Life span has little to do with genes, analysis of large ancestry database shows

Millions of amateur genealogists assembling family trees on Ancestry.com probably figure they’re just finding lost relatives and assessing their genetic proximity to Prince Harry, but in fact they have unintentionally made a significant contribution to science. An analysis of 54 million of the website’s public family trees finds that the heritability of life span, a hot research topic for decades, is considerably less than widely thought.

Scientists reported on Tuesday that genes accounted for well under 7 percent of people’s life span, versus the 20 to 30 percent of most previous estimates.

That low heritability “implies that it would be harder” to affect life span through genetic tinkering or other life extension ideas, said computational geneticist J. Graham Ruby of Calico Life Sciences, lead author of the study published in the journal Genetics. Google founded Calico in 2013 to find ways to combat aging.

Scholars who study the contributions of genes and environment to health weren’t surprised at the tiny heritability. “Welcome aboard!” said Dr. Sandro Galea, dean of the Boston University School of Public Health, who has studied the limits of genetic influences on complex outcomes like life span. “As long as you accept that genetic factors co-occur with environmental ones, it limits the influence that genetics can have.”

Calico and Ancestry researchers analyzed 54.43 million family trees, amassing birth and death dates for 406 million people (all de-identified, so no personal data were revealed) who were born from the 19th century to the mid-20th century. Most people born more recently are still alive, their life span TBD.

The researchers calculated the correlations of the life spans of spouses and of parent-child pairs, of siblings and first cousins, and of more distantly related pairs such as a person and his sibling’s spouse’s sibling or his spouse’s sibling’s spouse.

The first tip-off that genes aren’t the dominant influence came from the life spans of spouses: They were more similar than the life spans of sisters and brothers. Since spouses share relatively few DNA variants, that suggested a strong influence of non-genetic factors that they do share, like living far from disease outbreaks, having access to clean water, being literate, eating healthy food, and not smoking.

Even more striking were the high correlations of the life spans of people related even more remotely by marriage and not by blood, such as siblings-in-law, first-cousins-in-law, aunts- and uncles-in-law, and first cousins-once-removed-in-law. If a person lived to a ripe old age, or died young, so did his or her sibling’s spouse’s sibling, spouse’s sibling’s spouse, or other distant in-law.

“As we got out to more and more distant relatives, the correlations didn’t drop by as much” as they should have if genes determined life span, Ruby said. One way to explain that, as well as the high correlation between the life spans of in-laws, is if people are choosing spouses with whom they share traits that are important for life span. If income boosts life span (it does), and 1 percenters marry other 1 percenters (especially in the eras the scientists analyzed), then spouses’ life spans would be highly correlated for that reason, as would in-law life spans.

“You’re more likely to have a life span similar to that of your in-laws than to an equally unrelated stranger,” said Catherine Ball, Ancestry’s chief scientific officer and the study’s senior author. In addition, by choosing spouses like themselves, people in effect give their kids double doses of factors that affect life span but are not genetic, such as attitudes about risk-taking or consuming alcohol.

It’s rare “for a teetotaler to marry a party girl or an ultra-marathoner to marry a couch potato,” she said.

These hundreds of millions of birth and death dates let the Calico and Ancestry researchers calculate that factors passed down from generation to generation explain 7 percent of the difference between one person’s life span and another’s. The genetic contribution, however, is even smaller: In addition to genes, people in the U.S. and other societies with low (and declining) economic mobility also inherit non-genetic contributions to life span such as education, income, access to health care, and other sociocultural influences.

The finding that the heritability of life span from factors of all kinds is no more than 7 percent means that genes explain even less (but it still pays to pick your parents, because of the non-genetic inheritance of life span).

“This is a very nice work that again demonstrates the power of crowd sourcing genealogy,” said computational biologist Yaniv Erlich, who last year took a leave of absence from Columbia University to become chief science officer of genealogy site MyHeritage. “I also praise the authors for their scientific integrity. Most of them work in a company that tries to understand the genetics of longevity and they basically claim that the heritability of longevity is really low.”

Ancestry’s family trees contain the largest data set ever analyzed in a life span study, but smaller analyses have been ratcheting down the estimate of its heritability, too, from 20 percent or more in the 1990s. A 2011 study of Alpine populations came up with 15 percent, and a 2018 study of family trees on the genealogy site Geni.com, led by Erlich, came up with 15 percent.

Although genes seem to have only a small influence on life span, they might play a larger role in whether someone lives to a super-old age. Research on centenarians has reported a strong influence of genes, identifying a couple dozen variants that are especially common in those who live to 100.