University of Adelaide Press

Chapter Title: Genes for teeth — drawing inference from family data
Chapter Author(s): Toby Hughes and Grant Townsend

Book Title: New Directions in Dental Anthropology

Book Subtitle: Paradigms, methodologies and outcomes
Book Editor(s): Grant Townsend, Eisaku Kanazawa, Hiroshi Takayama
Published by: University of Adelaide Press. (2012)
Stable URL: [Link]

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@[Link].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
[Link]

This book is licensed under a Creative Commons Attribution-NonCommercial-

NoDerivatives 4.0 International. To view a copy of this license, visit
[Link]

University of Adelaide Press is collaborating with JSTOR to digitize, preserve and extend
access to New Directions in Dental Anthropology

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
 3
Genes for teeth — drawing
inference from family data
Toby Hughes1,2,3 and Grant Townsend1,2,3

1
School of Dentristry, The University of Adelaide, South Australia, Australia
2
Centre for Orofacial Research and Learning, The University of Adelaide
3
International Collaborating Network in Oro-facial Genetics and Development

Abstract
Development of the human dentition, a complex, self-organising system, is underpinned by a series
of reiterative steps involving a number of key gene pathways, supplemented by smaller influences of
a polygenic background. Modelling familial data of dental phenotypes can help to unravel genetic
and environmental influences. This paper presents a review of a number of model-based approaches
that can be useful analytically, with a focus on twins as the familial structure to elaborate genetic
complexity. Genetic modelling is methodologically robust, and provides a framework within which
to locate evidence of gene effects from modern, high-throughput genotyping approaches. The twin
family structure is particularly well-suited to this approach, and provides a number of distinct
advantages analytically, particularly in the presence of population stratification.

Introduction
The human dentition is of significant anthropological interest when considering variation
within and between modern populations. It is also a useful tool for examining evolutionary
change over time in response to changes in culture, diet, etc. Teeth provide a (relatively) stable
indirect source of information about processes occurring during pre- and early post-natal
development. They are also one of the most stable sources of information in the fossil record,
both morphologically, and as a repository of ancient DNA sequence information (Adler et
al., 2010). Furthermore, variation in tooth form and function can provide opportunities for
examining inter-individual variation as a means for forensic identification. More recently,

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
 Genes for Teeth 23

Figure 1. A scale free network in which the larger hubs are highlighted.

evolutionary models of oral microbial ecology have relied upon extraction of microbial DNA
from deposits on tooth surfaces (Preus et al., 2011).
Dental development, a function of many interactions between a range of factors at
multiple levels, has many network-like features. Many biological networks display substantial
non-trivial topological features, with patterns of connection between their elements that are
neither purely regular nor purely random. Figure 1 illustrates one conceptualization of such a
complex biological system, the scale-free network. A network is named scale-free if its degree
distribution, i.e., the probability that a node selected uniformly at random has a certain
number of links (degree), follows a particular mathematical function called a power law. The
most notable characteristic in a scale-free network is the relative commonness of vertices
with a degree that greatly exceeds the average. The highest-degree nodes are often called
‘hubs’, and are thought to serve specific purposes in their networks. The human dentition
can be considered a complex system (Brook and Brook O’Donnell, 2011) in which lower-
level, interacting components give rise to higher level, emergent phenomena. The system
is comprised of a hierarchical organisation of functional subunits, including cells, tissues,
organs, and, if the concept is taken to its logical end point, organisms and populations.
This ‘self-adaptive’ system has the capacity to react to change through time, both within an
individual and trans-generationally.
Understanding the developmental processes that give rise to morphological variation in
the human dentition is a goal of a diverse range of disciplines, including physical anthropology,
evolutionary biology, comparative anatomy, forensic odontology and clinical dentistry. The
human dentition demonstrates significant variation in development, form and function. This

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
24 New Directions in Dental Anthropology

variation exists within and between individuals, families, sexes, ethnic groups and populations.
It has been attributed to temporal effects acting at the level of the individual (within a lifetime)
and the population (across generations). Such variation poses three fundamental questions for
dental anthropologists:

1. How does the plasticity of the genome give rise to population adaptation to a
particular environment?
2. How do genes and the environment interact to produce a specific phenotype?
3. How does our understanding of the interaction between genes and the environment
fit within the broader context of the dentition as a complex system?

The first two questions can be addressed using family studies; the former through use
of population modelling of traits that exhibit familial aggregation; the latter through the use
of linkage and association analyses to elucidate the role of specific genes in trait development.
More recently, the role of the epigenome in dental development and patterns of trait
transmission has become of interest to the dental anthropologist. This, too, can be addressed
through judicious use of family data. The final question requires a conceptual framework in
which the actions of the genome, the epigenome and the environment are linked through a
scale-free network of genes and gene products, in which some factors (hubs) have a greater
significance in terms of the final phenotype, whilst others (nodes) are bit players.
This review will illustrate how studies of families, and particularly twin families, can
be used to partition population variation into genetic and environmental components using
mathematical models of the twin relationship. It will explore how such models can reveal
information about relationships between dental features, and how these family models can
advantageously incorporate molecular marker data to identify genes of major influence. The
review will examine how studies of monozygotic twins can be used to examine the role of the
epigenome in dental development. Finally, the author will attempt to consolidate the role of
genes and environment in a broader conception of the dentition as a complex system.

Phenotypic variation
At maturity, the human dentition is a highly organised, dynamic system that has a major
role in the maintenance of homeostasis. A self-organising system with multiple symmetries
(left-right; maxillary-mandibular), the dentition arises from a complex series of interactions
occurring at multiple organisational levels, initiated in the first trimester pre-natally, and
not maturing fully until around 20 years of age. Evidence of reiterative systems operating
throughout dental development is further evidence that it is a truly complex, as distinct from
a complicated, system (Jernvall and Thesleff, 2000).
In recent decades, the model of phenotypic variation arising as a result of conflict
between genome and environment (‘nature’ vs ‘nurture’) has been supplanted by a more
complex model that encapsulates the complex interaction (and in many cases association)
between genes and the external organismal environment (‘nature’ via ‘nurture’). Despite

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
 Genes for Teeth 25

evidence of strong genetic regulation of both developmental timing (Hughes et al., 2007)
and morphological characteristics (Dempsey and Townsend, 2001; Hughes et al., 2000;
Townsend and Martin, 1992), there is still a great deal of unexplained phenotypic variation
in the human dentition, both within and between populations.
The role of the genome in dental development is similar to that of many human
conditions. A number of features are influenced by only a single or a few genes and show a
very simple pattern of inheritance. These are most commonly disease states, and may be the
result of specific allelic variants ‘tipping’ an individual over a phenotypic threshold in the
presence of a polygenic background (Brook, 1984).
Most features of interest are due to the additive effects of many genes and/or the
environment (classical heritability). They characteristically show a distinct distribution (most
commonly normal) within a population, and can be considered multifactorial. These features
provide the most challenge to elaborate aetiologically.
Some genes act on multiple dental phenotypes pleiotropically. These are commonly
homeobox-like genes, which regulate expression of structural genes, and often play a role
reiteratively during development. They can be considered the hubs in a gene-centric form of
a scale-free network, and will be discussed in more detail later.
Other effects may complicate the outcome, including allele interactions at the same locus
(e.g. genetic dominance), allele interactions between loci (e.g. epistasis) and the interaction of
genes with their environment (e.g. epigenetics).

Experimental and analytic techniques

The dental anthropologist has a range of analytic techniques available to explore the factors
that contribute to observed variation in the form and function of the human dentition. Many
of the most powerful rely on a detailed understanding of the complex interactions between
the genome and the environment during development, maturity and senescence. Quantitative
genetics, the application of intensive statistical models to deconstruct the observed phenotypic
information from a population, is one such tool. It works hand-in-glove with modern, high-
throughput molecular genetics approaches to provide dental anthropologists with a detailed
understanding of the role of functionally significant genes in the dentition of modern human
populations. It also allows researchers to draw inferences about gene flow through populations,
and hence to elaborate putative models of dental evolution.

Quantitative genetics
The birth of quantitative genetics, the mathematical description of population variation in
a genetic framework, is widely credited to a 1918 paper by then student, Ronald Fisher
(Fisher, 1918). Elaborated by the likes of Falconer (Falconer and Mackay, 1996), it has been
further developed to deal with the complexities associated with studying the human organism
— namely, the requirement to capitalise on naturally-occurring mating systems. A ‘top-
down’ approach to describing complexity, quantitative genetics seeks to apportion observed
phenotypic variation into that due to the additive and non-additive effects of genes, and that

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
26 New Directions in Dental Anthropology

due to the environment. More powerful still, are those approaches that seek to understand the
interaction between these two ‘latent’ sources of variation.
Quantitative genetics is complementary to the domain of the molecular biologist,
seeking to provide a framework of genetic variation within which can be located models of
specific gene action. Together, the two disciplines provide an opportunity to better understand
the interplay between components of complex systems. Quantitative genetics relies on the
development of theoretical models from a sound understanding of the biological system under
analysis. These models may then be validated using real-world data, often using likelihood-
based approaches. This requires collection of both intensive and extensive phenotypic data in
order to substantiate the conceptual model.
A key feature of quantitative genetic analysis in humans is a reliance on known or
inferred familial relationships. Knowledge of these relationships, and the transmission of
alleles via meiosis, enables the dental anthropologist to develop models of trait transmission
that predict the phenotypic outcome of genes segregating in families. The models are then
compared to the observed trait transmission in the same families to estimate goodness-of-fit.

How can twins help us understand dental development and morphology?

First documented in the scientific literature in Francis Galton’s 1875 paper, ‘History of
Twins’ (Galton, 1875), the twin relationship is a particularly powerful familial structure for
quantitative genetic analysis. Twin pairs can be either monozygous (MZ), or dizygous (DZ).
Monozygous twin pairs arise from the fertilisation of one ova by one sperm, followed by
an extra round of post-meiotic division resulting in two genetically identical zygotes that
subsequently implant and grow. Dizygous twin pairs arise from the concurrent release of two
ova, which are fertilised independently by two sperm, and subsequently implant and grow. As

Figure 2. Path diagram representing the twin relationship.

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
 Genes for Teeth 27

a consequence of the phenomena of crossing-over of non-sister chromatids prior to the first

meiotic division, and of genes assorting independently during the second meiotic division,
DZ twins only share, on average, 50 percent of their alleles identical by descent, the same
as non-twin siblings. The dental anthropologist can capitalise on these two distinct genetic
relationships to draw inferences about possible genetic aetiologies of dental development.
The twin relationship, as distinct from other genetic relationships, is useful for a number
of reasons. Twins are matched for age and, when reared together, share similar pre-natal and
early post-natal environments (placental arrangement can complicate this generalisation).
There are a number of manifestations of the twin relationship that appear within
models of familial aggregation. These include the ‘traditional’ twin model, which assumes
that differences between MZ twin pairs reared together reflect environmental effects unique
to individual twins within a pair, whereas differences between DZ twin pairs reared together
reflect both genetic and environmental differences, and is the primary focus of this discussion.
Other designs of note include twin adoption studies (twins reared apart), twins and their
families, the MZ half-sib design, DZ opposite-sex studies (useful for drawing inference about
sexual dimorphism), and the MZ co-twin design. This last model is particularly useful for
examining environmental and epigenetic influences on trait mean and variance, and will be
examined later.
Figure 2 presents a simple path diagram of a structural equation model (SEM)
representing the twin relationship for a single trait. Variation in the observed twin phenotypes
(squares) is influenced by a number of latent (unmeasured) variables (circles). Broadly-
speaking, these are the additive effects of an individual’s genes (A), the non-additive effects
(dominance, epistasis) of an individual’s genes (D), the influence of the environment shared
by co-twins (C) and the unique environment experienced by an individual twin (E). This last
also encapsulates experimental error.
Using likelihood-based approaches, the model completely decomposes observed
variation into a number of discrete linear relationships between latent and measured variables,
related by a series of parameters (a, d, c and e). The ‘structural’ elements of the model (intra-
pair correlations, r) capitalise on the observer’s knowledge of biology underpinning the
relationships between latent variables. To this end, additive genetic effects have a correlation
(r) of 1 in MZ twins, and 0.5 in DZ twins, and, unsurprisingly, the correlation between
shared environments is 1 regardless of zygosity — twin pairs experience the same shared
environment.
Given an observed covariance matrix from the raw data, parameter estimates for the
model are derived using a multinomial implementation of the likelihood function, maximising
the likelihood iteratively to produce a model that best-approximates the real-world data (with
judicious use of good starting values). Structural equation modelling software such as Mx
(Neale et al., 2006), now implemented in R (R Development Core Team, 2011) is ideal for
this purpose. Invariably, models of this nature fit well, being essentially a transformation
of the data. The focus then switches to whether simpler models may also be fit to the data
without a significant decrease in model fit. Simpler models can be compared to more complex
models using appropriate statistics or information criteria to reach the most parsimonious

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
28 New Directions in Dental Anthropology

explanation of the observed data (Neale and Cardon, 1992).

For twins reared together, even this simple model is underidentified (too many
parameters for too few observations), necessitating that the observer settle first on a model
incorporating either ADE or ACE — this can be assisted by observing intra-pair correlations
within zygosity in the first instance.
The univariate model presented above can also be extrapolated to the multivariate case
to answer more explicit questions regarding the structure of the data – do genetic effects
change through time; is there sexual heterogeneity for trait variance; are there pleiotropic
influences of individual genes. Figure 3 illustrates a multivariate model of the mesiodistal
size of the permanent dentition in a cohort of male European twins from earlier work by our
group (Hughes et al., 2007). Of note is the increased complexity of the covariance structure.
Variation in adult male incisor tooth size is best described by a model incorporating a single
general genetic effect on all teeth, as well as specific genetic factors acting on lower antimeres,
and lateral incisors respectively (illustrated for one half of the twin pair). Sexual heterogeneity
in variance necessitates a subtly different structure for female incisor tooth size. Figure 4
illustrates a longitudinal model of arch breadth in the same cohort. The simplex model for
longitudinal data allows for innovation elements at each time point (ζa, ζe) and directional
transmission elements between time points (βa, βe), as well as an estimate of experimental

Figure 3. A multivariate model of the mesiodistal dimension of incisors in adult male twins of European descent.

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
 Genes for Teeth 29

Figure 4. A general AE simplex model of arch shape through time.

error (ε, constrained equal across time) that is now independent of unique environmental
effects, Ei. Factor loadings (λ) are fixed at 1 for model identification. This model allows for
genetic elements acting at discrete time-points, as well as transmissible elements that account
for variation through time.

Molecular genetics
One of the appeals of the structural equation modelling approach is that it is flexible enough
to enable incorporation of molecular genetic data in order to establish the putative influence
of key genes. This is true regardless of whether one is using a genetic linkage-based approach
or a genetic association-based approach.
These methods emphasize the utility of using familial data for modelling gene action.
Linkage analysis, by definition, requires information on the co-transmission of traits and
genetic markers between family members, and hence relies on family-based approaches.
Dizygotic twins are one such group who may be utilised for linkage. Monozygotic twins,
on the other hand, are uninformative for linkage unless data is available from other family
members. Linkage can localise complex trait loci with 1-10 Mbp resolution, however
the locus effect size needs to be more than ten percent of the trait genetic variance to be
detectable. Quantitative trait loci detected by linkage can be considered the hubs of a complex
This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
30 New Directions in Dental Anthropology

system due to their large influence on trait variation. Because of the natural randomization
induced by segregation during meiosis, linkage is robust to confounding. Figure 5 illustrates
a path diagram incorporating a putative quantitative trait locus (QTL). The intra-pair QTL
correlation (pi hat) is an estimate (not all relationships are fully informative) of allele sharing
identical by descent between DZ twins (Martin et al., 1997). Linkage can be tested by
dropping Q from the model and examining the change in model fit — a significant decrease
in model fit is suggestive of linkage.
Association (candidate gene) analysis extracts information from the co-occurrence of
traits and markers within individuals. These approaches have traditionally utilised unrelated
case/control (or similar) population samples. A key liability with this type of cohort is that
underlying population stratification may result in spurious association. Familial structures
(and particularly twins), whilst generally more expensive to genotype, allow for family-based
approaches (within/between transmission disequilibrium testing) which are robust to the
presence of population stratification. The localization of complex trait loci using whole-
genome approaches is usually at the 0.01-0.1 Mbp resolution, provided the locus effect size is
more than one percent of the genetic variance. Such loci can be considered nodes within the
complex system framework. Association analysis is less robust to confounding than linkage
analysis.
The flexibility of the SEM approach allows both linkage and association to be modelled
simultaneously in familial data sets. Figure 6 illustrates a combined model of linkage and
association in sibling pairs for a phenotype for which molecular marker data are available,
allowing for possible population stratification. Latent variables for family resemblance F, QTL
variance Q, and individual-specific variance E cause the phenotypes of two siblings, P1 and
P2. S represents half the sum of the sibling pair’s genotypic effects, and D represents half

Figure 5. A simple univariate model of the twin relationship, incorporating a genetic marker at a specific locus in order to
test for linkage.

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
 Genes for Teeth 31

Figure 6. A twin model incorporating both linkage and association.

their difference. These components contribute to between (B) and within pair effects (W) via
parameters b and w respectively. In the absence of stratification, b and w are expected to be
equal. Genuine association with observed genotypes G1 and G2 will decrease the size of the
linkage-based QTL effect, q.
Using appropriate model specification, SEM can be extrapolated to modern, whole-
genome approaches, which, in the case of association, can identify causal variants (Vieira et
al., 2008). This enables the dental anthropologist to capitalise on comprehensive marker data
arising from high-throughput, chip-based approaches to ascertaining large numbers of markers
simultaneously. There is, however, a, concomitant increase in the numbers of statistical tests
required, necessitating consideration of the experiment-wise error rate. MERLIN (Abecasis

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
32 New Directions in Dental Anthropology

et al., 2002), a Multipoint Engine for Rapid Likelihood Inference, is a purpose-built piece of
software whose capabilities include linkage analysis (variance components, non-parametric
linkage, parametric linkage, clustered marker data), association, haplotyping, information
content, error detection (most SNP typing errors are Mendelian consistent) and simulation.

Epigenetics
In the last ten years, with the completion of the Human Genome Project (Collins et al.,
2003), the generation of progressive iterations of the Human Hapmap (International Hapmap
Consortium, 2005), and with the rapid publication of many large-scale, high-powered whole-
genome association studies of human phenotypes, focus has shifted from variation in the
genetic code per se, to how gene expression is modulated. There is a growing appreciation that
epigenetic factors can have a major influence on trait expression, and have been implicated in
changes over life course (Poulsen, 2007). In its broadest sense, epigenetics refers to differential
modification of gene effects, due to stochastic variation in the local genetic milieu. A more
narrow interpretation is the influence of (potentially heritable) changes in local chemical
mediators of gene transcription or translation (CpG methylation, histone deactylation,
X inactivation, etc.). Monozygous twins provide an ideal model for the role of epigenetic
factors in trait variance, and there are numerous publications that have provided evidence of
epigenetic discordance between MZ twins. Our group is currently investigating the influence
of methylation on MZ discordance for tooth number (missing/extra teeth).

The dentition as a complex system

The concept of the development of the dentition as a complex system has been dealt with
explicitly elsewhere in this book (Brook et al., 2012). The methodologies outlined in this
article provide the mathematical framework within which to test associations and interactions
between different subunits that contribute to dental variation at both the organismal and
population level. Genes, signalling pathways, epigenetic factors and environmental factors
may each play one or more roles as a node, hub, or module in a complex system. Elaborating
the interconnections between these subunits will enable the development of predictive models
of dental development, and will improve the capacity to fine tune phylogenetic relationship
data, to aid in modelling evolutionary change.

Future directions
To-date, our research group has examined a large range of morphological and oral-health
phenotypes from a series of three Australian twin cohorts, some of which are reported
elsewhere in this special issue. We aim to continue intense phenotyping of cohort three,
whilst further extracting data from records already available for cohorts one and two. Having
already reported extensive heritability estimates for a range of phenotypes, we now plan to
develop robust multivariate models of orofacial variation. High density genetic profiling
of the three twin cohorts is currently underway, supplemented by data from collaborators.
We aim to integrate molecular marker data with our current models to identify putative
This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
 Genes for Teeth 33

QTLs for further fine-mapping and identification of causal variants, with the ultimate aim of
developing predictive models of oral phenotypes. Epigenetic profiling of discordant MZ pairs
has become a recent focus of our research group, and we are currently investigating the role
of differential methylation in tooth number discordance. Finally, we will seek to replicate our
initial findings in other datasets.

Conclusion
Genetic modelling offers a methodologically robust approach for exploring the complexities
of dental development and evolution, and articulates with a conception of the dentition as
a complex system. It provides a framework within which to locate evidence of gene effects
from high-throughput genotyping. It capitalises on familial structure, of which twins provide
a number of distinct advantages analytically.

Acknowledgements
The authors would like to acknowledge all twin participants and their families, funding bodies
(NHMRC, ADRF, the Financial Markets Foundation for Children and Colgate Australia),
and collaborators.

References
Abecasis G.R., Cherny S.S., Cookson W.O., and Cardon L.R. (2002) Merlin — Rapid
analysis of dense genetic maps using sparse gene flow trees. Nature Genetics, 30: 97-
101.
Adler C.J., Haak W., Donlon D., and Cooper A. (2010) Survival and recovery of DNA from
ancient teeth and bones. Journal of Archaeological Science, 38: 956-964.
Brook A.H. (1984) A unifying aetiological explanation for anomalies of tooth number and
size. Archives of Oral Biology, 29: 373-378.
Brook A.H., and Brook O’Donnell M. (2012) Modelling the complexity of the dentition.
In: New Directions in Dental Anthropology: paradigms, methodologies and outcomes.
Townsend G., Kanazawa E., Takyama H. (Eds.)  University of Adelaide Press, Adelaide.
Collins F.S., Morgan M., and Patrinos A. (2003) The Human Genome Project: lessons from
large-scale biology. Science, Apr 11 2003: 286
Dempsey P.J., and Townsend G.C. (2001) Genetic and environmental contributions to
variation in human tooth size. Heredity, 86: 685-693.
Falconer D.S., and Mackay T.F.C. (1996) Introduction to quantitative genetics. 4th ed. Essex,
England: Addison Wesley Longman Limited.
Fisher R.A. (1918) The correlation between relatives on the supposition of Mendelian
inheritance. Philosophical Transactions of the Royal Society of Edinburgh, 52: 399-
433.
Galton F. (1875) The history of twins, as a criterion of the relative powers of nature and
nurture. Fraser’s Magazine, 12: 566-576.
Hughes T.E., Bockman M.R., Seow K., Gully N., Richards L.C., and Townsend G.C. (2007)
This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]
34 New Directions in Dental Anthropology

Strong genetic control of emergence of human primary incisors. Journal of Dental

Research, 86: 1160-1165.
Hughes T., Dempsey P., Richards L., and Townsend G. (2000) Genetic analysis of deciduous
tooth size in Australian twins. Archives of Oral Biology, 45: 997-1004.
Jernvall J., and Thesleff I. (2000) Reiterative signalling and patterning in mammalian tooth
morphogenesis. Mechanisms of Development, 92: 19-29.
Martin N., Boomsma D., and Machin G. (1997) A twin-pronged attack on complex traits.
Nature Genetics, 17: 387-392.
Neale M.C., Boker S.M., Xie G., and Maes H.H. (2006) Mx: statistical modelling. 7th rev.
ed. Richmond, VA: Department of Psychiatry, Virginia Commonwealth University.
Neale M.C., and Cardon L.R. (1992) Methodology for genetic studies of twins and their
families. Dordrecht, The Netherlands: Kluwer Academic Publishers.
Poulsen P., Esteller M., Vaag A., and Fraga M.G. (2007) The epigenetic basis of twin
discordance in age-related diseases. Pediatric Research, 61: 38R-42R.
Preus H.R., Marvik O.J., Selvig K.A., and Bennike P. (2011) Ancient bacterial DNA (aDNA)
in dental calculus from archaeological human remains. Journal of Archaeological
Science, 38: 1827-1831.
R Development Core Team. (2011) R: A language and environment for statistical computing.
Vienna, Austria: R Foundation for Statistical Computing.
The International Hapmap Consortium. (2005) A haplotype map of the human genome.
Nature, 437: 1299-1320.
Townsend G.C., and Martin N.G. (1992) Fitting genetic models to Carabelli trait data in
South Australian twins. Journal of Dental Research, 71: 403-409.
Vieira A.R., Marazita M.L., and Goldstein-McHenry T. (2008) Genome wide scan finds
suggestive caries loci. Journal of Dental Research, 87: 435-439.

Correspondence
Doctor Toby Hughes
School of Dentistry, The University of Adelaide
Adelaide, South Australia, 5005
Email: [Link]@[Link]

This content downloaded from [Link] on Wed, 04 Apr 2018 [Link] UTC
All use subject to [Link]