Studies disagree on how much genes shape lifespan, and all leave room for habits

Key takeaways

• A 2018 study of Ancestry family trees put the heritability of lifespan “well below 10%” after accounting for the way spouses resemble each other¹².
• A January 2026 study in Science estimated about 50% once deaths from accidents, infections and other outside causes were set aside, though it rests on modeling assumptions³.
• In a 2025 UK Biobank study of 492,567 adults, genetic risk scores for 22 diseases explained under 2 percentage points of the variation in death, compared with about 17 points for exposures such as smoking and income⁵⁶.
• Among people at high genetic risk of heart disease, a healthy lifestyle went with about half the 10-year rate of coronary events: 5.1% versus 10.7% in one U.S. study group⁷.
• The American Heart Association’s 2022 statement says polygenic risk scores are only beginning to enter clinical practice, and an AHA-published commentary notes most are built on European-ancestry data¹⁰¹².

In 2018, researchers at Calico Life Sciences, an aging-research company, and the genealogy company Ancestry sifted family trees covering almost 500 million people². They were looking for a classic sign of inherited longevity: relatives whose life spans rise and fall together.

They found it. They also found it where it shouldn’t have been.

Siblings-in-law, cousins-in-law and other relatives by marriage had linked life spans, too². These people generally share neither genes nor a household². Lead author J. Graham Ruby said the in-laws didn’t line up quite as closely as blood relatives, but it was close².

The explanation was assortative mating: people tend to pair up with partners like themselves¹². Once the team accounted for it, the heritability of lifespan dropped from the usual 15% to 30% to “well below 10%”¹.

Then, in January 2026, a team at Israel’s Weizmann Institute of Science published an estimate of about 50% in the journal Science³⁴.

Both studies can be right, because they answer different questions. This guide covers what each kind of study measures, where genes clearly matter most, and what habits can and can’t do when your genes are stacked against you.

Heritability describes a population, not your personal forecast

Heritability is the share of the differences between people, in a given population, time and place, that traces to differences in their genes. The 2026 authors stress that it is specific to a population, an environment and an era³. It doesn’t tell you what share of your own lifespan your genes decide.

Twin studies estimate it by comparing identical twins, who share nearly all their genes, with fraternal twins, who share about half. In Danish twins born from 1870 to 1900, 2,872 pairs in all, the standard method gives about 23%³. Older twin studies generally landed between 20% and 30%, Boston University’s Thomas Perls has noted, but those twins mostly lived only into their mid-to-late 80s¹³.

Each method has a weak spot. This table is our own summary.

Type of study	Example	Heritability estimate	What can skew it
Family trees	Ancestry records, 201812	Under 10%	Data mix many eras and places; partners resemble each other13
Twins, standard method	Danish twins born 1870–19003	About 23%	Many early deaths came from accidents and infections3
Twins, corrected by modeling	Danish and Swedish twins, 20263	About 50%	Depends on the model; no cause-of-death records to check it3
Measured DNA scores	UK Biobank, 20255	Under 2 points of mortality variation	Scores include only known genetic markers (our reading)

One disclosure: the family-tree study came from two companies, Calico and Ancestry². The 2026 study was funded mainly by European research grants and a Weizmann longevity research center³.

The 2026 estimate counts only deaths from inside the body

The Weizmann team, led by doctoral student Ben Shenhar and senior author Uri Alon, began with a problem in the old twin records⁴. Many twins born in the 1800s died of infections, accidents or violence³⁴. Those causes have little to do with how a body ages, and Shenhar said such deaths were about 10 times as common before antibiotics as they are today⁴.

The researchers call these “extrinsic” deaths³. Old registries didn’t record causes of death, so the team used a mathematical model to set such deaths aside. They tested it on Danish and Swedish twins, Swedish twins raised apart and siblings of American centenarians³.

With extrinsic deaths removed, the heritability of “intrinsic” lifespan came out around 55%, with different models giving 45% to 62%³. That’s roughly double earlier estimates, the authors write³.

The authors list the limits themselves. The method relies on twin-study assumptions, the correction depends on the model, and about half the variation in lifespan is still unexplained by genes³. Shenhar put it plainly: lifespan reflects genes, lifestyle and chance⁴.

The two headline studies don’t fully answer each other. The 2026 paper argues that the family-tree data mixed eras with very different death risks, but it didn’t test the partner-resemblance problem the 2018 study raised³. Expect more back-and-forth.

When researchers measured genes and daily life together, daily life won

A 2025 study in Nature Medicine took a different route. Led by Austin Argentieri, it measured both DNA and the exposome, the sum of a person’s life exposures, in 492,567 British adults⁵.

The team screened 164 environmental factors, from smoking to home ownership⁶. Twenty-five stood out, linked both to early death and to faster aging on a blood-protein clock⁵.

After accounting for age and sex, genetic risk scores for 22 major diseases explained less than 2 percentage points of the variation in who died⁵. The exposures explained about 17⁵.

Smoking was tied to 21 diseases, and socioeconomic factors such as household income, home ownership and employment were tied to 19⁶. Of the 25 standouts, smoking, socioeconomic status, physical activity and living conditions mattered most for both death and biological aging⁶.

Genes didn’t lose everywhere. Genetic scores explained more than exposures for dementias and for breast, prostate and colorectal cancer⁵. Exposures dominated for lung, heart and liver disease⁵.

Our reading: today’s genetic scores capture only the markers scientists have found, so “under 2 points” doesn’t mean genes barely matter. Shenhar expects hundreds or thousands of genes to shape lifespan⁴. What the study does show is that the exposures we can change carry a large share of the risk.

That message fits natural-health values. It puts more weight on how you live, and on the community around you, than on a fixed inheritance. It also argues against blame: income, housing and work, factors no one fully chooses, carried real weight, and the British Heart Foundation’s Bryan Williams said income, postcode and background shouldn’t decide who lives long⁶.

High genetic risk wasn’t a verdict for heart disease or dementia

In 2016, a New England Journal of Medicine study by Khera and colleagues pooled 55,685 people from four studies⁷⁸. People in the top fifth for genetic risk had 91% more coronary events than those in the bottom fifth⁸.

The team also scored four habits: not smoking, avoiding obesity, regular physical activity and a healthy diet⁸. A favorable lifestyle meant at least three of the four⁸.

Among people at high genetic risk, a favorable lifestyle went with a 46% lower relative risk of coronary events⁷⁸. The absolute gaps over 10 years were large⁷:

• ARIC study: 10.7% with an unfavorable lifestyle, 5.1% with a favorable one.
• Women’s genome health study: 4.6% versus 2.0%.
• Malmö diet and cancer study: 8.2% versus 5.3%.

By our arithmetic, that’s about 6 fewer events per 100 high-risk people in the first group. In a fourth study group, people with healthier habits had less calcium buildup in their heart arteries⁸. These were observational studies, so other differences between people may explain part of the gap.

Diabetes shows a similar pattern. In an AHA commentary, Henry Lin and Jerome Rotter noted that the benefits of lifestyle change and the drug metformin didn’t appear to differ by genetic risk¹².

Dementia leans more on genes, as the UK Biobank study found⁵. Even so, a 2019 JAMA study by Lourida and colleagues followed 196,383 UK adults 60 and older for about eight years⁹.

Those with high genetic risk and the least healthy habits had 2.83 times the dementia risk of people with low genetic risk and healthy habits⁹. Within the high-genetic-risk group, 1.78% with unhealthy habits developed dementia, compared with 1.13% with healthy ones⁹. If the link is causal, the authors estimate, one case would be prevented for every 121 high-risk people who improved their habits over 10 years⁹.

Two cautions apply. All participants were of European ancestry, and the habit score counted moderate drinking as healthy, reflecting how it was built⁹. A companion article in this series covers what newer research says about alcohol.

Single genes are the exceptions, and polygenic scores aren’t routine yet

Some inherited conditions turn on one gene with a big effect. Familial hypercholesterolemia, an inherited form of high cholesterol, affects about 1 in 250 people, according to an American College of Cardiology summary¹¹.

Even here, genes interact. Among people with familial hypercholesterolemia, those with a lower polygenic score for LDL cholesterol had lower LDL and less coronary disease¹¹. Our article on family health history covers inherited cancer syndromes, which relatives’ diagnoses change your screening and when genetic counseling makes sense.

Polygenic risk scores add up many small genetic effects into one number¹⁰. In 2022, an American Heart Association scientific statement reviewed them for coronary disease, high cholesterol, type 2 diabetes, atrial fibrillation and blood clots¹⁰.

Its message was cautious. The scores are only beginning to enter clinical practice, and how to interpret and use them isn’t well understood¹⁰. Health systems should first ask whether a score improves on the risk tools doctors already use¹⁰.

For coronary disease, the scores add information, but how much that matters clinically is debated¹². Most are built on European-ancestry data, which could widen health gaps¹². The statement also calls for reporting them as absolute risk rather than percentiles, the ACC’s summary notes¹¹. Our article on consumer DNA kits covers what spit tests can and can’t tell you.

Epigenetics changes how genes are read, not the genes themselves

Epigenetics describes chemical tags that help switch genes on and off. CDC explains that behaviors and environment can change these tags, which don’t alter the DNA sequence itself and can be reversed¹⁴.

Smoking offers the clearest example. Smokers tend to have fewer tags, called methyl groups, on part of a gene called AHRR, and heavy smokers have the biggest change¹⁴. After quitting, those tags can drift back toward nonsmoker levels, sometimes within a year¹⁴.

Some effects last. People exposed in the womb to the Dutch Hunger Winter famine of 1944 and 1945 still showed altered tags about 60 years later, compared with their unexposed siblings¹⁴. Tags also shift with age: in one study, a newborn had the most methylation, a 103-year-old the least and a 26-year-old fell between¹⁴.

Tags can matter for disease, too. Extra methyl tags that turn down the BRCA1 gene raise the risk of breast and other cancers, CDC notes¹⁴.

So “you can change your genes” is half right. You can’t rewrite your DNA, but your habits can influence how some of it is read. Our articles on biological age tests and anti-aging claims explain why tests built on these tags shouldn’t be read as a precise personal verdict.

Centenarians are where genes matter most

Genes seem to matter more the older you get. Boston University’s New England Centenarian Study looked at more than 1,900 sibling pairs in which at least one sibling reached 90¹³.

Compared with others born around the same time, siblings of people who reached 90 were about 1.7 times as likely to do the same. The odds rose to 3.5 times at 95, about 9 times at 100 and 35 times at 105¹³.

In Nordic twins born from 1870 to 1910, the genetic influence on survival was minimal before age 60 and rose sharply into the late 90s, the Centenarian Study team wrote in 2014¹⁵. Twin studies suggest that environment and behaviors explain most of the variation in surviving to the mid-to-late 80s¹⁵. Shenhar adds that many centenarians appear to carry protective genes⁴.

For most of us, the practical goal is a healthy 80s, and that’s the range where habits and surroundings carry the most weight¹⁵. Our articles on purpose and on healthspan cover the debate over longevity hot spots and how centenarians stay healthy longer.

Genes, lifestyle or both? A condition-by-condition guide

This table is our own summary of the studies above. “Inherited share” means how much genes explain in the cited study, not your personal odds.

Condition	Inherited share or key genes	What lifestyle changed	Where genetic testing fits
Survival to the mid-to-late 80s	Modest; genetic influence minimal before 6015	Exposures explained far more mortality variation than genetic scores5	Disease scores added little prediction5
Living past 100	Siblings of people who reached 100 were about 9 times as likely to13	Less known; genes weigh more at extreme ages1315	Family history is the main clue13
Coronary heart disease	Heritability estimated at 40% to 60%11	Healthy habits went with about half the 10-year event rate at high genetic risk7	Polygenic scores are early and not routine1012
Familial hypercholesterolemia	One gene change, about 1 in 250 people11	Other genes modify it11	Single-gene findings can be traced through relatives12
Type 2 diabetes	Polygenic scores exist10	Lifestyle benefit didn’t differ by genetic risk12	A score’s value for finding high-risk people is unclear12
Dementia	Genetic scores explained more than exposures5	Healthy habits went with 0.65 fewer cases per 100 over about 8 years at high genetic risk9	Scores in these studies were built for research9
Breast, prostate, colorectal cancer	Genetic scores explained more than exposures5	Exposures explained less than genes5	See the family-history red flags in our article on family health history
Lung, heart and liver disease	Exposures explained more than genes5	Smoking was tied to 21 diseases6	Genetic scores explained less than exposures5

What the evidence shows

Claim	Rating	Why
Your genes alone decide how long you live	Evidence says no, or unsafe	Even the highest recent estimate leaves about half to other factors; family-tree data say under 10%13
Genes play a bigger role in reaching extreme old age	Strong evidence (observational)	Sibling odds rise sharply at 100 and 1051315
Healthy habits lower heart disease risk even with high genetic risk	Strong evidence (observational)	Four studies, 55,685 people78
Healthy habits lower dementia risk even with high genetic risk	Promising but early	One large observational study; small absolute gap9
Polygenic scores should guide routine care	Not enough evidence	AHA says use is only beginning and poorly understood10
Quitting smoking lets smoking-linked DNA tags drift back	Strong evidence (observational)	CDC summary of methylation research14
Lifestyle can rewrite your DNA sequence	Evidence says no, or unsafe	Epigenetic changes don’t alter the sequence14

Questions readers ask

The bottom line

How much genes shape lifespan depends on what you count. Family trees say under 10%, standard twin studies about a quarter, and a 2026 model that sets aside accidents and infections about half. Every estimate leaves a large share to how and where people live, and the genetic share grows mainly at the very oldest ages.

For heart disease, a high genetic score didn’t decide who had events; healthy habits roughly halved the rate. For dementia, genes weigh more, but habits still moved the odds a little. Polygenic tests aren’t ready for routine care, and single-gene conditions such as familial hypercholesterolemia are the clear exceptions to a habits-first story.

Related reading

• A relative’s diagnosis can move your screening up by a decade
• DNA kits read a sliver of your genes, and the data can be sold
• Biological age tests can disagree with themselves by years
• Purpose tracks with a longer life, while the Blue Zones story is less certain
• Most of your health is shaped outside the exam room

Sources

1. Ruby JG, Wright KM, Rand KA, et al., Ball C. “Estimates of the Heritability of Human Longevity Are Substantially Inflated due to Assortative Mating.” Genetics, 2018;210(3):1109–1124 (UC San Diego CARTA record). carta.anthropogeny.org
2. Bay S. “Family Tree” (on Ruby JG, et al., Genetics, 2018). Genes to Genomes, Genetics Society of America, November 6, 2018. genestogenomes.org
3. Shenhar B, Pridham G, Lopes De Oliveira T, Raz N, Yang Y, Deelen J, Hägg S, Alon U. “Heritability of Intrinsic Human Life Span Is About 50% When Confounding Factors Are Addressed.” Science, 2026;391(6784):504–510 (University of British Columbia copy). zoology.ubc.ca
4. Reuters, via Malay Mail. “Half of How Long You Live Could Come Down to DNA, Researchers Say.” January 30, 2026. malaymail.com
5. Argentieri MA, Amin N, Nevado-Holgado AJ, et al., van Duijn CM. “Integrating the Environmental and Genetic Architectures of Aging and Mortality.” Nature Medicine, 2025;31(3):1016–1025 (University of Oxford record). ora.ox.ac.uk
6. Oxford Population Health, University of Oxford. “Lifestyle and Environmental Factors Affect Health and Ageing More Than Our Genes” (on Argentieri MA, et al., Nature Medicine, 2025). News release, February 2025. ndph.ox.ac.uk
7. Khera AV, Emdin CA, Drake I, Natarajan P, Bick AG, Cook NR, et al. “Genetic Risk, Adherence to a Healthy Lifestyle, and Coronary Disease.” New England Journal of Medicine, 2016;375(24) (abstract reposted by Infomed, Cuba). especialidades.sld.cu
8. American College of Cardiology. “Genetic Risk, Healthy Lifestyle, and Coronary Disease” (Journal Scan). ACC.org, November 2016. acc.org
9. Lourida I, Hannon E, Littlejohns TJ, Langa KM, Hyppönen E, Kuźma E, Llewellyn DJ. “Association of Lifestyle and Genetic Risk With Incidence of Dementia.” JAMA, 2019;322(5):430–437 (University of Exeter repository). pubmed.ncbi.nlm.nih.gov
10. American Heart Association. “Top Things to Know: Polygenic Risk Scores for Cardiovascular Disease” (on O’Sullivan JW, et al., Circulation, July 18, 2022). professional.heart.org
11. American College of Cardiology. “Polygenic Risk Scores for CVD: AHA Scientific Statement” (Ten Points to Remember). ACC.org, July 27, 2022. acc.org
12. Lin HJ, Rotter JI. Commentary on “Polygenic Risk Scores for Cardiovascular Disease: A Scientific Statement From the American Heart Association.” American Heart Association, July 18, 2022. professional.heart.org
13. Boston University. “Want to Live to 105? Have a Sibling the Same Age” (on Sebastiani P, et al., Journal of Gerontology: Biological Sciences). BU Today, April 7, 2015. bu.edu
14. Centers for Disease Control and Prevention. “Epigenetics, Health, and Disease.” CDC, updated January 31, 2025. cdc.gov
15. Sebastiani P, Andersen SL, McIntosh AI, et al., Perls TT. “Contribution of Familial Longevity to Living to 100.” Society of Actuaries, Living to 100 Symposium, 2014. dr.soa.org