Researchers at 23andMe report that common variants in GLP1R and GIPR genes predict differential weight loss and gastrointestinal side effects among patients taking semaglutide or tirzepatide, according to a study published April 8, 2026 in Nature. The findings, drawn from a genome-wide association study of 27,885 GLP-1 receptor agonist users, identify the first potentially pharmacogenomically actionable loci for this drug class and lay groundwork for further research that could enable genotype-informed prescribing in obesity treatment.
The study's most statistically robust result centers on GLP-1-induced nausea, localized to the GLP1R locus, with genome-wide significant associations identified for nausea and vomiting. For efficacy, a missense variant in GLP1R (rs10305420, p.Pro7Leu) located in the receptor's signal peptide was associated with approximately 0.76 kg additional weight loss per allele copy, with the effect attributed to enhanced receptor trafficking to the cell surface rather than altered ligand binding. No genome-wide significant loci for overall BMI loss were identified, consistent with a polygenic architecture for GLP-1 efficacy.
The side-effect genetics yielded a mechanistically distinct finding specific to tirzepatide. A partial loss-of-function variant in GIPR (rs1800437, p.Glu354Gln) was associated with increased vomiting risk exclusively in tirzepatide users — not in those taking semaglutide, which lacks GIP receptor activity. The biological rationale is that GIP receptor co-activation in tirzepatide normally buffers GLP-1-induced nausea centrally; when GIPR function is impaired, that buffering is lost. Individuals homozygous for risk alleles at both GLP1R and GIPR loci showed 14.8-fold increased odds of tirzepatide-mediated vomiting, a finding that could have direct implications for drug selection prior to treatment initiation.
The study was conducted entirely within 23andMe's consumer genetics infrastructure, with no academic university co-affiliations listed among the authors. Participants were recruited from the company's customer base following a GLP-1 survey launched in August 2024. The primary GWAS for efficacy included 15,237 participants with percentage BMI loss as the outcome. Self-reported weight data showed strong correlation with electronic health record measurements in a 195-person overlap cohort (r ≈ 0.57), supporting the validity of the survey-based phenotyping approach. An independent replication cohort of 4,855 participants from the NIH's All of Us program showed directional consistency with discovery findings.
Predictive models incorporating both genetic and non-genetic variables — including baseline BMI, drug type, and treatment duration — outperformed genetics alone, though the genetic contribution was statistically detectable. Tirzepatide was associated with greater average BMI loss than semaglutide across the cohort, consistent with data from the SURMOUNT and STEP clinical trial programs.