not be attributed to Nature.

Independent reviewers of the Nature paper were measured. A Cambridge researcher said the genetic effects are “small in clinical terms,” that the evidence is “not yet sufficient to support using genetic information to guide treatment decisions” in routine care, and that self-reported data introduce measurement error and reporting bias⁵. The cohort was mostly female and of largely European ancestry⁵. All authors are current or former 23andMe employees³, and the company offers a related report exclusively through its Total Health service⁶⁷.

Earlier work on rare GLP1R variants

Functional biology gives the Helix result some plausibility. A 2023 Nature Metabolism study profiled 60 GLP1R variants across four signaling pathways and found phenotypes from defective cell-surface expression to gain of function; in UK Biobank data (about 200,000 participants), impaired cell-surface expression was associated with higher HbA1c and body mass index. In cells, defective insulin secretion from loss-of-function variants was rescued by high concentrations of semaglutide or exendin-4⁸.

A Danish study of 8,642 people found 36 non-synonymous GLP1R variants, 10 with significantly reduced GLP-1-induced cAMP signaling. Heterozygous carriers had cardiometabolic phenotypes similar to non-carriers, and no homozygous loss-of-signaling variants were found⁹. Together these studies show that rare variants do change receptor function, but that carrying one rarely produces an obvious metabolic phenotype on its own.

What a physician should take from it

• Treat the result as a hypothesis-generating signal about non-response to semaglutide, not as a test to order.
• No genotype-guided choice between semaglutide and tirzepatide is established in any guideline or trial we found; the Nature authors themselves frame their work as the foundation for future precision approaches³.
• Rare damaging GLP1R variants are, by definition, uncommon, so they cannot explain most poor responders; sex, age, other health conditions and the specific drug also influence response⁴.
• Wait for the full data, confidence intervals, the variant definition and peer review.