GG15CH04-Ponting ARI 15 July 2014 11:5

ANNUAL
REVIEWS Further No Gene in the Genome Makes
Click here for quick links to
Annual Reviews content online,
including:
Sense Except in the Light
• Other articles in this volume
of Evolution
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

• Top cited articles

Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

• Top downloaded articles

• Our comprehensive search
Wilfried Haerty and Chris P. Ponting
MRC Functional Genomics Unit, Department of Physiology, Anatomy, and Genetics,
University of Oxford, Oxford OX1 3PT, United Kingdom;
email: wilfried.haerty@dpag.ox.ac.uk, chris.ponting@dpag.ox.ac.uk

Annu. Rev. Genomics Hum. Genet. 2014. Keywords

15:71–92
selection, neutral evolution, noncoding, regulatory element, molecular
First published online as a Review in Advance on
April 24, 2014 function
The Annual Review of Genomics and Human Genetics Abstract
is online at genom.annualreviews.org
Evolutionary conservation has been an accurate predictor of functional el-
This article’s doi:
10.1146/annurev-genom-090413-025621 ements across the first decade of metazoan genomics. More recently, there
has been a move to define functional elements instead from biochemical an-
Copyright  c 2014 by Annual Reviews.
All rights reserved notations. Evolutionary methods are, however, more comprehensive than
biochemical approaches can be and can assess quantitatively, especially for
subtle effects, how biologically important—how injurious after mutation—
different types of elements are. Evolutionary methods are thus critical for
understanding the large fraction (up to 10%) of the human genome that
does not encode proteins and yet might convey function. These methods
can also capture the ephemeral nature of much noncoding functional se-
quence, with large numbers of functional elements having been gained and
lost rapidly along each mammalian lineage. Here, we review how different
strengths of purifying selection have impacted on protein-coding and non-
protein-coding loci and on transcription factor binding sites in mammalian
and fruit fly genomes.

71
 GG15CH04-Ponting ARI 15 July 2014 11:5

1. INTRODUCTION
Nothing makes sense in biology except in the light of evolution.

—Theodosius Dobzhansky (25)

Functional regions—coding exons, regulatory elements, and others—are sparsely distributed in
mammalian genomes, scattered widely in a sea of apparently inert sequence. We understand much
about the functions of protein-coding genes but little about the molecular mechanisms, locations,
and properties of mammalian non-protein-coding functional elements. The approach adopted
by the Encyclopedia of DNA Elements (ENCODE) project to shed light on this genomic “dark
matter” detected and categorized regions that participate in one or more biochemical processes
(32). However, this approach could not be comprehensive because it investigated only a limited
subset of cell types and could not consider all developmental stages. Moreover, it could not
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

effectively distinguish between functional sites and inconsequential sites, such as those involved
in low-occupancy transcription factor interactions (39), and it conflated epiphenomena (such as
random transcription events) with primary causes (biological function). Current experimental
approaches thus are unable to predict comprehensively or accurately how important particular
regulatory elements are to organismal biology. Experimental targeting of mutations in vivo can
reveal deleterious effects, but formally it is still necessary to demonstrate that these have effects in
natural settings, something that current protocols rarely cover.
An alternative approach to predict the biological importance and locations of noncoding func-
tional elements is to recognize the telltale signatures written over time into genomes by the sieving
of mutations by natural selection. This approach has the advantages of being comprehensive (be-
cause the scrutiny of selection extends to any element that is functional in any cell type and at
any developmental stage) and inexpensive (because of the availability of genome sequences and of
cheap sequencing technologies). Selection also will not have regarded as functional any sequence
that, despite being bound by a factor or being transcribed, remains of no biological consequence.
Finally, unlike biochemical assays, evolutionary analyses can predict an element’s biological im-
portance by estimating the strength of selection that acted on its sequence. This approach can
indicate the extent to which fitness has been affected by ancient mutations that landed in single
elements, and more generally in classes of functional elements, and thus can rank elements by
their proposed biological importance.
Here, we review the relative contributions to human biology of different classes of functional el-
ements inferred from comparisons of recently sequenced mammalian and human genomes. Owing
to space limitations, we mostly restrict the discussion to classes of functional elements rather than
individual examples, and we focus on conservation and constraint as opposed to positive selection
and adaptation, which have been reviewed elsewhere (41). As a consequence, the review focuses
on the approximately 10% of the human genome that is under significant selective constraints and
not on the vast majority that evolves neutrally (101, 103). The review is structured around three
counterpoints (Figure 1): (a) selection that acted on either non-protein-coding or protein-coding
sequence; (b) selection that acted on either the genomes of humans (or other mammals) or those
of other, more distantly related metazoans, such as Drosophila (fruit flies); and (c) selection that
was either ancient or more contemporary.

2. MUTATION AND SELECTION, CONSERVATION AND CONSTRAINT

Imagine if we could train a time-lapse camera on a single nucleotide position in your genome and,
by winding back time, watch how it changed by chance mutations as it was passed back through
the generations (and along the germline) over hundreds of millions of years. If this nucleotide

72 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

Weak S E LE C T IO N Strong

Non-protein-
coding genes
a
Protein-
coding genes

Fruit fly

Protein-
b
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

coding Human
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

genes

Ancient
human
c
Contemporary
human

Figure 1
Schematic illustration of the differing extents to which selection has acted on exonic sequences of genes. These genes have been
grouped according to the threefold organizational principle of the review: Comparisons are made between (a) selection acting on
non-protein-coding versus protein-coding genes; (b) selection acting on protein-coding genes in species such as Drosophila (fruit flies)
versus those in humans; and (c) selection acting on human protein-coding genes during ancient versus more contemporary evolution.
On average, selection has been weakest on transcribed non-protein-coding sequence and strongest on protein-coding sequence from
species (such as fruit flies) that have high effective population sizes.

was functional, we would observe it to have changed only very rarely. This is because change
would be mostly deleterious and thus would be negatively selected and less likely to have been
propagated to subsequent generations. For example, three-fourths of the coding bases of human
histone H4, a protein of crucial importance in packaging DNA, have remained unaltered since
we last shared an ancestor with plants. However, if the scrutinized nucleotide was not functional,
changes would not have been selected against and thus would have occurred more frequently. At
the end of these very long time periods, a neutrally evolving sequence would have experienced so
many changes that little trace of its ancestral sequence would remain in extant species. After this
amount of divergence time, alignment of DNA sequence has often become so inaccurate as to be
uninformative.
Comparisons of sequences of extant species whose common ancestors lived in this long-ago
time period often reveal short regions that are alignable and well conserved, separated by long
stretches that are unalignable and poorly conserved. For example, only 2.5% of nucleotides align
in the genomes of chickens and humans, species that last shared a common ancestor a little
over 300 million years ago (52). When an accurate alignment can be obtained, it permits an
evolutionary change to be assigned, usually unambiguously, to a lineage in a phylogeny. When
two such alignments are compared, different numbers of changes assigned to this lineage can
indicate differences in these two regions’ evolutionary rates. A region whose evolutionary rate
significantly exceeds that of another region, chosen specifically because it is believed to have
evolved neutrally (Table 1), is likely to have experienced episodes of positive selection, perhaps
owing to adaptive evolution. A region whose evolutionary rate is significantly lower than that of

www.annualreviews.org • Selection on Functional Elements 73

 GG15CH04-Ponting ARI 15 July 2014 11:5

Table 1 Genomic elements used as neutrally evolving sequences

Neutral background Advantages Disadvantages References
Fourfold degenerate sites Interdigitated Codon usage bias, regulatory 22, 33, 71, 74
elements
Small introns Neutral Applicable only in Drosophila species 96
Ancestral repeats Neutral Neutrally assessed only in 79
mouse–human
Unannotated nonconserved Widespread Background selection, hitchhiking 124
sites across the genome
Sampling within flanking Control for background Not necessarily fully neutral 5, 48
sequences selection
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

the putatively neutrally evolved region has been subject to constraint—in other words, negative
Purifying selection: (or purifying) selection. Conserved sequence is not necessarily constrained: Human and great ape
the process by which genomes are highly conserved because of their relatively recent common ancestry, yet most of their
deleterious mutations sequence is not constrained. Conversely, constrained sequence is not necessarily conserved if its
are preferentially
function has arisen only recently and thus is not shared with the other species under consideration.
removed from the
population Metazoan genomes differ substantially in their fractions of constrained sequence, depending on
their gene content and the size and extent of their regulatory sequence (103) (Figure 2).

Constrained

Genome Protein-coding
Coding Noncoding Unconstrained size (Mb) genes

H. sapiens 3,102 20,806

M. musculus 2,731 23,158

G. gallus 1,047 15,508

T. rubripes 393 18,523

D. melanogaster 169 13,937

0 0.2 0.4 0.6 0.8 1.0

Proportion of the genome

Figure 2
Proportion of the genome identified as being under constraint (purifying selection) when using the neutral indel model (79, 84).
Alignments between Homo sapiens and Macaca mulatta, Mus musculus and Rattus norvegicus, Gallus gallus and Taeniopygia guttata, Takifugu
rubripes and Tetraodon nigroviridis, and Drosophila melanogaster and Drosophila simulans were used to produce the fraction of the genome
under constraint. The proportions of sequence under constraint depend on the species pair analyzed and the alignment-quality
processing. Data from Meader et al. (84).

74 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

Detecting species-specific functional sequence that has only recently become constrained re-
quires comparisons not between species whose lineages have long been separated but rather within
a population—for example, that of modern humans. Over the several hundred thousand years since
dN : the number of
the ancestral population of modern humans arose, mutations have risen and fallen in population nonsynonymous
frequency: Those that are advantageous tend to rise in frequency because of positive selection, substitutions per
whereas those that are substantially deleterious tend toward extinction because of negative selec- nonsynonymous site
tion. Approaches to identify the imprint of selection thus compare the population frequencies of dS : the number of
alleles in the regions of interest against those in putatively neutrally evolved sequence: Higher synonymous
frequencies might reflect episodes of positive selection, whereas lower frequencies might reflect substitutions per
synonymous site
purifying selection. Importantly, even if a coding sequence region shows a signature of constraint,
it is not possible to infer definitively whether this discriminative sieve of selection has acted on
DNA or RNA or on protein molecular functions.
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

3. BETWEEN-SPECIES COMPARISONS

3.1. Spectrum of Selection Across Protein-Coding Gene Sequence

Winding back time by approximately 100 million years takes us to the Cretaceous period, just
before the separation of lineages that led to either modern primates (including humans) or rodents
(such as mice). The protein-coding DNA of humans and mice has been subject mostly to strong
purifying selection over this long time period, with approximately 85% of bases remaining the
same as they were in their last common ancestor (87, 109). Aligned non-protein-coding DNA,
however, has in the meantime changed considerably, with approximately twice as many bases
differing between the two species as compared with protein-coding sequence.
The variable impression left across coding sequence from the selective purging of deleterious
alleles is evident from a classic image of early vertebrate genomics, which Jim Kent produced by
sampling alignments between human and mouse genes (87, figure 25a). Instead of using pair-
wise sequence identity, as he did, we have recalculated the profile of purifying selection across
mammalian genes using conservation (phastCons) scores estimated across vertebrate evolution
(109) (Figure 3). Sequence conservation here directly indicates the relative importance of differ-
ent nucleotide positions along a generic gene model. Sites in exons, particularly those near exon
boundaries (Figure 3), tend to be the least changeable without impairing gene function, whereas
the bulk of intronic sequence is more accepting of changes because it contains a much lower pro-
portion of functional sequence; mammalian 5 and 3 untranslated regions (UTRs) show levels of
conservation that are intermediate between these two extremes.
The discriminative impact of selection is most apparent within codons. Owing to the three-
fold degeneracy of the genetic code, some nucleotide substitutions fail to alter the amino acid
and are therefore synonymous. Many such substitutions occur in the third codon position, and
there are eight amino acids whose third site—termed a fourfold degenerate (4D) site—is en-
tirely degenerate; substitutions at these sites are always synonymous. Purifying selection thus
acts discriminately on these different sites, with stronger selection on the first and second sites
than on the third (and thus 4D) site (Figure 3 insets). Nucleotide substitutions in codons that
alter the amino acid (nonsynonymous changes) are usually fixed far less frequently than synony-
mous changes, and under an assumption—challenged below—that synonymous changes are free
from selection, the strength of selection acting on amino acid sites can be inferred. More for-
mally, the ratio of dN (or KA , the number of nonsynonymous substitutions per nonsynonymous
site) to dS (or KS , the number of synonymous substitutions per synonymous site) is expected
to be 0 when amino acids are essential, 1 when the codons evolve neutrally, and significantly

www.annualreviews.org • Selection on Functional Elements 75

 GG15CH04-Ponting ARI 15 July 2014 11:5

5' flank 5' UTR CDS 3' UTR 3' flank

0.8 H. sapiens 0.9 Codon position

First
0.8 Second
Third
0.7
Fourfold
0.6 degenerate
0.6 site

0.5
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

0.4
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

0.2
Average phastCons score

0.9
D. melanogaster
0.8

0.8 0.7

0.6

0.5
0.6

0.4

0.2

Figure 3
Average nucleotide conservation (phastCons) score sampled across 5 untranslated region (UTR), coding DNA sequence (CDS), and
3 UTR exons and introns in Homo sapiens and Drosophila melanogaster at the first (orange), second (dark red ), and third (dark blue) codon
positions and fourfold degenerate sites (dark aqua). Only exons and introns that were at least 200 nucleotides long and did not overlap
other annotated features were used. The high conservation values observed for the intermediate 3 UTR exons in D. melanogaster are
likely stochastic noise, a consequence of their low number (20 exons larger than 200 nucleotides). The red and light blue shaded areas
represent the range of nucleotide conservation across long noncoding RNA gene models in H. sapiens and D. melanogaster, respectively.

76 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

BACKGROUND SELECTION

The local environment (background) of a variant can affect whether it is purged from a population. For example,
neutral variation in a UTR will be reduced because of its linkage with genomically proximal coding sequence that is
under negative selection. This is due to neutral changes propagating in a population for an extended period of time
only when gametes are free of deleterious alleles (19). Recombination disrupts genetic linkage and thus diminishes
background selection.

greater than 1 when positive selection has acted to preferentially fix amino acid changes in the
population (56). Selective pressure on synonymous sites, however, is not negligible. This is be-
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

cause for many species there is selection on codon usage and for the preservation of splicing
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

regulatory elements (18, 37). In mammals, approximately 20% of 4D sites are subject to a non-
negligible degree of selective constraint (33). Consequently, the use of 4D sites to model neu-
tral evolution in tests for selection can lead to significant biases (Table 1). Splicing regulatory
elements tend to lie near intron–exon boundaries (within 50 base pairs), and it is their conserva-
tion that appears to underlie the peak of conservation scores close to these boundaries (95, 125)
(Figure 3). The heterogeneity of conservation within and between coding exons (Figure 3) in-
dicates that a substantial amount of selection occurs not because of changes in protein function
but because of RNA and/or DNA function. This is important because ignoring selection on nu-
cleotide function in coding sequence leads to underestimation of neutral rates, which then leads
to higher rates of false predictions of positive selection and more false negatives for purifying
selection.
As we have observed, noncoding exonic bases of UTRs are better conserved than noncoding
intronic bases (Figure 3). This higher degree of conservation is due to its content of various types of
functional elements, such as G-quadruplex sequences, internal ribosome entry sites, and upstream
open reading frames (7, 33, 54, 94, 109); conservation may also reflect background selection (see
sidebar, Background Selection). Many mammalian 5 UTRs also contain spliced exons whose
intronic boundaries, similarly to those in coding sequence, show elevated nucleotide conservation
(107) (Figure 3), presumably reflecting the high density of splicing regulatory elements lying
near exon junctions. The increased sequence conservation of 3 UTRs over introns is attributable
at least in part to purifying selection acting to preserve or avoid microRNA (miRNA) binding
sites, also called miRNA response elements (MREs) (20, 112, 129). Mammalian miRNA seed
sequences bind via base-pairing with imperfectly complementary MREs within mRNAs, leading
to translational repression or target degradation (8). Based on the conservation of at least 45,000
MRE: microRNA
miRNA target sites within human 3 UTRs, the majority of human genes appear to have been response element
under selective pressure to maintain pairing to miRNAs (40). Mutations that create MREs can
Neutral proxy:
also be under negative selection because of their deleterious effect on gene expression regulation sequence regions that
(20, 112). are considered to be
Some investigators consider intronic sequences to have evolved neutrally and thus to provide a neutrally evolving and
useful neutral proxy against which selection on coding sequences can be assessed (Table 1). How- that are used as the
null expectation when
ever, in addition to sequences flanking splice acceptor and donor sites and other splicing regulatory
testing for selection
motifs that occur within 200 base pairs of exons, introns contain diverse and numerous functional that acted on
elements that all contribute to lower evolutionary rates (17, 81). Kim & Pritchard (63) reported sequences of interest
that approximately 37,000 conserved noncoding elements fall within introns, representing ap-
proximately 4.6 Mb, although a more recent estimate of intronic functional elements numbers
these at nearly 1.5 million by counting all DNase I–hypersensitive sites (118). Evolutionary rates

www.annualreviews.org • Selection on Functional Elements 77

 GG15CH04-Ponting ARI 15 July 2014 11:5

also vary depending on the position of an intron within a gene (45). For example, relative to other
introns within a gene, first introns tend to be better conserved, are on average twice as long, and
differ in their nucleotide composition (17, 45).
lncRNA: long
noncoding RNA

3.2. Spectrum of Selection Across Noncoding RNA Loci

In mammals, thousands of multi-exonic long noncoding RNA (lncRNA) genes have been identified
(12, 24, 51) for which cellular functions and molecular mechanisms remain largely unknown (104,
120). Sequence conservation levels of mammalian intergenic lncRNAs are low, much lower than
those of protein-coding sequences and only marginally higher than those of putatively neutral
sequences (82, 100). The variable impression that purifying selection has made on these exons
and introns of noncoding loci, however, is qualitatively similar to that on protein-coding genes
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

(Figures 3 and 4). Nevertheless, across mammalian evolution the strength of this selection on
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

mature lncRNA sequences has been only modest, and sequence conservation only marginally
exceeds that seen for intronic and untranscribed intergenic sequences (Figure 4). Indeed, based
on evidence for purifying selection on inserted or deleted sequences, only 5% of all bases in mouse
RNA transcripts are estimated to be functional (100).
This low level of conservation might imply that such lncRNAs are rarely functional. Instead,
they could be associated with transcriptional noise (114), perhaps emanating from transcription at
a neighboring locus (27). If biologically relevant, the lack of sequence conservation of transcribed
intergenic noncoding loci is likely to reflect the presence in RNA transcripts of only short patches of
functional sequence involved in base-pairing or protein interactions, similar to the limited MREs in
3 UTRs of protein-coding genes (13, 16). The observed low nucleotide conservation of lncRNAs
could also be associated with functional redundancy (24, 108). In contrast, rapid divergence could
result from compensatory mutations that are rapidly fixed when they mitigate the deleterious effect
of a second mutation at a functional site. This mechanism has been proposed to explain both the
accumulation of potentially deleterious mutations within protein-coding sequences or regulatory
elements (67, 130) and the maintenance of RNA secondary structures (98).

3.3. Spectrum of Selection Across Other Noncoding Sequence

Protein-coding sequence is thus the principal substrate of negative selection. However, a significant
minority of intergenic noncoding sequence is also subject to purifying selection (Figure 2). When
the mouse genome was first sequenced and compared with the human genome, investigators
realized that most of the constrained sequence lay outside of the protein-coding sequence. In
addition to the 1.2% of the human genome that encodes proteins, nearly 4% was estimated to be
constrained (78, 87, 109) (Figure 2). An analysis of 29 placental mammal genomes showed that
approximately 39% of the 3.6 million constrained elements lie within intergenic sequences, more
than 2 kb away from annotated gene models (78). Experiments in zebrafish and mice indicate that
many of these sequences are biologically functional, with potential roles as enhancers, insulators,
or promoters (78, 97, 122, 128).
Purifying selection on noncoding sequence acts most stringently on bases adjacent to transcrip-
tion start sites and rapidly declines in strength for approximately 200 nucleotides farther upstream
(14, 115) (Figure 3). These upstream regions of protein-coding genes have experienced unusually
high levels of nucleotide substitutions, short insertions and deletions, and transposable element
insertions during primate evolution (115, 126). Rather than being the result of pervasive positive
selection, this likely reflects elevated mutation rates in sequence that is made accessible and more
mutable by the act of transcription initiation. A higher mutability for the upstream regions of

78 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

5' 3'

0.18 H. sapiens

0.16

0.14

0.12
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

Average phastCons score

0.10

D. melanogaster

0.5

0.4

0.3

0.2

0.1

Figure 4
Average nucleotide conservation (phastCons) scores sampled across intergenic long noncoding RNA loci in Homo sapiens and Drosophila
melanogaster.

genes that are frequently transcribed in the germline also likely explains the otherwise curious
observation that promoters of non-protein-coding RNA genes tend to be better conserved than
promoters of protein-coding genes (3, 82, 100). At the 3 termini of genes, sequence in the vicinity
of the polyadenylation site (motif AWUAAA) also exhibits a trend for more stringent purifying
selection, presumably owing to its requirement by cleavage and polyadenylation specificity factors
(23, 90).
The effects of mutations in noncoding sequence can sometimes be predicted only by the
scrutiny of purifying selection and are not corroborated by the scrutiny of laboratory experiments.
For example, the deletion of long intergenic sequences containing more than 1,000 noncoding

www.annualreviews.org • Selection on Functional Elements 79

 GG15CH04-Ponting ARI 15 July 2014 11:5

elements, each conserved between human and rodents (>70% sequence identity over 100 base
pairs), in laboratory mice has failed to reveal phenotypic effects with respect to litter size, body
weight, or longevity (3). Similarly, the deletion of elements that are completely conserved in
Effective population
size (Ne ): the number sequence between humans, mice, and rats or the deletion of a conserved enhancer region involved
of individuals in an in limb patterning leads to viable and fertile mice lacking overt phenotypes (3, 23). Knockout
idealized population mutant mice for three intergenic lncRNA loci (Hotair, Malat1, and Neat1) also show no overt
that show a level of phenotype, even though these loci are among the most highly expressed and most highly conserved
loss of heterozygosity
of all intergenic lncRNAs (29, 89, 90, 108, 132).
due to genetic drift
equivalent to that of How can the absence of overt phenotypes be reconciled with the deep phylogenetic conser-
the actual population vation of these elements? Different hypotheses have been proposed to resolve this conundrum.
TFBS: transcription First, phenotypes not observable under laboratory conditions may be revealed under more natural
factor binding site conditions. Knockout mutants for the BC1 locus, for example, showed no phenotypic changes
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

except when assayed in natural conditions (75). Second, changes may be missed owing to shallow
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

phenotyping. Fewer than one-fifth of genes in either Caenorhabditis elegans or Saccharomyces cere-
visiae initially yielded phenotypic effects when disrupted, yet upon more extensive investigation,
larger proportions (42–60% and 97% of genes, respectively) were found to yield significant effects
(55, 105). Natural selection is thus better able to discriminate deleterious from neutral variants
than are laboratory experiments.

3.4. Evolutionary Turnover of Functional Sequence

Most analyses described thus far successfully associate deep phylogenetic conservation with se-
quence functionality (78, 87). However, methods that identify genomic regions by their reduced
rates of substitution sacrifice sensitivity for specificity, leading to a relatively high rate of false-
negative predictions. Sequences of regulatory regions that have been subject to weak selective
pressure or lineage-specific evolution will diverge rapidly (9, 38, 83). This may explain, in part,
why the pilot ENCODE project found that nearly half of all annotated functional elements were
not conserved across species (30).
A sequence may be constrained in one species yet not conserved in others if it has gained
functionality only recently; another sequence may be relatively well conserved yet be without
constraint if it has recently lost its functionality. These events represent turnover, occurring
either when environmental changes or sequence mutations alter the functionality of the locus,
or when a functional element is lost through genetic drift or gained when (for example) such an
element is acquired that compensates for prior losses. Turnover will occur most frequently for
elements that are subject to low constraint and for species or sequences that are associated with
relatively low effective population sizes. New functional elements will emerge most frequently
from previously functionally inert sequence when their lengths are short (57, 66). Consequently,
relatively long and highly constrained protein-coding sequence is only rarely turned over, whereas
short noncoding elements, such as transcription factor binding sites (TFBSs) and enhancers, turn
over rapidly.
Half of all functional noncoding sequence is predicted to have turned over in approximately
the past 130 million years of mammalian evolution (106). Experimental studies of TFBSs support
these evolutionary predictions, with 41–89% of such events not being conserved between humans
and mice for four liver transcription factors (91) and almost half of all intergenic lncRNA loci
having been gained or lost in the interval since the last common ancestor of mice and rats (72).
Target sites of miRNA have turned over more slowly, with 20% of such sites turning over among
diverse mammals (129). A prominent example of a sequence that has been the frequent substrate
of functional loss and gain is human HAR1F (99), a lncRNA that has gained 18 substitutions in

80 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

118 nucleotides since humans’ last common ancestor with chimpanzees. These substitutions are
very strongly biased toward G or C bases, indicating that, rather than being adaptive changes, they
are the consequences of a mutational process, specifically recombination-associated GC-biased
gene conversion (44).

3.5. Selection on Other Metazoan Genomes

The sequencing of the genomes of 12 Drosophila species enabled an assessment of whether the
signatures of selection found for mammalian genes and genomes are also evident for these inver-
tebrates (21, 47). In the main, the patterns of selection seen for D. melanogaster genes mirror well
those found for mammalian gene models: Their protein-coding exons are the most conserved,
and their introns are the least (21, 86) (Figure 3); a similar proportion (22%) of fruit fly genes’ 4D
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

sites appear to have been subject to selective constraints (22, 74); and sequences upstream of their
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

transcription start sites are diverged, a likely consequence of a lower degree of selection relative
to transcribed regions (80). Additionally, TFBSs, insulator sites, and RNA polymerase II binding
sites all exhibit moderately strong conservation patterns across these drosophilids, as (to a lesser
extent) do UTRs (86).
Nonetheless, several important differences are apparent. In contrast to mammalian coding se-
quence, codon usage is under selection in fruit flies, which strongly modulates protein sequence
evolution (53, 110). Moreover, introns in Drosophila species are subject to a substantial degree of
selection (46), excepting those that are short (<86 nucleotides) (22, 74) (Table 1). Perhaps the
most striking difference with mammalian sequence is that fruit fly lncRNA loci are moderately well
conserved across fly species (86, 131), particularly in their exons (48) (Figure 4). This could reflect
a more central functional role of these noncoding genes in fruit fly evolution and biology. Never-
theless, such differences are best explained through the application of the theoretical framework
provided by population genetics. Indeed, more generally, the ability to distinguish between tran-
scriptional noise and signal, or between random and consequential binding events, or between
ephemeral or long-lasting functional sequences requires methods that go beyond species-level
conservation. To achieve these distinctions requires evidence of constraint among individuals of
a species.

4. CONTEMPORARY SELECTION: POPULATION-BASED ESTIMATES

Comparisons between mammalian genomes can only infer the strength of selection that occurred
over long time periods, often tens of millions of years, that separate these species from their
last common ancestor. Moreover, interspecies analyses produce results that average across the
phylogenetic scope of the study and may not hold true for individual lineages or for extant species.
If the species being compared are too distant phylogenetically, then only the most conserved
elements will be detected, at the cost of low sensitivity (high false-negative rate). By contrast, if
the selected species are too closely related, regions that are neutral may wrongly be considered to
evolve under selective constraints, owing to selection acting at linked sites (see sidebar, Background
Selection), leading to a high false-positive discovery rate (10, 28, 113). Finally, sequencing and
alignment errors can also significantly inflate the neutral evolutionary rate for closely related
species.
The strength of selection can instead be inferred using population genetic approaches that
consider a shorter, more recent timescale, such as the past 35,000–45,000 years, which is the
average age of a new (i.e., derived) variant in the human population (42). Deleterious mutations
are expected to segregate at a lower population frequency than mutations occurring at neutral sites.

www.annualreviews.org • Selection on Functional Elements 81

 GG15CH04-Ponting ARI 15 July 2014 11:5

The converse is expected for advantageous mutations: They should become fixed or segregate at a
higher frequency than neutral mutations. Accordingly, the recent 1000 Genomes Project analysis
as well as an analysis of 6,515 exomes demonstrated a negative correlation between the age of a
DAF: derived allele
frequency mutation and its deleterious effect, with most deleterious mutations within the human populations
having arisen only within the past 5,000 years (2, 42, 62).
Genetic hitchhiking:
the process by which Population-based studies initially investigated the density of polymorphic sites in different se-
positive selection at quence categories, based on the idea that polymorphisms in functional sites will have been more
one site leads to an frequently purged if they are deleterious and more rapidly fixed if they are advantageous. For
increased allele example, both 5 and 3 sequences flanking genes show an excess of low-frequency polymorphisms
frequency at a linked
relative to intergenic sites, presumably owing to a combination of background selection origi-
neutral site
nating from linked protein-coding sequence and their cis-regulatory element content (73, 119).
Furthermore, within human populations, both 5 and 3 UTRs exhibit a significantly lower den-
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

sity of polymorphic sites than do introns or putatively neutral sequences (pseudogenes, ancestral
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

repeats) (73, 88). Although polymorphic sites tend to be evenly distributed across 3 UTRs, single-
nucleotide polymorphism (SNP) density is significantly lower at the UTR boundaries (and, more
specifically, at the polyadenylation sites) and within miRNA binding sites (20, 117).
Analyses of the site frequency spectrum, in particular minor allele frequencies or derived allele
frequencies (DAFs), can infer selection not for single functional elements but for many such
elements when they are considered together as a class. Low frequencies of alleles could reflect the
purging of deleterious variants from the population but could also reflect population structure or
past demographic events. Consequently, selection is best inferred by comparing allele frequencies
in the sequence class of interest to frequencies in a set of presumed neutrally evolving sequences
(Table 1). The strength of purifying selection acting on recently arisen deleterious variants can
be estimated from the excess density of low-frequency derived alleles in a sequence class (e.g.,
lncRNA exons) compared with such alleles within putatively neutral sequences (Table 1).
As expected, the different strengths of selection at various genomic elements, estimated using
this DAF test, mirror those derived from cross-species comparisons. In humans, the skew toward
rarer derived alleles, and thus stronger selective constraints, is strongest for variants within both
nonsynonymous and splice sites. In contrast, lower selective constraints act on variants in UTRs,
introns, and intergenic regions (1, 2, 88). Only 6% of synonymous substitutions are predicted
to be deleterious in humans (116), a lower proportion than the 20% inferred from cross-species
comparisons (33). As it does not rely on interspecies sequence conservation, the DAF test can
identify constraint that has arisen recently. For example, it was used to predict that 30–50% of
nonconserved MREs are functional when their mRNA and cognate miRNA are coexpressed (20).
The method has also been used in an important demonstration that sets of human noncoding
sequences are conserved not because they are subject to low mutation rates but rather because
mutations within them have been preferentially purged from the human population (6, 26, 58,
119).
The intersection of human sequence variation (2) and functional annotations (32) permits
population-based estimation of selection acting on open chromatin, transcription factor binding,
or more generally transcribed sequence (5, 48, 118). TFBSs, such as those for POU-, HOX-,
or FOX-domain-containing transcription factors, include both a lower density of SNPs and
an excess of low-frequency SNPs relative to putatively neutral sites or similar sequences in the
genome that are without evidence of binding (5, 111, 121). Using a more rigorous approach that
better corrects for selection’s effect of distorting the pattern of polymorphism at neutral sites
(background selection, genetic hitchhiking), Arbiza et al. (5) estimated that, on average, 33% of
the nucleotides in a set of high-confidence TFBSs are under selection. They also reported that
patterns of selection identified in TFBSs varied greatly depending on the transcription factor

82 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

H. sapiens D. melanogaster

Zerofold degenerate sites

0.6
lncRNAs
Ancestral repeats/small introns
Proportion of sites

Fourfold degenerate sites

0.4
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

0.2

0.75 0.50 0.25 0 0 0.25 0.50 0.75

Derived allele frequency

Figure 5
Comparison of derived allele frequency spectra in Homo sapiens and Drosophila melanogaster for zerofold (dark blue) and fourfold (light
blue) degenerate sites, intergenic long noncoding RNAs (lncRNAs) (red ), and neutrally evolving ancestral repeats (H. sapiens) or small
introns (D. melanogaster) (dark yellow). Adapted from Haerty & Ponting (48).

(5, 88) and that the strength of selection depended on both the size of the binding site and the
number of degenerate positions in the consensus motif.
It is striking that applying the DAF test to sequences and populations from other species yields
results that contrast with those obtained from the human population. Drosophila conserved non-
coding sequences, for example, appear to be under greater purifying selection than such sequences
in humans (15). This is most evident for intergenic lncRNAs, which in fruit flies, but not in hu-
mans, exhibit a clear excess of low-frequency alleles (48) (Figure 5). These differences extend to
other mammals because selective constraints on conserved noncoding sequences, including pro-
moters, in rodents appear to be almost twice as strong as those in the human population (0.53 ±
0.01 and 0.29 ± 0.02, respectively) (60, 115). These differences are explained below as a direct
effect of these species’ contrasting effective population sizes.
Comparison of DAF test results between species is complicated by their population size histories
and structures. The human population underwent multiple bottlenecks of various constrictions
that reduced its effective population size to approximately 1,200 individuals and were then followed
by demographic expansion (76). These changes strongly bias allele frequencies at neutral sites
and, if not properly accounted for, lead to false inferences of selection. This is because there is an
enrichment of rare variants at neutral sites in cases of population expansion (43) and an increased
frequency of intermediate variants in cases of population contraction or bottlenecks (123, 127).
Additionally, population range expansions associated with founder effects at the margin of the
population spatial distribution can lead to a strong bias in allele frequency that mimics the action
of positive selection (also called allele surfing) (65).
Methods that correct for the otherwise confounding demographic effects have been developed
that infer the distribution of fitness effects from allele frequencies (127). Three methods, developed

www.annualreviews.org • Selection on Functional Elements 83

 GG15CH04-Ponting ARI 15 July 2014 11:5

by Eyre-Walker et al. (36), Boyko et al. (11), and Keightley & Eyre-Walker (59), estimate demo-
graphic parameters from user-provided putatively neutral sites to correct the frequency spectrum
of sites of interest (11, 34, 36, 59, 127). Each of these methods is based on predictions from the
nearly neutral theory of evolution (64, 92, 93) together with assumptions that adaptive polymor-
phisms should be very rare, because they are fixed rapidly, and that the majority of nonneutral
mutations are deleterious. Consequently, based on the product of the effective population size
(Ne ) and the selection coefficient, mutations are classified as being effectively neutral (|Ne s| < 1),
weakly deleterious (1 < |Ne s| < 10), deleterious (10 < |Ne s| < 100), or strongly deleterious
(|Ne s| > 100). These algorithms have been applied to polymorphism data in humans, mice, and
fruit flies to assess the variable strength of selection at both coding and putatively functional non-
coding sites (Figure 6). In humans, 27–38% of mutations at nonsynonymous sites were deemed
effectively neutral, 21–30% were weakly deleterious, and the remainder were strongly deleterious
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

(11, 34). By comparison, the DAF spectrum of SNPs within intergenic lncRNAs was not different
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

from putatively neutrally evolving neighboring sequences, providing no evidence for the action
of purifying selection acting contemporarily on human lncRNAs. Nevertheless, the same anal-
ysis applied to intergenic lncRNAs from D. melanogaster predicted that approximately 30% of
mutations within these loci are weakly deleterious (48) (Figure 6).

5. EFFECTIVE POPULATION SIZE AND RELAXATION OF

SELECTION WITHIN THE PRIMATE LINEAGE
Many more substitutions at nonsynonymous sites are predicted to be deleterious in Mus musculus
castaneus or D. melanogaster (59, 68) than in humans. Similarly, there appears to be a greater strength
of purifying selection on codon usage, lncRNAs, UTRs, and conserved noncoding sequence in
fruit flies than in humans (48, 49, 61, 70, 111). These observations are likely to reflect the low value
of Ne for humans (1,200–15,000) (76) relative to D. melanogaster (1,450,000) (35) or M. musculus
castaneus (290,000–580,000) (49). This is because Ne conditions the probability that a mutation
under selection will be fixed. Any newly arising mutation can be neutral (selection coefficient s =
0), deleterious (s < 0), or advantageous (s > 1). Ohta (92, 93) and Kimura (64) showed that the
probability θ that a mutation with s = 0 will be fixed is θ = Ne μs, where μ is the mutation rate. If
|Ne s| > 1, the mutation will be effectively selected, whereas if |Ne s| < 1, it will evolve effectively
neutrally, under genetic drift. Consequently, the same mutation whose evolution is nearly neutral
in a species of low Ne would have been effectively selected (positively or negatively) in a species
with a much larger Ne .
This implies that the human population has accumulated weakly deleterious variants faster
than species with larger Ne values. Such a relaxation of selection likely explains the greater-than-
expected divergence in functional elements. A greater extent of positive selection in primates
would also be consistent with this increased divergence in functional elements (99). Nevertheless,
this is not a parsimonious explanation, because it requires positive selection to have acted relatively
indiscriminately on very large numbers of elements that in other species are under the greatest
constraint. As a consequence, functional elements are expected to turn over rapidly in species with
low Ne values, such as primates. Functional elements in species with high Ne values, in contrast, are
predicted to persist over longer evolutionary time and thus to have a greater chance of acquiring
more fundamental functions.
The evolutionary analyses described above compare the decreased divergence or diversity of
sequences of interest to others whose evolution is proposed to have been neutral. Whether these
neutral proxy sequences have always escaped selection thus affects quantitatively, although rarely
qualitatively, these analyses’ results. Of all the commonly used neutral proxies (Table 1), 4D

84 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

N EU T R A L D E LE T E RIO U S
|Nes|: 0–1 1 – 10 10 – 100 10 – ∞ 100 – ∞

H. sapiens M. musculus castaneus D. melanogaster

1.00
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

0.75
Proportion of sites

0.50

0.25

Zerofold 5’ UTR 3’ UTR lncRNA Zerofold 5’ UTR 3’ UTR CNE Zerofold 5’ UTR Intron 3’ UTR lncRNA
degenerate degenerate degenerate
Type of site

Figure 6
Comparison of the distribution of fitness effects of mutations occurring within 5 and 3 untranslated regions (UTRs), zerofold
degenerate sites, introns, conserved noncoding elements (CNEs), and long noncoding RNAs (lncRNAs) in Homo sapiens, Mus musculus
castaneus, and Drosophila melanogaster. For the CNEs in M. musculus castaneus, only three categories are present: |Ne s| < 1, 1 < |Ne s| <
10, and 10 < |Ne s| < ∞. Data from Haerty & Ponting (48), Eyre-Walker & Keightley (34), Halligan et al. (50), and Kousathanas et al.
(69).

www.annualreviews.org • Selection on Functional Elements 85

 GG15CH04-Ponting ARI 15 July 2014 11:5

sites, particularly in high-Ne species, will be the least applicable. Ancestral repeats, defined as
transposable element sequences that inserted prior to the last common ancestor of two species
(often humans and mice), are proposed to be the most reliable neutral proxy because virtually all
of them have been demonstrated to have evolved neutrally (79, 102). Only 0.2% and 1% of bases
in rat–mouse and human–mouse ancestral repeat sequences, respectively, depart from neutrality
(84).
Neutral sites are commonly collated across the whole genome, thereby forming a putatively
neutral reference against which test sequence is compared. However, this overlooks important
confounding issues. First, because test and neutral sites are most often not fully interdigitated,
they are likely to not share the same genealogical history (4, 85, 133), and comparisons will
not necessarily account for mutation rates, which vary greatly across the genome (31). Second,
selection acting at linked sites, via either background selection (see sidebar, Background Selection)
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

or hitchhiking, will distort polymorphism patterns at neutral proxy sites.

Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

The optimal choice of neutral proxy is sequence that has escaped selection and is interdigi-
tated or that lies adjacent to and is compositionally equivalent to sites of interest (for example,
nonsynonymous sites, MREs, and lncRNAs) (4, 133). These requirements are met by matching
candidate neutral sequence that lies in close physical proximity, and is compositionally similar,
to each test sequence. Nevertheless, matches are often not possible when, for example, ancestral
repeats of the required composition are absent from the genomic vicinity of the test sequence.
Consequently, a compromise procedure, adopted by Halligan et al. (50), Haerty & Ponting (48),
and Arbiza et al. (5), is to employ proxy neutral sequence that is noncoding and not conserved
in diverse species and that lies outside of annotations such as TFBSs or DNase I–hypersensitive
sites. Such sequence constitutes the majority of the reference human genome but only a small
proportion of the fruit fly genome.

SUMMARY POINTS
1. Sequences that have long retained functionality are characterized by an increased inter-
species sequence conservation that derives from the past purging of mutations that had
a deleterious impact on fitness.
2. Purifying selection is strongest on protein-coding sequences, more moderate on flank-
ing noncoding functional sequences (5 and 3 untranslated regions), and very weak on
intergenic transcribed or bound functional elements that tend to turn over rapidly in
evolution.
3. The most powerful approach to assess the current biological relevance of these rapidly
evolving elements is to analyze variants’ site frequency spectra.
4. Because a species’ effective population size directly affects the efficacy of natural selection
on mutations, classes of elements identified as evolving under selective constraints in one
species may be effectively neutral in another species of smaller effective population size.
5. The observation of a biochemical event at a locus is necessary but not sufficient to infer
that it is functional. Instead, indication of biochemical activity, such as transcription or
transcription factor binding, combined with the inference of selective constraint between
or within species provides the most compelling evidence for the functionality of elements
that can then be prioritized for subsequent experimental validation.

86 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

FUTURE ISSUES
1. With decreasing sequencing costs of large numbers of samples, population genomic ap-
proaches that assess the functionality of elements within the genome will spread. This will
allow the in silico assessment of the biological relevance of a locus that draws upon both
evolutionary and functional (for example, transcription or transcription factor binding)
evidence.
2. Not all transcriptional or binding events will be biologically relevant. However, we do
not have a clear understanding of the expected background levels of such biological
noise when interpreting functional genomic data. The use of more quantitative methods
to investigate the strength of transcription factor binding (77) will also greatly help to
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

identify genomic regions in which binding is only transient and likely inconsequential.
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

3. Many tests for selection were initially developed for protein-coding sequence and only
later adapted for non-protein-coding sequence. There is a considerable need for new tests
that consider the deleterious effects of mutation and are tailored specifically to functional
noncoding elements, because these are the primary substrate of rapid turnover along
different evolutionary lineages.
4. We know almost nothing about the relative impacts that disruption of classes of func-
tional noncoding sequence—including microRNA response elements, transcription fac-
tor binding sites, and long noncoding RNA loci—have on fitness and phenotypes. There
is thus a pressing need for large-scale in vivo and cellular mutagenesis and phenotyping
projects.

DISCLOSURE STATEMENT
The authors are not aware of any affiliations, memberships, funding, or financial holdings that
might be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS
We thank Jim Kent (University of California, Santa Cruz) for helpful discussions on the conser-
vation of mouse/human genes. We are grateful to Chris Rands for his comments on the early
versions of this review. W.H. and C.P.P. are supported by the Medical Research Council, and this
work was also supported by a European Research Council Advanced Grant to C.P.P.

LITERATURE CITED
1. 1000 Genomes Proj. Consort. 2010. A map of human genome variation from population-scale sequenc-
ing. Nature 467:1061–73
2. 1000 Genomes Proj. Consort. 2012. An integrated map of genetic variation from 1,092 human genomes.
Nature 491:56–65
3. Ahituv N, Zhu Y, Visel A, Holt A, Afzal V, et al. 2007. Deletion of ultraconserved elements yields viable
mice. PLoS Biol. 5:e234
4. Andolfatto P. 2008. Controlling type-I error of the McDonald–Kreitman test in genomewide scans for
selection on noncoding DNA. Genetics 180:1767–71
5. Arbiza L, Gronau I, Aksoy BA, Hubisz MJ, Gulko B, et al. 2013. Genome-wide inference of natural
selection on human transcription factor binding sites. Nat. Genet. 45:723–29

www.annualreviews.org • Selection on Functional Elements 87

 GG15CH04-Ponting ARI 15 July 2014 11:5

6. Asthana S, Noble WS, Kryukov G, Grant CE, Sunyaev S, Stamatoyannopoulos JA. 2007. Widely dis-
tributed noncoding purifying selection in the human genome. Proc. Natl. Acad. Sci. USA 104:12410–15
7. Barreiro LB, Laval G, Quach H, Patin E, Quintana-Murci L. 2008. Natural selection has driven popu-
lation differentiation in modern humans. Nat. Genet. 40:340–45
8. Bartel DP. 2004. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116:281–97
9. Blow MJ, McCulley DJ, Li Z, Zhang T, Akiyama JA, et al. 2010. ChIP-Seq identification of weakly
conserved heart enhancers. Nat. Genet. 42:806–10
10. Boffelli D, Nobrega MA, Rubin EM. 2004. Comparative genomics at the vertebrate extremes. Nat. Rev.
Genet. 5:456–65
11. Boyko AR, Williamson SH, Indap AR, Degenhardt JD, Hernandez RD, et al. 2008. Assessing the
evolutionary impact of amino acid mutations in the human genome. PLoS Genet. 4:e1000083
12. Cabili MN, Trapnell C, Goff L, Koziol M, Tazon-Vega B, et al. 2011. Integrative annotation of human
large intergenic noncoding RNAs reveals global properties and specific subclasses. Genes Dev. 25:1915–27
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

13. Calin GA, Liu CG, Ferracin M, Hyslop T, Spizzo R, et al. 2007. Ultraconserved regions encoding
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

ncRNAs are altered in human leukemias and carcinomas. Cancer Cell 12:215–29
14. Carninci P, Sandelin A, Lenhard B, Katayama S, Shimokawa K, et al. 2006. Genome-wide analysis of
mammalian promoter architecture and evolution. Nat. Genet. 38:626–35
15. Casillas S, Barbadilla A, Bergman CM. 2007. Purifying selection maintains highly conserved noncoding
sequences in Drosophila. Mol. Biol. Evol. 24:2222–34
16. Cesana M, Cacchiarelli D, Legnini I, Santini T, Sthandier O, et al. 2011. A long noncoding RNA controls
muscle differentiation by functioning as a competing endogenous RNA. Cell 147:358–69
17. Chamary JV, Hurst LD. 2004. Similar rates but different modes of sequence evolution in introns and at
exonic silent sites in rodents: evidence for selectively driven codon usage. Mol. Biol. Evol. 21:1014–23
18. Chamary JV, Parmley JL, Hurst LD. 2006. Hearing silence: non-neutral evolution at synonymous sites
in mammals. Nat. Rev. Genet. 7:98–108
19. Charlesworth B, Morgan MT, Charlesworth D. 1993. The effect of deleterious mutations on neutral
molecular variation. Genetics 134:1289–303
20. Chen K, Rajewsky N. 2006. Natural selection on human microRNA binding sites inferred from SNP
data. Nat. Genet. 38:1452–56
21. Clark AG, Eisen MB, Smith DR, Bergman CM, Oliver B, et al. 2007. Evolution of genes and genomes
on the Drosophila phylogeny. Nature 450:203–18
22. Clemente F, Vogl C. 2012. Evidence for complex selection on four-fold degenerate sites in Drosophila
melanogaster. J. Evol. Biol. 25:2582–95
23. Cretekos CJ, Wang Y, Green ED, Martin JF, Rasweiler JJ IV, Behringer RR. 2008. Regulatory divergence
modifies limb length between mammals. Genes Dev. 22:141–51
24. Derrien T, Johnson R, Bussotti G, Tanzer A, Djebali S, et al. 2012. The GENCODE v7 catalog of
human long noncoding RNAs: analysis of their gene structure, evolution, and expression. Genome Res.
22:1775–89
25. Dobzhansky T. 1964. Biology, molecular and organismic. Am. Zool. 4:443–52
26. Drake JA, Bird C, Nemesh J, Thomas DJ, Newton-Cheh C, et al. 2006. Conserved noncoding sequences
are selectively constrained and not mutation cold spots. Nat. Genet. 38:223–27
27. Ebisuya M, Yamamoto T, Nakajima M, Nishida E. 2008. Ripples from neighbouring transcription. Nat.
Cell Biol. 10:1106–13
28. Eddy SR. 2005. A model of the statistical power of comparative genome sequence analysis. PLoS Biol.
3:e10
29. Eißmann,Gutschner T, Hämmerle M, Günther S, Caudron-Herger M, et al. 2012. Loss of the abundant
nuclear non-coding RNA MALAT1 is compatible with life and development. RNA Biol. 9:1076–87
30. Elgar G, Vavouri T. 2008. Tuning in to the signals: noncoding sequence conservation in vertebrate
genomes. Trends Genet. 24:344–52
31. Ellegren H, Smith NG, Webster MT. 2003. Mutation rate variation in the mammalian genome. Curr.
Opin. Genet. Dev. 13:562–68
32. ENCODE Proj. Consort. 2012. An integrated encyclopedia of DNA elements in the human genome.
Nature 489:57–74

88 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

33. Eory L, Halligan DL, Keightley PD. 2010. Distributions of selectively constrained sites and deleterious
mutation rates in the hominid and murid genomes. Mol. Biol. Evol. 27:177–92
34. Eyre-Walker A, Keightley PD. 2009. Estimating the rate of adaptive molecular evolution in the presence
of slightly deleterious mutations and population size change. Mol. Biol. Evol. 26:2097–108
35. Eyre-Walker A, Keightley PD, Smith NG, Gaffney D. 2002. Quantifying the slightly deleterious mu-
tation model of molecular evolution. Mol. Biol. Evol. 19:2142–49
36. Eyre-Walker A, Woolfit M, Phelps T. 2006. The distribution of fitness effects of new deleterious amino
acid mutations in humans. Genetics 173:891–900
37. Fairbrother WG, Holste D, Burge CB, Sharp PA. 2004. Single nucleotide polymorphism-based valida-
tion of exonic splicing enhancers. PLoS Biol. 2:E268
38. Fisher S, Grice EA, Vinton RM, Bessling SL, McCallion AS. 2006. Conservation of RET regulatory
function from human to zebrafish without sequence similarity. Science 312:276–79
39. Fisher WW, Li JJ, Hammonds AS, Brown JB, Pfeiffer BD, et al. 2012. DNA regions bound at low
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

occupancy by transcription factors do not drive patterned reporter gene expression in Drosophila. Proc.
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

Natl. Acad. Sci. USA 109:21330–35

40. Friedman RC, Farh KKH, Burge CB, Bartel DP. 2009. Most mammalian mRNAs are conserved targets
of microRNAs. Genome Res. 19:92–105
41. Fu W, Akey JM. 2013. Selection and adaptation in the human genome. Annu. Rev. Genomics Hum. Genet.
14:467–89
42. Fu W, O’Connor TD, Jun G, Kang HM, Abecasis G, et al. 2013. Analysis of 6,515 exomes reveals the
recent origin of most human protein-coding variants. Nature 493:216–20
43. Fu YX, Li WH. 1993. Statistical tests of neutrality of mutations. Genetics 133:693–709
44. Galtier N, Duret L, Glémin S, Ranwez V. 2009. GC-biased gene conversion promotes the fixation of
deleterious amino acid changes in primates. Trends Genet. 25:1–5
45. Gazave E, Marqués-Bonet T, Fernando O, Charlesworth B, Navarro A. 2007. Patterns and rates of
intron divergence between humans and chimpanzees. Genome Biol. 8:R21
46. Haddrill PR, Charlesworth B, Halligan DL, Andolfatto P. 2005. Patterns of intron sequence evolution
in Drosophila are dependent upon length and GC content. Genome Biol. 6:R67
47. Haerty W, Jagadeeshan S, Kulathinal RJ, Wong A, Ravi Ram K, et al. 2007. Evolution in the fast lane:
rapidly evolving sex-related genes in Drosophila. Genetics 177:1321–35
48. Haerty W, Ponting CP. 2013. Mutations within lncRNAs are effectively selected against in fruitfly but
not in human. Genome Biol. 14:R49
49. Halligan DL, Oliver F, Eyre-Walker A, Harr B, Keightley PD. 2010. Evidence for pervasive adaptive
protein evolution in wild mice. PLoS Genet. 6:e1000825
50. Halligan DL, Oliver F, Guthrie J, Stemshorn KC, Harr B, Keightley PD. 2011. Positive and negative
selection in murine ultraconserved noncoding elements. Mol. Biol. Evol. 28:2651–60
51. Hangauer MJ, Vaughn IW, McManus MT. 2013. Pervasive transcription of the human genome produces
thousands of previously unidentified long intergenic noncoding RNAs. PLoS Genet. 9:e1003569
52. Hedges SB. 2002. The origin and evolution of model organisms. Nat. Rev. Genet. 3:838–49
53. Heger A, Ponting CP. 2007. Variable strength of translational selection among 12 Drosophila species.
Genetics 177:1337–48
54. Hellmann I, Zöllner S, Enard W, Ebersberger I, Nickel B, Pääbo S. 2003. Selection on human genes as
revealed by comparisons to chimpanzee cDNA. Genome Res. 13:831–37
55. Hillenmeyer ME, Fung E, Wildenhain J, Pierce SE, Hoon S, et al. 2008. The chemical genomic portrait
of yeast: uncovering a phenotype for all genes. Science 320:362–65
56. Hurst LD. 2002. The Ka/Ks ratio: diagnosing the form of sequence evolution. Trends Genet. 18:486
57. Kaessmann H. 2010. Origins, evolution, and phenotypic impact of new genes. Genome Res. 20:1313–26
58. Katzman S, Kern AD, Bejerano G, Fewell G, Fulton L, et al. 2007. Human genome ultraconserved
elements are ultraselected. Science 317:915
59. Keightley PD, Eyre-Walker A. 2007. Joint inference of the distribution of fitness effects of delete-
rious mutations and population demography based on nucleotide polymorphism frequencies. Genetics
177:2251–61

www.annualreviews.org • Selection on Functional Elements 89

 GG15CH04-Ponting ARI 15 July 2014 11:5

60. Keightley PD, Kryukov GV, Sunyaev S, Halligan DL, Gaffney DJ. 2005. Evolutionary constraints in
conserved nongenic sequences of mammals. Genome Res. 15:1373–78
61. Keightley PD, Lercher MJ, Eyre-Walker A. 2005. Evidence for widespread degradation of gene control
regions in hominid genomes. PLoS Biol. 3:e42
62. Kiezun A, Pulit SL, Francioli LC, van Dijk F, Swertz M, et al. 2013. Deleterious alleles in the human
genome are on average younger than neutral alleles of the same frequency. PLoS Genet. 9:e1003301
63. Kim SY, Pritchard JK. 2007. Adaptive evolution of conserved noncoding elements in mammals. PLoS
Genet. 3:1572–86
64. Kimura M. 1983. The Neutral Theory of Molecular Evolution. Cambridge, UK: Cambridge Univ. Press
65. Klopfstein S, Currat M, Excoffier L. 2006. The fate of mutations surfing on the wave of a range expansion.
Mol. Biol. Evol. 23:482–90
66. Knowles DG, McLysaght A. 2009. Recent de novo origin of human protein-coding genes. Genome Res.
19:1752–59
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

67. Kondrashov AS, Sunyaev S, Kondrashov FA. 2002. Dobzhansky-Muller incompatibilities in protein
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

evolution. Proc. Natl. Acad. Sci. USA 99:14878–83

68. Kousathanas A, Keightley PD. 2013. A comparison of models to infer the distribution of fitness effects
of new mutations. Genetics 193:1197–208
69. Kousathanas A, Oliver F, Halligan DL, Keightley PD. 2011. Positive and negative selection on noncoding
DNA close to protein-coding genes in wild house mice. Mol. Biol. Evol. 28:1183–91
70. Kryukov GV, Schmidt S, Sunyaev S. 2005. Small fitness effect of mutations in highly conserved non-
coding regions. Hum. Mol. Genet. 14:2221–29
71. Künstner A, Nabholz B, Ellegren H. 2011. Significant selective constraint at 4-fold degenerate sites in
the avian genome and its consequence for detection of positive selection. Genome Biol. Evol. 3:1381–89
72. Kutter C, Watt S, Stefflova K, Wilson MD, Goncalves A, et al. 2012. Rapid turnover of long noncoding
RNAs and the evolution of gene expression. PLoS Genet. 8:e1002841
73. Lachance J, Vernot B, Elbers CC, Ferwerda B, Froment A, et al. 2012. Evolutionary history and adapta-
tion from high-coverage whole-genome sequences of diverse African hunter-gatherers. Cell 150:457–69
74. Lawrie DS, Messer PW, Hershberg R, Petrov DA. 2013. Strong purifying selection at synonymous sites
in D. melanogaster. PLoS Genet. 9:e1003527
75. Lewejohann L, Skryabin BV, Sachser N, Prehn C, Heiduschka P, et al. 2004. Role of a neuronal small
non-messenger RNA: behavioural alterations in BC1 RNA-deleted mice. Behav. Brain Res. 154:273–89
76. Li H, Durbin R. 2011. Inference of human population history from individual whole-genome sequences.
Nature 475:493–96
77. Lickwar CR, Mueller F, Lieb JD. 2013. Genome-wide measurement of protein-DNA binding dynamics
using competition ChIP. Nat Protoc. 8:1337–53
78. Lindblad-Toh K, Garber M, Zuk O, Lin MF, Parker BJ, et al. 2011. A high-resolution map of human
evolutionary constraint using 29 mammals. Nature 478:476–82
79. Lunter G, Ponting CP, Hein J. 2006. Genome-wide identification of human functional DNA using a
neutral indel model. PLoS Comput. Biol. 2:e5
80. Main BJ, Smith AD, Jang H, Nuzhdin SV. 2013. Transcription start site evolution in Drosophila. Mol.
Biol. Evol. 30:1966–74
81. Majewski J, Ott J. 2002. Distribution and characterization of regulatory elements in the human genome.
Genome Res. 12:1827–36
82. Marques AC, Ponting CP. 2009. Catalogues of mammalian long noncoding RNAs: modest conservation
and incompleteness. Genome Biol. 10:R124
83. McGaughey DM, Vinton RM, Huynh J, Al-Saif A, Beer MA, McCallion AS. 2008. Metrics of sequence
constraint overlook regulatory sequences in an exhaustive analysis at phox2b. Genome Res. 18:252–60
84. Meader S, Ponting CP, Lunter G. 2010. Massive turnover of functional sequence in human and other
mammalian genomes. Genome Res. 20:1335–43
85. Messer PW, Petrov DA. 2013. Frequent adaptation and the McDonald–Kreitman test. Proc. Natl. Acad.
Sci. USA 110:8615–20
86. modENCODE Consort., Roy S, Ernst J, Kharchenko PV, Kheradpour P, et al. 2010. Identification of
functional elements and regulatory circuits by Drosophila modENCODE. Science 330:1787–97

90 Haerty · Ponting
 GG15CH04-Ponting ARI 15 July 2014 11:5

87. Mouse Genome Seq. Consort. 2002. Initial sequencing and comparative analysis of the mouse genome.
Nature 420:520–62
88. Mu XJ, Lu ZJ, Kong Y, Lam HYK, Gerstein MB. 2011. Analysis of genomic variation in non-coding
elements using population-scale sequencing data from the 1000 Genomes Project. Nucleic Acids Res.
39:7058–76
89. Nakagawa S, Ip JY, Shioi G, Tripathi V, Zong X, et al. 2012. Malat1 is not an essential component of
nuclear speckles in mice. RNA 18:1487–99
90. Nakagawa S, Naganuma T, Shioi G, Hirose T. 2011. Paraspeckles are subpopulation-specific nuclear
bodies that are not essential in mice. J. Cell Biol. 193:31–39
91. Odom DT, Dowell RD, Jacobsen ES, Gordon W, Danford TW, et al. 2007. Tissue-specific transcrip-
tional regulation has diverged significantly between human and mouse. Nat. Genet. 39:730–32
92. Ohta T. 1973. Slightly deleterious mutant substitutions in evolution. Nature 246:96–98
93. Ohta T, Gillespie JH. 1996. Development of neutral and nearly neutral theories. Theor. Popul. Biol.
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

49:128–42
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

94. Osada N, Hirata M, Tanuma R, Kusuda J, Hida M, et al. 2005. Substitution rate and structural divergence
of 5 UTR evolution: comparative analysis between human and cynomolgus monkey cDNAs. Mol. Biol.
Evol. 22:1976–82
95. Parmley JL, Urrutia AO, Potrzebowski L, Kaessmann H, Hurst LD. 2007. Splicing and the evolution
of proteins in mammals. PLoS Biol. 5:e14
96. Parsch J, Novozhilov S, Saminadin-Peter SS, Wong KM, Andolfatto P. 2010. On the utility of short
intron sequences as a reference for the detection of positive and negative selection in Drosophila. Mol.
Biol. Evol. 27:1226–34
97. Pennacchio LA, Ahituv N, Moses AM, Prabhakar S, Nobrega MA, et al. 2006. In vivo enhancer analysis
of human conserved non-coding sequences. Nature 444:499–502
98. Piskol R, Stephan W. 2008. Analyzing the evolution of RNA secondary structures in vertebrate introns
using Kimura’s model of compensatory fitness interactions. Mol. Biol. Evol. 25:2483–92
99. Pollard KS, Salama SR, Lambert N, Lambot MA, Coppens S, et al. 2006. An RNA gene expressed during
cortical development evolved rapidly in humans. Nature 443:167–72
100. Ponjavic J, Ponting CP, Lunter G. 2007. Functionality or transcriptional noise? Evidence for selection
within long noncoding RNAs. Genome Res. 17:556–65
101. Ponting CP, Hardison RC. 2011. What fraction of the human genome is functional? Genome Res.
21:1769–76
102. Ponting CP, Lunter G. 2006. Signatures of adaptive evolution within human non-coding sequence.
Hum. Mol. Genet. 15(Suppl. 2):R170–75
103. Ponting CP, Nellåker C, Meader S. 2011. Rapid turnover of functional sequence in human and other
genomes. Annu. Rev. Genomics Hum. Genet. 12:275–99
104. Ponting CP, Oliver PL, Reik W. 2009. Evolution and functions of long noncoding RNAs. Cell 136:629–
41
105. Ramani AK, Chuluunbaatar T, Verster AJ, Na H, Vu V, et al. 2012. The majority of animal genes are
required for wild-type fitness. Cell 148:792–802
106. Rands CM, Meader S, Ponting CP, Lunter G. 2014. 8.2% of the human genome is constrained: variation
in rates of turnover across functional element classes in the human lineage. PLoS Genet. In press
107. Roy SW, Penny D, Neafsey DE. 2007. Evolutionary conservation of UTR intron boundaries in Cryp-
tococcus. Mol. Biol. Evol. 24:1140–48
108. Schorderet P, Duboule D. 2011. Structural and functional differences in the long non-coding RNA
Hotair in mouse and human. PLoS Genet. 7:e1002071
109. Siepel A, Bejerano G, Pedersen JS, Hinrichs AS, Hou M, et al. 2005. Evolutionarily conserved elements
in vertebrate, insect, worm, and yeast genomes. Genome Res. 15:1034–50
110. Singh ND, Larracuente AM, Clark AG. 2008. Contrasting the efficacy of selection on the X and auto-
somes in Drosophila. Mol. Biol. Evol. 25:454–67
111. Spivakov M, Akhtar J, Kheradpour P, Beal K, Girardot C, et al. 2012. Analysis of variation at transcription
factor binding sites in Drosophila and humans. Genome Biol. 13:R49

www.annualreviews.org • Selection on Functional Elements 91

 GG15CH04-Ponting ARI 15 July 2014 11:5

112. Stark A, Brennecke J, Bushati N, Russell RB, Cohen SM. 2005. Animal microRNAs confer robustness
to gene expression and have a significant impact on 3 UTR evolution. Cell 123:1133–46
113. Stone EA, Cooper GM, Sidow A. 2005. Trade-offs in detecting evolutionarily constrained sequence by
comparative genomics. Annu. Rev. Genomics Hum. Genet. 6:143–64
114. Struhl K. 2007. Transcriptional noise and the fidelity of initiation by RNA polymerase II. Nat. Struct.
Mol. Biol. 14:103–5
115. Taylor MS, Kai C, Kawai J, Carninci P, Hayashizaki Y, Semple CAM. 2006. Heterotachy in mammalian
promoter evolution. PLoS Genet. 2:e30
116. Tennessen JA, Bigham AW, O’Connor TD, Fu W, Kenny EE, et al. 2012. Evolution and functional
impact of rare coding variation from deep sequencing of human exomes. Science 337:64–69
117. Thomas LF, Sætrom P. 2012. Single nucleotide polymorphisms can create alternative polyadenylation
signals and affect gene expression through loss of microRNA-regulation. PLoS Comput. Biol. 8:e1002621
118. Thurman RE, Rynes E, Humbert R, Vierstra J, Maurano MT, et al. 2012. The accessible chromatin
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

landscape of the human genome. Nature 489:75–82

Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

119. Torgerson DG, Boyko AR, Hernandez RD, Indap A, Hu X, et al. 2009. Evolutionary processes acting
on candidate cis-regulatory regions in humans inferred from patterns of polymorphism and divergence.
PLoS Genet. 5:e1000592
120. Ulitsky I, Bartel DP. 2013. lincRNAs: genomics, evolution, and mechanisms. Cell 154:26–46
121. Vernot B, Stergachis AB, Maurano MT, Vierstra J, Neph S, et al. 2012. Personal and population genomics
of human regulatory variation. Genome Res. 22:1689–97
122. Visel A, Prabhakar S, Akiyama JA, Shoukry M, Lewis KD, et al. 2008. Ultraconservation identifies a
small subset of extremely constrained developmental enhancers. Nat. Genet. 40:158–60
123. Wakeley J, Aliacar N. 2001. Gene genealogies in a metapopulation. Genetics 159:893–905
124. Ward LD, Kellis M. 2012. Evidence of abundant purifying selection in humans for recently acquired
regulatory functions. Science 337:1675–78
125. Warnecke T, Parmley JL, Hurst LD. 2008. Finding exonic islands in a sea of non-coding sequence:
Splicing related constraints on protein composition and evolution are common in intron-rich genomes.
Genome Biol. 9:R29
126. Warnefors M, Pereira V, Eyre-Walker A. 2010. Transposable elements: insertion pattern and impact
on gene expression evolution in hominids. Mol. Biol. Evol. 27:1955–62
127. Williamson SH, Hernandez R, Fledel-Alon A, Zhu L, Nielsen R, Bustamante CD. 2005. Simultaneous
inference of selection and population growth from patterns of variation in the human genome. Proc. Natl.
Acad. Sci. USA 102:7882–87
128. Woolfe A, Goodson M, Goode DK, Snell P, McEwen GK, et al. 2005. Highly conserved non-coding
sequences are associated with vertebrate development. PLoS Biol. 3:e7
129. Xu J, Zhang R, Shen Y, Liu G, Lu X, Wu CI. 2013. The evolution of evolvability of microRNA target
sites in vertebrates. Genome Res. 23:1810–16
130. Yokoyama KD, Thorne JL, Wray GA. 2011. Coordinated genome-wide modifications within proximal
promoter cis-regulatory elements during vertebrate evolution. Genome Biol. Evol. 3:66–74
131. Young RS, Marques AC, Tibbit C, Haerty W, Bassett AR, et al. 2012. Identification and properties of
1,119 lincRNA loci in the Drosophila melanogaster genome. Genome Biol. Evol. 4:427–42
132. Zhang B, Arun G, Mao YS, Lazar Z, Hung G, et al. 2012. The lncRNA Malat1 is dispensable for mouse
development but its transcription plays a cis-regulatory role in the adult. Cell Rep. 2:111–23
133. Zhen Y, Andolfatto P. 2012. Methods to detect selection on noncoding DNA. Methods Mol. Biol. 856:141–
59

92 Haerty · Ponting
 GG15-FrontMatter ARI 14 July 2014 13:54

Annual Review of
Genomics and
Human Genetics

Volume 15, 2014 Contents

DNA and Other Strands: The Making of a Human Geneticist

Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org

Leon E. Rosenberg p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 1
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

An Opportune Life: 50 Years in Human Cytogenetics

Patricia A. Jacobs p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p29
Determinants of Mutation Rate Variation in the Human Germline
Laure Ségurel, Minyoung J. Wyman, and Molly Przeworski p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p47
No Gene in the Genome Makes Sense Except in the Light of Evolution
Wilfried Haerty and Chris P. Ponting p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p71
Phenotypic Outcomes of Imprinted Gene Models in Mice: Elucidation
of Pre- and Postnatal Functions of Imprinted Genes
Mary A.M. Cleaton, Carol A. Edwards, and Anne C. Ferguson-Smith p p p p p p p p p p p p p p p p p p p93
The Pivotal Regulatory Landscape of RNA Modifications
Sheng Li and Christopher E. Mason p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 127
Age-Related Macular Degeneration: Genetics and Biology
Coming Together
Lars G. Fritsche, Robert N. Fariss, Dwight Stambolian, Gonçalo R. Abecasis,
Christine A. Curcio, and Anand Swaroop p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 151
Disorders of Cholesterol Metabolism and Their Unanticipated
Convergent Mechanisms of Disease
Frances M. Platt, Christopher Wassif, Alexandria Colaco, Andrea Dardis,
Emyr Lloyd-Evans, Bruno Bembi, and Forbes D. Porter p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 173
The Diverse Genetic Landscape of Neurodevelopmental Disorders
Wen F. Hu, Maria H. Chahrour, and Christopher A. Walsh p p p p p p p p p p p p p p p p p p p p p p p p p p p p 195
The Genetics of Microdeletion and Microduplication Syndromes:
An Update
Corey T. Watson, Tomas Marques-Bonet, Andrew J. Sharp,
and Heather C. Mefford p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 215
The Genetics of Skin Fragility
Cristina Has and Leena Bruckner-Tuderman p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 245

vi
 GG15-FrontMatter ARI 14 July 2014 13:54

Mendelian Disorders of the Epigenetic Machinery: Tipping the

Balance of Chromatin States
Jill A. Fahrner and Hans T. Bjornsson p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 269
Deep Sequencing of HIV: Clinical and Research Applications
Shiven B. Chabria, Shaili Gupta, and Michael J. Kozal p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 295
Noninvasive Prenatal Screening by Next-Generation Sequencing
Anthony R. Gregg, Ignatia B. Van den Veyver, Susan J. Gross,
Rajeevi Madankumar, Britton D. Rink, and Mary E. Norton p p p p p p p p p p p p p p p p p p p p p p p p 327
Personalized Pharmacogenomics: Predicting Efficacy and Adverse
Drug Reactions
Annu. Rev. Genom. Hum. Genet. 2014.15:71-92. Downloaded from www.annualreviews.org
Access provided by WIB6106 - University Munchen on 04/17/16. For personal use only.

Munir Pirmohamed p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 349

Therapeutics Based on Stop Codon Readthrough
Kim M. Keeling, Xiaojiao Xue, Gwen Gunn, and David M. Bedwell p p p p p p p p p p p p p p p p p p p 371
Translating Genomics for Precision Cancer Medicine
Sameek Roychowdhury and Arul M. Chinnaiyan p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 395
The Evolutionary Genomics of Cichlid Fishes: Explosive Speciation
and Adaptation in the Postgenomic Era
Frederico Henning and Axel Meyer p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 417
The Evolutionary Origin of the Vertebrate Body Plan: The Problem
of Head Segmentation
Takayuki Onai, Naoki Irie, and Shigeru Kuratani p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 443
Emerging Issues in Public Health Genomics
J. Scott Roberts, Dana C. Dolinoy, and Beth A. Tarini p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 461
The Ethical, Legal, and Social Implications Program of the National
Human Genome Research Institute: Reflections on an Ongoing
Experiment
Jean E. McEwen, Joy T. Boyer, Kathie Y. Sun, Karen H. Rothenberg,
Nicole C. Lockhart, and Mark S. Guyer p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 481
Professional Medical Education and Genomics
Laurie A. Demmer and Darrel J. Waggoner p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 507

Errata

An online log of corrections to Annual Review of Genomics and Human Genetics articles
may be found at http://www.annualreviews.org/errata/genom

Contents vii