TL;DR
- 1.A 27,000-person study published in Nature found a GLP1R gene variant that predicts who loses significantly more weight on semaglutide and tirzepatide.
- 2.People with one GLP1R variant lost weight 58% faster per month than those without it, on the same dose of semaglutide.
- 3.A separate GIPR variant predicts who gets hit hardest by tirzepatide nausea, useful if you are deciding between Ozempic and Mounjaro.
- 4.No FDA or clinical guideline currently requires genetic testing before prescribing GLP-1 drugs. The evidence exists. The system has not caught up.
- 5.Your existing 23andMe or AncestryDNA file likely contains the relevant variant data. You do not need a new test.
Thousands of people have spent over a thousand dollars a month on semaglutide and lost five pounds. Their doctors adjusted their diet. They increased their dose. They switched injection sites. Nobody told them that a 27,885-person study had already identified the gene that predicted they would lose weight nearly three times slower than average, and that the answer may have been sitting in their 23andMe file the entire time.
Number of GLP-1 receptor agonist users analyzed in the 23andMe Research Institute's April 2026 GWAS, published in Nature. The study identified the first pharmacogenomically actionable genetic variants predicting weight loss outcomes and side effects for semaglutide and tirzepatide at population scale.
Semaglutide (sold as Ozempic for diabetes and Wegovy for weight loss) mimics glucagon-like peptide-1, a hormone your gut naturally releases after eating. It binds GLP-1 receptors in the brain to reduce appetite, slows gastric emptying to keep you feeling full longer, and reduces the reward signal your brain generates from food. The mechanism is well-understood and the clinical trial data is genuinely strong: a 2021 trial in the New England Journal of Medicine found semaglutide 2.4 mg produced average weight loss of 14.9% of body weight over 68 weeks. But "average" hides an enormous distribution. Some people in that trial lost over 20%. Others lost less than 5%. The drug was identical. The doses were identical. The difference was biology.
The biological variable hiding in plain sight is your GLP-1 receptor itself. GLP1R, the gene that encodes this receptor, carries common variants that alter the receptor's structure, sensitivity, and downstream signaling efficiency. A drug that works by binding a receptor is going to work differently depending on what that receptor looks like at the molecular level. This is not a fringe idea in pharmacogenomics research. It is the most replicated finding in GLP-1 pharmacology, confirmed now across multiple large independent datasets.
Think of the GLP-1 receptor as a lock and semaglutide as the key. Most people have the same lock. Some people have a lock where the keyhole is slightly narrower. The key still fits, but it does not turn as cleanly. You can use a stronger key (higher dose), but the lock shape is genetic and does not change. Knowing you have that lock before you start the drug would let you set realistic expectations, choose the right dose, or consider a different key entirely.
What does a 27,000-person study show about semaglutide non-response?
In April 2026, the 23andMe Research Institute published the largest GLP-1 pharmacogenomics genome-wide association study ever conducted, in Nature. The study analyzed 27,885 people who had taken semaglutide or tirzepatide, using genetic data to identify variants that predicted how much weight each person lost and which side effects they experienced.
The headline finding: a missense variant in GLP1R, rs10305420 (p.Pro7Leu), was significantly associated with greater weight loss, reaching genome-wide significance with a P value of 2.9 times 10 to the negative 10. Each copy of the effect allele predicted an additional 0.76 kg of weight loss over the treatment period (roughly 0.64% BMI change per T allele). People carrying two copies lost approximately 3.3 pounds more over the median 8-month treatment window than non-carriers, on identical drug regimens. Separately, variants in both GLP1R and GIPR predicted nausea and vomiting, with the GIPR finding specific to tirzepatide. This was the first time genetic predictors of GLP-1 class drug side effects had been identified at this statistical confidence level.
That study provides the population-level view. A complementary study from 2025 by Phan et al. in Obesity provides the most striking individual-level data. The researchers followed 112 patients with severe obesity (BMI 40 or above) on semaglutide 2.4 mg for at least four months and compared weight loss velocity by GLP1R genotype at rs6923761.
The GLP1R rs6923761 A/A genotype was associated with a significantly higher monthly rate of weight loss (1.64% vs 1.04% in G-allele carriers), representing an approximately 58% difference in weight loss velocity on semaglutide 2.4 mg. These findings suggest GLP1R genotype should be considered a clinically meaningful predictor of GLP-1 receptor agonist outcomes.
Phan et al., Obesity (Wiley), 2025, PMID 40384505
A 58% difference in monthly weight loss velocity is not a marginal pharmacogenomic signal. It is the difference between losing 18 pounds in six months and losing 12 pounds in six months, on the same drug at the same dose. The G-allele is relatively common: estimates vary by ancestry, but a meaningful fraction of semaglutide users carry at least one copy. These are not rare-variant effects. They are common genetic differences that apply to a substantial portion of the population using these drugs right now.
What does the Lancet data add?
A 2023 pharmacogenomics study by Dawed et al. in Lancet Diabetes and Endocrinology added a third independent replication using randomized controlled trial data and observational cohorts. Dawed et al. identified GLP1R coding variants and loci in ARRB1 associated with HbA1c reduction on GLP-1 drugs. Critically, the study confirmed a longstanding finding: carriers of the TCF7L2 rs7903146 risk allele show blunted GLP-1-stimulated insulin release. This variant is particularly relevant for people using GLP-1 drugs to manage type 2 diabetes rather than weight loss alone, because the glucose-lowering mechanism depends on that insulin response. If you carry this variant and are not seeing adequate glycemic control on semaglutide, the limitation may be downstream of the GLP-1 receptor rather than at it.
Which genes actually matter for semaglutide response?
rs10305420 (p.Pro7Leu): identified in the 27,885-person Nature 2026 GWAS. Each T allele predicts approximately 0.76 kg of additional weight loss. rs6923761: A/A genotype produces 58% faster monthly weight loss than G-allele carriers (Phan et al., 2025). These two variants operate at the receptor level and represent the primary pharmacogenomic explanation for GLP-1 response variance between individuals on identical protocols.
The 2026 Nature GWAS found GIPR variants that predict nausea and vomiting on tirzepatide (Mounjaro, Zepbound) specifically. Tirzepatide is a dual GLP-1 and GIP receptor agonist, so GIPR genotype matters for side effect tolerance in ways that do not apply to semaglutide alone. If you switched from Ozempic to Mounjaro and the nausea was significantly worse, GIPR genotype may be the pharmacogenomic explanation.
TCF7L2 rs7903146 risk allele carriers show blunted GLP-1-stimulated insulin secretion, confirmed across multiple datasets (Dawed et al., Lancet Diabetes and Endocrinology, 2023). This variant is most relevant for type 2 diabetes patients who need glycemic control from semaglutide. For pure weight loss use cases, TCF7L2 has a smaller predicted impact. For metabolic syndrome management, it is the most important secondary gene to check alongside GLP1R.
MC4R variants disrupt the melanocortin-4 receptor pathway, which sits downstream of GLP-1 receptor signaling in the brain's appetite regulation cascade. Even when GLP-1 receptor binding is normal, a MC4R variant can blunt the hunger-reduction effect semaglutide is designed to produce. This explains a subset of people who show pharmacokinetically normal drug exposure but minimal appetite suppression. See the DNA-first decision framework for how MC4R fits into the full metabolic axis.
| Gene | Variant | Effect on semaglutide | Relevant drug | Consumer array coverage |
|---|---|---|---|---|
| GLP1R | rs6923761 (A/A vs G) | 58% faster weight loss for A/A genotype | Semaglutide, GLP-1 class | 23andMe v5 direct; AncestryDNA imputed |
| GLP1R | rs10305420 (p.Pro7Leu) | +0.76 kg weight loss per T allele | Semaglutide, tirzepatide | Uncertain; targeted panel recommended |
| GIPR | Multiple loci | Predicts nausea severity on tirzepatide | Tirzepatide specifically | Specialist panel recommended |
| TCF7L2 | rs7903146 | Blunted GLP-1-stimulated insulin release | GLP-1 class (glycemic effect) | Yes, all major consumer arrays |
| MC4R | rs17782313 | Blunted downstream satiety signal | GLP-1 class (appetite effect) | Yes, all major consumer arrays |
The difference in monthly weight loss velocity between GLP1R rs6923761 A/A homozygotes and G-allele carriers on semaglutide 2.4 mg, from the Phan et al. 2025 study in Obesity (PMID 40384505). This translates to several extra pounds of weight loss over a six-month course at identical doses.
Why has your doctor not mentioned any of this?
No FDA label update for semaglutide mentions pharmacogenomic testing. No CPIC (Clinical Pharmacogenomics Implementation Consortium) guideline recommends GLP1R genotyping before prescribing. A 2025 review in Expert Opinion on Pharmacotherapy concluded that evidence is promising but "insufficient for clinical genotyping recommendations" at this time. The regulatory and clinical infrastructure for GLP-1 pharmacogenomics is roughly a decade behind the research.
This is not unusual in pharmacogenomics. CYP2D6 and CYP2C19 testing is now standard for antidepressant prescribing, but it took over 20 years after the variants were characterized for guidelines to catch up. GLP-1 pharmacogenomics is probably at the equivalent early stage: the studies are replicating, the effect sizes are large enough to be clinically meaningful, but the translation to clinical practice has not happened yet. The gap matters most for the people paying out of pocket for a drug that is not working for their biology.
One company is moving ahead without waiting for guidelines: Mayo Clinic's Phenomix Sciences developed a test called MyPhenome, presented at the 2024 American Diabetes Association Scientific Sessions, claiming 75% accuracy in predicting semaglutide response based on a genetic and phenotypic profile. Whether this translates to widespread clinical adoption is unclear, but it signals that the diagnostic market is recognizing the clinical need before the guidelines do. For a detailed look at what consumer genetic data can already tell you about GLP-1 response without a specialist test, the 23andMe peptide guide covers which variants are reliably covered on existing consumer arrays.
How do you find your semaglutide genotype in existing DNA data?
GLP1R rs6923761 is directly genotyped on 23andMe v5, the chip version in use since 2017. If you tested with 23andMe after 2017, your raw data file contains your genotype at this specific position. GLP1R rs10305420 is less certain on consumer arrays: its presence on the v5 chip manifest has not been independently confirmed in public sources, meaning it may require statistical imputation even on v5 data.
For rs6923761, the lookup is straightforward. Download your raw data from 23andMe (Settings, then "23andMe Data", then "All DNA Raw Data"). Open the file and search for rs6923761. Your two-letter genotype call tells you whether you are A/A (faster response), AG (intermediate), or GG (slower monthly velocity on semaglutide). This single lookup takes under five minutes and answers the most actionable pharmacogenomic question before or during a GLP-1 course.
AncestryDNA and MyHeritage users typically have rs6923761 as an imputed variant rather than directly genotyped, adding statistical uncertainty to the result. For these users, or for anyone who wants genotyping on the newer GWAS variants including rs10305420, uploading your raw file to a dedicated peptide genetics service processes all current GLP-1 pharmacogenomic variants simultaneously. See the DNA-first peptide decision framework for the step-by-step process of applying genetic results to protocol selection.
What do you do if you are a semaglutide non-responder?
If your GLP1R genotype suggests reduced receptor sensitivity, the clinical options fall into three categories. The first is dose escalation: the G-allele effect is a partial receptor sensitivity reduction, not a complete block. Higher doses of semaglutide can partially compensate, and some G-allele carriers see adequate response at 2.4 mg who did not respond well at lower doses. This is worth a conversation with your prescribing physician before abandoning the drug class entirely.
The second option is switching within the GLP-1 class. Tirzepatide (Mounjaro, Zepbound) acts on both GLP-1 and GIP receptors. If your non-response is primarily at the GLP1R level and your GIPR pathway is normal, adding a second receptor mechanism may produce better outcomes. Clinical data supports meaningfully greater average weight loss with tirzepatide than semaglutide at their respective maximum doses. However, if your GIPR genotype predicts poor GIP signaling or severe nausea, this approach may also underperform or be poorly tolerated.
The third option is looking outside the GLP-1 class entirely. For weight loss specifically, compounds that work through the LEPR, ADRB3, or FTO pathways operate independently of GLP-1 receptor sensitivity. For peptide-based approaches, AOD-9604 (a fragment of growth hormone that does not affect growth hormone signaling itself) works through a fat-mobilization mechanism independent of the GLP-1 axis. MOTS-c, a mitochondrial peptide, targets metabolic flexibility rather than appetite suppression directly. If your non-response is receptor-level, these alternatives offer different mechanisms with different genetic determinants. See the semaglutide peptide page for a comparison of the full GLP-1 class and adjacent weight-loss options. For the full genetic framework covering all weight-related peptides, the semaglutide genetics guide covers GLP1R, MC4R, TCF7L2, and the related variants that together predict metabolic peptide response.
Your DNA shapes how you respond to every peptide in this report.
A personalized report scores 25+ peptides against your unique genetic profile — including the ones covered in this article.
Frequently asked questions
Why is semaglutide not working for me?
The most common genetic explanation is GLP1R rs6923761 genotype. G-allele carriers lose weight 58% slower per month than A/A homozygotes on identical semaglutide doses, according to a 2025 study in Obesity (Phan et al., PMID 40384505). Behavioral factors also matter: diet quality, dose adequacy, and injection technique. But if those are optimized and results are still below expectations after 3 to 4 months at the maximum dose, GLP1R genotype is the most likely biological explanation. Check your 23andMe raw data for rs6923761 or upload to a peptide genetics service for a full pharmacogenomic analysis.
What gene predicts semaglutide response?
Two GLP1R variants are the strongest predictors. GLP1R rs6923761 (A/A vs G-allele) was associated with a 58% difference in monthly weight loss velocity in a 2025 prospective study of 112 patients. GLP1R rs10305420 (p.Pro7Leu) was identified in a 27,885-person 2026 Nature GWAS as predicting approximately 0.76 kg additional weight loss per T allele. MC4R rs17782313 predicts blunted downstream appetite suppression even when GLP-1 receptor binding is normal. TCF7L2 rs7903146 predicts blunted glycemic response specifically.
Can I test my GLP1R gene before starting Ozempic or Wegovy?
Yes, through two routes. If you already have 23andMe v5 raw data (tested after 2017), GLP1R rs6923761 is directly genotyped and you can look it up by searching for the rsID in your downloaded raw file. Upload the raw data file to a dedicated analysis service for a full pharmacogenomic interpretation covering all relevant variants. Alternatively, order a targeted peptide genetics panel or a new saliva kit if you do not yet have genetic data. No FDA or clinical guideline currently requires this testing before prescribing, so you will need to request it proactively.
Does tirzepatide work better for semaglutide non-responders?
Sometimes, but not automatically. Tirzepatide (Mounjaro, Zepbound) adds GIP receptor activation on top of GLP-1 receptor activation, offering a second mechanism that could compensate for reduced GLP1R sensitivity. But the 2026 Nature GWAS identified GIPR variants that predict worse nausea on tirzepatide specifically, so switching is not guaranteed to help and can introduce new side-effect risk depending on your GIPR genotype. Knowing both your GLP1R and GIPR genotypes before switching makes the decision considerably clearer.
What should I do if semaglutide is not working genetically?
Three options in order: first, confirm adequate dose, as G-allele carriers sometimes respond to 2.4 mg who did not respond at lower doses. Second, consider switching to tirzepatide if your GIPR genotype is favorable. Third, look outside the GLP-1 class entirely toward weight-loss approaches that work through LEPR, ADRB3, or FTO pathways, or peptide options like AOD-9604 and MOTS-c that target metabolic flexibility through mechanisms independent of GLP-1 receptor sensitivity. A full genetic report helps identify which alternative pathway is best matched to your biology.
Is there a genetic test for semaglutide response?
Yes. Mayo Clinic's Phenomix Sciences developed a test called MyPhenome that claims 75% accuracy in predicting semaglutide response, presented at the 2024 American Diabetes Association Scientific Sessions. Standard consumer genetic data (23andMe v5, AncestryDNA) covers several key GLP1R and MC4R variants that can provide a first-pass pharmacogenomic assessment without a new test. No FDA-approved companion diagnostic for semaglutide response exists as of May 2026, and no clinical guideline yet requires genetic testing before prescribing.
How common is genetic non-response to semaglutide?
Precise population prevalence depends on ancestry and which variant you are measuring. GLP1R rs6923761 G-allele carriers are present across all ancestry groups, with frequency varying by population. The 2025 Phan et al. study found a statistically significant difference within 112 patients with severe obesity, suggesting the effect is detectable even in relatively small clinical populations. The 2026 Nature GWAS reaching genome-wide significance across 27,885 people confirms this is a common pharmacogenomic signal affecting a substantial fraction of current GLP-1 drug users.
This article is for informational and educational purposes only. It is not medical advice and does not diagnose, treat, cure, or prevent any disease. Consult a qualified healthcare professional before starting any peptide protocol. Individual results vary.