ArticlePDF Available

Genome-Wide Association Study of Personality Traits in the Long Life Family Study


Abstract and Figures

Personality traits have been shown to be associated with longevity and healthy aging. In order to discover novel genetic modifiers associated with personality traits as related with longevity, we performed a genome-wide association study (GWAS) on personality factors assessed by NEO-five-factor inventory in individuals enrolled in the Long Life Family Study (LLFS), a study of 583 families (N up to 4595) with clustering for longevity in the United States and Denmark. Three SNPs, in almost perfect LD, associated with agreeableness reached genome-wide significance (p < 10(-8)) and replicated in an additional sample of 1279 LLFS subjects, although one (rs9650241) failed to replicate and the other two were not available in two independent replication cohorts, the Baltimore Longitudinal Study of Aging and the New England Centenarian Study. Based on 10,000,000 permutations, the empirical p-value of 2 × 10(-7) was observed for the genome-wide significant SNPs. Seventeen SNPs that reached marginal statistical significance in the two previous GWASs (p-value <10(-4) and 10(-5)), were also marginally significantly associated in this study (p-value <0.05), although none of the associations passed the Bonferroni correction. In addition, we tested age-by-SNP interactions and found some significant associations. Since scores of personality traits in LLFS subjects change in the oldest ages, and genetic factors outweigh environmental factors to achieve extreme ages, these age-by-SNP interactions could be a proxy for complex gene-gene interactions affecting personality traits and longevity.
Content may be subject to copyright.
published: 08 May 2013
doi: 10.3389/fgene.2013.00065
Genome-wide association study of personality traits in the
Long Life Family Study
HaroldT. Bae1*, Paola Sebastiani 1, Jenny X. Sun1, Stacy L. Andersen2, E. Warwick Daw 3,
AntonioTerracciano4,5 , Luigi Ferrucci 4andThomasT. Perls2
1Department of Biostatistics, Boston University School of Public Health, Boston, MA, USA
2New England Centenarian Study, Section of Geriatrics, Department of Medicine, Boston University School of Medicine, Boston, MA, USA
3Division of Statistical Genomics, Washington University School of Medicine, St. Louis, MO, USA
4National Institute on Aging, National Institutes of Health, Baltimore, MD, USA
5Department of Geriatrics, College of Medicine, Florida State University,Tallahassee, FL, USA
Edited by:
Berit Kerner, University of California
Los Angeles, USA
Reviewed by:
Dimitrios Avramopoulos, Johns
Hopkins University, USA
Judith Ann Badner, University of
Chicago, USA
Darlene A. Kertes, University of
Florida, USA
Jaime Derringer, University of
Colorado, USA
Harold T. Bae, Department of
Biostatistics, Boston University
School of Public Health, 801
Massachusetts Avenue, Boston, MA
02118, USA.
Personality traits have been shown to be associated with longevity and healthy aging. In
order to discover novel genetic modifiers associated with personality traits as related with
longevity, we performed a genome-wide association study (GWAS) on personality factors
assessed by NEO-five-factor inventory in individuals enrolled in the Long Life Family Study
(LLFS), a study of 583 families (Nup to 4595) with clustering for longevity in the United
States and Denmark. Three SNPs, in almost perfect LD, associated with agreeableness
reached genome-wide significance (p<108) and replicated in an additional sample of
1279 LLFS subjects, although one (rs9650241) failed to replicate and the other two were
not available in two independent replication cohorts, the Baltimore Longitudinal Study of
Aging and the New England Centenarian Study. Based on 10,000,000 permutations, the
empirical p-value of 2 ×107was observed for the genome-wide significant SNPs. Sev-
enteen SNPs that reached marginal statistical significance in the two previous GWASs
(p-value <104and 105), were also marginally significantly associated in this study (p-value
<0.05), although none of the associations passed the Bonferroni correction. In addition, we
tested age-by-SNP interactions and found some significant associations. Since scores of
personality traits in LLFS subjects change in the oldest ages, and genetic factors outweigh
environmental factors to achieve extreme ages, these age-by-SNP interactions could be a
proxy for complex gene–gene interactions affecting personality traits and longevity.
Keywords: neo scores, GWAS, family study, gene-environment interaction, longevity
Personality traits have been shown to be associated with important
health outcomes and longevity (Terracciano et al., 2008). Previ-
ous findings suggest that low levels of neuroticism, high levels
of conscientiousness, and high levels of extraversion are associ-
ated with reduced mortality (Friedman et al., 1993;Wilson et al.,
2004, 2005;Weiss and Costa, 2005;Chapman et al., 2011). Cente-
narian offspring have lower neuroticism and higher extraversion
in comparison to published normative data (Givens et al., 2009).
We recently assessed domains of agreeableness, conscientiousness,
extraversion, neuroticism, and openness to experience using the
NEO-five-factor inventory (NEO-FFI) (Costa and McCrae, 1992)
in subjects enrolled in the Long Life Family Study (LLFS): a family-
based longitudinal study of longevity and healthy aging (Newman
et al., 2011). The analysis replicated the association of low neu-
roticism and high extraversion with longevity and also confirmed
differences in distributions of personality scores at different ages.
Heritability estimates of agreeableness, conscientiousness,
extraversion, neuroticism, and openness assessed by the NEO-FFI
in 6148 Sardinians ranged from 17 to 33% (Pilia et al., 2006), and a
recent genome-wide association study (GWAS) and meta-analysis
identified many genetic variants associated with personality traits,
although just a few reached levels of genome-wide significance
(Terracciano et al., 2010;de Moor et al., 2012). This enrichment
of associations suggests that personality traits are likely influenced
by many genes in a complex manner, each with small effects.
The association between some personality traits and longevity
however triggers the question as to whether additional or differ-
ent genetic variants in long-lived individuals may be associated
with longevity-promoting personality traits such as low neuroti-
cism and high extraversion, and as such could contribute to longer
life span and health-span. To test this hypothesis, we conducted a
GWAS of five domains of NEO-FFI in subjects from the LLFS
and examined the results in the context of other studies, Ter-
racciano et al. (2010) and de Moor et al. (2012), and tested
gene-by-environment interactions.
Long Life Family Study
The LLFS is a study of 583 families demonstrating clustering
for longevity and healthy aging living in the United States and
Denmark. Study eligibility criteria have been described in detail
elsewhere (Newman et al., 2011) and enrollment was conducted May 2013 | Volume 4 | Article 65 | 1
Bae et al. NEO GWAS
between 2006 and 2009 from three study centers in the United
States (Boston University, University of Pittsburgh, and Colum-
bia University) and in Denmark at the University of Southern
Denmark (Sebastiani et al., 2009;Newman et al., 2011). Potential
probands were screened for familial longevity using the Family
Longevity Selection Score (FLoSS), which scores a family accord-
ing to birth-year cohort survival probabilities of the proband and
siblings (Sebastiani et al., 2009). Family eligibility criteria were a
FLoSS >7.0, a proband and at least one living sibling, who did not
have dementia, and at least one offspring also willing to partic-
ipate. The spouses were enrolled as controls of subjects enrolled
for familial longevity. Spouses of the offspring generation were
recruited as controls while spouses of the proband generation
were recruited only if their biological children were enrolled in
the study. Enrollment was closed in 2009, and since 2010 sub-
jects have been followed with annual data collections. Phase 1
of the study refers to the initial data collection when the sub-
jects were enrolled and Phase 2 refers to subsequent annual data
NEO-five-factor inventory
The NEO-FFI is a shortened version of the Revised NEO Person-
ality Inventory (NEO PI-R) and the correlation between the two
versions range from 0.75 to 0.89 (Costa and McCrae, 1992). The
NEO-FFI consists of 60 items, with 12 items for each of the five
domains of personality as compared to the NEO PI-R which has
48 items per domain. Each item in the NEO-FFI is scored using
a five-item Likert scale of agreement with each statement. During
Phase 1 of the LLFS (2006–2009), only the domains neuroticism
and conscientiousness of the NEO-FFI were administered to 4938
participants during a phone interview or an in-home interview.
During Phase 2 (2010–present) of the LLFS, all five domains of
the NEO-FFI were administered by phone or mail to all living
and willing participants, and the data were distributed by the data
management and coordinating center in two batches: an initial
batch of data from 3032 participants (July 2011) (Andersen et al.,
2012a), and an additional batch of data from 1300 participants
(November 2012). Because of some loss at follow-up and deaths
of participants, the sample size for the three domains of agree-
ableness, extraversion, and openness was 4400 (Table 1). The
way in which we received the data from the data coordinating
center created an independent split of the entire data into two
non-overlapping data sets that will be used for internal repli-
cation, even though individuals in the two batches of the data
are correlated. Note that in the analysis of the two data sets, we
fully account for this family correlation. Note also that while the
proband generation receives a full follow-up every year, the off-
spring generation receives a full follow-up every 3 years. As a result
of this staggered follow-up window, 99.5% of the samples in the
second batch are primarily comprised of subjects in the offspring
New England Centenarian Study
The New England Centenarian Study (NECS) is an ongoing study
of exceptional longevity1that began in 1995 (Andersen et al.,
Table 1 | Characteristics of studies.
N Discovery (as
of February
4595 (N, C) 244 840
2628 (A)
2631 (E)
2612 (O)
Additional data
(as of
1279 (A)
1287 (E)
1276 (O)
Age 71 (SD 16) 79 (SD 7) 59 (SD 17)
Sex 45% Males 50% Males 54% Males
Array 2.5 Million 610 Quad 550K
Reported are the characteristics of each cohort (LLFS, NECS, and BLSA).
Agreeableness, conscientiousness, extraversion, neuroticism, and openness are
denoted by A, C, E, N, and O, respectively. Due to the difference in the follow-up
windows between the proband and offspring generations, data on additional sam-
ples in the domain of agreeableness, extraversion, and openness were released
on November 2012.
2012b). Approximately 1500 centenarians, 500 offspring, and 150
spouses of offspring have been enrolled and followed annually.
Personality data for this study were collected in 2008 from 244
unrelated offspring of centenarians using the NEO-FFI question-
naire for a study of personality traits and exceptional longevity
(Givens et al., 2009).
Baltimore Longitudinal Study of Aging
Started in 1958, the Baltimore Longitudinal Study of Aging (BLSA)
is an ongoing multidisciplinary study of aging2. The community-
dwelling volunteers are assessed at scheduled visits, which include
personality assessment. A total of 840 subjects were genotyped
(using the Illumina 550K array) and completed the NEO-FFI per-
sonality questionnaire at least once. The sample included 46%
women and had a mean age of 58.5 years (SD =17) at the baseline
personality assessment.
DNA samples were genotyped at CIDR, and genotypes calls were
determined using Illumina recalibrated clusters. The LLFS GWAS
SNP data (Illumina Omni 2.5 or 2.5 million SNPs) underwent
checking and quality correction centrally at the LLFS data coordi-
nating center (Washington University, St. Louis) where a series of
standard procedures were applied. By using GRR package (Abeca-
sis et al., 2001), the GWAS data were checked with the pedigree
structure data to verify that the relationships were correct and
to avoid sample mismatches. In some cases, direct comparisons
of Y and Mitochondrial markers were also applied to verify rela-
tionships. Mendelian consistency for all SNPs was assessed with
Loki (Heath, 1997) for autosomal markers, PedCheck (O’Connell
and Weeks, 1998) for X markers or direct comparison for Y and
mitochondrial markers. All Mendel inconsistencies were removed:
Frontiers in Genetics | Behavioral and Psychiatric Genetics May 2013 | Volume 4 | Article 65 | 2
Bae et al. NEO GWAS
a sliding threshold depending on minor allele frequency (MAF)
was set and if a SNP had inconsistencies below that threshold, the
SNP was set to missing in all families with an inconsistency. If
the number of families with inconsistencies was above the thresh-
old, the SNP was removed from the analysis. The threshold was
set to remove the “worst” 0.2% of markers, and ranged from
35 for a MAF of 0.45–0.5 to 1 for a MAF below 0.1. Filters
on call rates both by SNP and by individual were also applied
as follows: Individuals with a call rate below 97.5% and SNPs
with a call rate below 98% had genotype data removed. These
steps resulted in 18 individuals and 86,233 autosomal SNPs being
removed, as well as 153,363 Mendel inconsistencies set to miss-
ing in the families in which they occurred. Approximately two
million autosomal SNPs were judged fit for analysis after this
The NECS DNA samples were genotyped at Boston University
using the Illumina Human610-Quad SNP array, with 600,000
SNPs. All samples genotyped at Boston University were processed
according to the manufacturer’s protocol and BeadStudio Software
was used to make genotype calls utilizing the Illumina pre-defined
clusters. Samples with less than a 95% call rate were removed
and SNPs with a call rate <97.5% were re-clustered. After re-
clustering, SNPs with call rates >97.5%, cluster separation score
>0.25, excess heterozygosity between 0.10 and 0.10, and MAF
>5% were retained in the analysis. We also removed samples
with inconsistent sex defined by heterozygosity of the X chro-
mosome that was not consistent with the sex recorded in the
In LLFS, imputation of un-typed genotypes was performed using
MACH (version 1.0.16) (Li et al., 2010) for pre-phasing the geno-
typic data and MINIMAC (version May 29, 2012) (Howie et al.,
2012) for actual imputation with 1000HG genotypic data (ver-
sion 2010–2011 data freeze, 2012-03-04 haplotypes) including all
races as a reference panel. A number of filters before imputing
were implemented in the LLFS genotypic data by removing mark-
ers that had MAF <1%, HWE p-value <106, if LLFS SNPs alleles
mismatched with those of 1000HG, and not present in the 1000HG
panel. As a result, 38,045,518 variants were imputed and only those
with r2>0.3 were used for analysis.
Prior to all analyses, raw scores of NEO-FFI in the LLFS and NECS
subjects were transformed into sex-specific standardized T-scores
with mean =50 and standard deviation of 10 using the sex-specific
means and standard deviations in Table B-4 on page 78 of the
NEO PI-R manual (Table S1 in Supplementary Material) (Costa
and McCrae, 1992). T-scores between 45 and 55 represent normal
values, while T-scores <45 represent lower than normal, and T-
scores >55 represent higher than normal values. Lack of departure
of the T-score from normality was verified by using the Kurtosis
test as reported in Andersen et al. (2012a).
Heritability estimates (Table S2 in Supplementary Material)
were obtained using variance components analysis implemented
in Sequential Oligogenic Linkage Analysis Routines (SOLAR)
(Almasy and Blangero, 1998). Under variance components analy-
sis, the total phenotypic variance can be modeled as a sum of an
additive genetic component and a non-additive genetic compo-
nent consisting of environmental factors and measurement errors.
The narrow-sense heritability is estimated as the ratio of additive
genetic variance to the total phenotypic variance. For each domain,
covariates included sex, field centers, and significant polynomial
terms of age.
Only subjects with completed questionnaires and genotype data
(including spouses) and with Caucasian origins were included
in the discovery analysis in the LLFS (agreeableness: n=2628;
conscientiousness: n=4590; extraversion: n=2631; neuroticism:
n=4595; openness: n=2612). This discovery set comprised data
released in February 2012 (batch 1). Again, due to the difference
in the data collection between the proband and offspring gener-
ations, the additional data we received on November 2012 (batch
2) were an independent split of the entire data and were used to
replicate the top findings from the discovery set in the domains of
agreeableness, extraversion, and openness. There were 2759 par-
ticipants, who had repeated measures on conscientiousness and
neuroticism, and the average time interval between the repeated
measures was 2.6 years. The agreement between repeated mea-
sures of conscientiousness and neuroticism was estimated using
the Spearman correlation coefficient (0.67 and 0.66, respectively)
and repeated measures were summarized by the average scores
and the average ages. No significant trend between age and dif-
ference between the two domains was observed, as reported in
Andersen et al. (2012a). The association between the five domains
of personality and the genotype for each SNP was tested in a lin-
ear mixed model with random effects per subject. The random
effects were modeled as a multivariate normal distribution with
zero mean vector and variance-covariance matrix proportional to
the kinship matrix to fully account for familial relations. Covari-
ates included sex, field centers, and significant polynomial terms
of age. Significant polynomial terms of age were searched using
the model search strategy described in Andersen et al. (2012a).
Analyses incorporating the top 10 principal components were also
conducted, but the results did not change substantially with this
adjustment. All GWAS analyses were performed with R statistical
software (version R.2.14) using the kinship” package. The addi-
tive genetic model, which codes the SNP genotype as the number
of minor alleles (0, 1, 2), was assumed. SNPs with MAF greater 5%
and genotype count >2 were used. To correct for multiple testing,
the genome-wide significance threshold of 108was used. Table 2
shows the SNPs that reached genome-wide significance. To assess
the chance of a false positive association, 10,000,000 permuta-
tion tests were performed for the SNPs that reached genome-wide
Replication in NECS and BLSA
SNPs with p-value <105in the discovery set were sought for
replication in the NECS (n=197) and BLSA (n=848). We defined
replication as SNPs having p-value <0.05 in the replication cohort
and consistent direction of effects as in the discovery set. In the May 2013 | Volume 4 | Article 65 | 3
Bae et al. NEO GWAS
Table 2 | Genome-wide significant SNPs in the GWAS of LLFS.
Beta P Beta P Beta P Beta P Beta P
Agree rs9650241 8 G 0.086 2.89 1.65E09 1.35 4.11E02 2.40 8.12E10 0.4 0.21
Agree rs2701448 8 A 0.087 2.87 1.80E09 1.35 4.04E02 2.39 9.46E10 – – –
Agree kgp6080058 8 A 0.087 2.85 2.44E09 1.29 4.78E02 2.35 1.57E09 – – –
CA, coded allele.
CAF, coded allele frequency.
Reported are the three genome-wide significant (p<108) SNPs in the discovery (LLFS) and their results in the three replication cohorts (additional sample in LLFS,
NECS, and BLSA). Only rs9650241 is found in the BLSA. SNPs that could not be tested are indicated by “–.
NECS, the association between the five domains and the genotype
for each SNP was tested using linear regression analysis, adjust-
ing for gender and significant polynomial terms of age, where
appropriate, in PLINK (Purcell et al., 2007) software (Table S3 in
Supplementary Material). SNPs that were not in the 610 Illumina
array were replaced by the closest proxy SNPs in strongest linkage
disequilibrium (r2>0.8), within a region of 50 kb. Consistency of
effects for proxy SNPs was checked by examining the coded allele
frequencies of the original SNP and the corresponding proxy SNP.
For 53 SNPs we could not find a good proxy. In the BLSA the asso-
ciation analyses were conducted using MERLIN3, and age, sex,
and principal components were used as covariates (Table S4 in
Supplementary Material).
Replication of findings from other GWASs
All published results from Terracciano et al. (2010) and de Moor
et al. (2012) with p-value <104and 105(available through
online Supplementary Material), respectively, were tested in the
LLFS using the full sample data as of November 2012 and NECS
sets. The intention of this replication test was to examine whether
any variants that were shown to be significantly associated with
personality traits in the previous GWASs also exhibit significant
associations in the LLFS and/or NECS. In the case of non-matching
SNPs, imputed dosages were used for the LLLF set. The lowest r2,
a measure of correlation between the imputed genotype and true
genotype, was 0.74 for the SNPs we tested. Proxy SNPs in the
NECS set were searched within the region of up to ±50,000 base
pairs and r2>0.8. Again, consistency of effects for proxy SNPs was
checked by examining the coded allele frequencies of the original
SNP and the corresponding proxy SNP. Note that only one of the
SNPs from Terracciano et al.’s study reached marginal statistical
significance in the meta-analysis of de Moor et al.
SNP-by-age interaction
In order to identify SNPs whose effects change with partici-
pants’ age on each NEO domain, we tested the significance of a
SNP ×Age interaction term for those SNPs with significant main
effects (p-value <106), where age represents the participant’s
age at the assessment of NEO Genome-wide testing for inter-
action would require too much power, as reported in Thomas
(2010). Therefore, we chose the p-value of 106to limit the num-
ber of significant testing for interaction to 6, 7, 24, 37, and 7
SNPs respectively for agreeableness, conscientiousness, extraver-
sion, neuroticism, and openness, and have some statistical power.
Because of the smaller number of tests, a p-value <0.05 was used
for statistical significance of the interaction term.
Characteristics of each cohort are summarized in Table 1. The her-
itability estimates for the five domains and p-values for statistical
significance are presented in Table S2 in Supplementary Mater-
ial and show that all five domains of personality are heritable.
Openness was the most heritable (h2=49%), while Agreeable-
ness was the least heritable (h2=18%). The heritability estimates
for conscientiousness, extraversion, and neuroticism were 30, 32,
and 25%, respectively. With the exception of the higher heritabil-
ity of openness, the other estimates were comparable to those
reported in Pilia et al. (2006). The QQ-plots and Manhattan plots
of the GWAS from the LLFS are shown in Figure 1. Three SNPs
reached genome-wide significance (Table 2) in the initial GWAS
of the LLFS. The top findings (p-value <105) from the LLFS
can be found in Table S5 in Supplementary Material (21 SNPs
associated with agreeableness; 26 SNPs associated with conscien-
tiousness; 7 SNPs associated with extraversion; 12 SNPs associated
with neuroticism; 9 SNPs associated with openness). None of the
SNPs that could be tested in the NECS (22 SNPs) or BLSA (24
SNPs) reached statistical significance. Table 3 provides a sum-
mary of the replicated results, and Table 4 lists the SNPs in which
the genetic effect changes with age. Eighty-one SNPs had signif-
icant main effects (p-value <106), and were included in the
analysis of significant interactions. There were seven SNPs that
had significant interactions (four in extraversion, two in neuroti-
cism, and one in openness). Next, results of specific domains are
Three SNPs on chromosome 8, which are in almost perfect
LD, reached genome-wide significance in the first batch of data
Frontiers in Genetics | Behavioral and Psychiatric Genetics May 2013 | Volume 4 | Article 65 | 4
Bae et al. NEO GWAS
FIGURE 1 | (A) Q-Q plots for five domains of NEO-FFI.The y-axis is
the quantiles of observed p-values and the x-axis is the quantiles from
the expected distribution; (B) Manahattan plots for five domains of
NEO-FFI. The y-axis is the -log(p) and the x-axis is the genomic
locations for each SNP ordered by chromosome and base pair
positions. May 2013 | Volume 4 | Article 65 | 5
Bae et al. NEO GWAS
Table 3 | Summary of SNPs from previous GWAS that replicated in the LLFS and the NECS.
Domain SNP Chr Gene CA CAF Beta PCAF Beta P
Discovery in SardiNIA Replication in LLFS
Agreeableness rs7042201 9 T 0.22 0.14 5.80E05 0.26 0.51 0.04
Agreeableness rs382847 9 C 0.39 0.13 3.90E05 0.26 0.53 0.04
Agreeableness rs3204145 9 IKBKAP T 0.16 0.15 5.30E05 0.19 0.55 0.05
Agreeableness rs6484998 11 GALNTL4 C 0.18 0.15 2.30E05 0.18 0.66 0.02
Agreeableness rs7121652 11 GALNTL4 T 0.17 0.15 4.50E05 0.19 0.62 0.03
Extraversion rs17147371 11 PACS1 C 0.18 0.17 4.00E05 0.19 0.66 0.04
Neuroticism rs2039528 1 PTPRF G 0.39 0.14 1.60E05 0.35 0.46 0.02
Neuroticism rs11210864 1 PTPRF A 0.38 0.14 2.20E05 0.32 0.48 0.02
Neuroticism rs10890251 1 PTPRF C 0.38 0.14 2.20E05 0.33 0.44 0.03
Neuroticism rs6687571 1 PTPRF A 0.39 0.13 3.60E05 0.35 0.46 0.02
Neuroticism rs2926458 10 SORCS3 T 0.07 0.24 6.90E05 0.07 0.87 0.03
Neuroticism rs1606865 12 TMEM16D G 0.4 0.13 9.30E05 0.41 0.53 0.007
Discovery in SardiNIA Replication in NECS
Agreeableness rs2202069 15 T 0.35 0.12 3.40E05 0.32 2.62 0.02
Neuroticism rs1421989 5 C 0.08 0.24 1.94E05 0.16 3.43 0.03
Neuroticism rs7317522 13 T 0.45 0.12 7.28E05 0.45 2.39 0.02
Discovery in de Moor et al. (2012) Replication in NECS
Neuroticism rs12513013 4 SHROOM3 C 0.45 9.70E06 0.31 2.73 0.02
Neuroticism rs7212729 17 BCAS3 G 0.63 6.10E06 0.16 4.34 0.005
CA, coded allele.
CAF, coded allele frequency.
Table 4 | SNPs with significant interaction with age.
Domain SNP Chr Gene CA CAF SNP Age SNP ×Age
Beta SE P Beta SE P Beta SE P
Extra rs79926910 20 KIAA1755 G 0.37 7.48 1.46 3.02E07 0.16 0.019 4.54E17 0.092 0.019 1.01E06
Extra rs877600 20 KIAA1755 A 0.37 7.69 1.45 1.32E07 0.16 0.019 1.63E16 0.095 0.019 4.75E07
Extra rs1205452 20 KIAA1755 C 0.37 7.27 1.46 6.70E07 0.16 0.019 1.06E16 0.090 0.019 2.23E06
Extra rs11258100 10 CCDC3 T 0.09 11.86 2.42 9.85E07 0.26 0.015 1.46E65 0.148 0.031 2.42E06
Neuro rs60933298 7 A 0.09 7.47 1.49 5.42E07 1.48 0.490 0.0026 0.097 0.021 2.89E06
Neuro rs4728985 7 C 0.09 7.48 1.46 3.42E07 1.50 0.490 0.0022 0.098 0.020 1.34E06
Open rs7817266 8 A 0.34 6.55 1.33 8.96E07 2.83 0.902 0.0017 0.083 0.017 1.38E06
CA, coded allele.
CAF, coded allele frequency.
and consistent results in the second batch of data in the LLFS
with p=0.04 (Table 2). When analyzed using the combined
sample (n=3907), the statistical significance increased to a p-
value of 8.12 ×1010 for rs9650241 (the most significant SNP).
However, these SNPs did not replicate in the BLSA (p-value
for rs9650241 was 0.21, and the other two SNPs were not
found in their array). These three SNPs were not found in the
NECS. From Terracciano et al.’s (2010) findings on agreeable-
ness (114 SNPs with p<104), 109 were found in the LLFS
GWAS and 5 replicated, and 20 were found in the NECS GWAS
and 1 replicated (Table 3). From de Moor et al.’s (2012) find-
ings on agreeableness (14 SNPs with p<105), 14 were found
in the LLFS GWAS, and 2 in the NECS GWAS, but none
No SNP reached genome-wide significance. SNP rs79732200 on
chromosome 15 in the gene IGDCC3, reached almost genome-
wide significance (p-value 9 ×108). However, this SNP was not
found in either NECS or BLSA GWAS. From Terracciano et al.’s
(2010) findings on conscientiousness (35 SNPs with p<104),
31 were found in the LLFS GWAS, and 6 were found in the
NECS GWAS, but none replicated. de Moor et al. (2012) iden-
tified rs2576037 in KATNAL2 to be genome-wide significant
(4.9 ×108), but it did not replicate in the LLFS or NECS GWAS
(p=0.7 and p=0.5, respectively). From de Moor et al.’s findings
on conscientiousness (110 SNPs with p<105), 109 were found
in the LLFS GWAS, and 26 were found in the NECS GWAS, but
none replicated.
Frontiers in Genetics | Behavioral and Psychiatric Genetics May 2013 | Volume 4 | Article 65 | 6
Bae et al. NEO GWAS
No SNP reached genome-wide significance. From Terracciano
et al.’s (2010) findings on extraversion (56 SNPs with p<104),
55 were found in the LLFS GWAS and 7 were found in the NECS
GWAS, and 1 replicated in the LLFS GWAS (Table 3). From
de Moor et al.’s (2012) findings on extraversion (30 SNPs with
p<105reported), 30 were found in the LLFS GWAS and 7
were found in the NECS GWAS, but none replicated. Four SNPs
had significant main effects term (p<106) and interaction term
(p<0.05) (Table 4). Three of these SNPs are in strong LD and are
in the gene KIAA1755 and 1 SNP in CCDC3. The plots in Figure
S1 in Supplementary Material show that the changes of the effect
of the minor allele (A) in SNP rs877600 in KIAA1755 for differ-
ent ages:carriers of the AG and GG genotype (green and blue
lines) tend to score higher than carriers of the AA genotype (red
line) in extraversion at younger ages (approximately <80 years)
but this trend is reversed in older ages. However, the estimated
score remains within “average values” until age 100. The effect of
the age ×SNP interaction for rs11258100 in CCD3 is different:
while carriers of the GG genotype (red line) tend to score lower in
extraversion at older ages, in carriers of the genotypes TG and GG
the score for extraversion appear to be stable across a wide range
of ages.
No SNP associated with neuroticisms reached genome-wide sig-
nificance. From Terracciano et al.’s (2010) findings on neuroticism
(46 SNPs with p<104), 44 were found in the LLFS GWAS and
6 replicated; 12 were found in the NECS GWAS and 2 replicated
(Table 3). From de Moor et al.’s (2012) findings on neuroticism
(36 SNPs with p<105), 35 were found in the LLFS GWAS, but
none replicated; 3 were found in the NECS GWAS and 2 replicated
(Table 3). Two SNPs in strong LD had significant main effects
(p<106) and interaction (p<0.05) terms (Table 4). In Figure
S1 in Supplementary Material, subjects with CC genotype (two
minor alleles) for SNP rs4782985 have higher neuroticism scores
at earlier ages from 40 to 70 than subjects with CT or TT genotypes,
but their scores decrease below the normative values after the age
of 80 years, while subjects with CT or TT genotypes retain the
normative values at older ages.
No SNP associated with openness reached genome-wide signifi-
cance. From Terracciano et al.’s (2010) findings on openness (62
SNPs with p<104), 57 were found in the LLFS GWAS and 18
were found in the NECS GWAS, but none replicated. de Moor
et al. (2012) reported rs1477268 and rs2032794 in RASA1 to be
genome-wide significant (p=2.8 ×108and 3.1 ×108, respec-
tively). These SNPs did not replicate in the NECS GWAS with
p=0.83 and p=0.71, respectively, and did not replicate in the
LLFS GWAS with p=0.69 and p=0.73, respectively. From de
Moor et al.’s findings on openness (39 SNPs with p<105), 39
were found in the LLFS GWAS and 14 were found in the NECS
GWAS, but none replicated. In the interaction model, 1 SNP had
significant main effects term (p<106) and interaction term
(p<0.05) (Table 4). For this SNP (rs7817266), the three geno-
type groups show similar trend where their openness scores are
higher at earlier ages approximately from 50 to 75 and scores are
lower at older ages. However, subjects with AA genotype (two
minor alleles) remain within the normative values of openness at
all ages, while subjects with AG or GG genotype have openness
scores below the normal range after the age of 90.
Prior studies have shown that personality traits have a genetic
component (Loehlin and Martinb, 2001;Bouchard and McGue,
2003;Pilia et al., 2006) and the heritability estimates derived in the
LLFS data confirm these results. Compared to a twin study (Jang
and Ver non, 1996) that reported heritability estimates of 41, 53, 61,
41, and 44%, respectively for neuroticism, extraversion, openness,
agreeableness, and conscientiousness, LLFS estimates are lower in
every domain. Compared to the recent estimates reported by Pilia
et al. (2006) from 6148 Sardinians, the estimates of heritability in
the LLFS are higher in conscientiousness, extraversion, and open-
ness. The heritability estimate was lower in LLFS for agreeableness,
but comparable for neuroticism. These differences may be due to
the fact that samples in the two studies are ethnically different
(the Sardinian sample was from a genetically isolated population)
and phenotypically different. The LLFS is a study of longevity
and families were selected for evidence of familial longevity (New-
man et al., 2011). Andersen et al. (2012a) showed that all five
domains of NEO scores have different distributions at different
ages in the LLFS subjects, and this selected population may be
enriched for variants that are associated with longevity as well as
longevity-promoting personality traits that translate into different
heritability estimates.
Intriguingly, the variants associated with agreeableness that
reached genome-wide significance in the LLFS did not replicate
in the BLSA. The three genome-wide significant SNPs in agree-
ableness showed consistent results in the two batches of samples
of the LLFS. Even though the two data sets are not independent,
the fact that these SNPs show significant associations with con-
sistent effects strengthens the validity of this finding. Among the
10,000,000 permutation tests performed, there was 1 permutation
that achieved genome-wide significance, which yields an empiri-
cal p-value of 2 ×107(North et al., 2002). The analysis using the
combined samples yielded improved statistical significance for the
three SNPs and two additional SNPs that almost attained genome-
wide significance (rs2587559 with p=4.13 ×108and rs2587561
with p=5.27 ×108), located in the promoter region of TRPA1,
which is in a close proximity (within 40,000 bp) to the original
variants. This result further corroborates the association between
agreeableness and this particular region on chromosome 8, which
is linked to tolerance to pain (Doehring et al., 2011). The three
genome-wide significant SNPs in the initial GWAS of LLFS have a
MAF of around 9%, and there are only 18 subjects out of 2622 with
the GG genotype for rs9650241 (the most significant SNP). There-
fore, it is difficult to assess the true relationship with the additive
coding of SNP genotypes. Under dominant coding (dominant for
the G allele) for rs9650241, the statistical significance improves to
ap-value of 2.1 ×1010 , compared to 1.7×109in an additive
model. We obser ved that the median agreeableness in subjects with
the G allele is above the normal range (median T-score =56.1).
The association of this locus with agreeableness is novel in the May 2013 | Volume 4 | Article 65 | 7
Bae et al. NEO GWAS
LLFS, although it did not replicate in the BLSA and it could not
be tested in the NECS. Therefore, there is no evidence on whether
the same association exists in the general population or this locus
is linked to agreeableness through longevity.
There were a total of 75 SNPs (21 SNPs associated with agree-
ableness; 26 SNPs associated with conscientiousness; 7 SNPs asso-
ciated with extraversion; 12 SNPs associated with neuroticism; 9
SNPs associated with openness, Table S5 in Supplementary Mater-
ial) with p-value <105in the GWAS of LLFS. Some of these SNPs
were in interesting genes. For example,the most significant SNP in
neuroticism (rs177389) is a missense mutation in PAPLN, which
changes the aminoacid MET into ARG. This gene was shown to be
linked to suicidal ideation during anti-depressant treatment (Laje
et al., 2009). On average, carriers of the T allele scored lower in
neuroticism; the median scores for the three genotypes (GG, GT,
TT) were 43.8, 42.7, and 41.4, respectively, all of which are below
the normative value. This SNP could not be replicated in the NECS
and BLSA GWAS. Likewise, many of the top findings from LLFS
remain to be replicated by other investigators in a larger sample
with evidence of longevity.
Interestingly, the LLFS GWAS also replicated findings from
other studies (Terracciano et al., 2010;de Moor et al., 2012). Six
SNPs in agreeableness, 10 SNPs in neuroticism, and 1 SNP in extra-
version replicated in either the LLFS or NECS GWAS (Terracciano
et al., 2010;de Moor et al., 2012), although none of the replicated
associations passed the Bonferroni correction (p=0.00015). de
Moor et al. reported several SNPs in ARNTL to be associated with
agreeableness. In both the LLFS and NECS GWAS, a cluster of
SNPs in the same gene were significantly associated with agreeable-
ness at α=0.05 (Table S6 in Supplementary Material). However,
these SNPs were not in LD with those reported in de Moor et al.
(highest r2=0.234). ARNTL, one of the circadian clock genes, is
associated with winter depression and seasonal affective disorder
(Partonen et al., 2007) as well as alcohol use disorders (Kovanen
et al., 2010). These replicated associations strengthen the original
findings in Terracciano et al. (2010),de Moor et al. (2012). Lack of
replication of the other findings may be due to different ethnicities,
social, and environmental factors.
The significant SNP-by-age interaction terms in Table 4 suggest
an interesting hypothesis that it is depicted in Figure 2. It is well
known that aging is in part determined by genetic and environ-
mental factors and while genetic factors may explain only 25% of
the variability to survive to the mid 1980s (Herskind et al., 1996),
the contribution of genes to surviving to older ages is likely much
larger (Perls et al., 2000;Sebastiani et al., 2012). Being selected
for familial longevity, participants of the LLFS may be enriched
for genetic and non-genetic variants that promote longevity, and
LLFS subjects who have reached old ages may harbor varying
proportions of longevity associated variants. In this hypothetical
context, the SNP ×age interactions found in this study may actu-
ally be surrogates for several gene ×gene and gene ×environment
interactions. This hypothesis is consistent with work in McCrae
et al. (2010) that suggested the genetic basis of personality traits
is multifactorial, it is likely determined by many interacting genes
with individual small effects, and environmental factors are also
important factors. However, additional analysis that models the
FIGURE 2 | Graphical representation of hypothesis on gene-gene
interactions. Attributes inside the large rectangular box are the variables
included in the statistical model. Age, which is associated with NEO scores,
is reflective of both genetic and environmental factors associated with
aging. As genetics factors outweigh the environmental factors in individuals
enriched for longevity, SNP-by-age interaction term in the model may imply
gene-gene interaction which plays important roles in personality traits, as
related to longevity.
genetic influence on lifespan and personality traits simultaneously
is needed to begin to test this hypothesis.
Our ability to replicate the genome-wide significant associations
discovered in the LLFS GWAS was limited by the different arrays
that were used to genotype the DNA samples from the BLSA and
the NECS. In addition, given that five phenotypes are analyzed,
a genome-wide significance threshold of 108may be too liberal
and additional replication is necessary to confirm a role of these
discovered variants in personality traits.
This study is the first GWAS on five major domains of personal-
ity traits assessed by NEO-FFI in a sample enriched for longevity.
Our results replicated a few loci identified by others, and confirm
that effects of each common genetic variant are modest. As with
many complex polygenic traits, genetic framework of personality,
as related to longevity, seems multifactorial. In particular, genetic
variants that promote longevity may interact with other variants
in the establishment of personality traits.
Frontiers in Genetics | Behavioral and Psychiatric Genetics May 2013 | Volume 4 | Article 65 | 8
Bae et al. NEO GWAS
We are grateful to the participants and family members of
the Long Life Family Study and the New England Centenarian
Study for allowing us to study how they have achieved their
longevity. This work was funded by the National Institute on
Aging (NIA cooperative agreements U01-AG023755 and U19-
AG023122 to T.P. and Intramural Research Program to L.F.),
and the National Heart Lung Blood Institute (R21HL114237
to P.S).
The Supplementary Material for this article can be found online at
Abecasis, G. R., Cherny, S. S., Cookson,
W. O., and Cardon, L. R. (2001).
GRR: graphical representation of
relationship errors. Bioinformatics
17, 742–743.
Almasy, L., and Blangero, J. (1998).
Multipoint quantitative-traitlinkage
analysis in general pedigrees. Am. J.
Hum. Genet. 62, 1198–1211.
Andersen, S. L., Sun, J. X., Sebastiani,
P., Huntly, J., Gass, J. D., Feld-
man, L., et al. (2012a). Personal-
ity factors in the Long Life Fam-
ily Study. J. Gerontol. B Psychol. Sci.
Soc. Sci. doi:10.1093/geronb/gbs117.
[Epub ahead of print].
Andersen, S. L., Sebastiani, P., Dworkis,
D. A., Feldman, L., and Perls, T. T.
(2012b). Health span approximates
life span among many supercente-
narians: compression of morbidity
at the approximate limit of life span.
J. Gerontol. A Biol. Sci. Med. Sci. 67,
Bouchard, T. J. Jr., and McGue, M.
(2003). Genetic and environmental
influences on human psychological
differences. J. Neurobiol. 54, 4–45.
Chapman, B. P., Roberts, B., and Duber-
stein, P. (2011). Personality and
longevity: knowns, unknowns, and
implications for public health and
personalized medicine. J. Aging Res.
2011, 759170.
Costa, P. T. Jr.,and McCrae, R. R. (1992).
Revised NEO Personality Inventory
(NEO-PI-R) and NEO Five-Factor
(NEO-FFI) Inventory Professional
Manual. Odessa, FL: Psychological
Assessment Resources.
de Moor, M. H., Costa, P. T., Terrac-
ciano, A., Krueger, R. F., de Geus, E.
J., Toshiko, T., et al. (2012). Meta-
analysis of genome-wide association
studies for personality. Mol. Psychia-
try 17, 337–349.
Doehring, A., Kusener, N., Fluhr,
K., Neddermeyer, T. J., Schnei-
der, G., and Lotsch, J. (2011).
Effect sizes in experimental pain
produced by gender, genetic
variants and sensitization pro-
cedures. PLoS ONE 6:e17724.
Friedman, H. S., Tucker, J. S.,
Tomlinson-Keasey, C., Schwartz,
J. E., Wingard, D. L., and
Criqui, M. H. (1993). Does
childhood personality predict
longevity? J. Pers. Soc. Psychol. 65,
Givens, J. L., Frederick, M., Silverman,
L., Anderson, S., Senville, J., Silver,
M., et al. (2009). Personality traits
of centenarians’ offspring. J. Am.
Geriatr. Soc. 57, 683–685.
Heath, S. C. (1997). Markov chain
Monte Carlo segregation and link-
age analysis for oligogenic models.
Am. J. Hum. Genet. 61, 748–760.
Herskind, A. M., McGue, M., Holm, N.
V., Sorensen, T. I., Harvald, B., and
Vaupel,J. W. (1996). The heritability
of human longevity: a population-
based study of 2872 Danish twin
pairs born 1870-1900. Hum. Genet.
97, 319–323.
Howie, B., Fuchsberger, C., Stephens,
M., Marchini, J., and Abecasis, G. R.
(2012). Fast and accurate genotype
imputation in genome-wide associ-
ation studies through pre-phasing.
Nat. Genet. 44, 955–959.
Jang, K. L. L. W., and Vernon, P. A.
(1996). Heritability of the big five
personality dimensions and their
facets: a twin study. J. Pers. 64,
Kovanen, L., Saarikoski, S. T., Haukka,
J., Pirkola, S., Aromaa, A., Lonnqvist,
J., et al. (2010). Circadian clock gene
polymorphisms in alcohol use dis-
orders and alcohol consumption.
Alcohol Alcohol. 45, 303–311.
Laje, G., Allen, A. S., Akula, N.,
Manji, H., John Rush, A., and
McMahon, F. J. (2009). Genome-
wide association study of suicidal
ideation emerging during citalo-
pram treatment of depressed outpa-
tients. Pharmacogenet. Genomics 19,
Li, Y., Willer, C. J., Ding, J., Scheet, P.,
and Abecasis, G. R. (2010). MaCH:
using sequence and genotype data
to estimate haplotypes and unob-
served genotypes. Genet. Epidemiol.
34, 816–834.
Loehlin, J. C., and Martinb, N. G.
(2001). Age changes in personality
traits and their heritabilities during
the adult years: evidence from Aus-
tralian twin registry samples. Pers.
Individ. Dif. 30, 1147–1160.
McCrae, R. R., Scally, M., Terracciano,
A., Abecasis, G. R., and Costa, P.
T. Jr. (2010). An alternative to the
search for single polymorphisms:
toward molecular personality scales
for the five-factor model. J. Pers. Soc.
Psychol. 99, 1014–1024.
Newman, A. B., Glynn, N. W.,Taylor,C.
A., Sebastiani, P., Perls,T. T.,Mayeux,
R., et al. (2011). Health and function
of participants in the Long Life Fam-
ily Study: a comparison with other
cohorts. Aging 3, 63–76.
North, B. V., Curtis, D., and Sham, P. C.
(2002). A note on the calculation of
empirical P values from MonteCarlo
procedures. Am. J. Hum. Genet. 71,
O’Connell, J. R., and Weeks, D. E.
(1998). PedCheck: a program for
identification of genotype incom-
patibilities in linkage analysis. Am.
J. Hum. Genet. 63, 259–266.
Partonen, T., Treutlein, J., Alpman, A.,
Frank, J., Johansson, C., Depner, M.,
et al. (2007). Three circadian clock
genes Per2, Arntl, and Npas2 con-
tribute to winter depression. Ann.
Med. 39, 229–238.
Perls, T., Shea-Drinkwater, M., Bowen-
Flynn, J., Ridge, S. B., Kang, S., Joyce,
E., et al. (2000). Exceptional famil-
ial clustering for extreme longevity
in humans. J. Am. Geriatr. Soc. 48,
Pilia, G., Chen, W. M., Scuteri, A.,
Orru, M., Albai, G., Dei, M., et al.
(2006). Heritability of cardiovascu-
lar and personality traits in 6,148
Sardinians. PLoS Genet. 2:e132.
Purcell, S., Neale, B., Todd-Brown, K.,
Thomas, L., Ferreira, M. A., Bender,
D., et al. (2007). PLINK: a tool set
for whole-genome association and
population-based linkage analyses.
Am. J. Hum. Genet. 81, 559–575.
Sebastiani, P., Hadley, E. C., Province,
M., Christensen, K., Rossi, W., Perls,
T.T., et al. (2009). A family longevity
selection score: ranking sibships by
their longevity, size, and availabil-
ity for study. Am. J. Epidemiol. 170,
Sebastiani, P., Solovieff, N., Dewan, A.
T., Walsh, K. M., Puca, A., Hart-
ley, S. W., et al. (2012). Genetic
signatures of exceptional longevity
in humans. PLoS ONE 7:e29848.
Terracciano, A., Lockenhoff, C. E.,
Zonderman, A. B., Ferrucci, L.,
and Costa, P. T. Jr. (2008). Per-
sonality predictors of longevity:
activity, emotional stability, and
conscientiousness. Psychosom. Med.
70, 621–627.
Terracciano, A., Sanna, S., Uda, M.,
Deiana, B., Usala, G., Busonero, F.,
et al. (2010). Genome-wide associa-
tion scan for five major dimensions
of personality. Mol. Psychiatry 15,
Thomas, D. (2010). Gene –
environment-wide association
studies: emerging approaches. Nat.
Rev. Genet. 11, 259–272.
Weiss, A., and Costa, P. T. Jr. (2005).
Domain and facet personality pre-
dictors of all-cause mortality among
medicare patients aged 65 to 100.
Psychosom. Med. 67, 724–733.
Wilson, R. S., Krueger, K. R., Gu, L.,
and Bienias, J. L. (2005). Mendes
de Leon CF, Evans DA. Neu-
roticism, extraversion, and mor-
tality in a defined population of
older persons. Psychosom. Med. 67,
Wilson, R. S., Mendes de Leon, C. F.,
Bienias, J. L., Evans, D. A., and Ben-
nett, D. A. (2004). Personality and
mortality in old age. J. Gerontol. B
Psychol. Sci. Soc. Sci. 59, 110–116.
Conflict of Interest Statement: The
authors declare that the research was
conducted in the absence of any com-
mercial or financial relationships that
could be construed as a potential con-
flict of interest.
Received: 22 December 2012; accepted:
09 April 2013; published online: 08 May
Citation: Bae HT, Sebastiani P, Sun
JX, Andersen SL, Daw EW, Terrac-
ciano A, Ferrucci L and Perls TT
(2013) Genome-wide association study
of personality traits in the Long Life
Family Study. Front. Genet. 4:65. doi:
This article was submitted to Frontiers
in Behavioral and Psychiatric Genetics,
a specialty of Frontiers in Genetics.
Copyright © 2013 Bae, Sebastiani, Sun,
Andersen, Daw, Terracciano, Ferrucci
and Perls. This is an open-access arti-
cle distributed under the terms of the
Creative Commons Attribution License,
which permits use, distribution and
reproduction in other forums, provided
the original authors and source are cred-
ited and subject to any copyright notices
concerning any third-party graphics etc. May 2013 | Volume 4 | Article 65 | 9
... Openness to experience is the feature with the highest index of inheritance at 61% [7]. Genome-wide association study (GWAS) of personality traits showed higher values of heritability for conscientiousness (30%), extraversion (35%) and neuroticism (25%) [10]. Nevertheless, studies on human personality are not only limited to five dimensions, but additionally include other factors such as hereditary and environmental influence [11]. ...
... Although single-gene study by Uhart et al. found as association of GABRA6 gene polymorphism with personality trait, GWAS did not confirm it [10,[22][23][24][25]. The study of Uhart and co-workers was carried out on a small sample size-56 cases varied in terms of race (N = 40 Caucasian, N = 11 African American, N = 5 Asian) [22]. ...
... This locus is in close proximity of CNR1 gene. However, GWAS analyses did not confirm the relationship of CNR1 and NR3C1 loci with personality traits [10,[23][24][25]. ...
Full-text available
Background The most popular tool used for measuring personality traits is the Five-Factor Model (FFM). It includes neuroticism, extraversion, openness, agreeableness and conscientiousness. Many studies indicated the association of genes encoding neurotransmitter receptors/transporters with personality traits. The relationship connecting polymorphic DNA sequences and FFM features has been described in the case of genes encoding receptors of cannabinoid and dopaminergic systems. Moreover, dopaminergic system receives inputs from other neurotransmitters, like GABAergic or serotoninergic systems. Methods We searched PubMed Central (PMC), Science Direct, Scopus, Cochrane Library, Web of Science and EBSCO databases from their inception to November 19, 2020, to identify original studies, as well as peer-reviewed studies examining the FFM and its association with gene polymorphisms affecting the neurotransmitter functions in central nervous system. Results Serotonin neurons modulate dopamine function. In gene encoding serotonin transporter protein, SLC6A4, was found polymorphism, which was correlated with openness to experience (in Sweden population), and high scores of neuroticism and low levels of agreeableness (in Caucasian population). The genome-wide association studies (GWASs) found an association of 5q34-q35, 3p24, 3q13 regions with higher scores of neuroticism, extraversion and agreeableness. However, the results for chromosome 3 regions are inconsistent, which was shown in our review paper. Conclusions GWASs on polymorphisms are being continued in order to determine and further understand the relationship between the changes in DNA and personality traits.
... Second, associations between the vital personality score and biological aging may partly reflect genetic influences. For example, previous research suggests that genetic influences that contribute to individual differences in personality traits might also contribute to longevity (Bae et al., 2013). Third, individuals with higher vital personality scores were surrounded by parents, partners, and children with higher vital personality scores. ...
Objectives Personality traits are linked with healthy aging, but it is not clear how these associations come to manifest across the life-course and across generations. To study this question, we tested a series of hypotheses about (a) personality-trait prediction of markers of healthy aging across the life-course, (b) developmental origins, stability and change of links between personality and healthy aging across time, and (c) intergenerational transmission of links between personality and healthy aging. For our analyses we used a measure that aggregates the contributions of Big 5 personality traits to healthy aging: a “vital personality” score. Methods Our data come from two population-based longitudinal cohort studies, one based in New Zealand and the other in the UK, comprising over 6000 study members across two generations, and spanning an age range from birth to late life. Results Our analyses revealed three main findings: first, individuals with higher vital personality scores engaged in fewer health-risk behaviors, aged slower, and lived longer. Second, individuals’ vital personality scores were preceded by differences in early-life temperament and were relatively stable across adulthood, but also increased from young adulthood to midlife. Third, individuals with higher vital personality scores had children with similarly vital partners, promoted healthier behaviors in their children, and had children who grew up to have more vital personality scores themselves, for genetic and environmental reasons. Conclusion Our study shows how the health benefits associated with personality accrue throughout the life-course and across generations.
... England centenarian study (NECS; Sebastiani et al., 2012), the Long Life Family Study (LLFS; Bae et al., 2013), and the Framingham Heart Study (FHS; Lunetta et al., 2007); (3) ...
Full-text available
There is growing interest in studying the genetic contributions to longevity, but limited relevant genes have been identified. In this study, we performed a genetic association study of longevity in a total of 15,651 Chinese individuals. Novel longevity loci, BMPER (rs17169634; p = 7.91 × 10 −15) and TMEM43/XPC (rs1043943; p = 3.59 × 10 −8), were identified in a case-control analysis of 11,045 individuals. BRAF (rs1267601; p = 8.33 × 10 −15) and BMPER (rs17169634; p = 1.45 × 10 −10) were significantly associated with life expectancy in 12,664 individuals who had survival status records. Additional sex-stratified analyses identified sex-specific longevity genes. Notably, sex-differential associations were identified in two linkage disequilibrium blocks in the TOMM40/APOE region, indicating potential differences during meiosis between males and females. Moreover, polygenic risk scores and Mendelian randomization analyses revealed that longevity was genetically causally correlated with reduced risks of multiple diseases, such as type 2 diabetes, cardiovascular diseases, and arthritis. Finally, we incorporated genetic markers, disease status, and lifestyles to classify longevity
... Age represents a complex surrogate for a host of underlying phenomena, although its measurement is simple and accurate [40]. A previous study suggested that gene-age interactions may partially be surrogates for gene-gene and gene-environment interactions [41]. In a study investigating the efficacy of metronomic vinorelbine to treat patients with advanced unresectable NSCLC, age was an important factor that decreased treatment efficacy [42]. ...
Full-text available
DNA methylation changes during aging, but it remains unclear whether the effect of DNA methylation on lung cancer survival varies with age. Such an effect could decrease prediction accuracy and treatment efficacy. We performed a methylation-age interaction analysis using 1,230 early-stage lung adenocarcinoma patients from five cohorts. A Cox proportional hazards model was used to investigate lung adenocarcinoma and squamous cell carcinoma patients for methylation-age interactions, which were further confirmed in a validation phase. We identified one adenocarcinoma-specific CpG probe, cg14326354 PRODH , with effects significantly modified by age (HRinteraction = 0.989; 95% CI: 0.986-0.994; P = 9.18×10-7). The effect of low methylation was reversed for young and elderly patients categorized by the boundary of 95% CI standard (HRyoung = 2.44; 95% CI: 1.26-4.72; P = 8.34×10-3; HRelderly = 0.58; 95% CI: 0.42-0.82; P = 1.67×10-3). Moreover, there was an antagonistic interaction between low cg14326354 PRODH methylation and elderly age (HRinteraction = 0.21; 95% CI: 0.11-0.40; P = 2.20×10-6). In summary, low methylation of cg14326354PRODH might benefit survival of elderly lung adenocarcinoma patients, providing new insight to age-specific prediction and potential drug targeting.
... Patients are followed up annually to track vital and health status and approximately 80% of surviving participants consented to a second in-person evaluation between 2014 and 2017. Genome-wide genotype data were generated using Illumina SNP array (Bae et al. 2013), and genomewide genotype data are available from dbGaP (dbGaP Study Accession: phs000397.v1.p1). Approximately 40 circulating biomarkers were measured in all participants as described in ). ...
Maintaining good cognitive function at older age is important, but our knowledge of patterns and predictors of cognitive aging is still limited. We used Bayesian model-based clustering to group 5064 participants of the Long Life Family Study (ages 49–110 years) into clusters characterized by distinct trajectories of cognitive change in the domains of episodic memory, attention, processing speed, and verbal fluency. For each domain, we identified 4 or 5 large clusters with representative patterns of change ranging from rapid decline to exceptionally slow change. We annotated the clusters by their correlation with genetic and molecular biomarkers, non-genetic risk factors, medical history, and other markers of aging to discover correlates of cognitive changes and neuroprotection. The annotation analysis discovered both predictors of multi-domain cognitive change such as gait speed and predictors of domain-specific cognitive change such as IL6 and NTproBNP that correlate only with change of processing speed or APOE genotypes that correlate only with change of processing speed and logical memory. These patterns also suggest that cognitive decline starts at young age and that maintaining good physical function correlates with slower cognitive decline. To better understand the agreement of cognitive changes across multiple domains, we summarized the results of the cluster analysis into a score of cognitive function change. This score showed that extreme patterns of change affecting multiple cognitive domains simultaneously are rare in this study and that specific signatures of biomarkers of inflammation and metabolic disease predict severity of cognitive changes. The substantial heterogeneity of change patterns within and between cognitive domains and the net of correlations between patterns of cognitive aging and other aging traits emphasizes the importance of measuring a wide range of cognitive functions and the need for studying cognitive aging in concert with other aging traits.
... In a GWAS conducted in Europeans, rs12513013 and rs12509930 of the SHROOM3 were associated with neuroticism [39]. In the participants of the New England Centenarian Study, Bae et al replicated the association of rs12509930 with neuroticism [40]. In the present study, we found that SHROOM3-rs17319721 was associated with depressive symptoms. ...
Full-text available
Background: To explore the associations of several genetic variants identified in the genome-wide association studies (GWAS) of European ancestry with mild renal impairment glomerular filtration rate (GFR) in Chinese Han population. Methods: Data of 1788 community-dwelling elders from the baseline survey of the ageing arm of the Rugao Longevity and Ageing Study was used. Plasma creatinine based GFR was estimated using the eGFR-EPI equations. Results: Of the 10 common polymorphisms identified in GWAS of the European ancestry, rs17319721 located in the first intron of the SHROOM3, was associated with GFR. A allele was associated with both decreased GFR level and greater odds of mild renal impairment (OR 1.12, 95% CI 1.01-1.23, p=0.029) defined by GFR<90 mL/min/1.73 m² after adjusting for multiple confounds of chronic kidney disease. In addition, compared with rs17319721-GG genotype, AA was associated with both higher depressive score and greater risk of depression prevalence, showing a pleiotropic effects of rs17319721. However, we did not found significant association of GFR levels with another 42 common polymorphisms that was previously reported to be associated with the traditional risk factors of kidney diseases. Conclusions: SHROOM3-rs17319721 is associated with GFR levels, kidney impairment, and depressive symptoms in a Chinese population.
... Our findings that people with the creative-reliable profile were healthier than others confirm earlier work about how to describe a healthy human personality [70,[90][91][92][93]. We found that people with creative characters and reliable temperaments have greater well-being, including objective indicators of healthy longevity, such as optimal cardiovascular health, when compared with others. ...
Full-text available
Phylogenetic, developmental, and brain-imaging studies suggest that human personality is the integrated expression of three major systems of learning and memory that regulate (1) associative conditioning, (2) intentionality, and (3) self-awareness. We have uncovered largely disjoint sets of genes regulating these dissociable learning processes in different clusters of people with (1) unregulated temperament profiles (i.e., associatively conditioned habits and emotional reactivity), (2) organized character profiles (i.e., intentional self-control of emotional conflicts and goals), and (3) creative character profiles (i.e., self-aware appraisal of values and theories), respectively. However, little is known about how these temperament and character components of personality are jointly organized and develop in an integrated manner. In three large independent genome-wide association studies from Finland, Germany, and Korea, we used a data-driven machine learning method to uncover joint phenotypic networks of temperament and character and also the genetic networks with which they are associated. We found three clusters of similar numbers of people with distinct combinations of temperament and character profiles. Their associated genetic and environmental networks were largely disjoint, and differentially related to distinct forms of learning and memory. Of the 972 genes that mapped to the three phenotypic networks, 72% were unique to a single network. The findings in the Finnish discovery sample were blindly and independently replicated in samples of Germans and Koreans. We conclude that temperament and character are integrated within three disjoint networks that regulate healthy longevity and dissociable systems of learning and memory by nearly disjoint sets of genetic and environmental influences.
Full-text available
The divergence of Eastern and Western cultures signifies the two opposite, major branches that developed during human cultural evolution. While socioeconomic, historical, and geographical factors are thought to be responsible for this divergence, genetic factors may also contribute to the separation of Eastern and Western cultures. In an attempt to describe a potential biological basis for the differences between “Easterners” and “Westerners”, SNPs that were associated with personality/behavioral traits, were interrogated in different populations worldwide. For some but not all SNPs examined, a high correlation in their allelic frequencies in different racial groups was detected. Those that exhibited the highest difference in allelic frequencies between East Asians and European ancestry populations, were all highly correlated in pairwise comparisons and corresponded to traits that are aligned with typical characteristics that are thought to underscore Western and Eastern cultures. Genetic loci associated with these SNPs included CTNNA2 (rs7600563), OXTR (rs53576) LINC00461 (rs3814424) MTMR9 (rs2164273) and WSCD2 (rs1426371) that have been linked to excitement seeking, empathy, the perception of loneliness, conscientiousness, and extraversion. Among them, variations especially in LINC00461 in different populations correlated significantly with Hofstede’s cultural dimensions indices of the different countries. These findings highlight the potential role of genetic factors in cultural evolution and suggest that genetic differences may contribute to the divergence of Eastern and Western cultures.
Full-text available
The NIA Long Life Family Study (LLFS) is a longitudinal, multicenter, multi-national, population-based multi-generational family study of the genetic and non-genetic determinants of exceptional longevity and healthy aging. The Visit 1 in-person evaluation (2006-2009) recruited 4,953 individuals from 539 two-generation families, selected from the upper 1% tail of the Family Longevity Selection Score (FLoSS, which quantifies the degree of familial clustering of longevity). Demographic, anthropometric, cognitive, activities of daily living, ankle-brachial index, blood pressure, physical performancea and pulmonary function, along serum, plasma, lymphocytes, red cells, and DNA were collected. A GWAS (Ilumina Omni 2.5M chip) followed by imputation was conducted. Visit 2 (2014-2017) repeated all Visit 1 protocols and added carotid ultrasonography of atherosclerotic plaque and wall thickness, additional cognitive testing, and perceived fatigability. On average, LLFS families show healthier aging profiles than reference populations, such as the Framingham Heart Study, at all age/sex groups, for many critical healthy aging phenotypes. However, participants are not uniformly protected. There is considerable heterogeneity among the pedigrees, with some showing exceptional cognition, others showing exceptional grip strength, others exceptional pulmonary function, etc. with little overlap in these families. There is strong heritability for key healthy aging phenotypes, both cross sectionally and longitudinally, suggesting that at least some of this protection may be genetic. Little of the variance in these heritable phenotypes is explained by the common genome (GWAS + Imputation), which may indicate that rare protective variants for specific phenotypes may be running in selected families.
Sports psychogenetics is a novel field of scientific research that aims to use genetic methods to investigate the nature and origins of individual differences in cognitive abilities and personal traits of athletes. This chapter describes the heritability estimates of psychological traits (e.g., cognitive ability, memory, reaction time, temperament), their potential relationship with success in sport, as well as polymorphisms of genes expressed in the nervous system, which might be associated with athlete status and personality. Despite success in discovery of genetic variants associated with intelligence and personality in non-athletic cohorts (more than 7000 DNA polymorphisms have been identified in candidate-gene or genome-wide association studies, involving large sample sizes), there has been limited progress to date in the field of sports psychogenetics. To date, 16psychogenetics-specific genetic markers have been reported to be associated with predisposition to specific sports (via case-control designs), and 12 markers have been linked with personality traits (via genotype-phenotype designs) in athletes. Future genetic research with large cohorts of athletes, with further validation and replication, will substantially contribute to the discovery of causal genetic variants (i.e., mutations and DNA polymorphisms) that may partly explain the heritability of athlete status and related psychological phenotypes.
Full-text available
In family studies, phenotypic similarities between relatives yield information on the overall contribution of genes to trait variation. Large samples are important for these family studies, especially when comparing heritability between subgroups such as young and old, or males and females. We recruited a cohort of 6,148 participants, aged 14–102 y, from four clustered towns in Sardinia. The cohort includes 34,469 relative pairs. To extract genetic information, we implemented software for variance components heritability analysis, designed to handle large pedigrees, analyze multiple traits simultaneously, and model heterogeneity. Here, we report heritability analyses for 98 quantitative traits, focusing on facets of personality and cardiovascular function. We also summarize results of bivariate analyses for all pairs of traits and of heterogeneity analyses for each trait. We found a significant genetic component for every trait. On average, genetic effects explained 40% of the variance for 38 blood tests, 51% for five anthropometric measures, 25% for 20 measures of cardiovascular function, and 19% for 35 personality traits. Four traits showed significant evidence for an X-linked component. Bivariate analyses suggested overlapping genetic determinants for many traits, including multiple personality facets and several traits related to the metabolic syndrome; but we found no evidence for shared genetic determinants that might underlie the reported association of some personality traits and cardiovascular risk factors. Models allowing for heterogeneity suggested that, in this cohort, the genetic variance was typically larger in females and in younger individuals, but interesting exceptions were observed. For example, narrow heritability of blood pressure was approximately 26% in individuals more than 42 y old, but only approximately 8% in younger individuals. Despite the heterogeneity in effect sizes, the same loci appear to contribute to variance in young and old, and in males and females. In summary, we find significant evidence for heritability of many medically important traits, including cardiovascular function and personality. Evidence for heterogeneity by age and sex suggests that models allowing for these differences will be important in mapping quantitative traits.
Full-text available
Objectives: To evaluate personality profiles of Long Life Family Study participants relative to population norms and offspring of centenarians from the New England Centenarian Study. METHOD Personality domains of agreeableness, conscientiousness, extraversion, neuroticism, and openness were assessed with the NEO Five-Factor Inventory in 4,937 participants from the Long Life Family Study (mean age 70 years). A linear mixed model of age and gender was implemented adjusting for other covariates. Results: A significant age trend was found in all five personality domains. On average, the offspring generation of long-lived families scored low in neuroticism, high in extraversion, and within average values for the other three domains. Older participants tended to score higher in neuroticism and lower in the other domains compared with younger participants, but the estimated scores generally remained within average population values. No significant differences were found between long-lived family members and their spouses. Discussion: Personality factors and more specifically low neuroticism and high extraversion may be important for achieving extreme old age. In addition, personality scores of family members were not significantly different from those of their spouses, suggesting that environmental factors may play a significant role in addition to genetic factors.
Full-text available
Like most complex phenotypes, exceptional longevity is thought to reflect a combined influence of environmental (e.g., lifestyle choices, where we live) and genetic factors. To explore the genetic contribution, we undertook a genome-wide association study of exceptional longevity in 801 centenarians (median age at death 104 years) and 914 genetically matched healthy controls. Using these data, we built a genetic model that includes 281 single nucleotide polymorphisms (SNPs) and discriminated between cases and controls of the discovery set with 89% sensitivity and specificity, and with 58% specificity and 60% sensitivity in an independent cohort of 341 controls and 253 genetically matched nonagenarians and centenarians (median age 100 years). Consistent with the hypothesis that the genetic contribution is largest with the oldest ages, the sensitivity of the model increased in the independent cohort with older and older ages (71% to classify subjects with an age at death>102 and 85% to classify subjects with an age at death>105). For further validation, we applied the model to an additional, unmatched 60 centenarians (median age 107 years) resulting in 78% sensitivity, and 2863 unmatched controls with 61% specificity. The 281 SNPs include the SNP rs2075650 in TOMM40/APOE that reached irrefutable genome wide significance (posterior probability of association = 1) and replicated in the independent cohort. Removal of this SNP from the model reduced the accuracy by only 1%. Further in-silico analysis suggests that 90% of centenarians can be grouped into clusters characterized by different "genetic signatures" of varying predictive values for exceptional longevity. The correlation between 3 signatures and 3 different life spans was replicated in the combined replication sets. The different signatures may help dissect this complex phenotype into sub-phenotypes of exceptional longevity.
Full-text available
We analyze the relationship between age of survival, morbidity, and disability among centenarians (age 100-104 years), semisupercentenarians (age 105-109 years), and supercentenarians (age 110-119 years). One hundred and four supercentenarians, 430 semisupercentenarians, 884 centenarians, 343 nonagenarians, and 436 controls were prospectively followed for an average of 3 years (range 0-13 years). The older the age group, generally, the later the onset of diseases, such as cancer, cardiovascular disease, dementia, and stroke, as well as of cognitive and functional decline. The hazard ratios for these individual diseases became progressively less with older and older age, and the relative period of time spent with disease was lower with increasing age group. We observed a progressive delay in the age of onset of physical and cognitive function impairment, age-related diseases, and overall morbidity with increasing age. As the limit of human life span was effectively approached with supercentenarians, compression of morbidity was generally observed.
Full-text available
We review evidence for links between personality traits and longevity. We provide an overview of personality for health scientists, using the primary organizing framework used in the study of personality and longevity. We then review data on various aspects of personality linked to longevity. In general, there is good evidence that higher level of conscientiousness and lower levels of hostility and Type D or "distressed" personality are associated with greater longevity. Limited evidence suggests that extraversion, openness, perceived control, and low levels of emotional suppression may be associated with longer lifespan. Findings regarding neuroticism are mixed, supporting the notion that many component(s) of neuroticism detract from life expectancy, but some components at some levels may be healthy or protective. Overall, evidence suggests various personality traits are significant predictors of longevity and points to several promising directions for further study. We conclude by discussing the implications of these links for epidemiologic research and personalized medicine and lay out a translational research agenda for integrating the psychology of individual differences into public health and medicine.
Full-text available
Human behavioral genetic research aimed at characterizing the existence and nature of genetic and environmental influences on individual differences in cognitive ability, personality and interests, and psychopathology is reviewed. Twin and adoption studies indicate that most behavioral characteristics are heritable. Nonetheless, efforts to identify the genes influencing behavior have produced a limited number of confirmed linkages or associations. Behavioral genetic research also documents the importance of environmental factors, but contrary to the expectations of many behavioral scientists, the relevant environmental factors appear to be those that are not shared by reared together relatives. The observation of genotype-environment correlational processes and the hypothesized existence of genotype-environment interaction effects serve to distinguish behavioral traits from the medical and physiological phenotypes studied by human geneticists. Behavioral genetic research supports the heritability, not the genetic determination, of behavior.
The 1000 Genomes Project and disease-specific sequencing efforts are producing large collections of haplotypes that can be used as reference panels for genotype imputation in genome-wide association studies (GWAS). However, imputing from large reference panels with existing methods imposes a high computational burden. We introduce a strategy called 'pre-phasing' that maintains the accuracy of leading methods while reducing computational costs. We first statistically estimate the haplotypes for each individual within the GWAS sample (pre-phasing) and then impute missing genotypes into these estimated haplotypes. This reduces the computational cost because (i) the GWAS samples must be phased only once, whereas standard methods would implicitly repeat phasing with each reference panel update, and (ii) it is much faster to match a phased GWAS haplotype to one reference haplotype than to match two unphased GWAS genotypes to a pair of reference haplotypes. We implemented our approach in the MaCH and IMPUTE2 frameworks, and we tested it on data sets from the Wellcome Trust Case Control Consortium 2 (WTCCC2), the Genetic Association Information Network (GAIN), the Women's Health Initiative (WHI) and the 1000 Genomes Project. This strategy will be particularly valuable for repeated imputation as reference panels evolve.
Short versions of four Eysenck personality scales had been included in questionnaires given to several adult samples from the Australian Twin Registry, comprising altogether some 5400 pairs. Means and regressions with age are compared for three samples at average ages of 23, 37, and 61 years, and for two samples of retested individuals, one tested twice at average ages of 29 and 37 years, and one tested three times at average ages of 48, 56, and 62 years. For both males and females the trends for Psychoticism (P), Extraversion (E), and Neuroticism (N) were generally downward with age, and for Lie (L), upward. However, in the longitudinal sample between ages 56 and 62 the trends for P, E, and L stopped or reversed, although N continued downward. Heritabilities were reasonably stable across age for P, E, and N, and the effects of shared environments negligible, but L showed some influence of shared environment as well as genes in all but the oldest age group.
Multipoint linkage analysis of quantitative-trait loci (QTLs) has previously been restricted to sibships and small pedigrees. In this article, we show how variance-component linkage methods can be used in pedigrees of arbitrary size and complexity, and we develop a general framework for multipoint identity-by-descent (IBD) probability calculations. We extend the sib-pair multipoint mapping approach of Fulker et al. to general relative pairs. This multipoint IBD method uses the proportion of alleles shared identical by descent at genotyped loci to estimate IBD sharing at arbitrary points along a chromosome for each relative pair. We have derived correlations in IBD sharing as a function of chromosomal distance for relative pairs in general pedigrees and provide a simple framework whereby these correlations can be easily obtained for any relative pair related by a single line of descent or by multiple independent lines of descent. Once calculated, the multipoint relative-pair IBDs can be utilized in variance-component linkage analysis, which considers the likelihood of the entire pedigree jointly. Examples are given that use simulated data, demonstrating both the accuracy of QTL localization and the increase in power provided by multipoint analysis with 5-, 10-, and 20-cM marker maps. The general pedigree variance component and IBD estimation methods have been implemented in the SOLAR (Sequential Oligogenic Linkage Analysis Routines) computer package.