-
Title
-
DNA Revolution and the Social and Behavioral Sciences
-
Author
-
Trzaskowski, Maciej
-
Plomin, Robert
-
Research Area
-
Cognition and Emotions
-
Topic
-
Motivation
-
Abstract
-
A century of genetic research on the social and behavioral sciences has addressed the “how much” question, showing that genetic differences are important for nearly all traits. However, during the past few decades, research has moved beyond this rudimentary “how much” question to ask “how” questions about developmental change and continuity, about the relationship between traits, and about the interplay between genes (nature) and environment (nurture). We suggest that some of the most important and transformative findings in the social and behavioral sciences have come from this research. Moreover, the most revolutionary changes in genetic research are on their way with the rapid advances in DNA technology and methodology, which promise to transform the social and behavioral sciences. It is crucial that social and behavioral scientists stay on top of the DNA revolution. The purpose of our essay is to provide an overview of genetic research in the social and behavioral sciences.
-
Identifier
-
etrds0084
-
extracted text
-
DNA Revolution and the Social
and Behavioral Sciences
MACIEJ TRZASKOWSKI and ROBERT PLOMIN
Abstract
A century of genetic research on the social and behavioral sciences has addressed
the “how much” question, showing that genetic differences are important for nearly
all traits. However, during the past few decades, research has moved beyond this
rudimentary “how much” question to ask “how” questions about developmental
change and continuity, about the relationship between traits, and about the interplay between genes (nature) and environment (nurture). We suggest that some of
the most important and transformative findings in the social and behavioral sciences
have come from this research. Moreover, the most revolutionary changes in genetic
research are on their way with the rapid advances in DNA technology and methodology, which promise to transform the social and behavioral sciences. It is crucial that
social and behavioral scientists stay on top of the DNA revolution. The purpose of
our essay is to provide an overview of genetic research in the social and behavioral
sciences.
NATURE AND NURTURE: TWIN AND ADOPTION STUDIES
In general, genetic research can be divided into quantitative genetics
(statistical designs in which family members are used to estimate genetic
influence) and molecular genetics (research designs where genetic influence
is estimated by directly measured genotypes). For decades quantitative
geneticists used family, twin and adoption designs to estimate genetic and
environmental influences on many psychological and psychiatric traits. The
strength of these designs is that they can estimate the bottom line of genetic
influence regardless of how many genes are involved or how complex their
effects, unlike molecular genetics. All quantitative genetic designs have
limitations, but each design has different limitations, and, most importantly,
their results converge. For example, the adoption design, which typically
compares adopted children with their biological and adoptive parents, is
limited by possible prenatal factors and selective placement. In contrast, the
twin design, which compares monozygotic (MZ) twins who are genetically
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
100% identical, and dizygotic (DZ) twins who share on average 50% of their
segregating alleles, is limited by the assumption of equal environments
for the two types of twins. These limitations are specific to each design
and yet adoption and twin studies generally produce similar estimates of
genetic and environmental influences, especially for cognitive traits (Plomin,
DeFries, Knopik, & Neiderhiser, 2013). Adoption studies are less common
today than twin studies because of the sharp reduction in neonatal adoption,
but twin studies continue to proliferate.
Univariate twin studies have demonstrated consistently that virtually
every human trait is influenced in part by genetic variation. Multivariate
models go beyond this rudimentary finding about the relative influence
of nature and nurture on a single phenotype, to investigate the common
genetic architecture underlying multiple complex traits. Multivariate twin
designs decompose the covariance between multiple phenotypes to provide
insights into the common genetic and environmental relationship underlying different traits, as well as the genetic and environmental contributions
to their development over time (trait stability and change). Multivariate
genetic research has yielded three major findings consistent across many
complex traits: heritability increases across development, genetic effects
are developmentally stable—strong age-to-age genetic correlations—and
are highly pleiotropic, that is, one gene influences many traits. Genetic
pleiotropy is inferred from consistent findings of strong genetic covariance
between different traits at one measurement occasion: for example, this
is true for anxiety and depression (e.g., Thapar & McGuffin, 1997), and
for IQ and learning abilities (Davis, Haworth, & Plomin, 2009). Genetic
stability is inferred from high genetic correlations between the “same”
traits across time. Curiously, despite the observed genetic stability from
age to age, many traits appear to show increases in heritability across
development (Bergen, Gardner, & Kendler, 2007). Although genetic stability
and increasing heritability might appear paradoxical, there are several
mechanisms that might explain this. The phenomenon could be driven by a
correlation between genes and environments where, for example, children
choose environments that suit their genetic propensities (Plomin, DeFries,
et al., 2013). Imagine a young girl who has a propensity for reading; as she
grows up, she chooses environments in which she can be exposed to reading
as much as possible. This “tailored” selection will increase the importance
of these in explaining individual differences in reading and yet they will be
the same genes. It is also possible that heritability increases as new genes
come “online” over the lifetime explaining more of the variance (Kendler,
Gardner, & Lichtenstein, 2008). Unfortunately, twin designs cannot easily
tease out which of these mechanisms are responsible for the phenomena.
The most important advance in this field that will help to resolve this issue
DNA Revolution and the Social and Behavioral Sciences
3
and many others will come when specific genetic variants are identified that
account for these genetic effects.
LOW-HANGING FRUIT: MONOGENIC DISORDERS
Mendel, in his original work with peas, was fortunate in that he chose monogenic traits in which a single mutation is necessary and sufficient for the
trait to be expressed. For such single-gene traits the underlying genetic variants travel across generations according to predictions based on Mendel’s
first law of segregation, and they also assort independently from variants
of other traits according to Mendel’s second law. Unfortunately, complex
traits appeared not to conform to these laws, prompting some researchers
in the early twentieth century to argue that Mendel’s findings were limited
to peas, and not “higher-order” species. Although the dispute between the
two factions was at times fierce, it was resolved when it became apparent
that in complex traits Mendel’s laws hold for segregation of alleles at a single
locus but are masked by the additive influence of many such loci. The overall genetic influence is driven by many genes, called polygenic, and thus the
effects of individual genes is small and their overall effect on the phenotype
is complex and normally distributed in a population. This is the cornerstone
of quantitative genetics expounded by Fisher (1918), who described how
Mendel’s model could be extended to multiple genes in order to account for
inheritance of complex traits.
In molecular genetics, Mendel’s law of segregation culminated in huge
success in discovering genetic associations with monogenic disorders
mainly through implementation of linkage designs using just a few hundred
DNA markers across the genome. Linkage analysis uses samples drawn
from multigenerational families with affected and unaffected members,
where DNA of all family members is “tagged” with several markers whose
location on chromosomes is known. The aim is to detect a “link” between
one of these markers and a trait or disorder. Linkage successfully detected
genetic associations with many Mendelian (monogenic) traits because a
mutation in a single locus explained all the variance. However, monogenic
disorders affect a tiny proportion of clinical populations; most of the burden
of illness involves complex disorders influenced by many genes with
small effect sizes. Although linkage has been very successful in identifying
genes for single-gene disorders, the method does not have sufficient power
to detect genes of small effect size. A technological advance that greatly
improved attempts to find genes responsible for heritability of complex traits
is genome-wide association (GWA), which is based on allelic association
between DNA markers and a trait in unrelated individuals in the population
rather than linkage within families.
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
FINDING MANY NEEDLES IN THE HAYSTACK: GENOME-WIDE
ASSOCIATION STUDIES
The discovery that many complex traits are highly heritable motivated
researchers to search for the genetic variants associated with these traits.
The pre-GWA era relied mainly on candidate gene association studies,
where a specific region of genome was hypothesized to be important
for development of a particular trait. Unfortunately, a problem with the
candidate gene approach is that the hypotheses generated about the regions
we think should be important are often wrong. Many published candidate
gene association results is now known to be false positive results (Tabor,
Risch, & Myers, 2002). This is particularly telling when replication fails in
GWA analysis as this design is hypothesis free in the sense that it does not
“care” which region is associated with which trait. An example of such
failure is a GWA study of nearly 10,000 individuals where none of the 10
most frequently reported candidate regions for general cognitive ability
replicated (Chabris et al., 2012).
The genomic era began in the past decade when two major developments
provided cost-effective, thorough coverage of the genome. The first event
was the completion of the Human Genome Project (H.G.P., 2001), which
resulted in the first detailed maps of the human genome and the patterns of
linkage disequilibrium (the nonrandom association of alleles across neighboring genetic loci) for hundreds of thousands SNPs (single-nucleotide
polymorphism—variation in a single genomic base-pair). The second event
was the production of DNA arrays or “chips.” A single DNA chip could
genotypes hundreds of thousands of SNPs. For each SNP, the DNA chip
includes many probes, which is a short DNA sequence containing the
SNP. Understanding linkage disequilibrium patterns across the genome
was crucial because it showed that careful selection of a few hundred
thousand DNA markers was sufficient to comprehensively tag the genome’s
3 billion DNA base pairs for GWA studies. DNA variants close together
on a chromosome violate Mendel’s second law by segregating together
producing high between-SNP correlations. A single marker from DNA
variants in high linkage disequilibrium is sufficient to “tag” the region. This
knowledge was used to select SNPs for DNA chips. In addition, coverage
of the genome could be improved by imputing millions of additional SNPs
from existing reference maps, such as HapMap (Gibbs et al., 2003) or, more
recently, the 1000 Genomes Project (Siva, 2008) without actually genotyping
these additional SNPs. This meant that it was now possible to interrogate
the whole genome simultaneously for association between allele frequencies
of individual SNPs and any trait varying in unrelated individuals. This
DNA Revolution and the Social and Behavioral Sciences
5
method is known today as a GWA (e.g., Balding, 2006; Cardon & Bell, 2001;
Hirschhorn & Daly, 2005).
In GWA, associations are usually examined one SNP at the time, assuming
an additive model, although other models can also be implemented. Since
current genetic chips can assess millions of SNPs, the correction for multiple
testing is daunting, and the number of possible false positives is large; the
accepted p-value threshold for genome-wide significance is p < 5 × 10−8 ,
which can be thought of as p < 0.05 with a Bonferroni correction for one
million statistical tests. Despite such stringent correction, within the three
years between 2005 and 2008, more than 400 SNP associations for just
over 120 traits (www.genome.gov/GWAstudies) achieved this stringent
genome-wide threshold, and revealed unexpected insights about genetic
influences on complex traits (Visscher, Brown, McCarthy, & Yang, 2012).
Arguably the most useful biological insights in genetic associations have
been mainly limited to medical traits. For example, in Crohn’s disease,
many SNPs reported through GWA studies were found in and around genes
involved with autophagy, the cell’s maintenance process that breaks down
dysfunctional components of the cell (WTCCC, 2007). In addition, the same
study showed that type 2 diabetes was associated with loci encoding for
proteins relevant to insulin secretion, and not insulin signaling, as previously thought. Nevertheless, even if functionally not so obvious, many new
loci have also been identified in anthropometric (e.g., Speliotes et al., 2010;
Visscher, 2008) and some psychiatric traits (e.g., Gershon, Alliey-Rodriguez,
& Liu, 2011; Purcell et al., 2014). Unfortunately successes in cognitive, social,
and behavioral complex traits are yet to come.
Despite GWA successes, fewer associations were reported for complex
traits than expected, and associations that were identified could together
only account for a small fraction of twin-estimated heritability. The phenomenon is known as “missing heritability,” which is discussed in the
next section. The important lesson from GWA research is that the largest
effect sizes of individual loci of common SNPs are incredibly small and
require sample sizes in the hundreds of thousands to achieve genome-wide
significance. This realization led to the sudden emergence and proliferation
of world-wide consortia, which paid off by doubling the GWA hits to just
under 9000 SNPs in more than 700 traits (www.genome.gov/GWAstudies/).
Nonetheless, the yield was still not as substantial as expected in that the
largest effect sizes, such as the association between the FTO gene and body
mass index (BMI), are less than 1% of the variance (Speliotes et al., 2010),
which implies that the smallest effect sizes are likely to be infinitesimal.
This means that it will be difficult to detect and replicate associations with
complex traits in the social and behavioral sciences. However, once several
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
genes are found that are associated with these traits, their effects can be
aggregated as polygenic scores, as described in the following section.
TOGETHER WE ARE STRONGER: POLYGENIC SCORES
Even though the effects of individual SNPs are very small, their effects can
be aggregated to increase the total amount of variance explained and thus
increase power. For example, if we had 10 individual SNPs, each with an
effect size of 0.5%, a polygenic score comprising all of these would account
for 5% of the variance, which means that a sample of 150 individuals would
have 80% power to detect their cumulative effect. For this reason, polygenic
scores are now being used in a new generation of hypothesis-free candidate
gene association studies and are the way that the DNA revolution will come
to the social and behavioral sciences.
The creation of a polygenic score takes into account the direction of the
effect (i.e., which allele is the “increasing” allele) and it can also weight the
associations by the magnitude of their effect. To give priority to stronger associations, the scores can be weighted by the betas from the regression (i.e.,
their effect size). The associations can then be summed, similar to summing
items on a scale. Although summing scores is most commonly used, nonadditivity can also be incorporated. In addition, SNPs can either be selected
from previously reported significant “hits” (polygenic scores—PGS) or they
can be amassed by selecting all associations below an arbitrary p-value from
GWA studies, a genome-wide PGS (Wray et al., 2014). The former method is
preferred because of the trade-off between specificity and size but the choice
depends on the availability of previously reported robust “hits.” The tiny
effect sizes and moderate-to-large estimates of heritability suggest that the
more SNPs you “pool” together the more variance you should explain. However, adding a large number of SNPs with no effect or effects in the opposite
direction can attenuate the signal.
Application of polygenic predictors to medical disorders had some success
due at least in part to availability of financial resources and large sample
sizes. It has been shown that PGS of 150 SNPs accounted for 5% of the
variance in the liability for coronary artery disease (Deloukas et al., 2013).
For bipolar disorder, SNPs accounted for between 1% and 3% of the liability
(Psychiatric Gwas Consortium Bipolar Disorder Working & Group, 2011)
and a PGS of significant SNPs from discovery GWAS in schizophrenia,
accounted for approximately 1% of the liability in the independent sample
(Schizophrenia Psychiatric Genome-Wide Association Study, 2011). In the
schizophrenia study, extending the PGS to genome-wide PGS increased
variance explained to almost 6%. PGS research on quantitative traits has
largely been limited to weight and height, and other complex continuous
DNA Revolution and the Social and Behavioral Sciences
7
traits have had to resort to genome-wide PGS. GWA meta-analysis of BMI
revealed that 32 replicated SNPs accounted for only ∼1.5% of the variance
in BMI in independent samples, but a genome-wide PGS that included
thousands of SNPs explained ∼5% of the variance (Speliotes et al., 2010).
For height, 180 SNPs accounted for 10% of the variance, and genome-wide
PGS increased the variance explained to 13% (Lango Allen et al., 2010).
Importantly, investigation of behavioral traits of direct interest to social
and cognitive scientists are now achieving sample sizes that enable more
powerful genomic interrogation. Application of genome-wide PGS to these
traits is producing results that are similar to medical and anthropometric
traits (height and weight). Specifically, adding more SNPs increases amount
of variance explained up to a point. For example, an increase of variance
explained with an addition of more SNPs was reported for total years of
education (1% using 3500 SNPs and 2.5% using 2.5 million SNPs; Rietveld,
Medland, et al., 2013) and childhood IQ (using polygenic predictor at
p < 6 × 10−5 explained 0.5–1.2% increasing to 3.5% at p < 0.001; Benyamin
et al., 2014). These findings are important despite the small proportions of
variance explained, because they suggest that complex traits are indeed
highly polygenic.
It is now obvious that even though inclusion of a large number of SNPs
consistently increases the amount of variance explained, the gap between
twin-estimated heritability and these PGS estimates is still very wide. The
phenomenon known today as the “missing heritability” continues to pose
many questions, which are as yet unanswered. Twin studies suggest that
almost every human trait from our biology through cognition, behavior and
even the environment is heritable, with genetic influences usually explaining
moderate (30–40%) to high (80–90%) proportions of variance. GWA studies
have accounted for only a small fraction of this heritability. The most general
explanation is that the influence of each individual SNP is so small that most
of the GWA studies thus far have been greatly underpowered to detect them.
Another likely source of missing heritability is rarer variants—the markers
selected for the DNA arrays were limited to common variants only (minor
allele frequency>1%). This meant that many potential “true” associations
with SNPs of lower allele frequency could be missed owing to low linkage disequilibrium with the markers. Another possible source of missing
heritability is nonadditive effects, such as gene–environment or gene–gene
interactions because GWA is limited to additive genetic effects. Finally, it is
possible that heritability estimates derived from twin data are inflated. To
answer some of these questions, a new quantitative genetic technique has
emerged, called genome-wide complex trait analysis (GCTA), as described
in the following section.
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
HERITABILITY WITHOUT TWINS: GENOME-WIDE COMPLEX
TRAIT ANALYSIS (GCTA)
A recently developed method uses genome-wide genotyping data in a novel
way to address these questions. The linear mixed model implemented in
the GCTA package estimates the amount of phenotypic variance (Yang, Lee,
Goddard, & Visscher, 2011) and co-variance (Lee, Yang, Goddard, Visscher,
& Wray, 2012) that can be explained by additive effects of all common SNPs
tagged by DNA arrays. Several other methods have been developed with
the same intent (Wray et al., 2014) but GCTA is currently most widely used.
GCTA itself has been given different names, such as linear mixed model
and genomic-relationship-matrix restricted maximum likelihood. However,
clever usage of the four DNA base letters makes GCTA the catchiest acronym,
and this is what we call it henceforth.
GCTA uses genome-wide genotype data from unrelated individuals to estimate genetic influence on a trait that can be explained by all SNPs included
on DNA chips. Because GCTA is based on genome-wide DNA data alone,
it can be used to estimate genetic influence for unrelated individuals rather
than requiring special relatives such as MZ and DZ twins. For that reason, it
bypasses some of the assumptions of the twin method, although it has its own
set of assumptions (Plomin, Haworth, Meaburn, Price, & Davis, 2013). GCTA
can only detect genetic influence owing to the additive effects of common
SNPs that are included on currently available DNA arrays and cannot evaluate the contribution of any specific DNA locus. Nonetheless, GCTA provides
important information about the extent to which the genetic architecture of
complex traits includes additive effects of common SNPs, and sets the limit
for detecting associations in GWA studies.
GCTA analyses have shown that information captured by current DNA
arrays can explain a substantial amount of the variance in complex traits,
including human height (Yang et al., 2010), BMI (Llewellyn, Trzaskowski,
Plomin, & Wardle, 2013), psychiatric and medical disorders (Lee, Wray,
Goddard, & Visscher, 2011; Lee et al., 2012; Lubke et al., 2012), cognitive
traits (Deary et al., 2012; Plomin, Haworth, et al., 2013), and economic and
political preferences (Benjamin et al., 2012). All of these GTCA heritability
estimates are approximately half of twin-estimated heritability. In contrast,
initial research in psychopathology and personality is less consistent,
showing near zero variance explained for most analyses. For example,
analyses of a wide range of behavioral problems (symptoms of anxiety
and depression, hyperactivity, conduct problems) show negligible SNP
heritability despite moderate to high twin heritability estimated within the
same sample (Trzaskowski, Dale, & Plomin, 2013; Trzaskowski, Eley, et al.,
2013). SNP heritabilities of neuroticism and extraversion were reported as
DNA Revolution and the Social and Behavioral Sciences
9
0.06 (SE 0.03) and 0.12 (SE 0.03) respectively (Vinkhuyzen et al., 2012), and
similar near-zero GCTA estimates have also been shown for other aspects
of personality (Verweij et al., 2012), including wellbeing (Rietveld, Cesarini,
et al., 2013). The results are puzzling as all of the studies are adequately
powered to detect the expected effect sizes (∼40% of twin heritability)
and are not limited to self-report questionnaire. For example, one study
examined behavior problems reported by parents and teachers as well as
the children themselves; further discussion on this topic can also be found
in the same publication (Trzaskowski, Dale, et al., 2013).
As was the case for early twin studies, GCTA was initially applied in univariate analyses of one trait at a time, but like twin studies, the model was
extended to the bivariate analysis of covariance between traits or across age.
For example, bivariate GCTA was first applied to the remarkable phenotypic
stability of IQ across 60 years from childhood to later life (phenotypic correlation = 0.63) and suggested that the stability is largely due to genetic stability
(genetic correlation = 0.62) (Deary et al., 2012). Twin studies also suggested
that the genetic stability was present despite heritability of IQ increasing
across development. Bivariate GCTA supported this finding of genetic stability despite increasing heritability for IQ (Trzaskowski, Yang, Visscher, &
Plomin, 2013) and similar twin and GCTA results were reported for BMI
(Llewellyn, Trzaskowski, Plomin, & Wardle, 2014). The most likely explanation for these phenomena is gene–environment correlation (Plomin, DeFries,
et al., 2013). The high genetic correlation suggests that the same genes influence individual differences in IQ across time. However, as we grow older we
increasingly select environments that “match” our genetic predispositions
making genetic influence stronger even though the same genes are involved.
Although twin studies have reported for decades that most environments are nearly as heritable as behaviors, this work has been limited to
twin-specific environments. GCTA opens up the possibility of investigating
genetic influence on family-, neighborhood-, or even country-wide environmental measures that cannot be studied using the twin design because
they are shared in common by members of a twin pair (Trzaskowski et al.,
2014). This feature of GCTA should be particularly interesting to social
and behavioral scientists as it emphasizes the important interplay between
genes and environments. It shows that environments are not simply “out
there” that happen randomly to us, but that our genes shape our experiences
through our selection, modification and creation of our environments.
Another widely reported finding from twin analyses is the strong genetic
correlation (pleiotropy) across different aspects of cognition and across
diverse cognitive abilities (Davis et al., 2009; Kovas, Haworth, Dale, &
Plomin, 2007; Plomin, DeFries, et al., 2013). GCTA studies reported point
estimates for genetic correlations highly similar to those reported by twin
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
studies (Trzaskowski, Davis, et al., 2013). GCTA estimates of genetic correlation are similar to twin study estimates of genetic correlation, not only within
traditional domains such as verbal and nonverbal (Trzaskowski, Shakeshaft,
& Plomin, 2013) but also between intelligence and education-related skills
such as reading and mathematics (Trzaskowski, Davis, et al., 2013). We
can expect many more multivariate GCTA analyses in domains other than
cognition, most notably psychopathology where it increasingly appears that
there is a great deal of overlap among diverse symptoms (Caspi et al., 2013).
In summary, GCTA has shed some light on the nature of the discrepancy
between heritability estimates from twin studies and combined variance
explained from SNPs identified through GWA. Some of the “missing
heritability” is hiding in tiny influences of many common SNPs. We called
that gap between GWA and GCTA “missing GWA heritability,” because
GCTA represents the ceiling for GWA results. This gap can in theory be
filled with the additive effects of variants tagged by the SNPs on current
DNA chips if samples are large enough to detect nearly infinitesimal effect
sizes. GCTA results also suggest that even if the “missing GWA heritability”
were filled, there would still be a substantial chunk of genetic influence
not accounted for by additive effects of common SNPs. This part of the
“missing heritability” could be called “missing GCTA heritability,” which
falls short of twin study heritability because GCTA only reflect additive
effects of common SNPs. We delineated these two parts to emphasize a
distinction between missing heritability that is caused by lack of power (the
effect sizes of already captured common variants are too small to detect
with current sample sizes), and missing heritability that is due to genetic
variants that have not yet been captured (e.g., rare variants or nonadditive
genetic influences). The gap of missing GWA heritability can be narrowed
by the brute force of larger samples, but how can DNA variants responsible
for missing GCTA heritability be identified?
3 BILLION BASE PAIRS: WHOLE-GENOME SEQUENCING
Whole-genome sequencing involves genotyping all 3 billion base pairs of
DNA, rather than just a million or so SNPs as on current genetic chips.
Whole-genome sequencing is the end of the story of genetic variation in
the sense that all we inherit from our parents is differences in DNA base
pair sequences. Since current genetic chips only capture common SNP,
whole-genome sequencing could identify more genetic loci associated with
complex traits because it captures variants of any kind, not just common
SNPs. Rare variants may play a particularly important role in the extreme
tails of a trait’s distribution. For example, common SNPs influence BMI
across the distribution of the normal population, but very rare Mendelian
DNA Revolution and the Social and Behavioral Sciences
11
mutations may account for as much as 5% of extreme obesity (Farooqi &
O’Rahilly, 2006).
Other clinical phenotypes, such as intellectual disability, schizophrenia,
bipolar disorder, and autism, have been associated with an increased burden
of rare variants, including idiosyncratic mutations that first occur in that
individual and are not inherited from the individual’s parents, called de
novo mutations (Marioni et al., 2014; Neale et al., 2011; Purcell et al., 2014;
Stankiewicz & Lupski, 2010). Perhaps rare variants are responsible for a
puzzle in findings concerning mild and severe intellectual disability. One
study found that mild intellectual impairment was familial, but severe
impairment was not (Nichols, 1984). That is, siblings of severely mentally
retarded children showed no mental impairment, whereas siblings of children with “mild” mental disability did, suggesting that severe impairment
is not genetically related to common variation in mental ability. In general,
common (less severe) disorders are likely to be the quantitative extremes of
normal variation (Plomin, Haworth, & Davis, 2009), whereas extreme levels
of disability could be mainly a result of accumulation of much rarer variants,
including de novo mutations as well as environmental “mutations” such as
perinatal trauma.
Sequencing will also give us richer information about noncoding regions.
Not so long ago noncoding regions of the genome were thought of as an
evolutionary “junk,” but now these noncoding regions are known to play an
important role in regulation of genetic expression and even creation of new
genes (Mercer, Dinger, & Mattick, 2009; Muotri, Marchetto, Coufal, & Gage,
2007; Shimoni et al., 2007). Sequencing these regions will illuminate regulatory networks and thus contribute to our understanding of genetic responses
to changes in the environment.
Thus far the expense of whole-genome sequencing has slowed its progress.
However, as costs continue to fall, the availability of sequence data will
increase exponentially. It has been predicted by Francis Collins, former
director of the Human Genome Project and currently director of the US
National Institutes of Health, in his excellent book that: “I am almost certain
that complete genome sequencing will become part of newborn screening
in the next few years … . It is likely that within a few decades people will
look back on our current circumstances with a sense of disbelief that we
screened for so few conditions” (Collins, 2010). In fact, parents have already
begun paying for sequencing their children’s DNA (Rochman, 2012). If
this prediction is correct, future social and behavioral research, will see a
completely different world of data; a world where budgeting for DNA,
genotyping or sequencing will be a thing of the past. Genomic data on
nearly everyone will be widely available from centralized sources. For this
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
reason, it is crucial that social and behavioral scientists in the future are able
to capitalize on this opportunity to add genomics to their research.
REFERENCES
Balding, D. J. (2006). A tutorial on statistical methods for population association studies. Nature Reviews Genetics, 7(10), 781–791. doi:10.1038/nrg1916
Benjamin, D. J., Cesarini, D., van der Loos, M. J., Dawes, C. T., Koellinger, P. D., Magnusson, P. K., … , Visscher, P. M. (2012). The genetic architecture of economic and
political preferences. Proceedings of the National Academy of Sciences of the United
States of America, 109(21), 8026–8031. doi:10.1073/pnas.1120666109
Benyamin, B., Pourcain, B., Davis, O. S., Davies, G., Hansell, N. K., Brion, M. J.,
… , Visscher, P. M. (2014). Childhood intelligence is heritable, highly polygenic
and associated with FNBP1L. Molecular Psychiatry, 19(2), 253–258. doi:10.1038/
mp.2012.184
Bergen, S. E., Gardner, C. O., & Kendler, K. S. (2007). Age-related changes in heritability of behavioral phenotypes over adolescence and young adulthood: A
meta-analysis. Twin Research and Human Genetics, 10(3), 423–433. doi:10.1375/
twin.10.3.423
Cardon, L. R., & Bell, J. I. (2001). Association study designs for complex diseases.
Nature Reviews Genetics, 2(2), 91–99. doi:10.1038/35052543
Caspi, A., Houts, R. M., Belsky, D. W., Goldman-Mellor, S. J., Harrington, H., Israel,
S. … Moffitt, T. E. (2013). The p factor: One general psychopathology factor in
the structure of psychiatric disorders? Clinical Psychological Science, 2(2), 119–137.
doi:10.1177/2167702613497473
Chabris, C. F., Hebert, B. M., Benjamin, D. J., Beauchamp, J., Cesarini, D., van der
Loos, M. … Laibson, D. (2012). Most reported genetic associations with general
intelligence are probably false positives. Psychological Science, 23(11), 1314–1323.
doi:10.1177/0956797611435528
Collins, F. (2010). The language of life: DNA and the revolution in personalised medicine.
London, England: Profile Books.
Davis, O. S. P., Haworth, C. M. A., & Plomin, R. (2009). Learning abilities and disabilities: Generalist genes in early adolescence. Cognitive Neuropsychiatry, 14, 312–331.
doi:10.1080/13546800902797106
Deary, I. J., Yang, J., Davies, G., Harris, S. E., Tenesa, A., Liewald, D. … Corley, J.
(2012). Genetic contributions to stability and change in intelligence from childhood to old age. Nature, 482(7384), 212–215.
Deloukas, P., Kanoni, S., Willenborg, C., Farrall, M., Assimes, T. L., Thompson, J. R.
… Samani, N. J. (2013). Large-scale association analysis identifies new risk loci for
coronary artery disease. Nature Genetics, 45(1), 25–33. doi:10.1038/ng.2480
Farooqi, I. S., & O’Rahilly, S. (2006). Genetics of obesity in humans. Endocrine Reviews,
27(7), 710–718.
Fisher, R. A. (1918). The correlation between relatives on the supposition of
Mendelian inheritance. Transactions of the Royal Society of Edinburgh, 52, 399–433.
DNA Revolution and the Social and Behavioral Sciences
13
Gershon, E. S., Alliey-Rodriguez, N., & Liu, C. (2011). After GWAS: Searching for
genetic risk for schizophrenia and bipolar disorder. American Journal of Psychiatry,
168(3), 253–256. doi:10.1176/appi.ajp.2010.10091340
Gibbs, R. A., Belmont, J. W., Hardenbol, P., Willis, T. D., Yu, F., Yang, H. … Shen, Y.
(2003). The international HapMap project. Nature, 426(6968), 789–796.
H.G.P. (2001). The human genome. Nature, 409(6822), 745–964.
Hirschhorn, J. N., & Daly, M. J. (2005). Genome-wide association studies for common
diseases and complex traits. Nature Reviews Genetics, 6(2), 95–108.
Kendler, K. S., Gardner, C. O., & Lichtenstein, P. (2008). A developmental twin
study of symptoms of anxiety and depression: Evidence for genetic innovation and attenuation. Psychological Medicine, 38(11), 1567–1575. doi:10.1017/
S003329170800384X
Kovas, Y., Haworth, C. M. A., Dale, P. S., & Plomin, R. (2007). The genetic and environmental origins of learning abilities and disabilities in the early school years
7912. Monographs of the Society for Research in Child Development, 72, 1–144.
Lango Allen, H., Estrada, K., Lettre, G., Berndt, S. I., Weedon, M. N., Rivadeneira,
F. … Hirschhorn, J. N. (2010). Hundreds of variants clustered in genomic
loci and biological pathways affect human height. Nature, 467(7317), 832–838.
http://www.nature.com/nature/journal/v467/n7317/abs/nature09410.html supplementary-information
Lee, S. H., Wray, N. R., Goddard, M. E., & Visscher, P. M. (2011). Estimating missing
heritability for disease from genome-wide association studies. American Journal of
Human Genetics, 88(3), 294–305. doi:10.1016/j.ajhg.2011.02.002
Lee, S. H., Yang, J., Goddard, M. E., Visscher, P. M., & Wray, N. R. (2012). Estimation of pleiotropy between complex diseases using SNP-derived genomic
relationships and restricted maximum likelihood. Bioinformatics. doi:10.1093/
bioinformatics/bts474
Llewellyn, C. H., Trzaskowski, M., Plomin, R., & Wardle, J. (2013). Finding the missing heritability in pediatric obesity: The contribution of genome-wide complex
trait analysis. International Journal of Obesity, 37, 1506–1509.
Llewellyn, C. H., Trzaskowski, M., Plomin, R., & Wardle, J. (2014). From modelling
to measurement: Developmental trends in genetic influence on adiposity in childhood. Obesity (Silver Spring), 22, 1756–1761.
Lubke, G. H., Hottenga, J. J., Walters, R., Laurin, C., de Geus, E. J., Willemsen, G.
… Boomsma, D. I. (2012). Estimating the genetic variance of major depressive
disorder due to all single nucleotide polymorphisms. Biological Psychiatry, 72(8),
707–709. doi:10.1016/j.biopsych.2012.03.011
Marioni, R. E., Penke, L., Davies, G., Huffman, J. E., Hayward, C., & Deary, I. J. (2014).
The total burden of rare, non-synonymous exome genetic variants is not associated with childhood or late-life cognitive ability. Proceedings of the Royal Society B:
Biological Sciences, 281(1781), 20140117. doi:10.1098/rspb.2014.0117
Mercer, T. R., Dinger, M. E., & Mattick, J. S. (2009). Long non-coding RNAs: Insights
into functions. Nature Reviews Genetics, 10(3), 155–159.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Muotri, A. R., Marchetto, M. C. N., Coufal, N. G., & Gage, F. H. (2007). The necessary
junk: New functions for transposable elements. Human Molecular Genetics, 16(R2),
R159–R167.
Neale, B. M., Rivas, M. A., Voight, B. F., Altshuler, D., Devlin, B., Orho-Melander,
M. … Daly, M. J. (2011). Testing for an unusual distribution of rare variants. PLoS
Genetics, 7(3), e1001322. 10.1371/journal.pgen.1001322
Nichols, P. L. (1984). Familial mental retardation. Behavior Genetics, 14(3), 161–170.
Plomin, R., DeFries, J. C., Knopik, V. S., & Neiderhiser, J. M. (2013). Behavioral genetics
(6 ed.). New York, NY: Worth Publishers.
Plomin, R., Haworth, C. M., & Davis, O. S. (2009). Common disorders are quantitative
traits. Nature Reviews Genetics, 10, 872–878.
Plomin, R., Haworth, C. M., Meaburn, E. L., Price, T. S., & Davis, O. S. (2013). Common DNA markers can account for more than half of the genetic influence on cognitive abilities. Psychological Science, 24(4), 562–568. doi:10.1177/0956797612457952
Psychiatric Gwas Consortium Bipolar Disorder Working, & Group (2011). Largescale genome-wide association analysis of bipolar disorder identifies a new
susceptibility locus near ODZ4. Nature Genetics, 43(10), 977–983. doi:10.1038/
ng.943
Purcell, S. M., Moran, J. L., Fromer, M., Ruderfer, D., Solovieff, N., Roussos, P. …
Sklar, P. (2014). A polygenic burden of rare disruptive mutations in schizophrenia.
Nature, 506(7487), 185–190. doi:10.1038/nature12975
Rietveld, C. A., Cesarini, D., Benjamin, D. J., Koellinger, P. D., De Neve, J. E., Tiemeier,
H. … Bartels, M. (2013). Molecular genetics and subjective well-being. Proceedings
of the National Academy of Sciences of the United States of America, 110(24), 9692–9697.
doi:10.1073/pnas.1222171110
Rietveld, C. A., Medland, S. E., Derringer, J., Yang, J., Esko, T., Martin, N. W.
… Koellinger, P. D. (2013). GWAS of 126,559 individuals identifies genetic
variants associated with educational attainment. Science, 340(6139), 1467–1471.
doi:10.1126/science.1235488
Rochman, B. (2012). Want to know my future. Parents grapple with delving into their
kids’ DNA. Time, 180, 26.
Schizophrenia Psychiatric Genome-Wide Association Study, C (2011). Genome-wide
association study identifies five new schizophrenia loci. Nature Genetics, 43(10),
969–976. doi:10.1038/ng.940
Shimoni, Y., Friedlander, G., Hetzroni, G., Niv, G., Altuvia, S., Biham, O., & Margalit,
H. (2007). Regulation of gene expression by small non-coding RNAs: A quantitative view. Molecular Systems Biology, 3(1).
Siva, N. (2008). 1000 Genomes project. Nature Biotechnology, 26(3), 256–256.
Speliotes, E. K., Willer, C. J., Berndt, S. I., Monda, K. L., Thorleifsson, G., Jackson,
A. U. … Loos, R. J. F. (2010). Association analyses of 249,796 individuals reveal
18 new loci associated with body mass index. Nature Genetics, 42(11), 937–948.
http://www.nature.com/ng/journal/v42/n11/abs/ng.686.html-supplementaryinformation
DNA Revolution and the Social and Behavioral Sciences
15
Stankiewicz, P., & Lupski, J. R. (2010). Structural variation in the human genome
and its role in disease. Annual Review of Medicine, 61, 437–455. doi:10.1146/
annurev-med-100708-204735
Tabor, H. K., Risch, N. J., & Myers, R. M. (2002). Candidate-gene approaches for
studying complex genetic traits: Practical considerations. Nature Reviews Genetics,
3(5), 391–397.
Thapar, A., & McGuffin, P. (1997). Anxiety and depressive symptoms in childhood—
A genetic study of comorbidity. Journal of Child Psychology and Psychiatry and Allied
Disciplines, 38(6), 651–656.
Trzaskowski, M., Dale, P. S., & Plomin, R. (2013). No genetic influence for childhood
behavior problems from DNA analysis. Journal of the American Academy of Child
and Adolescent Psychiatry, 52(10), 1048–1056 e1043. doi:10.1016/j.jaac.2013.07.016
Trzaskowski, M., Davis, O. P., DeFries, J., Yang, J., Visscher, P., & Plomin, R. (2013).
DNA evidence for strong genome-wide pleiotropy of cognitive and learning abilities. Behavior Genetics, 1–7. doi:10.1007/s10519-013-9594-x
Trzaskowski, M., Eley, T. C., Davis, O. S. P., Doherty, S. J., Hanscombe, K. B.,
Meaburn, E. L. … Plomin, R. (2013). First genome-wide association study on
anxiety-related behaviours in childhood. PloS One, 8(4), e58676. 10.1371/journal.
pone.0058676
Trzaskowski, M., Harlaar, N., Arden, R., Krapohl, E., Rimfeld, K., McMillan, A. …
Plomin, R. (2014). Genetic influence on family socioeconomic status and children’s
intelligence. Intelligence, 42, 83–88. doi:10.1016/j.intell.2013.11.002
Trzaskowski, M., Shakeshaft, N. G., & Plomin, R. (2013). Intelligence indexes
generalist genes for cognitive abilities. Intelligence, 41(5), 560–565. doi:10.1016/
j.intell.2013.07.011
Trzaskowski, M., Yang, J., Visscher, P. M., & Plomin, R. (2013). DNA evidence for strong
genetic stability and increasing heritability of intelligence from age 7 to 12 (Vol. 19,
pp. 380–384). Molecular Psychiatry.
Verweij, K. J., Vinkhuyzen, A. A., Benyamin, B., Lynskey, M. T., Quaye, L., Agrawal,
A. … Medland, S. E. (2012). The genetic aetiology of cannabis use initiation: A
meta-analysis of genome-wide association studies and a SNP-based heritability
estimation. Addiction Biology. 10.1111/j.1369-1600.2012.00478.x
Vinkhuyzen, A. A., Pedersen, N. L., Yang, J., Lee, S. H., Magnusson, P. K., Iacono,
W. G. … Wray, N. R. (2012). Common SNPs explain some of the variation in the
personality dimensions of neuroticism and extraversion. Translational Psychiatry,
2, e102. 10.1038/tp.2012.27
Visscher, P. M. (2008). Sizing up human height variation. Nature Genetics, 40(5),
489–490. doi:10.1038/ng0508-489
Visscher, P. M., Brown, M. A., McCarthy, M. I., & Yang, J. (2012). Five years of
GWAS discovery. The American Journal of Human Genetics, 90(1), 7–24. doi:10.1016/
j.ajhg.2011.11.029
Wray, N. R., Lee, S. H., Mehta, D., Vinkhuyzen, A. A. E., Dudbridge, F., & Middeldorp, C. M. (2014). Research review: Polygenic methods and their application
to psychiatric traits. Journal of Child Psychology and Psychiatry, 55(10), 1068–1087.
doi:10.1111/jcpp.12295
16
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
WTCCC (2007). Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature, 447(7145), 661–678. doi:10.1038/
nature05911
Yang, J., Benyamin, B., McEvoy, B. P., Gordon, S. D., Henders, A. K., Nyholt, D. R.
… Visscher, P. M. (2010). Common SNPs explain a large proportion of the heritability for human height. Nature Genetics, 42(7), 565–569. http://www.nature.
com/ng/journal/v42/n7/suppinfo/ng.608_S1.html
Yang, J., Lee, S., Goddard, M., & Visscher, P. (2011). GCTA: A tool for genome-wide
complex trait analysis. American Journal of Human Genetics, 88(1), 76–82.
doi:10.1016/j.ajhg.2010.11.011
MACIEJ TRZASKOWSKI SHORT BIOGRAPHY
While studying for his PhD, Dr Maciej Trzaskowski has learnt and applied
an extensive selection of statistical methods for quantitative (structural
equation modeling) and molecular genetics (e.g., GWAS, and GREML),
using various programming languages (e.g., R, Python, Unix, and Shell
Scripts). Dr. Trzaskowski has a strong methodological propensity and high
interest in exploring new methods. He has been awarded Gottesman-Shields
Prize for the best PhD thesis 2013 and since the completion of his PhD, Dr.
Trzaskowski remained at the Social, Genetic and Developmental Psychiatry
Centre as a postdoctoral research worker. To date, he has published 23
papers, 11 of these as first/joint first author. He has built his own network
of collaborations with researchers in the Broad Institute, the Queensland
Institute of Medical Research, Queensland Brain Institute (the University of
Queensland), Vrije Universiteit (VU) Amsterdam, Harvard, amongst others.
His research is increasingly recognized (as evidenced by the Behavior
Genetics Association Thompson and Fulker Awards received in 2013 and
2014, respectively) and, most recently, by a successful award of a fellowship
from British Academy. He has given invited talks at the London School
of Economics (by Lord Richard Layard, Emeritus Professor of Economics)
and at the fourth annual meeting by the Social Science Genetic Association
Consortium (by Professor Phillip Koellinger).
ROBERT PLOMIN SHORT BIOGRAPHY
Since 1994, Professor Robert Plomin has been MRC Research Professor of
Behavioral Genetics at the Institute of Psychiatry, King’s College London.
In 1994, he cofounded and subsequently directed the MRC Social, Genetic
and Developmental Psychiatry Centre, whose goal is to bring together
genetic and environmental strategies to study behavioral development.
In 1995, he launched the Twins Early Development Study (TEDs) of all
DNA Revolution and the Social and Behavioral Sciences
17
twins born in England and Wales in 1994–1996, which focuses on developmental problems in cognition and behavior, and which Professor Plomin
continues to direct. He has published more than 500 papers and more
than a dozen books, including the major textbook in the field (“Behavioral
Genetics,” Worth Publishers, 6th edition, 2013). His most recent book,
coauthored with Kathryn Asbury, focuses on genetics and education (“G
is for Genes: the Impact of Genetics on Education and Achievement,” Wiley
Blackwell, 2013). He has received lifetime research achievement awards
from the three major international associations in his field. For details, see
https://kclpure.kcl.ac.uk/portal/robert.plomin.html.
RELATED ESSAYS
Telomeres (Psychology), Nancy Adler and Aoife O’Donovan
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
Biology and Culture (Psychology), Robert Peter Hobson
Genetics and the Life Course (Sociology), Evan Charney
Genetic and Environmental Approaches to Political Science (Political Science),
Zoltán Fazekas and Peter K. Hatemi
Genetics and Social Behavior (Anthropology), Henry Harpending and Gregory Cochran
A Gene-Environment Approach to Understanding Youth Antisocial Behavior (Psychology), Rebecca Waller et al.
Behavioral Heterochrony (Anthropology), Victoria Wobber and Brian Hare
-
DNA Revolution and the Social
and Behavioral Sciences
MACIEJ TRZASKOWSKI and ROBERT PLOMIN
Abstract
A century of genetic research on the social and behavioral sciences has addressed
the “how much” question, showing that genetic differences are important for nearly
all traits. However, during the past few decades, research has moved beyond this
rudimentary “how much” question to ask “how” questions about developmental
change and continuity, about the relationship between traits, and about the interplay between genes (nature) and environment (nurture). We suggest that some of
the most important and transformative findings in the social and behavioral sciences
have come from this research. Moreover, the most revolutionary changes in genetic
research are on their way with the rapid advances in DNA technology and methodology, which promise to transform the social and behavioral sciences. It is crucial that
social and behavioral scientists stay on top of the DNA revolution. The purpose of
our essay is to provide an overview of genetic research in the social and behavioral
sciences.
NATURE AND NURTURE: TWIN AND ADOPTION STUDIES
In general, genetic research can be divided into quantitative genetics
(statistical designs in which family members are used to estimate genetic
influence) and molecular genetics (research designs where genetic influence
is estimated by directly measured genotypes). For decades quantitative
geneticists used family, twin and adoption designs to estimate genetic and
environmental influences on many psychological and psychiatric traits. The
strength of these designs is that they can estimate the bottom line of genetic
influence regardless of how many genes are involved or how complex their
effects, unlike molecular genetics. All quantitative genetic designs have
limitations, but each design has different limitations, and, most importantly,
their results converge. For example, the adoption design, which typically
compares adopted children with their biological and adoptive parents, is
limited by possible prenatal factors and selective placement. In contrast, the
twin design, which compares monozygotic (MZ) twins who are genetically
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
100% identical, and dizygotic (DZ) twins who share on average 50% of their
segregating alleles, is limited by the assumption of equal environments
for the two types of twins. These limitations are specific to each design
and yet adoption and twin studies generally produce similar estimates of
genetic and environmental influences, especially for cognitive traits (Plomin,
DeFries, Knopik, & Neiderhiser, 2013). Adoption studies are less common
today than twin studies because of the sharp reduction in neonatal adoption,
but twin studies continue to proliferate.
Univariate twin studies have demonstrated consistently that virtually
every human trait is influenced in part by genetic variation. Multivariate
models go beyond this rudimentary finding about the relative influence
of nature and nurture on a single phenotype, to investigate the common
genetic architecture underlying multiple complex traits. Multivariate twin
designs decompose the covariance between multiple phenotypes to provide
insights into the common genetic and environmental relationship underlying different traits, as well as the genetic and environmental contributions
to their development over time (trait stability and change). Multivariate
genetic research has yielded three major findings consistent across many
complex traits: heritability increases across development, genetic effects
are developmentally stable—strong age-to-age genetic correlations—and
are highly pleiotropic, that is, one gene influences many traits. Genetic
pleiotropy is inferred from consistent findings of strong genetic covariance
between different traits at one measurement occasion: for example, this
is true for anxiety and depression (e.g., Thapar & McGuffin, 1997), and
for IQ and learning abilities (Davis, Haworth, & Plomin, 2009). Genetic
stability is inferred from high genetic correlations between the “same”
traits across time. Curiously, despite the observed genetic stability from
age to age, many traits appear to show increases in heritability across
development (Bergen, Gardner, & Kendler, 2007). Although genetic stability
and increasing heritability might appear paradoxical, there are several
mechanisms that might explain this. The phenomenon could be driven by a
correlation between genes and environments where, for example, children
choose environments that suit their genetic propensities (Plomin, DeFries,
et al., 2013). Imagine a young girl who has a propensity for reading; as she
grows up, she chooses environments in which she can be exposed to reading
as much as possible. This “tailored” selection will increase the importance
of these in explaining individual differences in reading and yet they will be
the same genes. It is also possible that heritability increases as new genes
come “online” over the lifetime explaining more of the variance (Kendler,
Gardner, & Lichtenstein, 2008). Unfortunately, twin designs cannot easily
tease out which of these mechanisms are responsible for the phenomena.
The most important advance in this field that will help to resolve this issue
DNA Revolution and the Social and Behavioral Sciences
3
and many others will come when specific genetic variants are identified that
account for these genetic effects.
LOW-HANGING FRUIT: MONOGENIC DISORDERS
Mendel, in his original work with peas, was fortunate in that he chose monogenic traits in which a single mutation is necessary and sufficient for the
trait to be expressed. For such single-gene traits the underlying genetic variants travel across generations according to predictions based on Mendel’s
first law of segregation, and they also assort independently from variants
of other traits according to Mendel’s second law. Unfortunately, complex
traits appeared not to conform to these laws, prompting some researchers
in the early twentieth century to argue that Mendel’s findings were limited
to peas, and not “higher-order” species. Although the dispute between the
two factions was at times fierce, it was resolved when it became apparent
that in complex traits Mendel’s laws hold for segregation of alleles at a single
locus but are masked by the additive influence of many such loci. The overall genetic influence is driven by many genes, called polygenic, and thus the
effects of individual genes is small and their overall effect on the phenotype
is complex and normally distributed in a population. This is the cornerstone
of quantitative genetics expounded by Fisher (1918), who described how
Mendel’s model could be extended to multiple genes in order to account for
inheritance of complex traits.
In molecular genetics, Mendel’s law of segregation culminated in huge
success in discovering genetic associations with monogenic disorders
mainly through implementation of linkage designs using just a few hundred
DNA markers across the genome. Linkage analysis uses samples drawn
from multigenerational families with affected and unaffected members,
where DNA of all family members is “tagged” with several markers whose
location on chromosomes is known. The aim is to detect a “link” between
one of these markers and a trait or disorder. Linkage successfully detected
genetic associations with many Mendelian (monogenic) traits because a
mutation in a single locus explained all the variance. However, monogenic
disorders affect a tiny proportion of clinical populations; most of the burden
of illness involves complex disorders influenced by many genes with
small effect sizes. Although linkage has been very successful in identifying
genes for single-gene disorders, the method does not have sufficient power
to detect genes of small effect size. A technological advance that greatly
improved attempts to find genes responsible for heritability of complex traits
is genome-wide association (GWA), which is based on allelic association
between DNA markers and a trait in unrelated individuals in the population
rather than linkage within families.
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
FINDING MANY NEEDLES IN THE HAYSTACK: GENOME-WIDE
ASSOCIATION STUDIES
The discovery that many complex traits are highly heritable motivated
researchers to search for the genetic variants associated with these traits.
The pre-GWA era relied mainly on candidate gene association studies,
where a specific region of genome was hypothesized to be important
for development of a particular trait. Unfortunately, a problem with the
candidate gene approach is that the hypotheses generated about the regions
we think should be important are often wrong. Many published candidate
gene association results is now known to be false positive results (Tabor,
Risch, & Myers, 2002). This is particularly telling when replication fails in
GWA analysis as this design is hypothesis free in the sense that it does not
“care” which region is associated with which trait. An example of such
failure is a GWA study of nearly 10,000 individuals where none of the 10
most frequently reported candidate regions for general cognitive ability
replicated (Chabris et al., 2012).
The genomic era began in the past decade when two major developments
provided cost-effective, thorough coverage of the genome. The first event
was the completion of the Human Genome Project (H.G.P., 2001), which
resulted in the first detailed maps of the human genome and the patterns of
linkage disequilibrium (the nonrandom association of alleles across neighboring genetic loci) for hundreds of thousands SNPs (single-nucleotide
polymorphism—variation in a single genomic base-pair). The second event
was the production of DNA arrays or “chips.” A single DNA chip could
genotypes hundreds of thousands of SNPs. For each SNP, the DNA chip
includes many probes, which is a short DNA sequence containing the
SNP. Understanding linkage disequilibrium patterns across the genome
was crucial because it showed that careful selection of a few hundred
thousand DNA markers was sufficient to comprehensively tag the genome’s
3 billion DNA base pairs for GWA studies. DNA variants close together
on a chromosome violate Mendel’s second law by segregating together
producing high between-SNP correlations. A single marker from DNA
variants in high linkage disequilibrium is sufficient to “tag” the region. This
knowledge was used to select SNPs for DNA chips. In addition, coverage
of the genome could be improved by imputing millions of additional SNPs
from existing reference maps, such as HapMap (Gibbs et al., 2003) or, more
recently, the 1000 Genomes Project (Siva, 2008) without actually genotyping
these additional SNPs. This meant that it was now possible to interrogate
the whole genome simultaneously for association between allele frequencies
of individual SNPs and any trait varying in unrelated individuals. This
DNA Revolution and the Social and Behavioral Sciences
5
method is known today as a GWA (e.g., Balding, 2006; Cardon & Bell, 2001;
Hirschhorn & Daly, 2005).
In GWA, associations are usually examined one SNP at the time, assuming
an additive model, although other models can also be implemented. Since
current genetic chips can assess millions of SNPs, the correction for multiple
testing is daunting, and the number of possible false positives is large; the
accepted p-value threshold for genome-wide significance is p < 5 × 10−8 ,
which can be thought of as p < 0.05 with a Bonferroni correction for one
million statistical tests. Despite such stringent correction, within the three
years between 2005 and 2008, more than 400 SNP associations for just
over 120 traits (www.genome.gov/GWAstudies) achieved this stringent
genome-wide threshold, and revealed unexpected insights about genetic
influences on complex traits (Visscher, Brown, McCarthy, & Yang, 2012).
Arguably the most useful biological insights in genetic associations have
been mainly limited to medical traits. For example, in Crohn’s disease,
many SNPs reported through GWA studies were found in and around genes
involved with autophagy, the cell’s maintenance process that breaks down
dysfunctional components of the cell (WTCCC, 2007). In addition, the same
study showed that type 2 diabetes was associated with loci encoding for
proteins relevant to insulin secretion, and not insulin signaling, as previously thought. Nevertheless, even if functionally not so obvious, many new
loci have also been identified in anthropometric (e.g., Speliotes et al., 2010;
Visscher, 2008) and some psychiatric traits (e.g., Gershon, Alliey-Rodriguez,
& Liu, 2011; Purcell et al., 2014). Unfortunately successes in cognitive, social,
and behavioral complex traits are yet to come.
Despite GWA successes, fewer associations were reported for complex
traits than expected, and associations that were identified could together
only account for a small fraction of twin-estimated heritability. The phenomenon is known as “missing heritability,” which is discussed in the
next section. The important lesson from GWA research is that the largest
effect sizes of individual loci of common SNPs are incredibly small and
require sample sizes in the hundreds of thousands to achieve genome-wide
significance. This realization led to the sudden emergence and proliferation
of world-wide consortia, which paid off by doubling the GWA hits to just
under 9000 SNPs in more than 700 traits (www.genome.gov/GWAstudies/).
Nonetheless, the yield was still not as substantial as expected in that the
largest effect sizes, such as the association between the FTO gene and body
mass index (BMI), are less than 1% of the variance (Speliotes et al., 2010),
which implies that the smallest effect sizes are likely to be infinitesimal.
This means that it will be difficult to detect and replicate associations with
complex traits in the social and behavioral sciences. However, once several
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
genes are found that are associated with these traits, their effects can be
aggregated as polygenic scores, as described in the following section.
TOGETHER WE ARE STRONGER: POLYGENIC SCORES
Even though the effects of individual SNPs are very small, their effects can
be aggregated to increase the total amount of variance explained and thus
increase power. For example, if we had 10 individual SNPs, each with an
effect size of 0.5%, a polygenic score comprising all of these would account
for 5% of the variance, which means that a sample of 150 individuals would
have 80% power to detect their cumulative effect. For this reason, polygenic
scores are now being used in a new generation of hypothesis-free candidate
gene association studies and are the way that the DNA revolution will come
to the social and behavioral sciences.
The creation of a polygenic score takes into account the direction of the
effect (i.e., which allele is the “increasing” allele) and it can also weight the
associations by the magnitude of their effect. To give priority to stronger associations, the scores can be weighted by the betas from the regression (i.e.,
their effect size). The associations can then be summed, similar to summing
items on a scale. Although summing scores is most commonly used, nonadditivity can also be incorporated. In addition, SNPs can either be selected
from previously reported significant “hits” (polygenic scores—PGS) or they
can be amassed by selecting all associations below an arbitrary p-value from
GWA studies, a genome-wide PGS (Wray et al., 2014). The former method is
preferred because of the trade-off between specificity and size but the choice
depends on the availability of previously reported robust “hits.” The tiny
effect sizes and moderate-to-large estimates of heritability suggest that the
more SNPs you “pool” together the more variance you should explain. However, adding a large number of SNPs with no effect or effects in the opposite
direction can attenuate the signal.
Application of polygenic predictors to medical disorders had some success
due at least in part to availability of financial resources and large sample
sizes. It has been shown that PGS of 150 SNPs accounted for 5% of the
variance in the liability for coronary artery disease (Deloukas et al., 2013).
For bipolar disorder, SNPs accounted for between 1% and 3% of the liability
(Psychiatric Gwas Consortium Bipolar Disorder Working & Group, 2011)
and a PGS of significant SNPs from discovery GWAS in schizophrenia,
accounted for approximately 1% of the liability in the independent sample
(Schizophrenia Psychiatric Genome-Wide Association Study, 2011). In the
schizophrenia study, extending the PGS to genome-wide PGS increased
variance explained to almost 6%. PGS research on quantitative traits has
largely been limited to weight and height, and other complex continuous
DNA Revolution and the Social and Behavioral Sciences
7
traits have had to resort to genome-wide PGS. GWA meta-analysis of BMI
revealed that 32 replicated SNPs accounted for only ∼1.5% of the variance
in BMI in independent samples, but a genome-wide PGS that included
thousands of SNPs explained ∼5% of the variance (Speliotes et al., 2010).
For height, 180 SNPs accounted for 10% of the variance, and genome-wide
PGS increased the variance explained to 13% (Lango Allen et al., 2010).
Importantly, investigation of behavioral traits of direct interest to social
and cognitive scientists are now achieving sample sizes that enable more
powerful genomic interrogation. Application of genome-wide PGS to these
traits is producing results that are similar to medical and anthropometric
traits (height and weight). Specifically, adding more SNPs increases amount
of variance explained up to a point. For example, an increase of variance
explained with an addition of more SNPs was reported for total years of
education (1% using 3500 SNPs and 2.5% using 2.5 million SNPs; Rietveld,
Medland, et al., 2013) and childhood IQ (using polygenic predictor at
p < 6 × 10−5 explained 0.5–1.2% increasing to 3.5% at p < 0.001; Benyamin
et al., 2014). These findings are important despite the small proportions of
variance explained, because they suggest that complex traits are indeed
highly polygenic.
It is now obvious that even though inclusion of a large number of SNPs
consistently increases the amount of variance explained, the gap between
twin-estimated heritability and these PGS estimates is still very wide. The
phenomenon known today as the “missing heritability” continues to pose
many questions, which are as yet unanswered. Twin studies suggest that
almost every human trait from our biology through cognition, behavior and
even the environment is heritable, with genetic influences usually explaining
moderate (30–40%) to high (80–90%) proportions of variance. GWA studies
have accounted for only a small fraction of this heritability. The most general
explanation is that the influence of each individual SNP is so small that most
of the GWA studies thus far have been greatly underpowered to detect them.
Another likely source of missing heritability is rarer variants—the markers
selected for the DNA arrays were limited to common variants only (minor
allele frequency>1%). This meant that many potential “true” associations
with SNPs of lower allele frequency could be missed owing to low linkage disequilibrium with the markers. Another possible source of missing
heritability is nonadditive effects, such as gene–environment or gene–gene
interactions because GWA is limited to additive genetic effects. Finally, it is
possible that heritability estimates derived from twin data are inflated. To
answer some of these questions, a new quantitative genetic technique has
emerged, called genome-wide complex trait analysis (GCTA), as described
in the following section.
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
HERITABILITY WITHOUT TWINS: GENOME-WIDE COMPLEX
TRAIT ANALYSIS (GCTA)
A recently developed method uses genome-wide genotyping data in a novel
way to address these questions. The linear mixed model implemented in
the GCTA package estimates the amount of phenotypic variance (Yang, Lee,
Goddard, & Visscher, 2011) and co-variance (Lee, Yang, Goddard, Visscher,
& Wray, 2012) that can be explained by additive effects of all common SNPs
tagged by DNA arrays. Several other methods have been developed with
the same intent (Wray et al., 2014) but GCTA is currently most widely used.
GCTA itself has been given different names, such as linear mixed model
and genomic-relationship-matrix restricted maximum likelihood. However,
clever usage of the four DNA base letters makes GCTA the catchiest acronym,
and this is what we call it henceforth.
GCTA uses genome-wide genotype data from unrelated individuals to estimate genetic influence on a trait that can be explained by all SNPs included
on DNA chips. Because GCTA is based on genome-wide DNA data alone,
it can be used to estimate genetic influence for unrelated individuals rather
than requiring special relatives such as MZ and DZ twins. For that reason, it
bypasses some of the assumptions of the twin method, although it has its own
set of assumptions (Plomin, Haworth, Meaburn, Price, & Davis, 2013). GCTA
can only detect genetic influence owing to the additive effects of common
SNPs that are included on currently available DNA arrays and cannot evaluate the contribution of any specific DNA locus. Nonetheless, GCTA provides
important information about the extent to which the genetic architecture of
complex traits includes additive effects of common SNPs, and sets the limit
for detecting associations in GWA studies.
GCTA analyses have shown that information captured by current DNA
arrays can explain a substantial amount of the variance in complex traits,
including human height (Yang et al., 2010), BMI (Llewellyn, Trzaskowski,
Plomin, & Wardle, 2013), psychiatric and medical disorders (Lee, Wray,
Goddard, & Visscher, 2011; Lee et al., 2012; Lubke et al., 2012), cognitive
traits (Deary et al., 2012; Plomin, Haworth, et al., 2013), and economic and
political preferences (Benjamin et al., 2012). All of these GTCA heritability
estimates are approximately half of twin-estimated heritability. In contrast,
initial research in psychopathology and personality is less consistent,
showing near zero variance explained for most analyses. For example,
analyses of a wide range of behavioral problems (symptoms of anxiety
and depression, hyperactivity, conduct problems) show negligible SNP
heritability despite moderate to high twin heritability estimated within the
same sample (Trzaskowski, Dale, & Plomin, 2013; Trzaskowski, Eley, et al.,
2013). SNP heritabilities of neuroticism and extraversion were reported as
DNA Revolution and the Social and Behavioral Sciences
9
0.06 (SE 0.03) and 0.12 (SE 0.03) respectively (Vinkhuyzen et al., 2012), and
similar near-zero GCTA estimates have also been shown for other aspects
of personality (Verweij et al., 2012), including wellbeing (Rietveld, Cesarini,
et al., 2013). The results are puzzling as all of the studies are adequately
powered to detect the expected effect sizes (∼40% of twin heritability)
and are not limited to self-report questionnaire. For example, one study
examined behavior problems reported by parents and teachers as well as
the children themselves; further discussion on this topic can also be found
in the same publication (Trzaskowski, Dale, et al., 2013).
As was the case for early twin studies, GCTA was initially applied in univariate analyses of one trait at a time, but like twin studies, the model was
extended to the bivariate analysis of covariance between traits or across age.
For example, bivariate GCTA was first applied to the remarkable phenotypic
stability of IQ across 60 years from childhood to later life (phenotypic correlation = 0.63) and suggested that the stability is largely due to genetic stability
(genetic correlation = 0.62) (Deary et al., 2012). Twin studies also suggested
that the genetic stability was present despite heritability of IQ increasing
across development. Bivariate GCTA supported this finding of genetic stability despite increasing heritability for IQ (Trzaskowski, Yang, Visscher, &
Plomin, 2013) and similar twin and GCTA results were reported for BMI
(Llewellyn, Trzaskowski, Plomin, & Wardle, 2014). The most likely explanation for these phenomena is gene–environment correlation (Plomin, DeFries,
et al., 2013). The high genetic correlation suggests that the same genes influence individual differences in IQ across time. However, as we grow older we
increasingly select environments that “match” our genetic predispositions
making genetic influence stronger even though the same genes are involved.
Although twin studies have reported for decades that most environments are nearly as heritable as behaviors, this work has been limited to
twin-specific environments. GCTA opens up the possibility of investigating
genetic influence on family-, neighborhood-, or even country-wide environmental measures that cannot be studied using the twin design because
they are shared in common by members of a twin pair (Trzaskowski et al.,
2014). This feature of GCTA should be particularly interesting to social
and behavioral scientists as it emphasizes the important interplay between
genes and environments. It shows that environments are not simply “out
there” that happen randomly to us, but that our genes shape our experiences
through our selection, modification and creation of our environments.
Another widely reported finding from twin analyses is the strong genetic
correlation (pleiotropy) across different aspects of cognition and across
diverse cognitive abilities (Davis et al., 2009; Kovas, Haworth, Dale, &
Plomin, 2007; Plomin, DeFries, et al., 2013). GCTA studies reported point
estimates for genetic correlations highly similar to those reported by twin
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
studies (Trzaskowski, Davis, et al., 2013). GCTA estimates of genetic correlation are similar to twin study estimates of genetic correlation, not only within
traditional domains such as verbal and nonverbal (Trzaskowski, Shakeshaft,
& Plomin, 2013) but also between intelligence and education-related skills
such as reading and mathematics (Trzaskowski, Davis, et al., 2013). We
can expect many more multivariate GCTA analyses in domains other than
cognition, most notably psychopathology where it increasingly appears that
there is a great deal of overlap among diverse symptoms (Caspi et al., 2013).
In summary, GCTA has shed some light on the nature of the discrepancy
between heritability estimates from twin studies and combined variance
explained from SNPs identified through GWA. Some of the “missing
heritability” is hiding in tiny influences of many common SNPs. We called
that gap between GWA and GCTA “missing GWA heritability,” because
GCTA represents the ceiling for GWA results. This gap can in theory be
filled with the additive effects of variants tagged by the SNPs on current
DNA chips if samples are large enough to detect nearly infinitesimal effect
sizes. GCTA results also suggest that even if the “missing GWA heritability”
were filled, there would still be a substantial chunk of genetic influence
not accounted for by additive effects of common SNPs. This part of the
“missing heritability” could be called “missing GCTA heritability,” which
falls short of twin study heritability because GCTA only reflect additive
effects of common SNPs. We delineated these two parts to emphasize a
distinction between missing heritability that is caused by lack of power (the
effect sizes of already captured common variants are too small to detect
with current sample sizes), and missing heritability that is due to genetic
variants that have not yet been captured (e.g., rare variants or nonadditive
genetic influences). The gap of missing GWA heritability can be narrowed
by the brute force of larger samples, but how can DNA variants responsible
for missing GCTA heritability be identified?
3 BILLION BASE PAIRS: WHOLE-GENOME SEQUENCING
Whole-genome sequencing involves genotyping all 3 billion base pairs of
DNA, rather than just a million or so SNPs as on current genetic chips.
Whole-genome sequencing is the end of the story of genetic variation in
the sense that all we inherit from our parents is differences in DNA base
pair sequences. Since current genetic chips only capture common SNP,
whole-genome sequencing could identify more genetic loci associated with
complex traits because it captures variants of any kind, not just common
SNPs. Rare variants may play a particularly important role in the extreme
tails of a trait’s distribution. For example, common SNPs influence BMI
across the distribution of the normal population, but very rare Mendelian
DNA Revolution and the Social and Behavioral Sciences
11
mutations may account for as much as 5% of extreme obesity (Farooqi &
O’Rahilly, 2006).
Other clinical phenotypes, such as intellectual disability, schizophrenia,
bipolar disorder, and autism, have been associated with an increased burden
of rare variants, including idiosyncratic mutations that first occur in that
individual and are not inherited from the individual’s parents, called de
novo mutations (Marioni et al., 2014; Neale et al., 2011; Purcell et al., 2014;
Stankiewicz & Lupski, 2010). Perhaps rare variants are responsible for a
puzzle in findings concerning mild and severe intellectual disability. One
study found that mild intellectual impairment was familial, but severe
impairment was not (Nichols, 1984). That is, siblings of severely mentally
retarded children showed no mental impairment, whereas siblings of children with “mild” mental disability did, suggesting that severe impairment
is not genetically related to common variation in mental ability. In general,
common (less severe) disorders are likely to be the quantitative extremes of
normal variation (Plomin, Haworth, & Davis, 2009), whereas extreme levels
of disability could be mainly a result of accumulation of much rarer variants,
including de novo mutations as well as environmental “mutations” such as
perinatal trauma.
Sequencing will also give us richer information about noncoding regions.
Not so long ago noncoding regions of the genome were thought of as an
evolutionary “junk,” but now these noncoding regions are known to play an
important role in regulation of genetic expression and even creation of new
genes (Mercer, Dinger, & Mattick, 2009; Muotri, Marchetto, Coufal, & Gage,
2007; Shimoni et al., 2007). Sequencing these regions will illuminate regulatory networks and thus contribute to our understanding of genetic responses
to changes in the environment.
Thus far the expense of whole-genome sequencing has slowed its progress.
However, as costs continue to fall, the availability of sequence data will
increase exponentially. It has been predicted by Francis Collins, former
director of the Human Genome Project and currently director of the US
National Institutes of Health, in his excellent book that: “I am almost certain
that complete genome sequencing will become part of newborn screening
in the next few years … . It is likely that within a few decades people will
look back on our current circumstances with a sense of disbelief that we
screened for so few conditions” (Collins, 2010). In fact, parents have already
begun paying for sequencing their children’s DNA (Rochman, 2012). If
this prediction is correct, future social and behavioral research, will see a
completely different world of data; a world where budgeting for DNA,
genotyping or sequencing will be a thing of the past. Genomic data on
nearly everyone will be widely available from centralized sources. For this
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
reason, it is crucial that social and behavioral scientists in the future are able
to capitalize on this opportunity to add genomics to their research.
REFERENCES
Balding, D. J. (2006). A tutorial on statistical methods for population association studies. Nature Reviews Genetics, 7(10), 781–791. doi:10.1038/nrg1916
Benjamin, D. J., Cesarini, D., van der Loos, M. J., Dawes, C. T., Koellinger, P. D., Magnusson, P. K., … , Visscher, P. M. (2012). The genetic architecture of economic and
political preferences. Proceedings of the National Academy of Sciences of the United
States of America, 109(21), 8026–8031. doi:10.1073/pnas.1120666109
Benyamin, B., Pourcain, B., Davis, O. S., Davies, G., Hansell, N. K., Brion, M. J.,
… , Visscher, P. M. (2014). Childhood intelligence is heritable, highly polygenic
and associated with FNBP1L. Molecular Psychiatry, 19(2), 253–258. doi:10.1038/
mp.2012.184
Bergen, S. E., Gardner, C. O., & Kendler, K. S. (2007). Age-related changes in heritability of behavioral phenotypes over adolescence and young adulthood: A
meta-analysis. Twin Research and Human Genetics, 10(3), 423–433. doi:10.1375/
twin.10.3.423
Cardon, L. R., & Bell, J. I. (2001). Association study designs for complex diseases.
Nature Reviews Genetics, 2(2), 91–99. doi:10.1038/35052543
Caspi, A., Houts, R. M., Belsky, D. W., Goldman-Mellor, S. J., Harrington, H., Israel,
S. … Moffitt, T. E. (2013). The p factor: One general psychopathology factor in
the structure of psychiatric disorders? Clinical Psychological Science, 2(2), 119–137.
doi:10.1177/2167702613497473
Chabris, C. F., Hebert, B. M., Benjamin, D. J., Beauchamp, J., Cesarini, D., van der
Loos, M. … Laibson, D. (2012). Most reported genetic associations with general
intelligence are probably false positives. Psychological Science, 23(11), 1314–1323.
doi:10.1177/0956797611435528
Collins, F. (2010). The language of life: DNA and the revolution in personalised medicine.
London, England: Profile Books.
Davis, O. S. P., Haworth, C. M. A., & Plomin, R. (2009). Learning abilities and disabilities: Generalist genes in early adolescence. Cognitive Neuropsychiatry, 14, 312–331.
doi:10.1080/13546800902797106
Deary, I. J., Yang, J., Davies, G., Harris, S. E., Tenesa, A., Liewald, D. … Corley, J.
(2012). Genetic contributions to stability and change in intelligence from childhood to old age. Nature, 482(7384), 212–215.
Deloukas, P., Kanoni, S., Willenborg, C., Farrall, M., Assimes, T. L., Thompson, J. R.
… Samani, N. J. (2013). Large-scale association analysis identifies new risk loci for
coronary artery disease. Nature Genetics, 45(1), 25–33. doi:10.1038/ng.2480
Farooqi, I. S., & O’Rahilly, S. (2006). Genetics of obesity in humans. Endocrine Reviews,
27(7), 710–718.
Fisher, R. A. (1918). The correlation between relatives on the supposition of
Mendelian inheritance. Transactions of the Royal Society of Edinburgh, 52, 399–433.
DNA Revolution and the Social and Behavioral Sciences
13
Gershon, E. S., Alliey-Rodriguez, N., & Liu, C. (2011). After GWAS: Searching for
genetic risk for schizophrenia and bipolar disorder. American Journal of Psychiatry,
168(3), 253–256. doi:10.1176/appi.ajp.2010.10091340
Gibbs, R. A., Belmont, J. W., Hardenbol, P., Willis, T. D., Yu, F., Yang, H. … Shen, Y.
(2003). The international HapMap project. Nature, 426(6968), 789–796.
H.G.P. (2001). The human genome. Nature, 409(6822), 745–964.
Hirschhorn, J. N., & Daly, M. J. (2005). Genome-wide association studies for common
diseases and complex traits. Nature Reviews Genetics, 6(2), 95–108.
Kendler, K. S., Gardner, C. O., & Lichtenstein, P. (2008). A developmental twin
study of symptoms of anxiety and depression: Evidence for genetic innovation and attenuation. Psychological Medicine, 38(11), 1567–1575. doi:10.1017/
S003329170800384X
Kovas, Y., Haworth, C. M. A., Dale, P. S., & Plomin, R. (2007). The genetic and environmental origins of learning abilities and disabilities in the early school years
7912. Monographs of the Society for Research in Child Development, 72, 1–144.
Lango Allen, H., Estrada, K., Lettre, G., Berndt, S. I., Weedon, M. N., Rivadeneira,
F. … Hirschhorn, J. N. (2010). Hundreds of variants clustered in genomic
loci and biological pathways affect human height. Nature, 467(7317), 832–838.
http://www.nature.com/nature/journal/v467/n7317/abs/nature09410.html supplementary-information
Lee, S. H., Wray, N. R., Goddard, M. E., & Visscher, P. M. (2011). Estimating missing
heritability for disease from genome-wide association studies. American Journal of
Human Genetics, 88(3), 294–305. doi:10.1016/j.ajhg.2011.02.002
Lee, S. H., Yang, J., Goddard, M. E., Visscher, P. M., & Wray, N. R. (2012). Estimation of pleiotropy between complex diseases using SNP-derived genomic
relationships and restricted maximum likelihood. Bioinformatics. doi:10.1093/
bioinformatics/bts474
Llewellyn, C. H., Trzaskowski, M., Plomin, R., & Wardle, J. (2013). Finding the missing heritability in pediatric obesity: The contribution of genome-wide complex
trait analysis. International Journal of Obesity, 37, 1506–1509.
Llewellyn, C. H., Trzaskowski, M., Plomin, R., & Wardle, J. (2014). From modelling
to measurement: Developmental trends in genetic influence on adiposity in childhood. Obesity (Silver Spring), 22, 1756–1761.
Lubke, G. H., Hottenga, J. J., Walters, R., Laurin, C., de Geus, E. J., Willemsen, G.
… Boomsma, D. I. (2012). Estimating the genetic variance of major depressive
disorder due to all single nucleotide polymorphisms. Biological Psychiatry, 72(8),
707–709. doi:10.1016/j.biopsych.2012.03.011
Marioni, R. E., Penke, L., Davies, G., Huffman, J. E., Hayward, C., & Deary, I. J. (2014).
The total burden of rare, non-synonymous exome genetic variants is not associated with childhood or late-life cognitive ability. Proceedings of the Royal Society B:
Biological Sciences, 281(1781), 20140117. doi:10.1098/rspb.2014.0117
Mercer, T. R., Dinger, M. E., & Mattick, J. S. (2009). Long non-coding RNAs: Insights
into functions. Nature Reviews Genetics, 10(3), 155–159.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Muotri, A. R., Marchetto, M. C. N., Coufal, N. G., & Gage, F. H. (2007). The necessary
junk: New functions for transposable elements. Human Molecular Genetics, 16(R2),
R159–R167.
Neale, B. M., Rivas, M. A., Voight, B. F., Altshuler, D., Devlin, B., Orho-Melander,
M. … Daly, M. J. (2011). Testing for an unusual distribution of rare variants. PLoS
Genetics, 7(3), e1001322. 10.1371/journal.pgen.1001322
Nichols, P. L. (1984). Familial mental retardation. Behavior Genetics, 14(3), 161–170.
Plomin, R., DeFries, J. C., Knopik, V. S., & Neiderhiser, J. M. (2013). Behavioral genetics
(6 ed.). New York, NY: Worth Publishers.
Plomin, R., Haworth, C. M., & Davis, O. S. (2009). Common disorders are quantitative
traits. Nature Reviews Genetics, 10, 872–878.
Plomin, R., Haworth, C. M., Meaburn, E. L., Price, T. S., & Davis, O. S. (2013). Common DNA markers can account for more than half of the genetic influence on cognitive abilities. Psychological Science, 24(4), 562–568. doi:10.1177/0956797612457952
Psychiatric Gwas Consortium Bipolar Disorder Working, & Group (2011). Largescale genome-wide association analysis of bipolar disorder identifies a new
susceptibility locus near ODZ4. Nature Genetics, 43(10), 977–983. doi:10.1038/
ng.943
Purcell, S. M., Moran, J. L., Fromer, M., Ruderfer, D., Solovieff, N., Roussos, P. …
Sklar, P. (2014). A polygenic burden of rare disruptive mutations in schizophrenia.
Nature, 506(7487), 185–190. doi:10.1038/nature12975
Rietveld, C. A., Cesarini, D., Benjamin, D. J., Koellinger, P. D., De Neve, J. E., Tiemeier,
H. … Bartels, M. (2013). Molecular genetics and subjective well-being. Proceedings
of the National Academy of Sciences of the United States of America, 110(24), 9692–9697.
doi:10.1073/pnas.1222171110
Rietveld, C. A., Medland, S. E., Derringer, J., Yang, J., Esko, T., Martin, N. W.
… Koellinger, P. D. (2013). GWAS of 126,559 individuals identifies genetic
variants associated with educational attainment. Science, 340(6139), 1467–1471.
doi:10.1126/science.1235488
Rochman, B. (2012). Want to know my future. Parents grapple with delving into their
kids’ DNA. Time, 180, 26.
Schizophrenia Psychiatric Genome-Wide Association Study, C (2011). Genome-wide
association study identifies five new schizophrenia loci. Nature Genetics, 43(10),
969–976. doi:10.1038/ng.940
Shimoni, Y., Friedlander, G., Hetzroni, G., Niv, G., Altuvia, S., Biham, O., & Margalit,
H. (2007). Regulation of gene expression by small non-coding RNAs: A quantitative view. Molecular Systems Biology, 3(1).
Siva, N. (2008). 1000 Genomes project. Nature Biotechnology, 26(3), 256–256.
Speliotes, E. K., Willer, C. J., Berndt, S. I., Monda, K. L., Thorleifsson, G., Jackson,
A. U. … Loos, R. J. F. (2010). Association analyses of 249,796 individuals reveal
18 new loci associated with body mass index. Nature Genetics, 42(11), 937–948.
http://www.nature.com/ng/journal/v42/n11/abs/ng.686.html-supplementaryinformation
DNA Revolution and the Social and Behavioral Sciences
15
Stankiewicz, P., & Lupski, J. R. (2010). Structural variation in the human genome
and its role in disease. Annual Review of Medicine, 61, 437–455. doi:10.1146/
annurev-med-100708-204735
Tabor, H. K., Risch, N. J., & Myers, R. M. (2002). Candidate-gene approaches for
studying complex genetic traits: Practical considerations. Nature Reviews Genetics,
3(5), 391–397.
Thapar, A., & McGuffin, P. (1997). Anxiety and depressive symptoms in childhood—
A genetic study of comorbidity. Journal of Child Psychology and Psychiatry and Allied
Disciplines, 38(6), 651–656.
Trzaskowski, M., Dale, P. S., & Plomin, R. (2013). No genetic influence for childhood
behavior problems from DNA analysis. Journal of the American Academy of Child
and Adolescent Psychiatry, 52(10), 1048–1056 e1043. doi:10.1016/j.jaac.2013.07.016
Trzaskowski, M., Davis, O. P., DeFries, J., Yang, J., Visscher, P., & Plomin, R. (2013).
DNA evidence for strong genome-wide pleiotropy of cognitive and learning abilities. Behavior Genetics, 1–7. doi:10.1007/s10519-013-9594-x
Trzaskowski, M., Eley, T. C., Davis, O. S. P., Doherty, S. J., Hanscombe, K. B.,
Meaburn, E. L. … Plomin, R. (2013). First genome-wide association study on
anxiety-related behaviours in childhood. PloS One, 8(4), e58676. 10.1371/journal.
pone.0058676
Trzaskowski, M., Harlaar, N., Arden, R., Krapohl, E., Rimfeld, K., McMillan, A. …
Plomin, R. (2014). Genetic influence on family socioeconomic status and children’s
intelligence. Intelligence, 42, 83–88. doi:10.1016/j.intell.2013.11.002
Trzaskowski, M., Shakeshaft, N. G., & Plomin, R. (2013). Intelligence indexes
generalist genes for cognitive abilities. Intelligence, 41(5), 560–565. doi:10.1016/
j.intell.2013.07.011
Trzaskowski, M., Yang, J., Visscher, P. M., & Plomin, R. (2013). DNA evidence for strong
genetic stability and increasing heritability of intelligence from age 7 to 12 (Vol. 19,
pp. 380–384). Molecular Psychiatry.
Verweij, K. J., Vinkhuyzen, A. A., Benyamin, B., Lynskey, M. T., Quaye, L., Agrawal,
A. … Medland, S. E. (2012). The genetic aetiology of cannabis use initiation: A
meta-analysis of genome-wide association studies and a SNP-based heritability
estimation. Addiction Biology. 10.1111/j.1369-1600.2012.00478.x
Vinkhuyzen, A. A., Pedersen, N. L., Yang, J., Lee, S. H., Magnusson, P. K., Iacono,
W. G. … Wray, N. R. (2012). Common SNPs explain some of the variation in the
personality dimensions of neuroticism and extraversion. Translational Psychiatry,
2, e102. 10.1038/tp.2012.27
Visscher, P. M. (2008). Sizing up human height variation. Nature Genetics, 40(5),
489–490. doi:10.1038/ng0508-489
Visscher, P. M., Brown, M. A., McCarthy, M. I., & Yang, J. (2012). Five years of
GWAS discovery. The American Journal of Human Genetics, 90(1), 7–24. doi:10.1016/
j.ajhg.2011.11.029
Wray, N. R., Lee, S. H., Mehta, D., Vinkhuyzen, A. A. E., Dudbridge, F., & Middeldorp, C. M. (2014). Research review: Polygenic methods and their application
to psychiatric traits. Journal of Child Psychology and Psychiatry, 55(10), 1068–1087.
doi:10.1111/jcpp.12295
16
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
WTCCC (2007). Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature, 447(7145), 661–678. doi:10.1038/
nature05911
Yang, J., Benyamin, B., McEvoy, B. P., Gordon, S. D., Henders, A. K., Nyholt, D. R.
… Visscher, P. M. (2010). Common SNPs explain a large proportion of the heritability for human height. Nature Genetics, 42(7), 565–569. http://www.nature.
com/ng/journal/v42/n7/suppinfo/ng.608_S1.html
Yang, J., Lee, S., Goddard, M., & Visscher, P. (2011). GCTA: A tool for genome-wide
complex trait analysis. American Journal of Human Genetics, 88(1), 76–82.
doi:10.1016/j.ajhg.2010.11.011
MACIEJ TRZASKOWSKI SHORT BIOGRAPHY
While studying for his PhD, Dr Maciej Trzaskowski has learnt and applied
an extensive selection of statistical methods for quantitative (structural
equation modeling) and molecular genetics (e.g., GWAS, and GREML),
using various programming languages (e.g., R, Python, Unix, and Shell
Scripts). Dr. Trzaskowski has a strong methodological propensity and high
interest in exploring new methods. He has been awarded Gottesman-Shields
Prize for the best PhD thesis 2013 and since the completion of his PhD, Dr.
Trzaskowski remained at the Social, Genetic and Developmental Psychiatry
Centre as a postdoctoral research worker. To date, he has published 23
papers, 11 of these as first/joint first author. He has built his own network
of collaborations with researchers in the Broad Institute, the Queensland
Institute of Medical Research, Queensland Brain Institute (the University of
Queensland), Vrije Universiteit (VU) Amsterdam, Harvard, amongst others.
His research is increasingly recognized (as evidenced by the Behavior
Genetics Association Thompson and Fulker Awards received in 2013 and
2014, respectively) and, most recently, by a successful award of a fellowship
from British Academy. He has given invited talks at the London School
of Economics (by Lord Richard Layard, Emeritus Professor of Economics)
and at the fourth annual meeting by the Social Science Genetic Association
Consortium (by Professor Phillip Koellinger).
ROBERT PLOMIN SHORT BIOGRAPHY
Since 1994, Professor Robert Plomin has been MRC Research Professor of
Behavioral Genetics at the Institute of Psychiatry, King’s College London.
In 1994, he cofounded and subsequently directed the MRC Social, Genetic
and Developmental Psychiatry Centre, whose goal is to bring together
genetic and environmental strategies to study behavioral development.
In 1995, he launched the Twins Early Development Study (TEDs) of all
DNA Revolution and the Social and Behavioral Sciences
17
twins born in England and Wales in 1994–1996, which focuses on developmental problems in cognition and behavior, and which Professor Plomin
continues to direct. He has published more than 500 papers and more
than a dozen books, including the major textbook in the field (“Behavioral
Genetics,” Worth Publishers, 6th edition, 2013). His most recent book,
coauthored with Kathryn Asbury, focuses on genetics and education (“G
is for Genes: the Impact of Genetics on Education and Achievement,” Wiley
Blackwell, 2013). He has received lifetime research achievement awards
from the three major international associations in his field. For details, see
https://kclpure.kcl.ac.uk/portal/robert.plomin.html.
RELATED ESSAYS
Telomeres (Psychology), Nancy Adler and Aoife O’Donovan
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
Biology and Culture (Psychology), Robert Peter Hobson
Genetics and the Life Course (Sociology), Evan Charney
Genetic and Environmental Approaches to Political Science (Political Science),
Zoltán Fazekas and Peter K. Hatemi
Genetics and Social Behavior (Anthropology), Henry Harpending and Gregory Cochran
A Gene-Environment Approach to Understanding Youth Antisocial Behavior (Psychology), Rebecca Waller et al.
Behavioral Heterochrony (Anthropology), Victoria Wobber and Brian Hare
DNA Revolution and the Social
and Behavioral Sciences
MACIEJ TRZASKOWSKI and ROBERT PLOMIN
Abstract
A century of genetic research on the social and behavioral sciences has addressed
the “how much” question, showing that genetic differences are important for nearly
all traits. However, during the past few decades, research has moved beyond this
rudimentary “how much” question to ask “how” questions about developmental
change and continuity, about the relationship between traits, and about the interplay between genes (nature) and environment (nurture). We suggest that some of
the most important and transformative findings in the social and behavioral sciences
have come from this research. Moreover, the most revolutionary changes in genetic
research are on their way with the rapid advances in DNA technology and methodology, which promise to transform the social and behavioral sciences. It is crucial that
social and behavioral scientists stay on top of the DNA revolution. The purpose of
our essay is to provide an overview of genetic research in the social and behavioral
sciences.
NATURE AND NURTURE: TWIN AND ADOPTION STUDIES
In general, genetic research can be divided into quantitative genetics
(statistical designs in which family members are used to estimate genetic
influence) and molecular genetics (research designs where genetic influence
is estimated by directly measured genotypes). For decades quantitative
geneticists used family, twin and adoption designs to estimate genetic and
environmental influences on many psychological and psychiatric traits. The
strength of these designs is that they can estimate the bottom line of genetic
influence regardless of how many genes are involved or how complex their
effects, unlike molecular genetics. All quantitative genetic designs have
limitations, but each design has different limitations, and, most importantly,
their results converge. For example, the adoption design, which typically
compares adopted children with their biological and adoptive parents, is
limited by possible prenatal factors and selective placement. In contrast, the
twin design, which compares monozygotic (MZ) twins who are genetically
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
100% identical, and dizygotic (DZ) twins who share on average 50% of their
segregating alleles, is limited by the assumption of equal environments
for the two types of twins. These limitations are specific to each design
and yet adoption and twin studies generally produce similar estimates of
genetic and environmental influences, especially for cognitive traits (Plomin,
DeFries, Knopik, & Neiderhiser, 2013). Adoption studies are less common
today than twin studies because of the sharp reduction in neonatal adoption,
but twin studies continue to proliferate.
Univariate twin studies have demonstrated consistently that virtually
every human trait is influenced in part by genetic variation. Multivariate
models go beyond this rudimentary finding about the relative influence
of nature and nurture on a single phenotype, to investigate the common
genetic architecture underlying multiple complex traits. Multivariate twin
designs decompose the covariance between multiple phenotypes to provide
insights into the common genetic and environmental relationship underlying different traits, as well as the genetic and environmental contributions
to their development over time (trait stability and change). Multivariate
genetic research has yielded three major findings consistent across many
complex traits: heritability increases across development, genetic effects
are developmentally stable—strong age-to-age genetic correlations—and
are highly pleiotropic, that is, one gene influences many traits. Genetic
pleiotropy is inferred from consistent findings of strong genetic covariance
between different traits at one measurement occasion: for example, this
is true for anxiety and depression (e.g., Thapar & McGuffin, 1997), and
for IQ and learning abilities (Davis, Haworth, & Plomin, 2009). Genetic
stability is inferred from high genetic correlations between the “same”
traits across time. Curiously, despite the observed genetic stability from
age to age, many traits appear to show increases in heritability across
development (Bergen, Gardner, & Kendler, 2007). Although genetic stability
and increasing heritability might appear paradoxical, there are several
mechanisms that might explain this. The phenomenon could be driven by a
correlation between genes and environments where, for example, children
choose environments that suit their genetic propensities (Plomin, DeFries,
et al., 2013). Imagine a young girl who has a propensity for reading; as she
grows up, she chooses environments in which she can be exposed to reading
as much as possible. This “tailored” selection will increase the importance
of these in explaining individual differences in reading and yet they will be
the same genes. It is also possible that heritability increases as new genes
come “online” over the lifetime explaining more of the variance (Kendler,
Gardner, & Lichtenstein, 2008). Unfortunately, twin designs cannot easily
tease out which of these mechanisms are responsible for the phenomena.
The most important advance in this field that will help to resolve this issue
DNA Revolution and the Social and Behavioral Sciences
3
and many others will come when specific genetic variants are identified that
account for these genetic effects.
LOW-HANGING FRUIT: MONOGENIC DISORDERS
Mendel, in his original work with peas, was fortunate in that he chose monogenic traits in which a single mutation is necessary and sufficient for the
trait to be expressed. For such single-gene traits the underlying genetic variants travel across generations according to predictions based on Mendel’s
first law of segregation, and they also assort independently from variants
of other traits according to Mendel’s second law. Unfortunately, complex
traits appeared not to conform to these laws, prompting some researchers
in the early twentieth century to argue that Mendel’s findings were limited
to peas, and not “higher-order” species. Although the dispute between the
two factions was at times fierce, it was resolved when it became apparent
that in complex traits Mendel’s laws hold for segregation of alleles at a single
locus but are masked by the additive influence of many such loci. The overall genetic influence is driven by many genes, called polygenic, and thus the
effects of individual genes is small and their overall effect on the phenotype
is complex and normally distributed in a population. This is the cornerstone
of quantitative genetics expounded by Fisher (1918), who described how
Mendel’s model could be extended to multiple genes in order to account for
inheritance of complex traits.
In molecular genetics, Mendel’s law of segregation culminated in huge
success in discovering genetic associations with monogenic disorders
mainly through implementation of linkage designs using just a few hundred
DNA markers across the genome. Linkage analysis uses samples drawn
from multigenerational families with affected and unaffected members,
where DNA of all family members is “tagged” with several markers whose
location on chromosomes is known. The aim is to detect a “link” between
one of these markers and a trait or disorder. Linkage successfully detected
genetic associations with many Mendelian (monogenic) traits because a
mutation in a single locus explained all the variance. However, monogenic
disorders affect a tiny proportion of clinical populations; most of the burden
of illness involves complex disorders influenced by many genes with
small effect sizes. Although linkage has been very successful in identifying
genes for single-gene disorders, the method does not have sufficient power
to detect genes of small effect size. A technological advance that greatly
improved attempts to find genes responsible for heritability of complex traits
is genome-wide association (GWA), which is based on allelic association
between DNA markers and a trait in unrelated individuals in the population
rather than linkage within families.
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
FINDING MANY NEEDLES IN THE HAYSTACK: GENOME-WIDE
ASSOCIATION STUDIES
The discovery that many complex traits are highly heritable motivated
researchers to search for the genetic variants associated with these traits.
The pre-GWA era relied mainly on candidate gene association studies,
where a specific region of genome was hypothesized to be important
for development of a particular trait. Unfortunately, a problem with the
candidate gene approach is that the hypotheses generated about the regions
we think should be important are often wrong. Many published candidate
gene association results is now known to be false positive results (Tabor,
Risch, & Myers, 2002). This is particularly telling when replication fails in
GWA analysis as this design is hypothesis free in the sense that it does not
“care” which region is associated with which trait. An example of such
failure is a GWA study of nearly 10,000 individuals where none of the 10
most frequently reported candidate regions for general cognitive ability
replicated (Chabris et al., 2012).
The genomic era began in the past decade when two major developments
provided cost-effective, thorough coverage of the genome. The first event
was the completion of the Human Genome Project (H.G.P., 2001), which
resulted in the first detailed maps of the human genome and the patterns of
linkage disequilibrium (the nonrandom association of alleles across neighboring genetic loci) for hundreds of thousands SNPs (single-nucleotide
polymorphism—variation in a single genomic base-pair). The second event
was the production of DNA arrays or “chips.” A single DNA chip could
genotypes hundreds of thousands of SNPs. For each SNP, the DNA chip
includes many probes, which is a short DNA sequence containing the
SNP. Understanding linkage disequilibrium patterns across the genome
was crucial because it showed that careful selection of a few hundred
thousand DNA markers was sufficient to comprehensively tag the genome’s
3 billion DNA base pairs for GWA studies. DNA variants close together
on a chromosome violate Mendel’s second law by segregating together
producing high between-SNP correlations. A single marker from DNA
variants in high linkage disequilibrium is sufficient to “tag” the region. This
knowledge was used to select SNPs for DNA chips. In addition, coverage
of the genome could be improved by imputing millions of additional SNPs
from existing reference maps, such as HapMap (Gibbs et al., 2003) or, more
recently, the 1000 Genomes Project (Siva, 2008) without actually genotyping
these additional SNPs. This meant that it was now possible to interrogate
the whole genome simultaneously for association between allele frequencies
of individual SNPs and any trait varying in unrelated individuals. This
DNA Revolution and the Social and Behavioral Sciences
5
method is known today as a GWA (e.g., Balding, 2006; Cardon & Bell, 2001;
Hirschhorn & Daly, 2005).
In GWA, associations are usually examined one SNP at the time, assuming
an additive model, although other models can also be implemented. Since
current genetic chips can assess millions of SNPs, the correction for multiple
testing is daunting, and the number of possible false positives is large; the
accepted p-value threshold for genome-wide significance is p < 5 × 10−8 ,
which can be thought of as p < 0.05 with a Bonferroni correction for one
million statistical tests. Despite such stringent correction, within the three
years between 2005 and 2008, more than 400 SNP associations for just
over 120 traits (www.genome.gov/GWAstudies) achieved this stringent
genome-wide threshold, and revealed unexpected insights about genetic
influences on complex traits (Visscher, Brown, McCarthy, & Yang, 2012).
Arguably the most useful biological insights in genetic associations have
been mainly limited to medical traits. For example, in Crohn’s disease,
many SNPs reported through GWA studies were found in and around genes
involved with autophagy, the cell’s maintenance process that breaks down
dysfunctional components of the cell (WTCCC, 2007). In addition, the same
study showed that type 2 diabetes was associated with loci encoding for
proteins relevant to insulin secretion, and not insulin signaling, as previously thought. Nevertheless, even if functionally not so obvious, many new
loci have also been identified in anthropometric (e.g., Speliotes et al., 2010;
Visscher, 2008) and some psychiatric traits (e.g., Gershon, Alliey-Rodriguez,
& Liu, 2011; Purcell et al., 2014). Unfortunately successes in cognitive, social,
and behavioral complex traits are yet to come.
Despite GWA successes, fewer associations were reported for complex
traits than expected, and associations that were identified could together
only account for a small fraction of twin-estimated heritability. The phenomenon is known as “missing heritability,” which is discussed in the
next section. The important lesson from GWA research is that the largest
effect sizes of individual loci of common SNPs are incredibly small and
require sample sizes in the hundreds of thousands to achieve genome-wide
significance. This realization led to the sudden emergence and proliferation
of world-wide consortia, which paid off by doubling the GWA hits to just
under 9000 SNPs in more than 700 traits (www.genome.gov/GWAstudies/).
Nonetheless, the yield was still not as substantial as expected in that the
largest effect sizes, such as the association between the FTO gene and body
mass index (BMI), are less than 1% of the variance (Speliotes et al., 2010),
which implies that the smallest effect sizes are likely to be infinitesimal.
This means that it will be difficult to detect and replicate associations with
complex traits in the social and behavioral sciences. However, once several
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
genes are found that are associated with these traits, their effects can be
aggregated as polygenic scores, as described in the following section.
TOGETHER WE ARE STRONGER: POLYGENIC SCORES
Even though the effects of individual SNPs are very small, their effects can
be aggregated to increase the total amount of variance explained and thus
increase power. For example, if we had 10 individual SNPs, each with an
effect size of 0.5%, a polygenic score comprising all of these would account
for 5% of the variance, which means that a sample of 150 individuals would
have 80% power to detect their cumulative effect. For this reason, polygenic
scores are now being used in a new generation of hypothesis-free candidate
gene association studies and are the way that the DNA revolution will come
to the social and behavioral sciences.
The creation of a polygenic score takes into account the direction of the
effect (i.e., which allele is the “increasing” allele) and it can also weight the
associations by the magnitude of their effect. To give priority to stronger associations, the scores can be weighted by the betas from the regression (i.e.,
their effect size). The associations can then be summed, similar to summing
items on a scale. Although summing scores is most commonly used, nonadditivity can also be incorporated. In addition, SNPs can either be selected
from previously reported significant “hits” (polygenic scores—PGS) or they
can be amassed by selecting all associations below an arbitrary p-value from
GWA studies, a genome-wide PGS (Wray et al., 2014). The former method is
preferred because of the trade-off between specificity and size but the choice
depends on the availability of previously reported robust “hits.” The tiny
effect sizes and moderate-to-large estimates of heritability suggest that the
more SNPs you “pool” together the more variance you should explain. However, adding a large number of SNPs with no effect or effects in the opposite
direction can attenuate the signal.
Application of polygenic predictors to medical disorders had some success
due at least in part to availability of financial resources and large sample
sizes. It has been shown that PGS of 150 SNPs accounted for 5% of the
variance in the liability for coronary artery disease (Deloukas et al., 2013).
For bipolar disorder, SNPs accounted for between 1% and 3% of the liability
(Psychiatric Gwas Consortium Bipolar Disorder Working & Group, 2011)
and a PGS of significant SNPs from discovery GWAS in schizophrenia,
accounted for approximately 1% of the liability in the independent sample
(Schizophrenia Psychiatric Genome-Wide Association Study, 2011). In the
schizophrenia study, extending the PGS to genome-wide PGS increased
variance explained to almost 6%. PGS research on quantitative traits has
largely been limited to weight and height, and other complex continuous
DNA Revolution and the Social and Behavioral Sciences
7
traits have had to resort to genome-wide PGS. GWA meta-analysis of BMI
revealed that 32 replicated SNPs accounted for only ∼1.5% of the variance
in BMI in independent samples, but a genome-wide PGS that included
thousands of SNPs explained ∼5% of the variance (Speliotes et al., 2010).
For height, 180 SNPs accounted for 10% of the variance, and genome-wide
PGS increased the variance explained to 13% (Lango Allen et al., 2010).
Importantly, investigation of behavioral traits of direct interest to social
and cognitive scientists are now achieving sample sizes that enable more
powerful genomic interrogation. Application of genome-wide PGS to these
traits is producing results that are similar to medical and anthropometric
traits (height and weight). Specifically, adding more SNPs increases amount
of variance explained up to a point. For example, an increase of variance
explained with an addition of more SNPs was reported for total years of
education (1% using 3500 SNPs and 2.5% using 2.5 million SNPs; Rietveld,
Medland, et al., 2013) and childhood IQ (using polygenic predictor at
p < 6 × 10−5 explained 0.5–1.2% increasing to 3.5% at p < 0.001; Benyamin
et al., 2014). These findings are important despite the small proportions of
variance explained, because they suggest that complex traits are indeed
highly polygenic.
It is now obvious that even though inclusion of a large number of SNPs
consistently increases the amount of variance explained, the gap between
twin-estimated heritability and these PGS estimates is still very wide. The
phenomenon known today as the “missing heritability” continues to pose
many questions, which are as yet unanswered. Twin studies suggest that
almost every human trait from our biology through cognition, behavior and
even the environment is heritable, with genetic influences usually explaining
moderate (30–40%) to high (80–90%) proportions of variance. GWA studies
have accounted for only a small fraction of this heritability. The most general
explanation is that the influence of each individual SNP is so small that most
of the GWA studies thus far have been greatly underpowered to detect them.
Another likely source of missing heritability is rarer variants—the markers
selected for the DNA arrays were limited to common variants only (minor
allele frequency>1%). This meant that many potential “true” associations
with SNPs of lower allele frequency could be missed owing to low linkage disequilibrium with the markers. Another possible source of missing
heritability is nonadditive effects, such as gene–environment or gene–gene
interactions because GWA is limited to additive genetic effects. Finally, it is
possible that heritability estimates derived from twin data are inflated. To
answer some of these questions, a new quantitative genetic technique has
emerged, called genome-wide complex trait analysis (GCTA), as described
in the following section.
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
HERITABILITY WITHOUT TWINS: GENOME-WIDE COMPLEX
TRAIT ANALYSIS (GCTA)
A recently developed method uses genome-wide genotyping data in a novel
way to address these questions. The linear mixed model implemented in
the GCTA package estimates the amount of phenotypic variance (Yang, Lee,
Goddard, & Visscher, 2011) and co-variance (Lee, Yang, Goddard, Visscher,
& Wray, 2012) that can be explained by additive effects of all common SNPs
tagged by DNA arrays. Several other methods have been developed with
the same intent (Wray et al., 2014) but GCTA is currently most widely used.
GCTA itself has been given different names, such as linear mixed model
and genomic-relationship-matrix restricted maximum likelihood. However,
clever usage of the four DNA base letters makes GCTA the catchiest acronym,
and this is what we call it henceforth.
GCTA uses genome-wide genotype data from unrelated individuals to estimate genetic influence on a trait that can be explained by all SNPs included
on DNA chips. Because GCTA is based on genome-wide DNA data alone,
it can be used to estimate genetic influence for unrelated individuals rather
than requiring special relatives such as MZ and DZ twins. For that reason, it
bypasses some of the assumptions of the twin method, although it has its own
set of assumptions (Plomin, Haworth, Meaburn, Price, & Davis, 2013). GCTA
can only detect genetic influence owing to the additive effects of common
SNPs that are included on currently available DNA arrays and cannot evaluate the contribution of any specific DNA locus. Nonetheless, GCTA provides
important information about the extent to which the genetic architecture of
complex traits includes additive effects of common SNPs, and sets the limit
for detecting associations in GWA studies.
GCTA analyses have shown that information captured by current DNA
arrays can explain a substantial amount of the variance in complex traits,
including human height (Yang et al., 2010), BMI (Llewellyn, Trzaskowski,
Plomin, & Wardle, 2013), psychiatric and medical disorders (Lee, Wray,
Goddard, & Visscher, 2011; Lee et al., 2012; Lubke et al., 2012), cognitive
traits (Deary et al., 2012; Plomin, Haworth, et al., 2013), and economic and
political preferences (Benjamin et al., 2012). All of these GTCA heritability
estimates are approximately half of twin-estimated heritability. In contrast,
initial research in psychopathology and personality is less consistent,
showing near zero variance explained for most analyses. For example,
analyses of a wide range of behavioral problems (symptoms of anxiety
and depression, hyperactivity, conduct problems) show negligible SNP
heritability despite moderate to high twin heritability estimated within the
same sample (Trzaskowski, Dale, & Plomin, 2013; Trzaskowski, Eley, et al.,
2013). SNP heritabilities of neuroticism and extraversion were reported as
DNA Revolution and the Social and Behavioral Sciences
9
0.06 (SE 0.03) and 0.12 (SE 0.03) respectively (Vinkhuyzen et al., 2012), and
similar near-zero GCTA estimates have also been shown for other aspects
of personality (Verweij et al., 2012), including wellbeing (Rietveld, Cesarini,
et al., 2013). The results are puzzling as all of the studies are adequately
powered to detect the expected effect sizes (∼40% of twin heritability)
and are not limited to self-report questionnaire. For example, one study
examined behavior problems reported by parents and teachers as well as
the children themselves; further discussion on this topic can also be found
in the same publication (Trzaskowski, Dale, et al., 2013).
As was the case for early twin studies, GCTA was initially applied in univariate analyses of one trait at a time, but like twin studies, the model was
extended to the bivariate analysis of covariance between traits or across age.
For example, bivariate GCTA was first applied to the remarkable phenotypic
stability of IQ across 60 years from childhood to later life (phenotypic correlation = 0.63) and suggested that the stability is largely due to genetic stability
(genetic correlation = 0.62) (Deary et al., 2012). Twin studies also suggested
that the genetic stability was present despite heritability of IQ increasing
across development. Bivariate GCTA supported this finding of genetic stability despite increasing heritability for IQ (Trzaskowski, Yang, Visscher, &
Plomin, 2013) and similar twin and GCTA results were reported for BMI
(Llewellyn, Trzaskowski, Plomin, & Wardle, 2014). The most likely explanation for these phenomena is gene–environment correlation (Plomin, DeFries,
et al., 2013). The high genetic correlation suggests that the same genes influence individual differences in IQ across time. However, as we grow older we
increasingly select environments that “match” our genetic predispositions
making genetic influence stronger even though the same genes are involved.
Although twin studies have reported for decades that most environments are nearly as heritable as behaviors, this work has been limited to
twin-specific environments. GCTA opens up the possibility of investigating
genetic influence on family-, neighborhood-, or even country-wide environmental measures that cannot be studied using the twin design because
they are shared in common by members of a twin pair (Trzaskowski et al.,
2014). This feature of GCTA should be particularly interesting to social
and behavioral scientists as it emphasizes the important interplay between
genes and environments. It shows that environments are not simply “out
there” that happen randomly to us, but that our genes shape our experiences
through our selection, modification and creation of our environments.
Another widely reported finding from twin analyses is the strong genetic
correlation (pleiotropy) across different aspects of cognition and across
diverse cognitive abilities (Davis et al., 2009; Kovas, Haworth, Dale, &
Plomin, 2007; Plomin, DeFries, et al., 2013). GCTA studies reported point
estimates for genetic correlations highly similar to those reported by twin
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
studies (Trzaskowski, Davis, et al., 2013). GCTA estimates of genetic correlation are similar to twin study estimates of genetic correlation, not only within
traditional domains such as verbal and nonverbal (Trzaskowski, Shakeshaft,
& Plomin, 2013) but also between intelligence and education-related skills
such as reading and mathematics (Trzaskowski, Davis, et al., 2013). We
can expect many more multivariate GCTA analyses in domains other than
cognition, most notably psychopathology where it increasingly appears that
there is a great deal of overlap among diverse symptoms (Caspi et al., 2013).
In summary, GCTA has shed some light on the nature of the discrepancy
between heritability estimates from twin studies and combined variance
explained from SNPs identified through GWA. Some of the “missing
heritability” is hiding in tiny influences of many common SNPs. We called
that gap between GWA and GCTA “missing GWA heritability,” because
GCTA represents the ceiling for GWA results. This gap can in theory be
filled with the additive effects of variants tagged by the SNPs on current
DNA chips if samples are large enough to detect nearly infinitesimal effect
sizes. GCTA results also suggest that even if the “missing GWA heritability”
were filled, there would still be a substantial chunk of genetic influence
not accounted for by additive effects of common SNPs. This part of the
“missing heritability” could be called “missing GCTA heritability,” which
falls short of twin study heritability because GCTA only reflect additive
effects of common SNPs. We delineated these two parts to emphasize a
distinction between missing heritability that is caused by lack of power (the
effect sizes of already captured common variants are too small to detect
with current sample sizes), and missing heritability that is due to genetic
variants that have not yet been captured (e.g., rare variants or nonadditive
genetic influences). The gap of missing GWA heritability can be narrowed
by the brute force of larger samples, but how can DNA variants responsible
for missing GCTA heritability be identified?
3 BILLION BASE PAIRS: WHOLE-GENOME SEQUENCING
Whole-genome sequencing involves genotyping all 3 billion base pairs of
DNA, rather than just a million or so SNPs as on current genetic chips.
Whole-genome sequencing is the end of the story of genetic variation in
the sense that all we inherit from our parents is differences in DNA base
pair sequences. Since current genetic chips only capture common SNP,
whole-genome sequencing could identify more genetic loci associated with
complex traits because it captures variants of any kind, not just common
SNPs. Rare variants may play a particularly important role in the extreme
tails of a trait’s distribution. For example, common SNPs influence BMI
across the distribution of the normal population, but very rare Mendelian
DNA Revolution and the Social and Behavioral Sciences
11
mutations may account for as much as 5% of extreme obesity (Farooqi &
O’Rahilly, 2006).
Other clinical phenotypes, such as intellectual disability, schizophrenia,
bipolar disorder, and autism, have been associated with an increased burden
of rare variants, including idiosyncratic mutations that first occur in that
individual and are not inherited from the individual’s parents, called de
novo mutations (Marioni et al., 2014; Neale et al., 2011; Purcell et al., 2014;
Stankiewicz & Lupski, 2010). Perhaps rare variants are responsible for a
puzzle in findings concerning mild and severe intellectual disability. One
study found that mild intellectual impairment was familial, but severe
impairment was not (Nichols, 1984). That is, siblings of severely mentally
retarded children showed no mental impairment, whereas siblings of children with “mild” mental disability did, suggesting that severe impairment
is not genetically related to common variation in mental ability. In general,
common (less severe) disorders are likely to be the quantitative extremes of
normal variation (Plomin, Haworth, & Davis, 2009), whereas extreme levels
of disability could be mainly a result of accumulation of much rarer variants,
including de novo mutations as well as environmental “mutations” such as
perinatal trauma.
Sequencing will also give us richer information about noncoding regions.
Not so long ago noncoding regions of the genome were thought of as an
evolutionary “junk,” but now these noncoding regions are known to play an
important role in regulation of genetic expression and even creation of new
genes (Mercer, Dinger, & Mattick, 2009; Muotri, Marchetto, Coufal, & Gage,
2007; Shimoni et al., 2007). Sequencing these regions will illuminate regulatory networks and thus contribute to our understanding of genetic responses
to changes in the environment.
Thus far the expense of whole-genome sequencing has slowed its progress.
However, as costs continue to fall, the availability of sequence data will
increase exponentially. It has been predicted by Francis Collins, former
director of the Human Genome Project and currently director of the US
National Institutes of Health, in his excellent book that: “I am almost certain
that complete genome sequencing will become part of newborn screening
in the next few years … . It is likely that within a few decades people will
look back on our current circumstances with a sense of disbelief that we
screened for so few conditions” (Collins, 2010). In fact, parents have already
begun paying for sequencing their children’s DNA (Rochman, 2012). If
this prediction is correct, future social and behavioral research, will see a
completely different world of data; a world where budgeting for DNA,
genotyping or sequencing will be a thing of the past. Genomic data on
nearly everyone will be widely available from centralized sources. For this
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
reason, it is crucial that social and behavioral scientists in the future are able
to capitalize on this opportunity to add genomics to their research.
REFERENCES
Balding, D. J. (2006). A tutorial on statistical methods for population association studies. Nature Reviews Genetics, 7(10), 781–791. doi:10.1038/nrg1916
Benjamin, D. J., Cesarini, D., van der Loos, M. J., Dawes, C. T., Koellinger, P. D., Magnusson, P. K., … , Visscher, P. M. (2012). The genetic architecture of economic and
political preferences. Proceedings of the National Academy of Sciences of the United
States of America, 109(21), 8026–8031. doi:10.1073/pnas.1120666109
Benyamin, B., Pourcain, B., Davis, O. S., Davies, G., Hansell, N. K., Brion, M. J.,
… , Visscher, P. M. (2014). Childhood intelligence is heritable, highly polygenic
and associated with FNBP1L. Molecular Psychiatry, 19(2), 253–258. doi:10.1038/
mp.2012.184
Bergen, S. E., Gardner, C. O., & Kendler, K. S. (2007). Age-related changes in heritability of behavioral phenotypes over adolescence and young adulthood: A
meta-analysis. Twin Research and Human Genetics, 10(3), 423–433. doi:10.1375/
twin.10.3.423
Cardon, L. R., & Bell, J. I. (2001). Association study designs for complex diseases.
Nature Reviews Genetics, 2(2), 91–99. doi:10.1038/35052543
Caspi, A., Houts, R. M., Belsky, D. W., Goldman-Mellor, S. J., Harrington, H., Israel,
S. … Moffitt, T. E. (2013). The p factor: One general psychopathology factor in
the structure of psychiatric disorders? Clinical Psychological Science, 2(2), 119–137.
doi:10.1177/2167702613497473
Chabris, C. F., Hebert, B. M., Benjamin, D. J., Beauchamp, J., Cesarini, D., van der
Loos, M. … Laibson, D. (2012). Most reported genetic associations with general
intelligence are probably false positives. Psychological Science, 23(11), 1314–1323.
doi:10.1177/0956797611435528
Collins, F. (2010). The language of life: DNA and the revolution in personalised medicine.
London, England: Profile Books.
Davis, O. S. P., Haworth, C. M. A., & Plomin, R. (2009). Learning abilities and disabilities: Generalist genes in early adolescence. Cognitive Neuropsychiatry, 14, 312–331.
doi:10.1080/13546800902797106
Deary, I. J., Yang, J., Davies, G., Harris, S. E., Tenesa, A., Liewald, D. … Corley, J.
(2012). Genetic contributions to stability and change in intelligence from childhood to old age. Nature, 482(7384), 212–215.
Deloukas, P., Kanoni, S., Willenborg, C., Farrall, M., Assimes, T. L., Thompson, J. R.
… Samani, N. J. (2013). Large-scale association analysis identifies new risk loci for
coronary artery disease. Nature Genetics, 45(1), 25–33. doi:10.1038/ng.2480
Farooqi, I. S., & O’Rahilly, S. (2006). Genetics of obesity in humans. Endocrine Reviews,
27(7), 710–718.
Fisher, R. A. (1918). The correlation between relatives on the supposition of
Mendelian inheritance. Transactions of the Royal Society of Edinburgh, 52, 399–433.
DNA Revolution and the Social and Behavioral Sciences
13
Gershon, E. S., Alliey-Rodriguez, N., & Liu, C. (2011). After GWAS: Searching for
genetic risk for schizophrenia and bipolar disorder. American Journal of Psychiatry,
168(3), 253–256. doi:10.1176/appi.ajp.2010.10091340
Gibbs, R. A., Belmont, J. W., Hardenbol, P., Willis, T. D., Yu, F., Yang, H. … Shen, Y.
(2003). The international HapMap project. Nature, 426(6968), 789–796.
H.G.P. (2001). The human genome. Nature, 409(6822), 745–964.
Hirschhorn, J. N., & Daly, M. J. (2005). Genome-wide association studies for common
diseases and complex traits. Nature Reviews Genetics, 6(2), 95–108.
Kendler, K. S., Gardner, C. O., & Lichtenstein, P. (2008). A developmental twin
study of symptoms of anxiety and depression: Evidence for genetic innovation and attenuation. Psychological Medicine, 38(11), 1567–1575. doi:10.1017/
S003329170800384X
Kovas, Y., Haworth, C. M. A., Dale, P. S., & Plomin, R. (2007). The genetic and environmental origins of learning abilities and disabilities in the early school years
7912. Monographs of the Society for Research in Child Development, 72, 1–144.
Lango Allen, H., Estrada, K., Lettre, G., Berndt, S. I., Weedon, M. N., Rivadeneira,
F. … Hirschhorn, J. N. (2010). Hundreds of variants clustered in genomic
loci and biological pathways affect human height. Nature, 467(7317), 832–838.
http://www.nature.com/nature/journal/v467/n7317/abs/nature09410.html supplementary-information
Lee, S. H., Wray, N. R., Goddard, M. E., & Visscher, P. M. (2011). Estimating missing
heritability for disease from genome-wide association studies. American Journal of
Human Genetics, 88(3), 294–305. doi:10.1016/j.ajhg.2011.02.002
Lee, S. H., Yang, J., Goddard, M. E., Visscher, P. M., & Wray, N. R. (2012). Estimation of pleiotropy between complex diseases using SNP-derived genomic
relationships and restricted maximum likelihood. Bioinformatics. doi:10.1093/
bioinformatics/bts474
Llewellyn, C. H., Trzaskowski, M., Plomin, R., & Wardle, J. (2013). Finding the missing heritability in pediatric obesity: The contribution of genome-wide complex
trait analysis. International Journal of Obesity, 37, 1506–1509.
Llewellyn, C. H., Trzaskowski, M., Plomin, R., & Wardle, J. (2014). From modelling
to measurement: Developmental trends in genetic influence on adiposity in childhood. Obesity (Silver Spring), 22, 1756–1761.
Lubke, G. H., Hottenga, J. J., Walters, R., Laurin, C., de Geus, E. J., Willemsen, G.
… Boomsma, D. I. (2012). Estimating the genetic variance of major depressive
disorder due to all single nucleotide polymorphisms. Biological Psychiatry, 72(8),
707–709. doi:10.1016/j.biopsych.2012.03.011
Marioni, R. E., Penke, L., Davies, G., Huffman, J. E., Hayward, C., & Deary, I. J. (2014).
The total burden of rare, non-synonymous exome genetic variants is not associated with childhood or late-life cognitive ability. Proceedings of the Royal Society B:
Biological Sciences, 281(1781), 20140117. doi:10.1098/rspb.2014.0117
Mercer, T. R., Dinger, M. E., & Mattick, J. S. (2009). Long non-coding RNAs: Insights
into functions. Nature Reviews Genetics, 10(3), 155–159.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Muotri, A. R., Marchetto, M. C. N., Coufal, N. G., & Gage, F. H. (2007). The necessary
junk: New functions for transposable elements. Human Molecular Genetics, 16(R2),
R159–R167.
Neale, B. M., Rivas, M. A., Voight, B. F., Altshuler, D., Devlin, B., Orho-Melander,
M. … Daly, M. J. (2011). Testing for an unusual distribution of rare variants. PLoS
Genetics, 7(3), e1001322. 10.1371/journal.pgen.1001322
Nichols, P. L. (1984). Familial mental retardation. Behavior Genetics, 14(3), 161–170.
Plomin, R., DeFries, J. C., Knopik, V. S., & Neiderhiser, J. M. (2013). Behavioral genetics
(6 ed.). New York, NY: Worth Publishers.
Plomin, R., Haworth, C. M., & Davis, O. S. (2009). Common disorders are quantitative
traits. Nature Reviews Genetics, 10, 872–878.
Plomin, R., Haworth, C. M., Meaburn, E. L., Price, T. S., & Davis, O. S. (2013). Common DNA markers can account for more than half of the genetic influence on cognitive abilities. Psychological Science, 24(4), 562–568. doi:10.1177/0956797612457952
Psychiatric Gwas Consortium Bipolar Disorder Working, & Group (2011). Largescale genome-wide association analysis of bipolar disorder identifies a new
susceptibility locus near ODZ4. Nature Genetics, 43(10), 977–983. doi:10.1038/
ng.943
Purcell, S. M., Moran, J. L., Fromer, M., Ruderfer, D., Solovieff, N., Roussos, P. …
Sklar, P. (2014). A polygenic burden of rare disruptive mutations in schizophrenia.
Nature, 506(7487), 185–190. doi:10.1038/nature12975
Rietveld, C. A., Cesarini, D., Benjamin, D. J., Koellinger, P. D., De Neve, J. E., Tiemeier,
H. … Bartels, M. (2013). Molecular genetics and subjective well-being. Proceedings
of the National Academy of Sciences of the United States of America, 110(24), 9692–9697.
doi:10.1073/pnas.1222171110
Rietveld, C. A., Medland, S. E., Derringer, J., Yang, J., Esko, T., Martin, N. W.
… Koellinger, P. D. (2013). GWAS of 126,559 individuals identifies genetic
variants associated with educational attainment. Science, 340(6139), 1467–1471.
doi:10.1126/science.1235488
Rochman, B. (2012). Want to know my future. Parents grapple with delving into their
kids’ DNA. Time, 180, 26.
Schizophrenia Psychiatric Genome-Wide Association Study, C (2011). Genome-wide
association study identifies five new schizophrenia loci. Nature Genetics, 43(10),
969–976. doi:10.1038/ng.940
Shimoni, Y., Friedlander, G., Hetzroni, G., Niv, G., Altuvia, S., Biham, O., & Margalit,
H. (2007). Regulation of gene expression by small non-coding RNAs: A quantitative view. Molecular Systems Biology, 3(1).
Siva, N. (2008). 1000 Genomes project. Nature Biotechnology, 26(3), 256–256.
Speliotes, E. K., Willer, C. J., Berndt, S. I., Monda, K. L., Thorleifsson, G., Jackson,
A. U. … Loos, R. J. F. (2010). Association analyses of 249,796 individuals reveal
18 new loci associated with body mass index. Nature Genetics, 42(11), 937–948.
http://www.nature.com/ng/journal/v42/n11/abs/ng.686.html-supplementaryinformation
DNA Revolution and the Social and Behavioral Sciences
15
Stankiewicz, P., & Lupski, J. R. (2010). Structural variation in the human genome
and its role in disease. Annual Review of Medicine, 61, 437–455. doi:10.1146/
annurev-med-100708-204735
Tabor, H. K., Risch, N. J., & Myers, R. M. (2002). Candidate-gene approaches for
studying complex genetic traits: Practical considerations. Nature Reviews Genetics,
3(5), 391–397.
Thapar, A., & McGuffin, P. (1997). Anxiety and depressive symptoms in childhood—
A genetic study of comorbidity. Journal of Child Psychology and Psychiatry and Allied
Disciplines, 38(6), 651–656.
Trzaskowski, M., Dale, P. S., & Plomin, R. (2013). No genetic influence for childhood
behavior problems from DNA analysis. Journal of the American Academy of Child
and Adolescent Psychiatry, 52(10), 1048–1056 e1043. doi:10.1016/j.jaac.2013.07.016
Trzaskowski, M., Davis, O. P., DeFries, J., Yang, J., Visscher, P., & Plomin, R. (2013).
DNA evidence for strong genome-wide pleiotropy of cognitive and learning abilities. Behavior Genetics, 1–7. doi:10.1007/s10519-013-9594-x
Trzaskowski, M., Eley, T. C., Davis, O. S. P., Doherty, S. J., Hanscombe, K. B.,
Meaburn, E. L. … Plomin, R. (2013). First genome-wide association study on
anxiety-related behaviours in childhood. PloS One, 8(4), e58676. 10.1371/journal.
pone.0058676
Trzaskowski, M., Harlaar, N., Arden, R., Krapohl, E., Rimfeld, K., McMillan, A. …
Plomin, R. (2014). Genetic influence on family socioeconomic status and children’s
intelligence. Intelligence, 42, 83–88. doi:10.1016/j.intell.2013.11.002
Trzaskowski, M., Shakeshaft, N. G., & Plomin, R. (2013). Intelligence indexes
generalist genes for cognitive abilities. Intelligence, 41(5), 560–565. doi:10.1016/
j.intell.2013.07.011
Trzaskowski, M., Yang, J., Visscher, P. M., & Plomin, R. (2013). DNA evidence for strong
genetic stability and increasing heritability of intelligence from age 7 to 12 (Vol. 19,
pp. 380–384). Molecular Psychiatry.
Verweij, K. J., Vinkhuyzen, A. A., Benyamin, B., Lynskey, M. T., Quaye, L., Agrawal,
A. … Medland, S. E. (2012). The genetic aetiology of cannabis use initiation: A
meta-analysis of genome-wide association studies and a SNP-based heritability
estimation. Addiction Biology. 10.1111/j.1369-1600.2012.00478.x
Vinkhuyzen, A. A., Pedersen, N. L., Yang, J., Lee, S. H., Magnusson, P. K., Iacono,
W. G. … Wray, N. R. (2012). Common SNPs explain some of the variation in the
personality dimensions of neuroticism and extraversion. Translational Psychiatry,
2, e102. 10.1038/tp.2012.27
Visscher, P. M. (2008). Sizing up human height variation. Nature Genetics, 40(5),
489–490. doi:10.1038/ng0508-489
Visscher, P. M., Brown, M. A., McCarthy, M. I., & Yang, J. (2012). Five years of
GWAS discovery. The American Journal of Human Genetics, 90(1), 7–24. doi:10.1016/
j.ajhg.2011.11.029
Wray, N. R., Lee, S. H., Mehta, D., Vinkhuyzen, A. A. E., Dudbridge, F., & Middeldorp, C. M. (2014). Research review: Polygenic methods and their application
to psychiatric traits. Journal of Child Psychology and Psychiatry, 55(10), 1068–1087.
doi:10.1111/jcpp.12295
16
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
WTCCC (2007). Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature, 447(7145), 661–678. doi:10.1038/
nature05911
Yang, J., Benyamin, B., McEvoy, B. P., Gordon, S. D., Henders, A. K., Nyholt, D. R.
… Visscher, P. M. (2010). Common SNPs explain a large proportion of the heritability for human height. Nature Genetics, 42(7), 565–569. http://www.nature.
com/ng/journal/v42/n7/suppinfo/ng.608_S1.html
Yang, J., Lee, S., Goddard, M., & Visscher, P. (2011). GCTA: A tool for genome-wide
complex trait analysis. American Journal of Human Genetics, 88(1), 76–82.
doi:10.1016/j.ajhg.2010.11.011
MACIEJ TRZASKOWSKI SHORT BIOGRAPHY
While studying for his PhD, Dr Maciej Trzaskowski has learnt and applied
an extensive selection of statistical methods for quantitative (structural
equation modeling) and molecular genetics (e.g., GWAS, and GREML),
using various programming languages (e.g., R, Python, Unix, and Shell
Scripts). Dr. Trzaskowski has a strong methodological propensity and high
interest in exploring new methods. He has been awarded Gottesman-Shields
Prize for the best PhD thesis 2013 and since the completion of his PhD, Dr.
Trzaskowski remained at the Social, Genetic and Developmental Psychiatry
Centre as a postdoctoral research worker. To date, he has published 23
papers, 11 of these as first/joint first author. He has built his own network
of collaborations with researchers in the Broad Institute, the Queensland
Institute of Medical Research, Queensland Brain Institute (the University of
Queensland), Vrije Universiteit (VU) Amsterdam, Harvard, amongst others.
His research is increasingly recognized (as evidenced by the Behavior
Genetics Association Thompson and Fulker Awards received in 2013 and
2014, respectively) and, most recently, by a successful award of a fellowship
from British Academy. He has given invited talks at the London School
of Economics (by Lord Richard Layard, Emeritus Professor of Economics)
and at the fourth annual meeting by the Social Science Genetic Association
Consortium (by Professor Phillip Koellinger).
ROBERT PLOMIN SHORT BIOGRAPHY
Since 1994, Professor Robert Plomin has been MRC Research Professor of
Behavioral Genetics at the Institute of Psychiatry, King’s College London.
In 1994, he cofounded and subsequently directed the MRC Social, Genetic
and Developmental Psychiatry Centre, whose goal is to bring together
genetic and environmental strategies to study behavioral development.
In 1995, he launched the Twins Early Development Study (TEDs) of all
DNA Revolution and the Social and Behavioral Sciences
17
twins born in England and Wales in 1994–1996, which focuses on developmental problems in cognition and behavior, and which Professor Plomin
continues to direct. He has published more than 500 papers and more
than a dozen books, including the major textbook in the field (“Behavioral
Genetics,” Worth Publishers, 6th edition, 2013). His most recent book,
coauthored with Kathryn Asbury, focuses on genetics and education (“G
is for Genes: the Impact of Genetics on Education and Achievement,” Wiley
Blackwell, 2013). He has received lifetime research achievement awards
from the three major international associations in his field. For details, see
https://kclpure.kcl.ac.uk/portal/robert.plomin.html.
RELATED ESSAYS
Telomeres (Psychology), Nancy Adler and Aoife O’Donovan
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
Biology and Culture (Psychology), Robert Peter Hobson
Genetics and the Life Course (Sociology), Evan Charney
Genetic and Environmental Approaches to Political Science (Political Science),
Zoltán Fazekas and Peter K. Hatemi
Genetics and Social Behavior (Anthropology), Henry Harpending and Gregory Cochran
A Gene-Environment Approach to Understanding Youth Antisocial Behavior (Psychology), Rebecca Waller et al.
Behavioral Heterochrony (Anthropology), Victoria Wobber and Brian Hare
