Download "An Evolutionary Framework for Common Disease".

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Hardy–Weinberg principle wikipedia , lookup

Designer baby wikipedia , lookup

Frameshift mutation wikipedia , lookup

Epistasis wikipedia , lookup

Behavioural genetics wikipedia , lookup

Tay–Sachs disease wikipedia , lookup

Fetal origins hypothesis wikipedia , lookup

Quantitative trait locus wikipedia , lookup

Dominance (genetics) wikipedia , lookup

Polymorphism (biology) wikipedia , lookup

Nutriepigenomics wikipedia , lookup

Genome (book) wikipedia , lookup

Genetic drift wikipedia , lookup

Neuronal ceroid lipofuscinosis wikipedia , lookup

Human genetic variation wikipedia , lookup

Epigenetics of neurodegenerative diseases wikipedia , lookup

Genome-wide association study wikipedia , lookup

Population genetics wikipedia , lookup

Medical genetics wikipedia , lookup

Microevolution wikipedia , lookup

Public health genomics wikipedia , lookup

Transcript
An Evolutionary
Framework for Common
Disease
Francesca Luca, University of Chicago, Chicago, Illinois, USA
Anna Di Rienzo, University of Chicago, Chicago, Illinois, USA
Advanced article
Article Contents
. Introduction
. Evolutionary Models of Common Diseases
. The Empirical Evidence about Susceptibility Alleles for
Common Diseases
. Conclusions
doi: 10.1002/9780470015902.a0020758
Susceptibility alleles for common diseases are a subset of all natural genetic variation; as
such they are jointly shaped by random drift and natural selection. Computer simulation
and empirical studies show that models of purifying selection and adaptive evolution
apply to risk alleles for common diseases. These models have important implications for
the design of disease-mapping studies.
Introduction
A major challenge in contemporary human genetics is to
dissect the genetic basis of human diseases and other phenotypes. While great success has been achieved in mapping
genes for Mendelian diseases, the search for the genetic
susceptibility to common diseases with complex patterns of
inheritance (e.g. asthma, hypertension, type 2 diabetes) has
met with considerable difficulties. Mendelian diseases tend
to be rare traits with simple patterns of transmission and
are usually due to highly penetrant and deleterious alleles
that segregate in families. They lend themselves to mapping
by linkage analysis, an approach that requires that markers
flanking the disease gene segregate with the disease phenotype in families. Importantly, the co-segregation of
marker alleles and disease phenotype is not affected by the
heterogeneity of disease mutations at a given locus.
Although genome-wide linkage scans have been carried
out for many common diseases, such studies often produced weak and inconsistent signals and the few genes
identified by this approach account for a small amount of
the total heritability of the disease. There are many reasons
for this difference in the success of linkage analysis for
Mendelian compared to common disease. Perhaps the
most important factor is that common disease susceptibility is the result of many genetic and environmental factors
and of the epistatic interactions between them. Therefore,
any given allele is expected to have only moderate to small
effects on disease risk so that linkage analysis does not
provide sufficient statistical power.
An alternative to linkage analysis for mapping common
disease susceptibility is association, which is based on
detecting a difference in allele or genotype frequencies between cases and controls. Disease association studies may
be performed on candidate genes or on a genome-wide
scale; in the latter, no prior assumptions are made about the
physiological defects underlying a disease and about the
genomic location of the risk variants. The recent development of high-density single nucleotide polymorphism
(SNP) genotyping platforms has allowed the application
of genome-wide association (GWA) studies to a range of
common diseases and has revealed many signals that are
consistently replicated across studies. Association studies
are particularly suitable for the case in which the disease
variants are common in the population and the allelic
heterogeneity at disease loci is low. This is because for intermediate frequency variants there is greater statistical power
to detect an excess of susceptibility alleles in cases compared
to controls. Unlike linkage analysis, association studies are
hampered by high allelic heterogeneity because each susceptibility variant contributes to disease risk only in a small
fraction of patients. Because of the crucial requirement that
disease variants be common in the population, association
studies provide little information on disease susceptibility
variants that are rare in the population.
In considering the properties of disease susceptibility
alleles (e.g. frequency spectrum, allelic heterogeneity, interethnic differentiation, etc.), it is important to recognize that
these alleles, like all natural genetic variation, are the outcome of the evolutionary process. This means that disease
alleles are influenced jointly by the history and strength of
the selective pressures acting in human populations as well
as by the demographic history, including population size
changes and geographic structure. Because the power of
different disease-mapping approaches depends on the
properties of disease alleles, insights into the evolutionary
models that apply to common diseases are of interest for
the development and optimization of disease-mapping
strategies. This motivation has stimulated a number of
modelling studies by computer simulations as well as
empirical population genetics studies.
Evolutionary Models of Common
Diseases
Mutation–selection balance
With regard to oligogenic Mendelian diseases, relatively
simple models of mutation–selection balance are likely to
ENCYCLOPEDIA OF LIFE SCIENCES & 2007, John Wiley & Sons, Ltd. www.els.net
1
An Evolutionary Framework for Common Disease
provide a reasonably good explanation for many disorders.
Under these models, disease alleles have strong effects on
fitness and are, therefore, quickly removed by purifying
selection. The mutation process continually introduces
new disease alleles into the population ultimately leading to
a balance between mutation and selection. This model
predicts high allelic heterogeneity and low disease prevalence, which are both observed features of Mendelian
diseases. Although negative selection is likely to be the
main evolutionary paradigm for alleles responsible for rare
Mendelian diseases, it is likely that a small number of
Mendelian conditions result from the action of balancing
selection due to the higher fitness of heterozygous disease
carriers. In addition to sickle cell anaemia, similar balancing selection scenarios have been proposed for several
Mendelian diseases characterized by relatively high prevalence, e.g. cystic fibrosis, but the evidence in these cases is
less clear. See also: Mutation–Selection Balance
The mutation–selection balance framework has also
been investigated in the context of common diseases. These
analyses were based on the assumption that, although most
common diseases, e.g. type 2 diabetes or hypertension,
have post-reproductive age of onset, they may still have
small effects on fitness due to the differential survival or
reproduction of the affected individuals themselves or of
their offspring. Pritchard (2001) performed computer
simulations of an evolutionary model for a disease susceptibility locus in a randomly mating population of constant
size incorporating the effects of mutation, genetic drift and
selection (Pritchard, 2001). An important aspect of this
analysis is that the total frequency of the susceptibility allele class was treated as an outcome of the evolutionary
process and was not fixed at a specified value. Simulations
were run under the assumption that susceptibility alleles
were either evolutionarily neutral or slightly deleterious.
Under neutrality and for a broad range of mutation rates,
susceptibility alleles tended to be either rare or near fixation. A consequence of this skewed frequency spectrum is
that these alleles are expected to make a minor contribution
to the total genetic variance for the disease. The results of
the neutral simulations were compared to those in which
susceptibility alleles were assumed to be slightly deleterious
and the mutation rate was relatively high. In this case, the
probability that the susceptibility class had total frequency
between 1 and 99% was much higher than in the neutral
simulations. A possible interpretation of these results is
that common diseases are unlikely to be due to selectively
neutral susceptibility alleles.
While Pritchard (2001) considered only the simplest
model of population history, i.e. a constant size and randomly mating population, Reich and Lander (2001) investigated the mutation–selection balance framework under
more complex, possibly more realistic, demographic
models. In this case, however, the frequency of the disease
susceptibility allele class at a given locus was not treated as
an outcome of the evolutionary process, rather it was fixed
at a specified level (20%). Assuming that susceptibility alleles are slightly deleterious, it was shown that, for a ‘typical’
2
disease mutation rate, there is a single predominant allele
within the susceptibility class (Reich and Lander, 2001). In
the case of an exponentially growing population, disease
alleles tend to be more heterogeneous. However, allelic
heterogeneity increases more quickly if the frequency of the
susceptibility class before the onset of growth is low, while at
disease loci with intermediate frequencies of susceptibility
alleles the increase in heterogeneity is much slower. Because
the onset of population growth in humans is likely to have
been relatively recent, the increase in allelic heterogeneity
for loci with intermediate equilibrium frequency is expected
to be negligible. Hence, if indeed common diseases are due
to susceptibility loci with intermediate equilibrium frequencies (e.g. 20%), the simulation study of Reich and Lander
predicts that the allelic heterogeneity is relatively low.
These two modelling studies differed in several regards,
including the parameterization of the population genetics
model and the choice of values for several crucial parameters. For example, an important difference between the
two studies concerns the assumptions about the rate at which
new susceptibility alleles are created by the mutation process.
While Pritchard (2001) considered a broad range of
mutation rates (i.e. 2.5 1026–1.25 1024), Reich and
Lander (2001) considered a single and low mutation rate
(i.e. 3.2 1026). Clearly, the rate at which susceptibility alleles are generated and the rate at which they are repaired or
compensated for depend on the functional effect of each
mutation, which in turn depends on its physical location
(e.g. coding versus regulatory sequences) and a variety of
unknown factors including the disease pathophysiology and
the biological function of the gene. For example, the mutation rate for variants leading to loss of function is likely to be
higher than the rate for variants resulting in gain of function.
An important feature of the study by Reich and Lander
(2001) is the comparison between the expected frequency
spectra of disease alleles under different assumptions about
population history. Clearly, human populations have complex histories and demography may differ across major
ethnic groups with important implications for modelling
the evolution of susceptibility alleles. Several lines of evidence point to a history of rapid recent growth from an
equilibrium population for populations of Sub-Saharan
African ancestry while genetic variation data from
non-African populations are compatible with bottleneck
models and, possibly, serial founder events. In addition to
population expansion, human variation data are also consistent with significant geographic structure on a continental scale as well as on a finer scale. It is currently unknown if
more complex scenarios of human population history (with
selective pressures being the same) would make different
predictions about allelic heterogeneity and frequency spectrum of susceptibility alleles.
Spatial and temporal variation in selective
pressures
The mutation–selection balance models discussed so far
envision selection regimes in which selective pressures do
ENCYCLOPEDIA OF LIFE SCIENCES & 2007, John Wiley & Sons, Ltd. www.els.net
An Evolutionary Framework for Common Disease
not change over time. However, the environment in which
we now live is radically different from the one that ancestral
human populations adapted to. A number of transitions
during human evolution have resulted in shifts in the
selective pressures acting on the biological processes that,
in contemporary populations, underlie major classes of
common diseases. Such transitions include the dispersal of
modern humans out of Africa, the end of the Last Glacial
Maximum, the shift away from a hunting and gathering
lifestyle to a subsistence based on agriculture and animal
husbandry, and the more recent transitions associated with
industrialization and globalization.
The impact of the selective shifts resulting from some of
these transitions on disease susceptibility genes has been
formalized in a series of hypotheses, among which the
most popular is the ‘thrifty genotype’ hypothesis for type 2
diabetes and obesity. This hypothesis posits that since
ancestral hunter-gatherer populations underwent seasonal
cycles of feast and famine, they would have benefited from
being ‘thrifty’, i.e. having extremely efficient fat and carbohydrate storage (Neel, 1962). With the transition away
from a hunting and gathering subsistence, this ‘thriftiness’
became detrimental because of more reliable food availability and of changes in dietary patterns. Similarly, it was
recently proposed that adaptations in energy metabolism,
glucose homeostasis and thermogenesis may account for
the low prevalence of type 2 diabetes in Europeans
(Fridlyand and Philipson, 2006). According to this hypothesis, adaptations to cold climates would favour alleles
leading to increased thermogenesis and lower risk to type 2
diabetes. A related framework, sometime referred to as the
‘sodium retention’ hypothesis, has been proposed to explain the inter-ethnic differences in the prevalence of hypertension (Gleibermann, 1973). Under this scenario,
ancestral populations living in equatorial Africa adapted
to hot and humid climates by increasing the rate of sodium/
water retention; as populations moved out of Africa into
cooler, drier environments this high level of sodium retention might have been deleterious. In addition to energy
metabolism and sodium homeostasis, there are other biological processes that are associated with currently common diseases and that are likely to have experienced major
selective shifts. Parasite loads and pathogen exposure are
likely to have increased dramatically with the switch to a
sedentary lifestyle, higher population densities and greater
proximity with domesticated animals brought about by the
transition to agriculture and animal husbandry. The reduced exposure to childhood infections and the overall
higher sanitary conditions typical of industrial societies are
thought to be responsible for the rapid increase in inflammatory and allergic diseases, such as, for example, asthma
and inflammatory bowel disease. See also: Thrifty Gene
Hypothesis: Challenges; Thrifty Genotype Hypothesis and
Complex Genetic Disease
A common feature of the earlier scenarios is that the
observed increase in disease prevalence associated with
the environmental shift is too rapid to be due to a change in
the gene pool. Rather, it is likely that the transition simply
uncovers the latent genetic susceptibility of human populations to those diseases that account for a disproportionate fraction of the public health burden. A possible
evolutionary framework for this class of common diseases
is that susceptibility alleles are ancestral adaptations to the
lifestyle, environment and diet of ancient human populations. Therefore, unlike most alleles causing Mendelian
diseases, ancestral alleles would increase risk to common
diseases. In the ancestral environment, these alleles were
maintained by purifying selection while the derived alleles
were slightly deleterious (Di Rienzo and Hudson, 2005).
Upon the environmental transition and the concomitant
shift in selective pressures, the ancestral alleles would
increase risk to disease while the derived alleles would
become either neutral or advantageous.
The Empirical Evidence about
Susceptibility Alleles for Common
Diseases
The earlier models make different predictions for the
frequency spectrum and heterogeneity of disease alleles.
These predictions can now be compared to a rapidly
growing body of empirical evidence generated by diseasemapping efforts.
Common alleles
A systematic re-examination and meta-analysis of 25 reported disease associations with common variants revealed
a significant excess of replications of the original reports
suggesting that some of these associations were real. This
constituted early evidence in support of common susceptibility variants with modest effects on disease risk (odds
ratios between 1.07 and 2.28) (Lohmueller et al., 2003).
The recent development of high-density platforms
for SNP genotyping coupled with the collection of large
patients cohorts has enabled disease association studies on
a genome-wide scale. These studies have already yielded a
number of disease alleles strongly supported by the primary and replication studies, indicating that common variants make an important contribution to common diseases.
The average minor allele frequency of disease susceptibility
variants identified through genome-wide association studies is approximately 30%, probably reflecting the high
power of association studies to detect intermediate frequency variants, and this average value varies little across
diseases (e.g. breast cancer, Crohn disease, type 2 diabetes,
type 1 diabetes, prostate cancer) (Easton et al., 2007;
Hunter et al., 2007; Rioux et al., 2007). However, some
preliminary patterns are beginning to emerge suggesting
different spectra of disease allele frequencies across diseases. For example, very few risk variants for type 2 diabetes and type 1 diabetes have minor allele frequency lower
than 10% in Europeans (i.e. the ethnic group in which the
studies were performed). Conversely, a substantial
ENCYCLOPEDIA OF LIFE SCIENCES & 2007, John Wiley & Sons, Ltd. www.els.net
3
An Evolutionary Framework for Common Disease
proportion of variants associated with breast cancer,
prostate cancer and Crohn disease in genome-wide studies have minor allele frequency lower than 10%. Whether
the frequency spectrum of susceptibility alleles truly varies
across diseases cannot be determined based on the limited
preliminary results of genome wide association studies.
However, if these differences are confirmed by larger studies, they may suggest different evolutionary histories for
different diseases.
Rare and heterogeneous alleles
Because association studies are adequately powered to
detect only relatively common disease variants, they
provide only limited information regarding the relative
contributions of common and rare variants to the genetic
susceptibility to common diseases.
A different strategy has been developed to test for
heterogeneous and rare susceptibility variants. This strategy relies on performing variation discovery in either
affected and unaffected subjects or in the extremes of a
distribution (e.g. top and bottom 5%) for a disease-related
phenotype, followed by testing for an excess of variants
with predicted functional effects (e.g. nonsynonymous
variants) in one group compared to the other.
This approach was successfully applied to genes influencing plasma lipid levels (Cohen et al., 2004), which are
risk factors for cardiovascular disease, and to candidate
genes for colorectal adenomas and colorectal cancer
(Fearnhead et al., 2004). Interestingly, in the case of the
ANGPTL4 gene which influences variation in triglyceride
levels, a significant excess of nonsynonymous variants
was observed in European Americans, but not in African
Americans (Romeo et al., 2007). This observation suggests
that the selective pressures acting on ANGPTL4 variation
differ across ethnic groups. Moreover, as the excess of
ANGPTL4 missense variants is likely to reflect slightly
deleterious mutations, this implies that the models of spatially or temporally varying selection discussed earlier may
apply within the context of both negative and positive
selection scenarios.
Because the mutation–selection balance model predicts
that disease alleles are relatively rare in the population, a
large number of alleles is necessary to explain the genetic
susceptibility to diseases, such as type 2 diabetes and
hypertension, which are common in human populations.
A recent analysis of coding sequence data for mutations
causing Mendelian diseases, fixed substitutions between
human and chimpanzee, and polymorphism surveys aimed
to address exactly this question (Kryukov et al., 2007).
Using a context-dependent mutation model, it was estimated that, under neutrality, the expected ratio of missense
to nonsense substitutions among de novo mutations is
approximately 19.7. The observed ratio for Mendelian disease mutations, which are likely to be strongly detrimental,
is 3.9. This led to the conclusion that 20% (i.e. 3.9 of 19.7)
of missense mutations are also strongly detrimental.
A similar result was obtained when splice site mutations
4
were considered instead of nonsense mutations. To
estimate the fraction of de novo missense mutations that
are neutral, the ratio of missense to synonymous mutations
expected under neutrality was compared to that observed
in human–chimpanzee sequence divergence (2.23 and 0.60,
respectively), leading to the estimate that 27% (i.e. 0.6 of
2.23) of missense mutations are functionally similar to
synonymous mutations and, hence, likely to be neutral.
Therefore, it was estimated that the proportion of detrimental, slightly deleterious variants, namely those that are
neither strongly detrimental nor neutral, is 53%. Analysis
of human polymorphism data for coding regions led to the
estimate that mildly deleterious missense mutations occur
at a frequency lower than 1% in human populations.
Therefore, the fraction of missense variants estimated to
be evolutionarily deleterious might be large enough to
account for the genetic susceptibility to common diseases
even though these variants are rare.
Ancestral susceptibility alleles and spatially
varying selection
Consistent with the idea that common disease susceptibility may be the result of ancient human adaptations to
a long-term steady environment, a number of alleles that
increase risk to common diseases are indeed ancestral (Di
Rienzo and Hudson, 2005). In this framework, the new
(derived) allele at these variants confers protection against
the disease. The results of GWA studies have largely confirmed this observation made initially on the basis of mapping studies that did not use similarly systematic
approaches. In addition, because many GWA studies use
the same SNP genotyping platforms, they allow comparing
the proportion of ancestral susceptibility alleles across
diseases. For example, an interesting pattern emerging
from GWA studies is that the proportion of risk alleles that
are ancestral is greater for type 2 and type 1 diabetes compared to most other diseases for which GWA data are
available and for which a substantial number of disease
alleles have been identified. Further disease association
studies will need to be performed to determine the range of
disease to which the ancestral susceptibility model applies.
In addition to susceptibility alleles for common diseases,
some mutations responsible for diseases with simpler patterns of inheritance were found to be identical to the allele
found at the orthologous position in chimpanzee (Azevedo
et al., 2006), macaque (Gibbs et al., 2007) or mouse (Gao
and Zhang, 2003). This apparent paradox may be
explained by the ‘compensatory mutation hypothesis’,
which postulates that the absence of strongly deleterious
effects in the other organisms is due to the buffering effect of
epistatic interactions between the human disease mutation
and alleles at linked sites that compensate for the functional
alteration. Additional, nonmutually exclusive explanations are also possible; for example, the basic metabolic
machinery of humans may be significantly different from
that of the species considered so far and/or it is possible that
some human diseases (particularly the late onset ones) do
ENCYCLOPEDIA OF LIFE SCIENCES & 2007, John Wiley & Sons, Ltd. www.els.net
An Evolutionary Framework for Common Disease
not manifest in other species as a result of their shorter
lifespan.
Interestingly, for some common diseases (i.e. heart
disease, diabetes), a signature of recent positive selection
has been reported for several derived alleles that have protective effects; this observation is somewhat surprising because, given the late age of onset of these diseases,
protective alleles are expected to have little or no fitness
consequences. A particularly intriguing case was described
at the TCF7L2 gene, which contains an ancestral allele
strongly associated with increased risk to type 2 diabetes.
The derived protective allele at this variant is in strong
linkage disequilibrium with the derived allele at another
variant within the same gene; together these alleles define a
haplotype that carries a signature of positive natural selection. This haplotype was shown to be associated with
variation in body mass and in levels of hunger-satiety hormones, namely ghrelin and leptin. It was proposed that the
nature of the selective advantage is related to these phenotypes rather than type 2 diabetes itself (Helgason et al.,
2007).
An additional scenario to explain the selective advantage
associated with alleles that protect against disease phenotypes is that they also protect against phenotypes with a
more direct effect on reproductive function. For example,
the ancestral Thr235 allele in the angiotensinogen gene was
reported to increase risk to pre-eclampsia in addition to
hypertension; accordingly, the haplotype carrying the
derived (protective) allele exhibits a haplotype structure
consistent with the action of recent positive selection.
Moreover, as predicted by the sodium retention hypothesis, the worldwide distribution of allele frequency at the
AGT Thr235Met variant as well as other variants influencing risk to hypertension is consistent with the action of
spatially varying selection: the protective alleles decrease
gradually in frequency with increasing distance from the
equator (Thompson et al., 2004; Young et al., 2005).
These findings open the appealing prospect that
approaches based on detecting the signature of selection
may be added to the arsenal of disease-mapping strategies.
Because population genetics studies and disease association studies are based on independent theoretical frameworks, evidence converging on an overlapping set of
candidate susceptibility variants may greatly increase the
confidence that the true causative variation has been
identified. A major advantage of population genetics
studies over disease-mapping studies is that they can be
performed on random samples of unrelated individuals
and do not require the costly and time-consuming collection of extensive phenotypic data. For example, to the extent that susceptibility to common diseases may be related
to phenotypes for which selection varies according to
measurable aspects of geography – e.g. sodium homeostasis or energy metabolism – mapping susceptibility alleles could be conducted by genotyping individuals from
geographically diverse populations and testing for correlations between population allele frequencies and environmental variables.
Conclusions
Although the development of evolutionary models for
common diseases is still in its infancy, it is already clear that
no single model can explain the evolution of susceptibility
alleles even for the same disease or the same susceptibility
gene. For example, analyses aimed at detecting rare susceptibility variants for a given phenotype showed that there
is strong evidence for such variants in some candidate
genes, but not in others (Ahituv et al., 2007; Cohen et al.,
2004; Fearnhead et al., 2004). Moreover, both common
and rare variants have been shown to contribute to the
susceptibility to the same common disease. This may apply
even to variants within the same gene; for example, the
ABCA1 gene, which codes for a cholesterol efflux pump,
was shown to harbour multiple rare variants as well as a
common variant (Ile883Met) that influence plasma high
density lipoprotein (HDL) cholesterol levels (Cohen et al.,
2004). Given the complex history of selective pressures
acting on humans and the composite pathways underlying
the pathophysiology of common diseases, it is not surprising that a diverse spectrum of plausible evolutionary
models is emerging. Modelling studies rooted in population genetics theory and data are likely to increase our
understanding of the evolutionary causes of common
diseases.
References
Ahituv N, Kavaslar N, Schackwitz W et al. (2007) Medical
sequencing at the extremes of human body mass. American
Journal of Human Genetics 80: 779–791.
Azevedo L, Suriano G, Van Asch B, Harding RM and Amorim A
(2006) Epistatic interactions: how strong in disease and evolution? Trends in Genetics 22: 581–585.
Cohen JC, Kiss RS, Pertsemlidis A et al. (2004) Multiple rare
alleles contribute to low plasma levels of HDL cholesterol. Science 305: 869–872.
Di Rienzo A and Hudson RR (2005) An evolutionary framework
for common diseases: the ancestral-susceptibility model. Trends
in Genetics 21: 596–601.
Easton DF, Pooley KA, Dunning AM et al. (2007) Genome-wide
association study identifies novel breast cancer susceptibility
loci. Nature 447: 1087–1093.
Fearnhead NS, Wilding JL, Winney B et al. (2004) Multiple
rare variants in different genes account for multifactorial inherited susceptibility to colorectal adenomas. Proceedings of the
National Academy of Sciences of the USA 101: 15992–15997.
Fridlyand LE and Philipson LH (2006) Cold climate genes and the
prevalence of type 2 diabetes mellitus. Medical Hypotheses 67:
1034–1041.
Gao L and Zhang J (2003) Why are some human diseaseassociated mutations fixed in mice? Trends in Genetics 19:
678–681.
Gibbs RA, Rogers J, Katze MG et al. (2007) Evolutionary and
biomedical insights from the rhesus macaque genome. Science
316: 222–234.
Gleibermann L (1973) Blood pressure and dietary salt in human
populations. Ecology of Food and Nutrition 2: 143–156.
ENCYCLOPEDIA OF LIFE SCIENCES & 2007, John Wiley & Sons, Ltd. www.els.net
5
An Evolutionary Framework for Common Disease
Helgason A, Palsson S, Thorleifsson G et al. (2007) Refining
the impact of TCF7L2 gene variants on type 2 diabetes and
adaptive evolution. Nature Genetics 39: 218–225.
Hunter DJ, Kraft P, Jacobs KB et al. (2007) A genome-wide
association study identifies alleles in FGFR2 associated with
risk of sporadic postmenopausal breast cancer. Nature Genetics
39: 870–874.
Kryukov GV, Pennacchio LA and Sunyaev SR (2007) Most rare
missense alleles are deleterious in humans: implications for
complex disease and association studies. American Journal of
Human Genetics 80: 727–739.
Lohmueller KE, Pearce CL, Pike M, Lander ES and Hirschhorn
JN (2003) Meta-analysis of genetic association studies supports
a contribution of common variants to susceptibility to common
disease. Nature Genetics 33: 177–182.
Neel JV (1962) Diabetes mellitus: a ‘thrifty’ genotype rendered
detrimental by ‘progress’? American Journal of Human Genetics
14: 353–362.
Pritchard JK (2001) Are rare variants responsible for susceptibility to complex diseases? American Journal of Human Genetics
69: 124–137.
Reich DE and Lander ES (2001) On the allelic spectrum of human
disease. Trends in Genetics 17: 502–510.
Rioux JD, Xavier RJ, Taylor KD et al. (2007) Genome-wide
association study identifies new susceptibility loci for Crohn
disease and implicates autophagy in disease pathogenesis.
Nature Genetics 39: 596–604.
Romeo S, Pennacchio LA, Fu Y et al. (2007) Population-based
resequencing of ANGPTL4 uncovers variations that reduce
triglycerides and increase HDL. Nature Genetics 39: 513–516.
6
Thompson EE, Kuttab-Boulos H, Witonsky D et al. (2004)
CYP3A variation and the evolution of salt-sensitivity variants.
American Journal of Human Genetics 75: 1059–1069.
Young JH, Chang YP, Kim JD et al. (2005) Differential
susceptibility to hypertension is due to selection during the
out-of-Africa expansion. Public Library of Science Genetics
1: e82.
Further Reading
Jobling MA, Hurles ME and Tyler-Smith C (2004) Health
implications of our evolutionary heritage. In: Human Evolutionary Genetics, pp. 439–471. New York/Abingdon: Garland
Science.
Ramachandran S, Deshpande O, Roseman CC et al. (2005) Support from the relationship of genetic and geographic distance in
human populations for a serial founder effect originating in
Africa. Proceeding of the National Academy of Sciences of the
USA 102: 15942–15947.
Sabeti P (2006) Positive natural selection in the human lineage.
Science 312: 1614–1620.
The Chimpanzee Sequencing and Analysis Consortium (2005)
Initial sequence of the chimpanzee genome and comparison
with the human genome. Nature 437: 69–87.
The Wellcome Trust Case Control Consortium (2007) Genomewide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 447: 661–678.
ENCYCLOPEDIA OF LIFE SCIENCES & 2007, John Wiley & Sons, Ltd. www.els.net