How “genetic” are BMI, blood sugar, blood pressure and cholesterol? It sounds like a question with one number as the answer. A new 2026 study shows why it is not.
Researchers combined data from 2,548 participants in the Chinese National Twin Registry with SNP-based heritability analyses based on GWAS summary data from Asian cohorts of up to 92,615 participants. The two methods both detected inherited variation — but they did not capture the same amount of it.
Quick answer
If cardiometabolic traits are heritable, why do twin and GWAS estimates differ?
They estimate genetic variation through different data and assumptions. In this study, twin-based heritability ranged from 34% to 72%, while SNP-based estimates from GWAS data ranged from 10% to 22%. The gap does not mean one method is “right” and the other “wrong”, and simple genetic dominance did not broadly explain it.
Key takeaways
- The study analysed ten cardiometabolic traits using 2,548 Chinese twin-registry participants and GWAS summary data from Asian cohorts of up to 92,615 participants.
- Twin-based heritability estimates ranged from 0.34 to 0.72; SNP-based heritability estimates ranged from 0.10 to 0.22.
- Statistically significant dominant genetic variance was reported only for BMI, at about 0.42, and dominance did not provide a broad explanation for the twin–GWAS gap.
- Twelve gene–environment interaction effects involving factors such as age, smoking, drinking, education and mental health were statistically significant but small.
- The exact estimates come from Asian datasets. They are not Swedish-population estimates and do not predict one person’s future BMI, glucose, blood pressure or cholesterol.
Two genetic lenses on the same traits
Classical twin models use the different genetic similarity of monozygotic and dizygotic twins to estimate how much population variation is consistent with genetic and environmental components. SNP-based GWAS analyses instead use measured common DNA variants across large samples.
Those approaches overlap, but they do not measure inherited variation in exactly the same way. Comparing them can therefore reveal both agreement and a gap rather than one definitive “genetic percentage”.
The twin estimates were higher than the SNP-based estimates
Across BMI, waist-to-hip ratio, HbA1c, fasting glucose, systolic and diastolic blood pressure, total cholesterol, triglycerides, LDL-C and HDL-C, twin-based heritability estimates ranged from 0.34 to 0.72. SNP-based estimates derived from GWAS data were lower, ranging from 0.10 to 0.22.
Heritability here describes variation between people in the studied populations under the studied conditions. It does not mean, for example, that one person’s cholesterol is “72% genetic”.
Did genetic dominance explain the gap?
The researchers tested whether non-additive or dominant genetic effects could help explain why twin estimates were higher. Statistically significant dominant genetic variance was detected only for BMI, at about 0.42 in the twin modelling.
The GWAS summary-statistic analysis did not show broad evidence of non-additive effects. The study therefore does not support the simple conclusion that genetic dominance solves the so-called missing-heritability gap.
Twelve interactions — but not twelve genes being “switched on”
The study reported 12 small but statistically significant gene–environment interaction effects involving age, smoking, drinking, education and mental health. An interaction means that genetic and environmental variation were statistically related in how a trait varied across the studied population.
It does not mean that researchers found 12 genes that lifestyle switches on or off. It also does not prove that changing one exposure will override inherited susceptibility for an individual.
The traits also shared genetic architecture
Pairwise genetic correlations between cardiometabolic traits ranged from about 0.09 to 0.90. That suggests some traits share more inherited architecture than others.
A genetic correlation is not the same as saying two conditions are the same disease or that one trait causes another. It describes overlap in genetic influences at population level.
Why the Asian population matters
The authors explicitly framed the work around Asian populations. The twin participants came from the Chinese National Twin Registry and the GWAS summary data came from Asian cohorts.
The study therefore adds population-specific evidence from Asian datasets rather than supplying a universal heritability estimate. It also means TwinPare should not transplant the exact estimates into a Swedish or global population as if they were universal constants.
What this means — and does not mean
The study strengthens the evidence that cardiometabolic traits are influenced by inherited biology while also showing that one genetic method does not capture every component estimated by another. Genes matter, but “genetic” is not synonymous with fixed, inevitable or individually predictable.
TwinPare Research can explain these population patterns and connect them with related evidence. TwinPare Health & Fitness must not turn them into a personal cardiometabolic diagnosis, genetic-risk score or treatment recommendation.
Source notes
The source has been verified and editorially reviewed for this article. The limitations below show which level of conclusion the sources support.
- [hong-cardiometabolic-twin-gwas-2026] Revealing genetic foundations underlying cardiometabolic traits: integrating twin and GWAS data in Asian populations Xuanming Hong; Ming Li; Weihua Cao; Jun Lv; Canqing Yu; Tao Huang; Dianjianyi Sun; Chunxiao Liao; Yuanjie Pang; Runhua Hu; Ruqin Gao; Min Yu; Jinyi Zhou; Xianping Wu; Yu Liu; Shengli Yin; Wenjing Gao; Liming Li. International Journal of Obesity, 2026. Evidence type: Primary study integrating Chinese twin data with SNP-based heritability analyses based on Asian GWAS summary data Limitation: The study combines 2,548 Chinese National Twin Registry participants with GWAS summary data from up to 92,615 participants in Asian cohorts. Exact heritability and interaction estimates are population- and method-specific and must not be treated as personal genetic percentages or Swedish-population estimates. PubMed DOI
Editorial source review
This section shows how the article's key factual claims are linked to the source.
Phrasings that require caution
- Heritability is a population variance estimate, not a personal genetic percentage.
- Use “SNP-based heritability analyses based on GWAS data”; do not substitute a broader “polygenic analysis” claim for the reported method.
- Do not describe the 12 statistical interactions as genes being switched on by smoking, alcohol, stress or other exposures.
- Do not present the exact Asian-population estimates as Swedish-population estimates.
- Do not use the study to predict an individual’s BMI, glucose, blood pressure, cholesterol, diagnosis or treatment needs.
| ID | Claim | Source support | Caution |
|---|