236
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Published by Oxford University Press on behalf of the International Epidemiological Association
ß The Author 2012; all rights reserved.
International Journal of Epidemiology 2012;41:236–247
doi:10.1093/ije/dys016
Epigenesis for epidemiologists: does evo-devo
have implications for population health
research and practice?
George Davey Smith
MRC Centre for Causal Analyses in Translational Epidemiology, University of Bristol, Bristol, UK. E-mail: [email protected]
Epidemiologists engage with a wide range of other disciplines, as reflected in many of the topics that would
now be considered essential for any comprehensive
grounding in the subject. Thus, ‘‘lifecourse’’ was a
word I was only familiar with from sociological writings
when I was completing an epidemiology master’s
degree at the London School of Hygiene and Tropical
Medicine a quarter of a century ago, but now there are
Lifecourse Epidemiology research units, a plethora of
textbooks, and diagnosticos of the health situation from
national and international agencies often appeal directly to the notion. Socio-economic inequalities in
health were of more interest to political activists than
to academic epidemiologists, but (perhaps in the spirit
of Herbert Marcuse’s ‘repressive tolerance’1) are now an
utterly mainstream concern. Psychological notions are
pervasive, and some debates in the field have pitted
primarily social notions of disease origins2 against
more psychologically inclined approaches,3 bypassing
traditional epidemiological explanatory frameworks.
From econometrics, our discipline has imported instrumental variables analyses, among other borrowings.
Genetics used to seem so distant from epidemiological
concerns that, shamefully, during recruitment of participants for one of the studies I was involved with,
I threw away large volumes of buffy coats from which
DNA could have been extracted. Developmental biology
became of greater epidemiological concern with the
seminal studies of David Barker’s group on indicators
of fetal growth and later life health outcomes.4 More
recently, epigenetics has come to the fore and it is now
unimaginable that a referee’s report would state ‘I do
not know what the term epigenetics means’, a sentence
included in a BMJ review I received when I included the
term in a 1997 submission.5
In the current context, the excellent short book by
Patrick Bateson and Peter Gluckman, Plasticity,
Robustness, Development and Evolution6 (henceforth
PRDE), which the authors précis for IJE readers,7
must be seen as dealing with issues of considerable
concern to epidemiologists. The IJE published Patrick
Bateson’s early exposition of the epidemiological relevance of some of the ideas in the book8 and also the
first review of epigenetic epidemiology.9 Recent developments at the intersection of developmental and
evolutionary biology—now generally given the moniker evo-devo—are covered in this book, and implications for human health and disease are touched upon.
Here I am not going to outline the contents of the
book, since this has been done by the authors,7 rather
I will discuss several issues of particular relevance for
epidemiologists and other public health scientists.
Plasticity: epidemiologists’ bread
and butter
The notion that development is plastic—i.e. malleable
in response to environmental, genetic or stochastic perturbations—is not something that will surprise epidemiologists, although the terminology may not be a
familiar one. Indeed, epidemiology is concerned almost
exclusively with ‘plastic’ processes, with the reliable
identification of exposures that lead to modification
of disease risk through such processes being of central
concern. The concept of critical periods during which
plasticity is enhanced has been formalized within expositions of lifecourse epidemiology,10 and the particular relevance of the fetal period recognized. The
introduction of epidemiologists to developmental biology approaches to plasticity is certainly of value and
may lead to concretize the above notions and perhaps
to lead to clearer specification regarding mechanisms
of disease aetiology. Black box thinking, whilst having
its obvious successes in epidemiology, has limitations
in terms of formulating how to intervene within pathways to illness. Epidemiological thinking incorporates
the balance between exposure intensity and susceptibility; perhaps we can consider ‘degree of plasticity’ as
a susceptibility indicator. Thus, whereas the fetus
EPIGENESIS FOR EPIDEMIOLOGISTS
Figure 1 Maternal smoking during pregnancy – is it
harming the mother or fetus most?
may be highly susceptible (or plastic), exposures are
buffered by the mother and are thus often of much
lower intensity. Directly smoking 20 cigarettes a day
could be expected to lead to a larger exposure load on
the adult doing the smoking than on the fetus passively (as it were) exposed (Figure 1). The lungs of the
fetus do not get hit with the tar, for example. Thus,
whereas the cells of the adult might be much less
‘plastic’ than those of the fetus, the balance of exposure and susceptibility may be greater in the former. A
primary mechanism of plasticity discussed in PRDE—
unsurprisingly given the zeitgeist—is epigenetics, the
focus of this special issue of the IJE. Epigenetic processes that ensure the transmission of cellular traits, in
particular gene expression, across mitotic cell division
are clearly essential for development from the pluripotent zygote to the formed organism; in trivial terms,
why kidney cells on division produce more kidney, not
brain. Disease will often depend on changes in cellular
physiology that are transmitted across cell division—
this is how an exposure acting at one time, but not
persisting across time can lead to the initiation of a
lasting disease state. Social epidemiologists
refer to the
–
embodiment of disease risk11 14—exposures getting
‘under the skin’15,16—and permanently influencing
bodily state. Birthweight, body proportions, height,
blood pressure, intima media thickness, endocrine
and exocrine activity, skin tone, musculature and
lung function are just a few measures of interest to
epidemiologists, the development of which presumably
reflects epigenetic profiles that will be influenced by
internal and external signals that we consider to be
‘exposures’. The epigenetic dimension may be the
one in which genetic and environmental influences
come together to influence development and disease
risk. The forgotten aspect of disease causation—
chance—may also work partially at the epigenetic
level.17
In this light, the increased molecular understanding
of, and ability to assess, epigenetic states must be
welcome to epidemiologists. Without reviewing key
237
epidemiological issues covered elsewhere18,19—including tissue specificity, component of the epigenome assessed (e.g. DNA methylation, histone modifications,
microRNAs), measurement platform, validity and reliability of assessments—it is clear that epigenetic
evaluation can enhance exposure assessment, characterization of timing of exposure, mediation between
exposure and disease outcome, evaluation of causality
and identification of targets for potential preventative
or therapeutic intervention. Whilst the current
hyping of epigenetics may elicit antibodies in some
epidemiologists (including, it must be said, in me),
my prediction is that most epidemiologists will be
incorporating epigenetics in their empirical or theoretical work within the next few years. To tool up for
this requires broad understanding before specific implementation, and towards this end PRDE is an important resource. It joins a few other excellent texts
covering overlapping material from different perspectives that even
epidemiologists might be able to
–
understand.20 22 PRDE does, however, raise some
problematic issues which, whilst not all representing
central foci of the book, are of considerable epidemiological concern. It is to the epidemiological implications of these, rather than a reiteration of what the
authors have summarized earlier in the symposium7
and is covered fully in their book,6 that I now turn
my attention.
On learning not to be afraid of
genetics
One noticeable feature of some current literature is
the counterpoising of (here a deliberate caricature)
reductionist, old-school, deterministic, pro-eugenicist,
asocial, boring genetics with its apparent antithesis:
epigenetics. There is a nod in this direction in PRDE.
In a chapter entitled ‘Clarifications’ heritability is
given short shrift, for conventional reasons—it tells
us little about individuals, being a population measure; it is population-specific; it says nothing about the
genetic or non-genetic basis of a trait (with the familiar example of the heritability of number of legs in a
population of humans being zero, despite having two
legs clearly being due to genes acting during development)—and, all in all, is said to deserve (and get) no
further consideration in the book. One of the authors
has written elsewhere in similar mode with respect to
the heritability of a particular trait, obesity.23
Twin studies have suggested a strong genetic component to obesity,24 but such studies do not easily
distinguish between genetic effects and intrauterine epigenetic effects; epigenetic changes alter patterns of gene expression—not by modifying a
gene’s DNA sequence but through DNA methylation and enzymatic modifications of histone proteins that package genomic DNA. Studies of
238
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
weight loss and weight gain under controlled conditions in monozygotic twins25 in fact reveal considerable variation between members of a pair,
suggesting that non-genetic factors are also operating even when the genetic background is the
same. Although single-gene mutations have been
shown to underlie some cases of severe familial
morbid obesity, including genetic aberrations
that affect appetite control,26 for moderate obesity
and obesity-associated diseases, the size of the attributable risk derived from genome-wide association studies [GWAS] has been disappointingly
modest:27 Genetic variation may account for only
10–15% of relative risk.28
This dismissal of heritability might feel like good
news to epidemiologists—it gives more room for the
environmentally modifiable risk processes we want to
identify in order to leverage public health improvement—but is, I think, short-sighted. Most importantly, it is empirically poorly founded. Whilst it is
possible to generate a list of potential fallacies in
the estimation of heritability—such as challenging
the assumption that monozygotic and dizygotic
twins have equally similar environments—these
apply to only one method of estimation (classical
twin studies), and different designs, such as studies
of twins reared apart, extended twin-family studies,
adoption studies (including quasi-randomized adoption) and extended pedigrees generally yield similar
estimates of heritability. All of these designs are susceptible to bias, but they are different biases, and it is
unlikely that they would all distort the findings in the
same direction and to the same extent.17,29
Heritabilities for various traits are generally similar
in animals studied in captivity and in the wild,30
something difficult to envisage if they were generated
by artefact. With regard to the above quotation,
observed differences in weight change between monozygotic twins in experimental conditions says nothing
about the heritability of obesity in the population, and
neither the cited review (where 10–15% figure relates to type II diabetes, not obesity, and in any case
only relates to currently identified genetic markers)
nor other evidence supports the dismissal.
Of relevance to epidemiologists interested in the
Developmental Origins of Health and Disease
(DOHaD) field, and contrary to the above quote, genetic and intrauterine effects are distinguishable in extended pedigree studies. Furthermore, the prediction
based on intrauterine influences on heritability (i.e.
association between the same trait across generations)
is that maternal–offspring correlations would be
greater than paternal–offspring correlations. This is
generally not seen,31 and the exceptions, such as
type II diabetes,28 probably reflect fetal exposure to
the maternal metabolome.
The much trumpeted ‘missing heritability’32 relates
to such established variants, and yet for some diseases
a considerable proportion of the heritability is already
explained33 and use of whole-genome data show that
genetic variants tagged by GWAS chip single nucleotide polymorphisms (SNPs) account for a substantial
proportion of the heritability of many conditions, of
which height is the paradigmatic example.34
Evolutionary biologists will not be surprised that for
many traits there appear to be a very large number of
Mendelian contributions, all of small effect. Consider
intelligence, for example; it is likely that there would
have been substantial selective advantage for greater
cognitive abilities over the period of hominid evolution, and that genetic variants related to on-average
poorer performance in this domain would be selected
against. Writing on intellectual ability in 1927, the
polymath biologist and (latterly) anti-eugenicist,
Raymond Pearl,35,36 reported that he adhered ‘firmly
to Galton’s view that heredity plays the principal role’,
but with an ‘almost infinite manifold of germ-plasmic
contributions’.37 Pearl’s prediction (based on biological reasoning) has proved spectacularly prescient:
molecular genetics suggests that a myriad of
Mendelian influences of individually tiny effect contribute to the heritability of intelligence.38 The antieugenicist in Pearl welcomed this—attempts at eugenic ‘improvement’ made no biological (or ethical)
sense in this context.39 Indeed, the shuffling of such
tiny Mendelian effects could, Pearl said, ‘be relied on,
I think, to produce in the future, as it has in the past,
Shakespeares, Lincolns, and Pasteurs, from socially
and economically humble origins’.37 Straying from
academic language, Pearl considered that the eugenicist inability to see that ‘the economic element is perhaps the most significant biologically’ was ‘stupid’.37
We should learn to be as unafraid of genetics and
heritability as Pearl evidently was.
Epidemiologists are connoisseurs of change; the root
of the term in ‘epidemic’ gives this away. The
co-existence of high heritabilities with substantial
and rapid fluctuations in levels of obesity, and sustained increases in intelligence scores (‘the Flynn effect’40), demonstrates that [100-heritability]% is not
the proportion of disease due to the environment,17
despite the persistent misunderstanding that this is
the case. As Geoffrey Rose pointed out, in a population where everyone smoked, lung cancer would
appear to be a genetic disease, yet virtually all the
cases could be removed by the cessation of smoking.41
High heritability says nothing which limits the potential to alleviate disease; consider the highly heritable
phenylketonuria compared with Parkinson’s disease,
which has low heritability, in terms of success in prevention and treatment. In more general terms, heritability provides a route to strengthening the
knowledge base for improving population health.
The genetic variants that are being reliably identified
as influencing traits can be utilized to generate more
robust evidence regarding causality of environmentally modifiable risk factors than is provided by
EPIGENESIS FOR EPIDEMIOLOGISTS
conventional observational epidemiology.42,43 This approach (‘Mendelian randomization’) is now being
applied to the study of transgenerational influences
on disease through the intrauterine environment31
and to establishing the causal influence of particular
epigenetic modifications.44–46
Mendelian randomization is essentially the application of the concepts of phenocopy and genocopy to
population-based research settings. The term phenocopy is attributed to Goldschmidt,47 describing how
an environmental exposure can produce the same
outcome as does a genetic mutation. As Goldschmidt
explained, ‘different causes produce the same end effect, presumably by changing the same developmental
processes in an identical way’. 47 In human genetics,
the term phenocopy refers to an environmentally produced disease state that is similar to a genetic syndrome. For example, the niacin-deficiency disease
pellagra is clinically similar to the autosomal recessive
condition Hartnup disease,48 and pellagra has been
referred to as a phenocopy of the genetic disorder.49,50
Hartnup disease is due to reduced neutral amino acid
absorption from the intestine and re-absorption from
the kidney, leading to low levels of blood tryptophan
which in turn leads to a biochemical anomaly which
is similar to that seen when the diet is deficient in
niacin.51,52 Genocopy, a less familiar term, attributed
to Schmalhausen,53 is the mirror image of phenocopy—i.e. when genetic variation generates an outcome that could be produced by an environmental
exposure.54 Hartnup disease is a genocopy of pellagra,
whereas pellagra is a phenocopy of Hartnup disease. Mendelian randomization can, therefore, be
viewed as an application of the phenocopy–genocopy
dialectic that allows causation to be separated
from association through the common outcome produced by environmental or genetic predisposing factors. In their classic early 20th century studies of
pellagra, Goldberger and Sydenstricker identified a
dietary deficieny as the cause. They could not move
beyond the hypothetical pellagra preventative factor
to the actual nutrient which was deficient, but if
they had been armed with evidence on genetic investigations of Hartnup disease, this would have been a
simple step.
The scope of phenocopy–genocopy has been discussed by Zuckerlandl and Villett,55 who advance
mechanisms through which there can be equivalence
between environmental and genotypic influences.
Indeed, they state that ‘no doubt all environmental
effects can be mimicked by one or several mutations’.
The notion that genetic and environmental influences
can be both equivalent and interchangeable has
received considerable attention in developmental biology.56,57 Furthermore, population genetic analyses of
correlations between different traits suggest there are
common pathways of genetic and environmental influences, with Cheverud concluding that ‘most environmentally caused phenotypic variants should have
239
genetic counterparts and vice-versa’.58 Epigenetic processes can be influenced by both germline genetic
variation and the environment, and the outcome for
the same degree of epigenetic perturbation may be
expected to be the same. In this sense, the insights
from developmental biology—of ‘gene/environment
equivalence’ or ‘gene–environment interchangability’
(to use West-Eberhard’s terminology)56—may offer
more to epidemiologists than the much lauded (but
often spurious) claims made for gene–environment
interactions. The early dismissal of ‘heritability’ and
all it entails in PRDE is something epidemiologists
may want to ignore.
Transgenerational epigenetic
inheritance: fad or fate?
The short shrift given to germline genetic inheritance
in PRDE is compensated for by an interest in epigenetic inheritance. The essence of epigenetics is mitotic
cellular inheritance, yet from some popular presentations it is possible to get the impression that epigenetic processes stably transmitted across meiosis (i.e.
during gamete formation) are the central concern.
Undoubtedly, the apparently heretical echoes of the
Lamarckian ‘inheritance of acquired characteristics’
feeds into this interest.59 The implications of so-called
‘soft’ inheritance for human health has been discussed by one of the authors of PRDE in several
other places,60,61 and this issue is clearly of interest
to epidemiologists, concerned as they are with the
determinants of disease-related phenotypic variance
within populations. Some enthusiastic commentators
consider that epigenetic transgenerational inheritance
is well established as a major cause of obesity, for
example, and that obesity prevention policies should
be refocused to recognize this.62 This is surely premature; first we need to know how important, quantitatively, are such processes compared with other
influences on disease risk.
A very useful (if partisan) review by Jablonka and
Raz (referred to in PRDE) details the evidence for
transgenerational epigenetic inheritance,63 presents
potential mechanisms (including cytoplasmic metabolic loops, templating of 3D protein structure,
microRNAs and, probably the most discussed and
researched, DNA methylation) and provides a helpful
figure (reproduced here as Figure 2) summarizing the
processes leading to the induction of such changes.
To become ‘inheritance’ in the way that is usually
understood by the term, an epigenetically influenced
parental phenotype must successfully reproduce itself
across generations (i.e. across meiosis), and this
common-sense notion is the end-product of all of
the different induction mechanisms represented in
Figure 2. Whilst there are examples of what could
be referred to as phenotypically consistent transgenerational epigenetic inheritance,63 these relate
240
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Figure 2 The induction of epigenetic inheritance.63 (A) Direct germline induction: an external stimulus induces a germline
change from G0–G1 with no effect on the parental soma, which remains S0. The G1 state is inherited, and leads to the
development of an S1 soma. (B) Parallel induction: an external stimulus induces a change in the parent’s soma from S0–S1
and in its germline from G0–G1. The G1 state is inherited and causes the development of an S1 soma in descendants.
(C) Somatic induction: an external stimulus induces a change in the parent, altering its somatic phenotype from S0–S1.
The effect is transmitted from the S1 soma to the germline, where G0 is changed into G1; G1 is consequently inherited
and results in the development of an S1 soma. (D) Parallel induction with non-parallel effects: an external stimulus
alters the soma from S0–S2, and the germline from G0–G1. The germline modification is inherited and leads to the
development of S1 soma in subsequent generations. With all four types of induction, S1 could have an effect on G1 in
all descendants of the original induced parents (data not shown). Jablonka E, Raz G. Transgenerational epigenetic
inheritance: prevalence, mechanisms, and implications for the study of heredity and evolution. The Quarterly Review of
Biology, 84:2 (2009). Figure p. 157. Copyright ß 2009 by The University of Chicago. All rights reserved. Reprinted with
permission
mainly to plants, and their quantitative importance is
uncertain. Indeed, there is a largely ignored but substantial body of evidence from the classical genetics
literature which suggests that even in plants the
quantitative contribution of such inheritance mechanisms to the resemblance of phenotype across generations is low.64 Such evidence dates back to the
foundations of Mendelian genetics, and the ‘pure
line’ experiments in beans by the originator of the
term ‘gene’, Wilhelm Johannsen (Figure 3).65,66
Inbreeding, self-pollination (in plants), parthenogenesis and cloning produce groups of genetically identical or very similar organisms that nevertheless show
phenotypic variation. Indeed, such variation featured
in the celebrated early 20th century debates between
Mendelians (led, in the UK at least, by William
Bateson) and biometricians (led by Francis Galton’s
protégé, Karl Pearson and his close colleague WFR
Weldon).67–70 The key issue was that phenotypic difference in pure lines was not transmitted to their descendants: selecting by extreme of phenotype produced
groups whose descendants regressed completely to the
mean for the pure line with respect to the phenotype
that selection was by. What are now called epigenetic
mechanisms will, almost by definition, generate
these phenotypic differences: yet they show no
inter-generational transmission. By 1917, Raymond
Pearl could catalogue an extensive series of studies,
‘wholly negative so far as the production of any
change in type is concerned’, in many organisms.71
Over subsequent decades, the litany of such studies
continued in an ever-expanding range of plants and
animals. The consistency of findings across many different organisms and many different phenotypes is,
indeed, what allowed a confident dismissal of the
quantitative importance of Lamarckian inheritance
EPIGENESIS FOR EPIDEMIOLOGISTS
Figure 3 Wilhelm Johannsen 1857–1927
by the founders of population genetics. For example,
in 1931 JBS Haldane wrote, with respect to a ‘nearly
genetically homogeneous population in a nearly constant environment’, that:
Most of the differences which occur are not
inherited, as Johannsen first showed. If the environment is variable more differences may occur.
And in general, these are not inherited. This important negative fact is the basis for the denial by
most geneticists of Lamarck’s theory that acquired
characters are inherited, or more accurately that
differences in one generation due to diversity of
environment give rise to differences in another
generation due to diversity of ancestry. The
denial of Lamarckism also rests on the positive
fact that most genetically determined differences
are due to differences in genes, distributed according to Mendel’s laws, and inconsistently with
Lamarck’s hypothesis.72
Such ‘pure line’ studies were utilized to identify the
effect and quantify the low-frequency occurrence of
new mutations (which against a previously shared
genetic background could produce heritable phenotypic perturbations).73 The conclusion from over 100
years of research must be that epigenetic inheritance
is not a major contributor to phenotypic resemblance
across generations, yet strangely—and perhaps because of the unexceptional nature of the findings—
this vast literature has, in some circles, been forgotten. Instead, occasional examples of phenotypically
241
consistent epigenetic inheritance relating to a particular phenotype in a particular organism are given considerable attention, with the implication that they
represent a general phenomenon. The transgenerational inheritance of longevity in Caenorhabditis elegans74 appears established, and there is good
evidence of transmission of leaf hairiness in response
to injury in monkey-flowers.75 Transmission of
stress-induced DNA methylation in asexual dandelions76 and of spontaneously arising ‘epialleles’
(as such methylation variants are sometimes called)
in other plants77,78 have been demonstrated although
there are no clear phenotypic effects of these transmissible epigenetic states. In many cases the epigenetic basis of transgenerational inheritance cannot be
established. For example, a celebrated example of
apparent epigenetic inheritance in mouse-ear cress,
of which Jablonka and Raz said ‘DNA methylation
[is] assumed to be involved’, has now been shown
to be due to gene duplication.79 Haldane referred to
Victor Jollos’ Dauermodifikationen (semi-heritable environmentally influenced and adaptive traits mediated
by the cytoplasm,80 which feature in Jablonka and
Raz’s review),63 but, even if real, these effects
(which dissipated over a few generations) were of
very minor importance compared with the influence
of germline genetic variation.72
Interested as they are in the determinants of health
of populations, epidemiologists will want to focus on
the major determinants of variation in disease risk in
human populations, and currently transgenerational
epigenetic inheritance is not an established player.
In mammals, there is almost complete reprogramming
of the epigenome in both the primordial germ cells
and the pre-implantation embryo which would
mitigate against such transmission.81 Even plant
researchers interested in such processes point out
that unequivocal examples ‘have remained elusive’82
or that reference is to ‘isolated and spectacular cases’,83 and in humans and other mammals evidence of
quantitatively important effects is thin. There is some
confusion in the literature in that on occasion
processes that do not produce phenotypic similarity
between parents and offspring, or that do so but
not through transmissible cellular states, are subsumed under the heading ‘epigenetic inheritance’.
Clearly, cultural transmission can lead to phenotypic
similarity of parents and offspring, and an initially
single-generation effect of maternal physiology on
the developing fetus may lead to cross-generation
replication of phenotype (as in the case of type II
diabetes) without transgressing conventional germline
genetic inheritance mechanisms. The influence of the
non-transmitted parental allele on offspring phenotype, including of the paternal Y chromosome on
daughters’ phenotype, have been demonstrated,84
and whilst such putative genetic effects which are
not transmitted by germline sequence are fascinating,
they do not represent mechanisms for the
242
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
transmission of environmentally induced parental
phenotype to the offspring. In the epidemiological
arena, many of the best known examples that are
sometimes referred to as demonstrating epigenetic inheritance—such as the apparent influence of grandpaternal dietary challenge on grandchild mortality or
paternal pre-pubertal smoking on offspring obesity85—relate to exposure-phenotypic associations
without any mechanism being demonstrated, and
are liable to at least the same degree of confounding
and other biases as are conventional epidemiological
studies. An important issue with the incorporation of
epigenetics into epidemiological studies is that, in
contrast to the investigation of germline genetic variation,86 epigenetic markers are phenotypic, with all
the problems that entails. Understanding how phenotypic resemblance is recreated across generations, and
how exposures in previous generations may influence
the current generation without generating transgenerational phenotypic similarity, are challenging
tasks. Epidemiology requires retooling to be up to
the job, but it is a job worth doing.
From FOAD to DOHaD
For epidemiologists, the Fetal Origins of Adult Disease
(FOAD) was an exciting concept, that rapidly developed from the seminal 1986 ecological analysis by
David Barker and Clive Osmond87 through the 1989
study of birthweight and cardiovascular mortality,88
Hales and Barker’s 1992 proposal of the thrifty phenotype,89 the First World Congress on FOAD in Mumbai
in 2001,90 to the Second World Congress on FOAD in
Brighton in 2003.91 Then FOAD disappeared, with
little fanfare, and we had the Third World Congress
on the Developmental Origins of Health and
Disease.92 At the time of the name change, I cynically
thought this was about image: I had Googled the
‘FOAD’ congress in 2003, and was directed to a
vastly more entertaining website—now sadly defunct—‘Fuck off and die’, which informed me about
the nefarious activities of Courtney Love and related
icons. Googling DOHaD gets you straight to where
you want to be. Also, as a one-time affiliate of a
groupuscule of the Trotskyist Fourth International (I
seem to remember we traded in the newspaper
Socialist Drinker) I thought perhaps the FOAD to
DOHaD name change was the result of a Kremlin
coup. The switch was, however, profound, emphasizing as it did the long sweep of ‘development’, from
well before birth (indeed from before one’s parents’
birth) to the time of life when ‘development’ becomes
a euphemism for decline and decay.
PRDE again proves highly stimulating when thinking
about DOHaD. Over what stages of development can
environmental cues provide information that allow for
predictive adaptive responses: the developing organism
preparing itself for the future it faces? PRDE largely
(but not exclusively) discusses the fetal period and
cues mediated through the mother. Others disagree:
Jonathan Wells, for example, argues that maternal
adaptation to both the environment and a
semi-parasitic fetus (whose genetic future the mother
partly—but only partly—shares) is key.93 Cues acting
during any particular pregnancy relate to a short time
period, and as Chris Kuzawa suggests a longer time
period may be more informative about the future,
and maternal measures relating to this extended
time course could have more motive force.94
Translating this into epidemiologically tractable questions, maternal height—a marker of under-nutrition,
infection and other potential causes of growth interruption—reflects a longer time window than the duration of a single pregnancy, and the potential influence
of maternal height can also be directly compared with
paternal height as a control exposure which is patterned in the same way as maternal height by
socio-economic and other confounders.95 Such a
design finds that in a low-income country setting maternal height is strongly associated with mortality of
offspring in infancy and early childhood, with no comparable association seen for paternal height, providing
support for the notion that embodied maternal capital
importantly shapes the life chances of offspring.96
Measures like maternal height can be considered to
be averaging apparatuses, providing cumulative indications of conditions prior to the period in which an
organism encounters the world. This might provide
more valuable cues than a snapshot of what is happening during the fetal period. Indeed, rapidly fluctuating environments may not provide useful
information regarding a favourable direction for
phenotypic change, and instead promote a generalized
increase in offspring plasticity (to cope with a changing
environment)97 and phenotypic variance (increasing
the probability that at least a proportion of offspring will be suited for the actual environment
experienced).17
In PRDE, the influence of early postnatal stressors
on timing of puberty is discussed—if the environment
is tough, better get started with breeding early—but
perhaps for chronic adulthood disease in general the
responses to cues provided in infancy and childhood
may be as important, if not more so, than fetal
experiences. Birthweight is a convenient measure
available from routine records for large proportions
of the populations of high-income countries; there is
no equivalent indicator of infancy and childhood experience. This could explain why there is a much
richer epidemiological literature on fetal experiences
(indexed by birthweight) than the early postnatal environment. The mother, however, provides a remarkable buffer between the environment and the fetus
she carries—consider how small the influences of
major shifts in maternal nutrition, such as during
famines, are on offspring birthweight—and thus the
direct experience of the environment by an infant or
child might provide considerably richer information
EPIGENESIS FOR EPIDEMIOLOGISTS
243
about the potential future they face than cues
mediated through the mother. This is an issue
crying out for serious epidemiological engagement.
For population health researchers, an implication of
DOHaD is that exposures which act at a particular
developmental period and are changing over time
should generate observable birth cohort effects. This
is not the place to review the substantial evidence on
this issue, but my reading98 is that where there is
good evidence that an exposure has its effects at a
particular early life period—for example, infancyacquired Helicobacter pylori infection—such birth
cohort effects in the related diseases, peptic
ulcer and stomach cancer in this case, can be seen.
For blood pressure, haemorrhagic stroke, bronchitis
and respiratory TB there is evidence of cohort effects
in population health data. However, for obesity, coronary heart disease, type II diabetes and ischaemic
stroke—which receive much attention in the DOHaD
literature—there are no clear cohort effects, and very
clear period effects, with all age groups at a given
time showing the same secular trend. This suggests
that changing exposure patterns in the population are
influencing disease risk independent of the period of
life (or development) during which they are experienced. The developmental origins of disease may be of
more population health importance in societies before
than after the transition to the so-called ‘noncommunicable disease’ period.98
‘Totally constant and highly
variable’: genetics, epigenetics and
the epidemiologist
In Wilhelm Johannsen’s seminal 1903 monograph, he
borrowed a term from Hugo De Vries, to describe genetically pure lines as ‘totally constant and highly variable’.65 Figure 4, from Johannsen’s beautiful paper
‘The genotype conception of heredity’,99 illustrates
how variance in a population (the lower panel) can
be considered as a summation of the variance between the pure lines illustrated in the upper five
panels and the variance within the pure lines. Both
the constancy (the means of the pure lines within a
given environment) and the high variability (the
variance within a pure line, due to environment and
chance) need to be interrogated to understand what
creates the distribution of a disease-related trait
within a population. As discussed above, the constant
average genetic effects can provide, paradoxically,
one of the most reliable ways of understanding the
environmental causes of variation. Thus whereas
epigenetics and plasticity are attractive—indeed
essential—concepts for epidemiologists, they should
be additional to, not a replacement for, continuing
efforts to understand germline genetic influences on
development and health.
Figure 4 The lengths of five different pure lines of beans
and a ‘population’ formed by their union99
244
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Development as an active and responsive generator
of phenotype is central to evo-devo, and Conrad
Waddington is seen, in PRDE and elsewhere, as the
hero who linked genetics to evolution and development. Name-checking Waddington is now de rigueur
in overviews of plasticity, development and epigenetics20–22 and it is appropriate that his work is recognized, but it is surprising that the contribution of
Sewall Wright is not acknowledged, in PRDE or in
many other recent presentations. This is perhaps because Wright is seen as one of the three founders
(with Haldane and Fisher) of population genetics,
which provided robust foundations for the so-called
‘modern synthesis’ of Darwinism with Mendelian
genetics in the 1930s and 1940s. This synthesis is
one which evo-devo is apparently transcending. Yet
Wright was committed to linking genetics with development (to the extent of contemplating a salary cut in
order to further such work100), and there is much in
his extensive writings that can help epidemiologists
think about what determines variation of disease
and disease-related phenotypes in human populations.100,101 Consider his celebrated study of otocephaly (a deformity of the face and head) in guinea
pigs.102 This developmental anomaly shows gradation,
from minimal effect to almost entire loss of the head
(see Wright’s diagrammatic schema in Figure 5), and
also clearly has a genetic basis. In a high-risk highly
inbred line of guinea pigs 1.5% were born with the
condition to some variable degree, but whether the
condition was experienced at all, and if so the severity
of the condition, did not appear to have a straightforward genetic basis. At the level of the individual
guinea pig, the probability that a littermate of an otocephalic would itself be otocephalic was not ‘appreciably greater than in the case of a sibling from a
different litter, or in the case of a random animal
from the whole strain. The precipitating factor must
be peculiar to individuals’.103 This is strikingly similar
to the many human characteristics of which Plomin
and Daniels asked the question ‘Why are children in
the same family so different from each other?’,104 as
recently considered from an epidemiological perspective in the IJE.17,105–109 For the individual guinea pig,
essentially chance (and what might now be considered epigenetic) irregularities at a crucial developmental stage were considered key,100,102 although the
season of birth, with conditions in the guinea pig
colony being poor in the winter, influenced the proportion of guinea pigs affected. The same pattern is
seen with respect to human disease risk;17 at a group
level external factors that influence the rate of a
disease—indeed, may be responsible for virtually all
cases—can be identified, even when at an individual
level it is not possible to clearly predict who will and
who will not develop the disease.
The gradation of anomalies in Wright’s guinea pigs
displayed in Figure 5 is reminiscent of the distribution
Figure 5 Sewall Wright’s semidiagrammatic representation
of the main grades of otocephaly102
of many human traits that underlie disease, such as
blood pressure and cholesterol level. In similar vein,
Wilhelm Johannsen pointed out that ‘theoretically, as
well as practically, there are no sharp limits between
‘normal’ and ‘pathological’ manifestations of life’,99
and the graphs of the distribution of the size of his
beans (in Figure 4) could be re-labelled as the size
(body mass index, say) of members of a human population. Those who are labelled as being obese are in
the upper tail of the distribution. The lessons of recent
genome-wide association studies is that common disorders behave as quantitative traits; such disorders
are either the extreme of a distribution or the probabilistic outcomes of an underlying liability with such
a continuous distribution.110 Trait variation will subsume the influence of the germline genome, the environment and chance, and epigenetic processes may
mediate all of these. The task for public health is to
shift such distributions by environmental change.41
Understanding the developmental processes influencing phenotype, and the gene–environment interchangability underlying this, can help in this task.
PRDE conveys the excitement of a rapidly developing field which epidemiologists will want to get to
grips with. I agree with the authors’ concluding statement that ‘with a whole array of promising new research areas and techniques emerging, integrative
biologists have a lot to be excited about’. The shock
of the new should not, however, obscure the fact that
we still have much to learn from the old biology, and
that the relative quantitative importance of different
processes in determining health outcomes within a
population do not correlate with the extent to which
they represent cutting-edge technologies or ideas.
EPIGENESIS FOR EPIDEMIOLOGISTS
Acknowledgements
20
I would like to thank Peter Donnelly for translating
parts of Wilhelm Johannsen’s 1903 text65 and Ezra
Susser for comments on an earlier draft.
21
22
Conflict of interest: None declared.
23
References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Marcuse H. Repressive tolerance. In: Wolff RP, Moore B
Jr, Marcuse H (eds). A Critique of Pure Tolerance. Boston:
Beacon Press, 1965, pp. 95–137.
Lynch J, Davey Smith G, Kaplan G, House J. Income inequality and mortality: importance to health of individual
income, psychosocial environment, or material conditions. BMJ 2000;320:1200–04.
Marmot M, Wilkinson RG. Psychosocial and material
pathways in the relation between income and health: a
response to Lynch et al. BMJ 2001;322:1233.
Barker DJP (ed.). Fetal and Infant Origins of Adult Disease.
London: British Medical Journal, 1992.
Davey Smith G, Hart C, Ferrell C et al. Birth weight of
offspring and mortality in the Renfrew and Paisley Study:
prospective observational study. BMJ 1997;315:1189–93.
Bateson P, Gluckman P. Plasticity, Robustness, Development
and Evolution. Cambridge: Cambridge University Press,
2011.
Bateson P, Gluckman P. Plasticity and robustness in development and evolution. Int J Epidemiol 2012;41:219–23.
Bateson P. Fetal experience and good adult design. Int J
Epidemiol 2001;30:928–34.
Jablonka E. Epigenetic epidemiology. Int J Epidemiol 2004;
33:1–7.
Ben-Shlomo Y, Kuh D. A life course approach to chronic
disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives. Int J Epidemiol
2002;31:285–93.
Najman JM, Davey Smith G. The embodiment of
class-related and health inequalities: Australian policies.
Aust N Z J Public Health 2000;24:3–4.
Krieger N, Davey Smith G. ‘Bodies count’, and body
counts: social epidemiology and embodying inequality.
Epidemiol Rev 2004;26:92–103.
Krieger N. Embodiment: a conceptual glossary for
epidemiology. J Epidemiol Community Health 2005;59:
35055.
Gravlee CC. How race becomes biology: embodiment of
social inequality. Am J Phys Anthropol 2009;139:47–57.
Ferraro KF, Shippee TP. Aging and cumulative inequality:
how does inequality get under the skin? Gerontologist
2009;49:333–43.
Shonkoff JP. Building a foundation for Prosperity on
the science of early childhood development. Pathways
2011;Winter:10–15.
Davey Smith G. Epidemiology, epigenetics and the ‘Gloomy
Prospect’: embracing randomness in population health research and practice. Int J Epidemiol 2011;40:537–62.
Heijmans BT, Mill J. The seven plagues of epigenetic
epidemiology. Int J Epidemiol 2012;41:74–78.
Relton CL, Davey Smith G. Is epidemiology ready
for epigenetics? Int J Epidemiol 2012;41:5–9.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
245
Gilbert SF, Epel D. Ecological Developmental Biology 2009:
Sunderland, Mass: Sinauer Associates, 2009.
Jablonka E, Lamb M. Evolution in Four Dimensions: Genetic,
Epigenetic, Behavioral, and Symbolic Variation in the History of
Life. USA: MIT, 2005.
Carey N. The Epigenetics Revolution: How Modern Biology is
Rewriting Our Understanding of Genetics, Disease and
Inheritance. UK: Icon Books Ltd, 2011.
Gluckman PD, Hanson M, Zimmet P, Forrrester T. Losing
the war against obesity: the need for a developmental
perspective. Sci Transl Med 2011;3: 93cm19.
Stunkard AJ, Foch TT, Hrubec Z. A twin study of human
obesity. JAMA 1986;256:51–54.
Bouchard C. Gene-environment interactions in the etiology of obesity: defining the fundamentals. Obesity
2008;16(Suppl. 3):S5–10.
Rankinen T, Zuberi A, Chagnon YC et al. The human
obesity gene map: the 2005 update. Obesity 2006;14:
529–644.
Manolio TA, Collins FS, Cox NJ et al. Finding the
missing heritability of complex diseases. Nature 2009;
461:747–53.
Ahlqvist E, Ahluwalia TS, Groop L. Genetics of type 2
diabetes. Clin Chem 2011;57:241–54.
Plomin R. Child development and molecular genetics: 13
years later. Child Dev 2012; in press.
Weigensberg I, Roff DA. Heritabilities: can they be reliably estimated in the laboratory? Evolution 1996;50:
2149–57.
Davey Smith G. Assessing intrauterine influences on offspring health outcomes: can epidemiological findings
yield robust results? Basic Clin Pharmacol Toxicol 2008;
102:245–56.
Maher B. Personal genomes: The case of the missing heritability. Nature 2008;456:18–21.
Orozco G, Barrett JC, Zeggini E. Synthetic associations in
the context of genome-wide association scan signals.
Hum Mol Genet 2011;19:137–44.
Yang J, Benyamin B, McEvoy BP et al. Common SNPs
explain a large proportion of the heritability for human
height. Nat Genet 2010;42:565–69.
Davey Smith G. Pearls of wisdom: eat, drink, have sex
(using condoms), abstain from smoking and be merry.
Int J Epidemiol 2010;39:941–47.
Hendricks M. Raymond Pearl’s ‘‘mingled mess’’. Johns
Hopkins Mag 2006;58:50–56.
Pearl R. Differential fertility. Q Rev Biol 1927;2:102–18.
Davies G, Tenesa A, Payton A et al. Genome-wide
association studies establish that human intelligence is
highly heritable and polygenic. Mol Psychiatry 2011;16:
996–1005.
Pearl R. The biology of superiority. Am Mercury 1927;12:
257–66.
Flynn JR. What is Intelligence?: Beyond the Flynn Effect.
Cambridge University Press, 2007.
Rose G. Sick individuals and sick populations. Int J
Epidemiol 1985;14:32–8; (Reprinted Int J Epidemiol 2001;
30:427–32).
Davey Smith G, Ebrahim S. ‘Mendelian randomization’:
can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol
2003;32:1–22.
Davey Smith G. Use of genetic markers and genediet interactions for interrogating population-level
246
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
causal influences of diet on health. Genes Nutr 2011;6:
27–43.
Relton CL, Davey Smith G. Epigenetic epidemiology of
common complex disease: prospects for prediction,
prevention and treatment. PLoS Med 2010;7:e1000356.
Relton CL, Davey Smith G. Two step epigenetic
Mendelian randomization: a strategy for establishing
the causal role of epigenetic processes in pathways to
disease. Int J Epidemiol 2012;41:161–76.
Groom A, Potter C, Swan DC et al. Postnatal growth and
DNA methylation are associated with differential gene
expression of the TACSTD2 gene and childhood fat
mass. Diabetes 2012;61:391–400.
Goldschmidt RB. Physiological Genetics. New York: McGraw
Hill, 1938.
Baron DN, Dent CE, Harris H, Hart EW, Jepson JB.
Hereditary pellagra-like skin rash with temporary cerebellar ataxia, constant renal amino-aciduria, and other bizarre biochemical features. Lancet 1956;271:421–28.
Snyder LH. Fifty years of medical genetics. Science 1959;
129:7–13.
Guy JT. Oral manifestations of systematic disease. In:
Cummings CW et al. (ed) Otolaryngology – Head and Neck
Surgery. Mosby-Year Book Inc., 1993.
Kraut JA, Sachs G. Hartnup disorder: unravelling the
mystery. Trends Pharmacol Sci 2005;26:53–55.
Broer S, Cavanaugh JA, Rasko JEJ. Neutral amino acid
transport in epithelial cells and its malfunction in
Hartnup disorder. Biochem Soc Trans 2004;33:233–36.
Gause GF. The relation of adaptability to adaption. Q Rev
Biol 1942;17:99–114.
Jablonka-Tavory E. Genocopies and the evolution of
interdependence. Evol Theor 1982;6:167–70.
Zuckerkandl E, Villet R. Concentration – affinity equivalence in gene regulation: convergence and environmental
effects. Proc Natl Acad Sci U S A 1988;85:4784–88.
West-Eberhard MJ. Developmental Plasticity and Evolution.
Oxford University Press, 2003.
Leimar O, Hammerstein P, Van Dooren TJM. A new
perspective on developmental plasticity and the principles
of adaptive morph determination. Am Nat 2006;167:
367–76.
Cheverud JM. A comparison of genetic and phenotypic
correlations. Evolution 1988;42:958–68.
Haig D. Weismann rules! ok? epigenetics and the
Lamarckian temptation. Biol Philos 2007;22:415–28.
Gluckman PD, Hanson MA, Beedle AS. Non-genomic
transgenerational inheritance of disease risk. Bioessays
2007;29:145–54.
Hanson M, Low FM, Gluckman PD. Epigenetic epidemiology: the rebirth of soft inheritance. Ann Nutr Metab
2011;58:8–15.
Niculescu MD. Epigenetic transgenerational inheritance:
should obesity prevention policies be reconsidered? Synesis
J Sci Technol Ethics Policy 2011;2:G18–G26.
Jablonka E, Raz G. Transgenerational epigenetic inheritance: prevalence, mechanisms, and implications for the
study of heredity and evolution. Q Rev Biol 2009;84:
131–76.
Charlesworth D, Charlesworth B. Not quite a revolution.
The Guardian Review; 3 September 2011, p. 15.
Johannsen W. Uber Erblichkeit in populationen und in
reinen linien, Gustav Fischer, Jena. 1903 (Partial English
translation available in Gall H, Putschar E (eds). Selected
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
readings in biology for natural sciences, volume 3.
Chicago: University of Chicago Press, 1955).
Rolls-Hansen N. The crucial experiment of Wilhelm
Johnanssen. Biology and Philosophy 1989;4: 303–29.
MacKenzie D. Statistics in Britain, 1865-1930: The Social
Construction of Scientific Knowledge. Edinburgh University
Press, 1981.
Provine WB. The Orgins of Theoretical Population Genetics
(Chicago History of Science & Medicine). Chicago: The
University of Chicago Press, 1971.
Rolls-Hansen N. Sources of Wilhelm Johannsen’s genotype theory. J Hist Biol 2009;42:457–93.
Churchill FB. William Johanssen and the genotype concept. J Hist Biol 1974;7:5–30.
Pearl R. The selection problem. Am Nat 1917;51:65–91.
Haldane JBS. Some Principles of causal analysis in genetics. Erkenntnis 1936;6:346–57.
Lynch M. The rate of polygenic mutation. Genet Res 1988;
51:137–48.
Greer EL, Maures TJ, Ucar D et al. Transgenerational epigenetic inheritance of longevity in Caenorhabditis elegans. Nature 2011;479:365–71.
Scoville AG, Barnett LL, Bodbyl-Roels S, Kelly JK,
Hileman LC. Differential regulation of a MYB transcription factor is correlated with transgenerational epigenetic
inheritance of trichome density in Mimulus guttatus. New
Phytol 2011;191:251–63.
Verhoeven KJ, Jansen JJ, van Dijk PJ, Biere A.
Stress-induced DNA methylation changes and their heritability in asexual dandelions. New Phytol 2010;185:
1108–18.
Schmitz RJ, Schultz MD, Lewsey MG et al. Transgenerational epigenetic instability is a source of novel methylation variants. Science 2011;334:369–73.
Johannes F, Porcher E, Teixeira FK, Saliba-Colombani V,
Simon M et al. Assessing the impact of transgenerational
epigenetic variation on complex traits. PLoS Genet 2009;5:
e1000530.
Yi H, Richards EJ. Gene duplication and hypermutation
of the pathogen resistance gene SNC1 in the Arabidopsis
bal variant. Genetics 2009;183:1127–34.
Harwood J. Geneticists and the evolutionary synthesis in
interwar Germany. Ann Sci 1985;42:279–301.
Kota SK, Feil R. Epigenetic transitions in germ cell development and meiosis. Dev Cell 2010;19:675–86.
Richard CL, Wendel JF. The hairy problem of epigenetics
in evolution. New Phytol 2011;191:7–9.
Angers B, Castonguay E, Massicotte R. Environmentally
induced phenotypes and DNA methylation: how to deal
with unpredictable conditions until the next generation
and after. Mol Ecol 2010;19:1283–95.
Nelson VR, Nadeau JH. Transgenerational genetic effects.
Epigenomics 2010;2:797–806.
Pembrey ME, Bygren LO, Kaati G et al. Sex-specific,
male-line transgenerational responses in humans. Eur J
Hum Genet 2006;14:159–66.
Davey Smith G, Lawlor DA, Harbord R, Timpson NJ,
Day I, Ebrahim S. Clustered environments and randomized genes: a fundamental distinction between conventional and genetic epidemiology. PLoS Med 2008;4:
1985–92.
Barker DJP, Osmond C. Infant mortality, childhood nutrition, and ischaemic heart disease in England and
Wales. Lancet 1986;327:1077–81.
SOMETHING OLD, SOMETHING NEW
88
Barker DJP, Osmond C, Winter PD, Margetts B,
Simmonds SJ. Weight in infancy and death from ischaemic heart disease. Lancet 1989;334:577–80.
89
Hales CN, Barker DJP. Type 2 (non-insulin-dependent)
diabetes mellitus: the thrifty phenotype hypothesis.
Diabetologia 1992;35:595–601.
90
Robinson R. The fetal origins of adult disease. BMJ 2001;
322:375.
91
Hanson M, Gluckman P, Bier D et al. Report on the 2nd
World Congress on Fetal Origins of Adult Disease,
Brighton, U.K., June 7-10, 2003. Pediatr Res 2004;55:
894–97.
92
Gillman MW, Barker D, Bier D et al. Meeting report on
the 3rd International Congress on Developmental
Origins of Health and Disease (DOHaD). Pediatr Res
2007;61:625–29.
93
Wells JCK. The thrifty phenotype as an adaptive maternal effect. Biol Rev 2007;82:143–72.
94
Kuzawa CW. Fetal origins of developmental plasticity:
are fetal cues reliable predictors of future nutritional
environments? Am J Hum Biol 2005;17:5–21.
95
Davey Smith G. Assessing intrauterine influences on
offspring health outcomes: can epidemiological studies
yield robust findings? Basic Clin Pharmacol Toxicol 2008;
102:245–56.
96
Subramanian SV, Ackerson LK, Davey Smith G,
John NA. Association of maternal height with child
mortality, anthropometric failure and anemia in India.
JAMA 2009;301:1691–701.
97
Beldade P, Mateus AR, Keller RA. Evolution and molecular mechanisms of adaptive developmental plasticity.
Mol Ecol 2011;20:1347–63.
98
Davey Smith G, Lynch J. Commentary: Social capital,
social epidemiology and disease aetiology. Int J
Epidemiol 2004;33:691–70.
99
100
101
102
103
104
105
106
107
108
109
110
247
Johannsen W. The Genotype Conception of Heredity. Am
Nat 1911;45:129–59.
Provine WB. Sewall Wright and Evolutionary Biology.
Chicago: University of Chicago Press, 1986.
Wright S. Evolution and the Genetics of Populations, Volume
4: Variability Within and Among Natural Populations.
Chicago: University of Chicago Press, 1978.
Wright S. On the genetics of subnormal development of
the head (otocephaly) in the guinea pig. Genetics 1934;
19:471–505.
Wright S. The First Meckel Oration: on the causes of
morphological differences in a population of guinea
pigs. Am J Med Genetics 1984;18:591–616.
Plomin R, Daniels D. Why are children in the same
family so different from each other? Behav Brain Sci
1987;10:1–16.
Plomin R, Daniels D. Why are children in the same
family so different from each other? Int J Epidemiol
2011;40:582–92.
Plomin R. Commentary. Why are children in the same
family so different? Non-shared environment three decades later. Int J Epidemiol 2011;40:592–96.
Sesardic N. Commentary: An explosion without a bang.
Int J Epidemiol 2011;40:592–96.
Turkheimer E. Commentary: Variation and causation in
the environment and genome. Int J Epidemiol 2011;40:
598–601.
Conley D. Commentary: Reading Plomin and Daniels
in the post-genomic age. Int J Epidemiol 2011;40:
596–98.
Plomin R, Haworth CMA, Oliver SP. Common disorders are quantitative traits. Nat Rev Genet 2009;10:
872–78.
Published by Oxford University Press on behalf of the International Epidemiological Association
ß The Author 2012; all rights reserved.
International Journal of Epidemiology 2012;41:247–249
doi:10.1093/ije/dys021
Something old, something new, something
false but much that’s true
Patrick Bateson* and Peter Gluckman
*Corresponding author. Sub-Department of Animal Behaviour, Department of Zoology, Madingley, Cambridge CB23 8AA, UK.
E-mail: [email protected]
Accepted
30 January 2012
We are grateful to Christopher Kuzawa, not only for
reading our book carefully and setting its messages in
a broad context, but also for drawing attention to the
implications of evolutionary biology for medicine. This
is an area in which we are both deeply interested, but
we did not present it as a major feature of our book.
We are glad that Kuzawa makes an issue of it in his
review. He points out that the comparative approach
encouraged by evolutionary biology should make biomedical scientists rather more careful than has often