This paper was retracted on 22 July 2011.

As denim trumps other fabrics in durability and longevity, genes may beat the pants out of diet and lifestyle in determining whether you’ll live to 100. In a study published online yesterday in ScienceXpress, researchers scoured the genomes of more than a thousand centenarians and identified 150 gene variants that could predict, with 77 percent accuracy, whether a person survived to very old age. “Exceptional longevity is not this vacuous entity that no one can figure out,” said lead investigator Thomas Perls, Boston University School of Medicine, in a press teleconference about the new findings. “I think we've made some inroads here in demonstrating a pretty important genetic component to this wonderful trait.” Consistent with other research, the scientists found that people tend to push toward the century mark not because they lack gene variants that drive up risk of dementia and other age-related diseases. Rather, their genomes teem with longevity genes that seem to guard against the effects of disease-associated genes.

Granted, lifestyle choices do matter for reaching one’s eightieth birthday—a milestone now attainable by average folk in developed countries. Research on Seventh-Day Adventists, for example, attributes their heightened life expectancy (88 years) to the healthy habits associated with their religion. In one study (Fraser and Shavlik, 2001), lifestyle factors including diet, exercise, and past smoking habits accounted for differences of up to 10 years of life expectancy among Adventists. Genetics play a role, too, determining 20 to 30 percent of the variation in average human lifespan in studies of twins (Herskind et al., 1996). However, making it all the way to 100 is harder (only one in 6,000 people gets there) and seems to run strongly in families (Perls et al., 2002).

“Centenarians are indeed a model of aging well,” Perls told reporters. “They very much compress their disability to the end of their lives.” About 90 percent of centenarians are disability-free at an average age of 93 (Christensen et al., 2008; Terry et al., 2008), and 15-25 percent have no discernible cognitive impairment (Perls, 2004).

To explore the genetics underlying these remarkable qualities of the oldest old, first author Paola Sebastiani and colleagues performed a genomewide association study (GWAS) of 801 unrelated Caucasians enrolled in the New England Centenarian Study (NECS), and 926 controls. The NECS subjects—all born around the turn of the last century (median age 103)—represented 16 ethnic groups. Most of the controls were ethnically matched subjects from a repository maintained by Illumina, Inc., which sells the microarrays that were used to analyze some 300,000 single nucleotide polymorphisms (SNPs) in the genomes of participants in the current study.

In the initial phase, the scientists identified 70 genomewide significant SNPs that associated with exceptional longevity in the NECS cohort, and replicated 33 in an independent set of 254 centenarians and 341 controls. However, given the stringency of the statistical significance thresholds, the researchers suspected some variants had fallen through the cracks. Loosening the criteria somewhat in the second phase of analysis, the team found 150 SNPs that could predict chances of achieving extreme longevity with 77 percent accuracy in the separate cohort of centenarians and controls.

How well people aged depended on their genetic signature—in other words, how many of the 150 life-lengthening variants they had, and which ones. Based on these signatures, most centenarians could be grouped into one of 19 clusters, some correlating with reduced prevalence of age-related diseases (e.g., stroke, diabetes, dementia), others with delayed age of onset of disease. “The signatures represent different paths to exceptional longevity,” Sebastiani told reporters.

One might expect that people who survive to very old age have fewer disease-risk alleles compared with those who die younger, but the researchers found that centenarians and controls hardly differed in this regard. “What makes people live very long lives is not a lack of genetic variants that predispose them to disease, but rather an enrichment of longevity-associated variants that may be protective,” Sebastiani said. Other studies have reached similar conclusions (e.g., Bergman et al., 2007).

Among the 150 SNPs (covering 77 genes) used to predict exceptional longevity in the current analysis, a handful of variants linked to AD and/or neuronal signaling turned up. These include the infamous ApoE/TOMM40 region (aka apolipoprotein E and translocase of the outer mitochondrial membrane 40), which clearly boosts late-onset AD risk, as well as CTNNA3 (catenin alpha 3), STX8 (syntaxin 8), MSI2 (musashi homolog 2), PLCB3 (phospholipase C beta3), and CELSR1 (cadherin EGF LAG seven-pass G-type receptor 1)—all of which reached genomewide significance in the first phase of analysis. Variants of SORCS1 (sortilin-related VPS10 domain containing receptor 1) and SORCS2—type 2 diabetes genes that may also be associated with AD—showed up in the list of 150 but just missed the bar for genomewide significance in the initial analysis.

Of the AD/neuronal signaling genes, ApoE has already been linked to longevity (Christensen et al., 2006). Whether this will hold true for the others will depend on replication in future studies, noted Bruce Yankner, who studies the molecular basis of brain aging at Harvard Medical School in Boston. In the past, this has been a tall order for most disease-linked genes coming out of GWASs, though just recently a few new ones have passed this test (see ARF related news story). Alzheimer disease geneticist John Hardy, University College London, believes the current study was underpowered, and notes that the replication study was an order of magnitude smaller than recent GWASs in AD (Harold et al., 2009; Lambert et al., 2009). (See full comment below.)

Nevertheless, the analysis by Perls and colleagues may help settle a long-standing debate over the number of genes it takes to survive to 100. “We all agreed it was few, but the definition of ‘few’ was problematic,” said longevity expert Nir Barzilai of Albert Einstein College of Medicine in New York, noting that recent speculations have ranged from four to several thousand. “This paper gives you a number,” he said. “It tells you that in the general population, there are probably 150 genes that could be involved in exceptional longevity.”

This jibes with Barzilai’s own research on centenarians, which also pins the number of genes needed to distinguish century-olds from common folk somewhere between one and 200. Barzilai and colleagues have analyzed 520 centenarians and about 700 controls—all Ashkenazi Jews living in the United States. Perls’s study, on the other hand, recruited a diverse population of centenarians but adjusted for this by choosing control groups with similar proportions of the represented ethnicities. Another key difference between the studies: Perls and colleagues used Illumina technology to analyze 300,000 genomewide markers, whereas Barzilai’s team genotyped its study participants on an Affymetrix platform containing two million markers. Because the markers on different platforms “rarely overlap,” Barzilai said, comparing gene data from the two studies will require additional statistical techniques. Therefore, Barzilai could not say at this point whether any of the AD-related genes from the Boston University study overlap with those found in his analysis of Jewish centenarians. He presented preliminary data from that study last month at the 2010 American Federation for Aging Research Grantee Conference in Santa Barbara, California. In addition to finding longevity genes, Barzilai’s team wants to figure out “which aging genes they’re protecting against,” he told ARF, noting that each longevity gene may protect against several aging genes.

For example, Barzilai and colleagues have identified a functional variant of the CETP (cholesterylester transfer protein) gene that protects against cognitive decline, metabolic diseases, heart disease, and hypertension (Barzilai et al., 2006; Sanders et al., 2010; Schechter et al., 2010). Furthermore, a recent study suggests that a longevity-associated CETP variant protects against a genetic polymorphism that confers unfavorable lipoprotein profiles (Bergman et al., 2007).

Dissecting these protective pathways may be the key to giving people better health over the long haul, Barzilai noted. In his view, the importance of slowing aging in general, rather than treating individual diseases, became clear with advances in treatments for heart disease, which is now largely manageable with surgery and medications. “What happens to people who survive heart disease is that within a year they are getting AD, diabetes, or cancer,” Barzilai said. “Unless we find protective mechanisms that change the rate of aging, all we’re going to do is switch one disease for another.” A recent review proposes that DNA damage caused by long interspersed nuclear elements (LINEs)—mobile segments that hop around the genome, potentially interrupting genes—might be a common pathway linking multiple mechanisms of aging (St. Laurent III et al., 2010).

Of course, a person’s lifespan does not depend entirely on genetics. Though just one in 6,000 makes it into the centenarian club, a much higher proportion of the population seems to have the potential to do so, “in terms of having what it takes genetically,” said Perls, noting that a good 15 percent of controls in his study had genetic signatures predisposing them to exceptional longevity. What keeps the vast majority from getting there? “Maybe they also need to not smoke, not be obese, not get hit by a bus or be in a war,” Perls suggested. “There are a range of other factors determining whether they’ll have the opportunity to live to 100.”

In the future, Perls and colleagues hope to identify functional variants of the longevity genes that came out of the current study, extend their analysis into non-Caucasian centenarians, and look at how genes and environmental factors may interact to influence likelihood of extreme longevity.—Esther Landhuis


  1. My view of this study is that it is generally underpowered and that, apart from alleles with large effect sizes, it will be not reliable. The Manhattan plot shows that only six SNPs exceeded the generally accepted level of significance of 10-7 to 10-8 (Figure S5), and none of these had the comforting “climbing tower” of less well-defined SNPs to indicate that they are part of a haplotype block showing association (as one of many of the examples of this, see the Simón-Sánchez et al., 2009 SNCA “climbing towers” in their Nature Genetics study of Parkinson’s disease). This does not mean the associations are wrong, but it does mean they may not be reliable. The replication study is an order of magnitude too small.

    A general rule is, the more complex the statistical analysis, the less reliable the outcome, because one does not know how many other complex statistical algorithms were tried. The way to do GWAS is to do simple chi squareds across the SNPs, apply a conservative Bonnferroni (10-7) and believe that, given no genotyping confounders. “Real” hits usually (but by no means always) have “climbing towers of less associated SNPs on each side based on the local haplotype structure.

    In this GWAS, some hits are likely to be real (ApoE, e.g., which has such an enormous effect on AD risk that it is likely, and has previously been shown, to have an effect on longevity). In general, AD genes like, for example, CLU, will have such a small effect on AD risk (odds ratio 1.2) that their effect on longevity will be really small unless they have other effects on other diseases, too.


    . Genome-wide association study reveals genetic risk underlying Parkinson's disease. Nat Genet. 2009 Dec;41(12):1308-12. PubMed.

    View all comments by John Hardy

Make a Comment

To make a comment you must login or register.


News Citations

  1. Research Brief: Genes and Longevity Study Retracted
  2. Repeat Offenders—CLU, CR1, PICALM Hold Up in Association Studies

Paper Citations

  1. . Ten years of life: Is it a matter of choice?. Arch Intern Med. 2001 Jul 9;161(13):1645-52. PubMed.
  2. . The heritability of human longevity: a population-based study of 2872 Danish twin pairs born 1870-1900. Hum Genet. 1996 Mar;97(3):319-23. PubMed.
  3. . Life-long sustained mortality advantage of siblings of centenarians. Proc Natl Acad Sci U S A. 2002 Jun 11;99(12):8442-7. PubMed.
  4. . Exceptional longevity does not result in excessive levels of disability. Proc Natl Acad Sci U S A. 2008 Sep 9;105(36):13274-9. PubMed.
  5. . Disentangling the roles of disability and morbidity in survival to exceptional old age. Arch Intern Med. 2008 Feb 11;168(3):277-83. PubMed.
  6. . Centenarians who avoid dementia. Trends Neurosci. 2004 Oct;27(10):633-6. PubMed.
  7. . Buffering mechanisms in aging: a systems approach toward uncovering the genetic component of aging. PLoS Comput Biol. 2007 Aug;3(8):e170. PubMed.
  8. . The quest for genetic determinants of human longevity: challenges and insights. Nat Rev Genet. 2006 Jun;7(6):436-48. PubMed.
  9. . Genome-wide association study identifies variants at CLU and PICALM associated with Alzheimer's disease. Nat Genet. 2009 Oct;41(10):1088-93. PubMed.
  10. . Genome-wide association study identifies variants at CLU and CR1 associated with Alzheimer's disease. Nat Genet. 2009 Oct;41(10):1094-9. PubMed.
  11. . A genotype of exceptional longevity is associated with preservation of cognitive function. Neurology. 2006 Dec 26;67(12):2170-5. PubMed.
  12. . Association of a functional polymorphism in the cholesteryl ester transfer protein (CETP) gene with memory decline and incidence of dementia. JAMA. 2010 Jan 13;303(2):150-8. PubMed.
  13. . Cholesteryl ester transfer protein (CETP) genotype and reduced CETP levels associated with decreased prevalence of hypertension. Mayo Clin Proc. 2010 Jun;85(6):522-6. PubMed.
  14. . A LINE-1 component to human aging: do LINE elements exact a longevity cost for evolutionary advantage?. Mech Ageing Dev. 2010 May;131(5):299-305. PubMed.

External Citations

  1. New England Centenarian Study

Further Reading


  1. . A LINE-1 component to human aging: do LINE elements exact a longevity cost for evolutionary advantage?. Mech Ageing Dev. 2010 May;131(5):299-305. PubMed.
  2. . Life-long sustained mortality advantage of siblings of centenarians. Proc Natl Acad Sci U S A. 2002 Jun 11;99(12):8442-7. PubMed.
  3. . Buffering mechanisms in aging: a systems approach toward uncovering the genetic component of aging. PLoS Comput Biol. 2007 Aug;3(8):e170. PubMed.
  4. . Exceptional longevity does not result in excessive levels of disability. Proc Natl Acad Sci U S A. 2008 Sep 9;105(36):13274-9. PubMed.
  5. . Disentangling the roles of disability and morbidity in survival to exceptional old age. Arch Intern Med. 2008 Feb 11;168(3):277-83. PubMed.

Primary Papers

  1. . Genetic signatures of exceptional longevity in humans. Science. 2010 Jul 1;2010 PubMed. RETRACTED