Body Mass Index and Your Exercise Response
By the ExomeDNA Research Team | Last reviewed May 29, 2026
Body mass index (BMI) is a population-level metric that researchers use to track weight relative to height across large groups — and for a subset of the genes linked to it, how much you exercise determines whether that genetic signal even shows up. A 2017 meta-analysis by Graff and colleagues analyzed gene-by-physical-activity interactions across more than 200,000 individuals, identifying variants whose effect on BMI changes depending on activity level. The genes implicated include BDNF, ADCY3, ADCY9, and BBS4 — molecules that operate in the brain's appetite circuits, in the primary cilia of hypothalamic neurons, and in the signaling pathways that connect movement to energy balance. Below: what each gene does, what the interaction research actually shows, and what that means for you.
This page contains general information only. For personal health decisions, consult a qualified clinician.
What is BMI?
BMI is a number calculated from height and weight — weight in kilograms divided by height in meters squared. Public health researchers use it as a population-level screening tool because it is easy to measure at scale and correlates broadly with body fat percentage in large studies. It is not a diagnostic measurement of body composition, health, or fitness for any individual person.
The important limitation is well established: BMI does not distinguish between muscle mass and fat mass, varies across ancestral groups, and tells you nothing about where body fat is stored. A person with substantial muscle mass can register a high BMI with little metabolic risk; a person with a "normal" BMI can carry metabolically active visceral fat. Researchers use BMI because it is measurable and heritable across millions of participants — not because it is a precise individual health indicator. The genetics of BMI, including the variants on this page, reflect population-level tendencies, not individual outcomes.
Heritability studies suggest that roughly 40–70% of BMI variation across a population can be attributed to genetic differences. The 2017 Graff et al. meta-analysis studied a specific layer of that genetic architecture: not which variants are associated with higher BMI on average, but which variants interact with physical activity such that their effect on BMI differs between active and sedentary people (Graff M et al., 2017, PMID: 28448500).
The genetics behind BMI and physical activity
Four genes appear among the variants implicated in the physical activity interaction analysis.
BDNF — brain-derived neurotrophic factor. BDNF is the primary neurotrophic factor for hippocampal and cortical neurons, meaning it supports their survival, growth, and synaptic plasticity. In the context of body weight, its most important role is in the hypothalamus: hypothalamic BDNF reduces food intake by activating TrkB receptors on POMC neurons, a population of cells that signal satiety. When hypothalamic BDNF signaling is low, appetite increases. When it is high, appetite is suppressed.
The connection to physical activity is direct and well-characterized. Aerobic exercise reliably increases BDNF levels in both the brain and blood plasma — this is one of the primary molecular mechanisms by which exercise improves mood and cognition, and it may partly explain why regular exercise supports a lower BMI. A commonly studied BDNF variant, the Val66Met polymorphism (rs6265), reduces activity-dependent BDNF secretion. Carriers release less BDNF in response to exercise, meaning one of the primary channels through which exercise suppresses appetite is partially attenuated — potentially reducing the BMI benefit of a given activity dose.
ADCY3 — adenylyl cyclase 3. ADCY3 is expressed in olfactory sensory neurons and in the primary cilia of hypothalamic neurons. Its job is to generate cyclic AMP (cAMP) in response to signals from leptin and melanocortin — the hormones that tell the brain the body is sufficiently fed. This cAMP signal propagates appetite suppression. ADCY3-deficient mice develop severe obesity, establishing it as one of the more compelling single-gene obesity loci. In humans, ADCY3 variants have been associated with obesity across multiple ancestral populations.
The The primary cilia on hypothalamic neurons act as metabolic sensor antennae. ADCY3 translates leptin and melanocortin receptor activation into cAMP. When ADCY3 function is reduced, metabolic sensing is impaired and the appetite-suppression circuit receives a weaker signal.
BBS4 — Bardet-Biedl syndrome 4. BBS4 is a component of the BBSome, a protein complex required for the trafficking of specific receptors to and from the surface of primary cilia. The most important BBSome cargo for metabolic regulation is the leptin receptor: for the leptin receptor to sit at the ciliary tip where it can detect circulating leptin, it needs the BBSome to transport it there. When BBSome function is disrupted, leptin receptors are not properly trafficked, the cilia cannot sense leptin, and the brain enters a state that resembles leptin resistance — hunger signals continue despite adequate fat stores. This is why mutations in BBS4 and other BBS genes cause the severe obesity seen in Bardet-Biedl syndrome. Common BBS4 variants that fall short of causing the syndrome may modestly impair ciliary leptin sensing, contributing to the population-level BMI variation detected in GWAS.
ADCY9 — adenylyl cyclase 9. ADCY9 generates cAMP in cardiac and metabolic tissues and is involved in beta-adrenergic signaling — the pathway through which epinephrine (adrenaline) activates during exercise to increase energy expenditure. The connection to physical activity interaction is conceptually coherent: if ADCY9 variants affect how efficiently beta-adrenergic activation translates into energy expenditure during exercise, that could alter how much BMI benefit a person derives from a given exercise bout.
What the research says
Research base: Moderate. The Graff et al. 2017 meta-analysis examined gene-by-physical-activity interactions across more than 200,000 individuals from studies participating in the CHARGE consortium and related consortia. This is a large sample for an interaction analysis — detecting gene-by-environment effects requires substantially larger cohorts than detecting main effects, because the signal is the difference between how a variant behaves in active versus sedentary people rather than the variant's average effect across everyone. The meta-analytic design, pooling data across multiple cohorts with harmonized physical activity and BMI measurements, was the right approach for this question (Graff M et al., 2017, PMID: 28448500).
The study found that physical activity modifies the BMI effects of several genetic variants — in some cases substantially attenuating genetic risk, in others showing that the variant's effect on BMI appears primarily or exclusively in sedentary individuals. This is the key scientific finding: for interaction variants, sedentary behavior is not a neutral baseline. It is the condition under which the genetic signal is expressed.
Study scale: The meta-analysis incorporated data from more than 200,000 individuals across multiple ancestry groups, making it among the largest gene-by-environment interaction studies for adiposity at the time of publication.
Interaction design: Gene-by-physical-activity interaction GWAS differ from standard BMI GWAS. The statistic of interest is whether the variant's effect on BMI differs between active and inactive individuals — a test of effect modification, not just association. This requires joint analysis of main effects and interaction terms, which the Graff et al. design executed across multiple contributing cohorts.
Important limitations: physical activity was self-reported in most contributing studies, introducing measurement error that can attenuate interaction effects. Self-reported activity is also subject to social desirability bias. The variants identified explain a modest portion of the genetic contribution to BMI; no single variant is determinative. For full methodological detail — see our methodology page for how we evaluate interaction GWAS and assign confidence tiers.
How this trait affects you
This GWAS studied gene-by-physical-activity interaction on BMI. What that means practically is that your result reflects variants whose BMI signal is modulated by how physically active you are — not just variants that predict a fixed BMI tendency regardless of behavior.
For a person with higher-loading variants on this particular interaction analysis: the genetic contribution to elevated BMI is more likely to be expressed when physical activity is low, and more likely to be attenuated when physical activity is consistent. This is different from BMI genetics in general. Many BMI-associated variants have effects that are relatively stable across activity levels. The variants in this analysis are specifically those where activity modifies the outcome.
The non-shaming framing matters here. BMI is a population metric, not a character judgment. Having variants that interact with physical activity does not mean you are destined for a particular weight — it means the magnitude of a genetic tendency depends partly on a behavior those with some degree of control over. It also does not mean exercise is a moral obligation or that people who struggle to exercise are at fault for any weight-related health outcomes. Weight is governed by genetics, environment, access, mental health, sleep, medications, stress, and dozens of other factors.
What the interaction genetics tell you is that, for these specific variants, the gap between active and sedentary outcomes appears larger than for average BMI variants. The lever of physical activity is somewhat larger for you than for someone whose BMI genetics operate independently of activity level.
Working with your BMI physical-activity interaction result
The mechanism biology of BDNF, ADCY3, ADCY9, and BBS4 suggests specific activity strategies that are more likely to activate the relevant pathways.
Prioritize aerobic exercise consistently. Aerobic exercise is the primary stimulus for exercise-induced BDNF release. Walking, running, cycling, swimming, and rowing all qualify. A single session provides an acute BDNF rise; chronic aerobic training produces lasting increases in baseline BDNF and associated receptor sensitivity. For BDNF-pathway variants, consistency over weeks and months matters more than intensity of individual sessions.
Include high-intensity intervals. Maximal aerobic effort produces the largest acute BDNF release. This does not require sustained high intensity — even 20-minute sessions that include several hard intervals trigger substantially larger BDNF responses than the same duration at moderate pace. If your schedule limits total exercise time, intervals give more BDNF per minute.
Add resistance training alongside cardio. Muscle mass increases basal metabolic rate through mechanisms that do not run through the BDNF or ADCY pathways. Building metabolic mass via resistance training provides a complement to the appetite-suppression effects of aerobic exercise, and the combination outperforms either alone for body composition in most research.
Protect sleep. Sleep deprivation reduces circulating BDNF and simultaneously increases ghrelin — the hunger-stimulating hormone — and decreases leptin. For someone with variants that already modestly reduce leptin-receptor signaling (BBS4) or appetite-suppression circuit efficiency (ADCY3, BDNF), a sleep debt compounds the genetic tendency. Consistent sleep of 7–9 hours protects BDNF levels and the appetite-regulation circuits these genes sit within.
Make exercise social or structured. The primary challenge with aerobic exercise consistency is not knowledge — it is follow-through. Group classes, team sports, exercise partners, and scheduled sessions with a trainer all produce better long-term adherence than individual unscheduled exercise intentions. If the genetic lever is larger for you, the adherence infrastructure around exercise matters more.
Think in months, not weeks. The gene-by-activity interaction observed in the Graff et al. meta-analysis reflects chronic activity exposure — how active participants generally were over time — not whether they exercised last Tuesday. Acute BDNF spikes from a single workout are real, but the BMI-relevant signal is accumulated regular activity.
Related traits and genes
The BDNF, ADCY3, ADCY9, and BBS4 pathways overlap with several other traits measured by ExomeDNA.
BDNF is also implicated in cognitive function, mood regulation, and exercise-related brain adaptations. If your ExomeDNA results include a page on working memory, depression tendency, or cognitive aging, BDNF may appear there as well — the same gene that helps regulate appetite via hypothalamic signaling also drives synaptic plasticity throughout the brain.
ADCY3 and BBS4 are part of the hypothalamic cilia biology that underlies several metabolic traits. Leptin signaling, insulin sensitivity, and appetite regulation all converge on the primary cilia of hypothalamic neurons. Related traits in the ExomeDNA library that touch this pathway include waist-to-hip ratio, fasting glucose, and metabolic rate.
ADCY9 and beta-adrenergic signaling intersect with cardiovascular response to exercise — heart rate variability, resting heart rate, and aerobic capacity all partly reflect how efficiently the sympathetic nervous system activates during and recovers from exertion.
For the broader BMI genetics picture beyond the physical-activity interaction layer, see the BMI genetic risk page, which covers main-effect BMI variants identified in standard GWAS without the activity-interaction design. These two pages together give a more complete picture of the genetic architecture underlying body weight regulation.
Explore related traits: Waist-to-Hip Ratio Genetics, Fasting Glucose Regulation, Resting Heart Rate, BDNF and Cognitive Function, Aerobic Capacity (VO2 Max).
Frequently asked questions
Does this result mean exercise won't work for me?
The opposite. The gene-by-physical-activity interaction research (Graff M et al., 2017) found that for the variants on this page, physical activity attenuates BMI genetic risk. Your result means these specific variants have a larger effect in sedentary people than in active people — which is a finding that favors exercise, not one that dismisses it. No genetic result eliminates the benefits of physical activity for cardiovascular health, metabolic function, and longevity, regardless of BMI outcome.
What type of exercise matters most?
Aerobic exercise is the primary driver of BDNF release, which is one of the key mechanisms linking physical activity to appetite suppression and the BMI-relevant signaling of these variants. Walking, running, cycling, rowing, and swimming all count. High-intensity intervals produce the largest acute BDNF response per minute of exercise. Adding resistance training complements the BDNF pathway by increasing metabolic mass through separate mechanisms.
Is BMI a reliable measure of my health?
BMI is a population screening tool, not an individual diagnostic. It does not distinguish muscle from fat, does not capture fat distribution, and has different predictive validity across ancestral groups. ExomeDNA reports BMI genetics because BMI is heritable, well-studied, and carries population-level risk signals — not because BMI is a complete picture of any individual's metabolic health. Your clinician can order body composition measurements that provide more individual-level precision.
What is the BDNF Val66Met variant?
Val66Met (rs6265) is a commonly studied BDNF polymorphism where a valine-to-methionine substitution at position 66 reduces activity-dependent BDNF secretion. Met allele carriers release less BDNF in response to neural activity, including exercise-induced neural activity. In the physical activity interaction context, this matters because exercise-induced BDNF is part of the mechanism by which aerobic activity suppresses appetite. Val66Met is among the BDNF variants studied in the interaction literature.
Does this mean my weight is genetic?
Partially. Twin and family studies suggest BMI heritability of roughly 40–70% across populations, meaning both genetics and environment contribute substantially. The interaction architecture here adds a layer: some of that genetic contribution is itself conditional on behavior. The Graff et al. research found variants that express their BMI effect more strongly under sedentary conditions — which means the genetic contribution to BMI for these variants is not fixed; it depends partly on how active a person is.
Wellness Information. ExomeDNA provides educational interpretation of genetic variants for general wellness purposes only. This is not a clinical finding, treatment recommendation, or clinical genetic test. Consult a healthcare provider before making medical decisions. See our methodology and test limitations for details.
References
- Graff M, Scott RA, Justice AE, et al. (2017). Genome-wide physical activity interactions in adiposity — A meta-analysis of 200,444 adults. PLOS Genetics, 13(4), e1006528. PMID: 28448500. DOI: 10.1371/journal.pgen.1006528.
This page is published by the ExomeDNA Research Team. Last reviewed: 2026-05-29.
ExomeDNA genetic results are for wellness and educational purposes only. Consult a clinician for personalized health guidance.