Blood-Based Epigenetic Test Predicts GLP-1 Response and Tracks Treatment Effects

Prescriptions for GLP-1 medicines for weight loss are expanding rapidly, yet clinicians still lack scalable tools to predict biological response before treatment or monitor drug-driven changes beyond the scale. Noninvasive assays that can identify likely responders and quantify pharmacodynamic effects remain an unmet need in obesity and liver disease care. New data now demonstrate a blood-based epigenetic assay and AI platform that identify semaglutide responders before treatment and track drug-associated biological changes over time.

Hepta’s (Foster City, CA, USA) blood-based assay analyzes cell-free DNA (cfDNA) methylation from a standard blood draw and is interpreted by LiquidTransformer, an attention-based artificial intelligence model purpose-built for liquid biopsy. The model reads hundreds of millions of cfDNA fragments per sample simultaneously to capture distributed, pathway-level biological signals across the methylome. This platform underpins both response prediction and longitudinal monitoring from plasma.

The findings, presented at Digestive Disease Week 2026, derive from SAMARA, a real-world study of semaglutide in obesity and MASLD. Participants were randomized to semaglutide up to 2.4 mg subcutaneous weekly or placebo, with blood collected at baseline and at approximately one year. Response was defined as 10% or greater annualized weight loss. Baseline cfDNA methylation profiles differentiated future semaglutide responders from non-responders before treatment began, achieving statistical significance (p < 0.01). Semaglutide-treated participants also showed longitudinal methylation shifts not observed in the placebo arm, indicating that cfDNA captures treatment-associated biological changes distinct from natural variation.

Pathway enrichment analysis revealed two distinct signatures. For pre-treatment response prediction, the predominant signal involved hepatic triglyceride export and lipid metabolism pathways, as well as adipocyte differentiation. For treatment monitoring, longitudinal changes mapped to de novo lipogenesis, incretin signaling, cholesterol homeostasis, and anti-fibrotic tissue remodeling, indicating a molecular footprint that extends beyond weight reduction alone.

The results are announced as the first known demonstration that pre-treatment epigenetic signals in circulating cfDNA can distinguish semaglutide responders from non-responders, and that therapy-associated biology can be tracked directly in blood over time. The abstract, “A Cell-Free DNA Methylation-Based Liquid Biopsy for Semaglutide Weight-Loss Response in At-Risk MASH,” was presented on May 5, 2026, in the session “Weight Loss: Pharmacotherapy & Precision Medicine.” Hepta reports it is engaging with pharmaceutical partners across the metabolic disease landscape and pursuing expanded clinical programs to validate cfDNA-based response prediction across additional mechanisms.

“The important signal in this study is baseline biology,” said Soheil Damangir, Ph.D., CTO and co-founder of Hepta. “Future responders were already biologically different before treatment began, including signals tied to how the fat is synthesized and how it is stored. That suggests response is not solely created by the drug, it is partly written into a patient's biology before treatment begins. Some patients may begin treatment in a biological state that makes them more likely to respond, and we can read that state from blood.”

“GLP-1s are being prescribed at a scale the healthcare system has never managed before, but clinicians still cannot tell how patients are biologically positioned to respond before they start treatment, or how these drugs are affecting their physiology once treatment begins,” said Hamed Amini, Ph.D., CEO and co-founder of Hepta.

“The ability to identify response-linked signatures from a blood draw, based on the body’s current biological state, before committing a patient to a long course of treatment, is exactly the kind of precision medicine capability this field has so far lacked. What this study shows is that response biology is visible in blood before treatment begins, and that the biological effects of treatment can be tracked the same way. That moves GLP-1 prescribing from trial and error toward biology-guided treatment,” added Amini.

Related Links
Hepta