TIGS-1186; No.

of Pages 5

Review

Genes from scratch – the evolutionary

fate of de novo genes
Christian Schlötterer
Institut für Populationsgenetik, Vetmeduni, Veterinärplatz 1, 1210 Wien, Austria

Although considered an extremely unlikely event, many sequence similarity of orphans with other genes is permit-
genes emerge from previously noncoding genomic ted, processes like exaptation of transposable elements,
regions. This review covers the entire life cycle of such gene duplication, and horizontal gene transfer emerge as
de novo genes. Two competing hypotheses about the potential forces underlying the generation of orphan genes
process of de novo gene birth are discussed as well as [6]. Genes originating from such processes with detectable
the high death rate of de novo genes. Despite the high sequence similarities are better characterized as young
death rate, some de novo genes are retained and remain genes and should be clearly distinguished from orphan
functional, even in distantly related species, through genes sensu stricto. Mechanisms resulting in true orphans
their integration into gene networks. Further studies can be placed into four categories, which I outline here. (i)
combining gene expression with ribosome profiling in Origin of new genes from previously noncoding DNA –
multiple populations across different species will be these genes have also been called de novo genes indicating
instrumental for an improved understanding of the evo- that the ancestral sequence was not functional. (ii) Gene
lutionary processes operating on de novo genes. duplication and rapid divergence: either gene duplications
or insertions of reverse transcribed mRNA sequences into
Are orphan genes a dated concept? the genome result in duplications of already existing genes.
For many years, it had been considered extremely unlikely, It has been proposed that duplicated copies may undergo
if not impossible, that genes with no detectable homology phases of rapid evolution in a combination of neutral and
could emerge (e.g., [1]). With the availability of the full adaptive changes [4]. This rapid evolution erases the se-
genomic sequence of yeast, however, this picture changed. quence similarity with the other copies, generating an
About one third of the entire set of genes in baker’s yeast orphan gene. Despite being conceptually appealing, this
has no sequence similarity to genes from other organisms class of orphan genes is difficult to distinguish from de novo
[2]. Because nothing was known about their ancestors, genes because it is very challenging to identify historically
these new genes were termed orphans (or ORFans in rapidly evolving sequences. Hence, I treat this class jointly
the microbial world [3]). with de novo genes. (iii) Horizontal gene transfer: integra-
It has become common practice to identify orphan genes tion of foreign DNA from bacteria or viruses into the host
based on sequence similarity searches (e.g., BLAST) using genome may result in the acquisition of hitherto absent
a very relaxed significance cutoff: those genes with no hit in genes. Given the vast number of viral sequences, it is very
other species are classified as orphans [4]. The term orphan likely that the source of the acquired gene has not yet been
was not only appealing but also precise as long as only a sequenced. Although this mechanism is prevalent in pro-
few sequenced genomes were available. With an increasing karyotes, based on the current surveys of orphan genes in
number of sequenced genomes, the taxonomic sampling eukaryotes, very little support for horizontal gene transfer
became denser and the definition of orphans lost its preci- has been found [6]. (iv) Frameshift mutations (overprint-
sion: orphans could now be detected in related species, ing): N-terminal frameshifts could generate an entirely
leading to a violation of the definition. To account for this, it different protein with almost no change in the protein
has been proposed that orphans be renamed as taxonomi- coding DNA sequence (CDS) [7]. In viruses, de novo genes
cally restricted genes [5], but this concept requires an often are frequently generated without frameshifts in the ances-
arbitrary definition of the taxonomic depth to distinguish tral gene [8]. Although up to 7% of the orphan genes may
the relevant units. originate by this process [9], I suggest their evolutionary
dynamics be treated separately because their emergence is
Mechanisms giving rise to orphans frequently coupled with the loss of the progenitor gene.
Given this imprecision, it may be more informative to focus
on the biological processes generating orphan genes. When Shifting the focus from orphans to de novo genes
the definition of orphan genes is relaxed such that some Given the diversity of processes underlying orphan births
and the uncertainty surrounding orphan definition, I
Corresponding author: Schlötterer, C. ([email protected]).
propose that future studies describing the patterns of
Keywords: orphans; de novo genes; transcription; population genetics. molecular evolution focus solely on de novo genes. The
0168-9525/
unambiguous definition of de novo genes will be of key
ß 2015 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.tig.2015.02.007 importance for informative meta analyses providing a
general picture of the evolutionary dynamics of these
Trends in Genetics xx (2015) 1–5 1
TIGS-1186; No. of Pages 5

Review Trends in Genetics xxx xxxx, Vol. xxx, No. x

genes. The importance of separating novel genes according studies use mRNA expression as an indicator for functional
to the underlying molecular mechanism is emphasized de novo genes. Given that a large fraction of the genome is
by their previously documented different evolutionary transcribed [19], several researchers additionally validat-
dynamics [10]. ed the translation of these mRNAs into proteins by the
presence of the corresponding peptides in databases. Such
Are de novo genes real? databases are biased towards larger proteins, however,
De novo genes arise from previously noncoding DNA, are and de novo genes are short. This bias has motivated the
short, and are expressed at low levels [10–12]. These fea- use of other methods, such as ribosome profiling to study the
tures frequently raise doubts about the biological signifi- translation of putative de novo genes [14]. Overall, function-
cance of de novo genes. In light of these concerns, several al importance of de novo genes is well-supported by the
approaches have been used to distinguish true de novo combined evidence from mRNA and protein expression
genes from random noise. [14,15,20,21].

Neutrality tests Regulation of gene expression

Molecular evolutionary theory provides an excellent theo- Another method for assessing whether or not a de novo
retical framework for the identification of functionally gene is functional rests on the assumption that the modu-
important sequences [13]. As purifying selection operates lation of gene expression patterns reflects functional
against deleterious mutations, functionally important requirements. To this end, several studies have shown
genes have either a low frequency of or even no such that de novo genes are not constitutively expressed, but
mutations, but this is not the case in stretches of neutrally exhibit clear patterns of regulated gene expression (e.g.,
evolving sequences. Protein coding sequences provide a [9,21–24]). Liu et al. [25] not only studied differential
particularly powerful method to detect purifying selection: regulation of de novo genes during the development of
if the number of putatively deleterious nonsynonymous Drosophila melanogaster embryos, but also identified some
mutations is significantly smaller than the number of developmental stages that were enriched for the expres-
approximately neutral synonymous substitutions, this sion of de novo genes, suggesting that de novo genes may
indicates a functional gene [13]. Although old genes show preferentially acquire functional roles during some devel-
a more pronounced signal of purifying selection, de novo opmental stages.
genes differ significantly from noncoding sequences in
interspecific [12,14] and intraspecific [12,15] analyses, Reverse genetics
strongly suggesting that de novo genes are subjected to The most stringent proof of the functional relevance of
purifying selection. The codon usage of de novo genes is de novo genes comes from reverse genetics. In Drosophila,
another feature that has been attributed to selection. about 30% of young genes are essential, and constitutive
Contrary to neutrally evolving sequences, several studies silencing them mainly affected the pupal stage [26]. Con-
have demonstrated that preferred codons are enriched in stitutive knockdown of de novo genes also resulted in
de novo genes (Box 1). With selection on codon usage being lethality: out of 11 genes tested, six had an effect on
weak [16,17] and optimization of codon usage being a slow viability [22,26]. Using a tissue-specific knockdown, three
process [18], it appears unlikely that codon usage has been out of 33 D. melanogaster-specific de novo genes induced a
optimized after the de novo gene emerged. Rather, it may bristle-related phenotype [25,27]. Similarly, reverse genet-
be that the preferred usage of optimal codons facilitates the ics has also validated the functional importance of de novo
emergence of de novo genes, specifically their translation. genes in mice [23]. Taken together, these reverse genetics
experiments provide the ultimate proof that de novo genes
Gene expression: RNA and protein are functional entities rather than a random pattern
The presence of open reading frames (ORFs) alone is not occurring by chance only.
sufficient evidence for a functional gene. Therefore, many
Birth of de novo genes
The emerging picture across different species is that
Box 1. Features of de novo genes de novo genes emerge at high rates [4,6,12]. The birth of
 Short open reading frames (ORFs) [7,12,46]: the ORFs of de novo de novo genes encoding functional proteins involves two
genes are shorter than those of old genes, but longer than important steps: the acquisition of an ORF and the addi-
expected by chance [33]. tion of regulatory signals needed for transcription. The
 Fewer exons [7,12]: de novo genes have fewer exons than old sequence of events is not clear, however, and evidence for
genes.
 Microsatellites [11,12,47]: de novo genes are enriched for
both models can be found in the literature (Figure 1).
repetitive sequences.
 Usage of preferred codons [11,12]: compared to noncoding Expression first
sequences, de novo genes use the preferred codons more often. It is now widely accepted that a large fraction of the
 Low expression level [9,12,33]: on average, the expression level of
genome is being transcribed, with many long-noncoding
de novo genes is lower than that of old genes.
 Higher tissue specificity [9,11,35]: the expression of de novo RNA molecules being generated [28,29]. Interestingly, a
genes is more tissue-specific than the expression of old genes. considerable fraction of these RNAs are also associated
 Chromosomal location [12,33,35]: de novo genes are enriched on with ribosomes [28,30,31], suggesting they are actively
the X-chromosome, while very young ones that still segregate in a translated. Such short peptides form proto genes [14],
population are under-represented on the X-chromosome.
which can be subject to selection. Through the acquisition
2
TIGS-1186; No. of Pages 5

Review Trends in Genetics xxx xxxx, Vol. xxx, No. x

(A) (B)

TF

TF

(C)

TF

TRENDS in Genetics

Figure 1. Two competing models of de novo gene birth. Open reading frames (ORFs) are shown as colored blocks. Active transcription is symbolized by an arrow and
the presence of translation by a peptide. Non-neutral phases are indicated by a broken box. (A) and (B) illustrate two versions of the expression first model. (A) The
protogene model assumes that several short peptides are expressed and during the course of evolution they are combined into a larger de novo gene. (B) the ORF contains
premature stop codons (yellow circles), which prevent the translation of the expressed mRNA; only after new mutations generate a full-length ORF is the functional de novo
gene obtained. (C) The ORF first model states that a fully functional ORF is present but not expressed because the necessary regulatory signals are missing. Once new
mutations generate functional transcription factor (TF) binding sites, the de novo gene is expressed and translated.

of new mutations, proto genes can grow and result in lines of evidence support this interpretation. (i) More
functional de novo genes [14] (Figure 1A). Alternatively, strains expressed the de novo genes than expected under
it is also possible that the full-length transcript is initially neutrality. (ii) Consistent with selectively favored spread
interrupted by stop codons, but new mutations generate of the expressed de novo genes, the amount of polymor-
the full-length ORF of the de novo gene (Figure 1B). The phism around them was lower in individuals carrying the
appeal of this model is that it builds on the ubiquity of expressed variant than in those with the non-expressed
expressed genomic regions and also circumvents the im- copy. Importantly, because only the expression of a func-
plausibility problem of de novo genes, noted by [1]. This tional gene could confer a fitness advantage, this pattern
model is strongly supported by a range of studies that suggests that a new mutation resulting in the expression of
found transcription preceded the emergence of an ORF and a pre-existing ORF leads to these de novo genes becoming
translation [14,22,32]. functional.

ORF first Death of de novo genes

This model assumes that ORFs are abundant and only The high rate of de novo gene birth [4,6,12] in combination
await the acquisition of regulatory elements that control with a rather constant number of genes over time [4]
transcription and translation. Indeed, in Drosophila, about predicts that many of the de novo genes have only a short
60% of 800-bp intergenic sequences harbor ORFs of at least lifetime. Testing this prediction, however, requires a phy-
150 bp [33]. Experimental support for this hypothesis logenetic framework, which includes a range of species
comes from the analysis of a de novo gene in mice, which with different evolutionary distances [12,14,22]. Starting
suggested that all the essential functional features of the from one focal species, the origin of de novo genes can be
gene pldi were already present, but only the acquisition of dated by applying the parsimony principle to the presence
the transcription has resulted in a functional gene [23]. Be- of the de novo genes in the species studied (Figure 2). Once
cause pldi is most likely not a de novo protein-coding gene the birth of the de novo gene has been dated, its evolution-
but a noncoding RNA gene, this example may not be ary dynamics can be studied in species that diverged
representative for protein coding genes. Stronger support subsequently (Figure 2). Although lineage-specific muta-
comes from an elegant, population genetics-based ap- tion patterns and rates are certainly interesting, the ability
proach in Drosophila [33]. The authors analyzed de novo to study loss-of-function mutations (premature termina-
genes that were not expressed in closely related species, tion codons) and thus the death of de novo genes is the
but had polymorphic expression within a D. melanogaster greatest benefit of this analysis [12]. Using this approach it
population. As the statistical power of neutrality tests is has been shown that the probability of loss-of-function
low for short genes, the authors were not able to provide a mutations is higher for de novo genes than for old genes
formal proof for a selective spread of the expressed de novo [12,34]. This high death rate of young de novo genes
genes in the D. melanogaster population. Nevertheless, two explains why the total number of genes remains relatively
3
TIGS-1186; No. of Pages 5

Review Trends in Genetics xxx xxxx, Vol. xxx, No. x

t6 t5 t4 t3 t2 t1 Integrating de novo genes into already existing

Focal species
networks
Sister species The roles of de novo genes that were discussed above are
Sister species mostly related to functions that require rapid change.
Sister species Thus, the short persistence times of de novo genes nicely
Sister species fits their functional role. Nevertheless, some de novo genes
quickly become essential [22,26] and persist for longer time
Sister species
spans. This raises the question of how de novo genes could
Outgroup
become essential. The prevailing hypothesis is that de novo
genes become integrated into already existing networks.
TRENDS in Genetics
The first step is the integration into regulatory networks,
Figure 2. Phylogenetic analysis of de novo genes: de novo genes are identified in primarily through acquisition of promoters [41]. The anal-
one focal species and their age is determined by the presence of an ortholog in
sister taxa (red line). Using the parsimony criterion, the origin of the de novo gene ysis of retroposed genes indicated that regulatory elements
is set to the most recent common ancestor of the focal species and the most can be acquired rapidly from nearby genes or more distant
diverged sister species. The evolutionary stability of de novo genes can be studied positions in the genome [42]. With increasing age of the
in those lineages that diverged after the origin of the de novo gene (green lines).
regulatory landscape of de novo genes, a higher level of
complexity is developed through the gradual acquisition of
constant despite the well-documented high rate of de novo regulatory motifs [41]. The integration of de novo genes
gene birth [12]. By contrasting conserved de novo genes to into protein–protein interaction networks is significantly
those that acquired disabling mutations it was found that slower [41]. It has been proposed that protein promiscuity
GC content, gene length, and expression level were posi- (i.e., non-specific interaction) provides the basis for novel
tively correlated, and microsatellite number negatively protein–protein interactions [43]. Once established, natural
correlated, with sequence conservation [12]. Particularly selection will favor beneficial protein–protein interactions
striking was the observation that de novo genes with male- and incorporate changes stabilizing them. Interestingly,
biased gene expression were less likely to acquire prema- de novo genes were found to interact preferentially, but
ture termination codons. This differential conservation not exclusively, with genes of the same age [10]. Most likely,
may explain why previous studies identified a high number de novo genes do not acquire catalytic functions, suggesting
of de novo genes based on gene expression in testis [35,36] that they serve primarily regulatory functions in their
or showed an excess of de novo genes with male-biased gene networks [44].
expression [37].
Concluding remarks: next steps towards understanding
De novo genes in action the evolution of de novo genes
Several strategies have been pursued to explore the func- While past research has proven that genes can originate
tional contribution of de novo genes. Potentially, the most de novo and may even acquire essential functions, the
rewarding approach has been the analysis of gene expres- process of de novo gene genesis deserves more attention,
sion. Putative de novo genes were found to show a higher as does their functional characterization.
gene expression response to abiotic and biotic stressors in Until recently, the evolution of de novo genes had been
Arabidopsis thaliana than young genes with a different mainly studied in the framework of comparative geno-
evolutionary origin [9]. Surprisingly, this signal was re- mics. However, because the processes of de novo gene birth
stricted to de novo genes originating before the A. thaliana and death occur on the population level, population ge-
and A. lyrata split [9]. Similarly, putative de novo genes in netic approaches will be central to understanding these
Daphnia magna are twice as likely to be differentially processes. Population genetic theory provides an analyti-
expressed under biotic and abiotic stress than old genes cal framework for the interpretation of the selective forces
[24]. Comparing genes of different ages in yeast, de novo operating on nascent genes. Thus, the combined popula-
genes, and their precursors were enriched for binding of tion genetic analysis of DNA sequences, gene expression,
transcription factors related to stress and mating [14]. Fi- and ribosomal profiling data in multiple individuals will
nally, de novo genes had more pronounced expression shed light on the selective pressures exerted on each of
differences in a comparison of two D. melanogaster popula- these levels. Extending this analysis to multiple popula-
tions collected from different environments [33]. This strik- tions from ecologically distinct habitats as well as addi-
ing similarity across different species strongly suggests tional closely related species holds great promise to
that de novo genes are particularly important for popula- determine the evolutionary forces determining the birth
tion-specific responses to biotic and abiotic stresses. and death of de novo genes. Experimental evolution in
The function of de novo goes beyond stress response, combination with whole genome re-sequencing (evolve
however, as they were also shown to serve a vital role and re-sequence, E&R [45]), may provide an opportunity
during developmental processes [25]. One particularly to test the selective advantage of de novo genes under
interesting role was found in Drosophila, where many controlled laboratory conditions.
de novo genes are related to male reproductive processes
[33,35–38]. This class of genes shares an interesting fea- Acknowledgments
I am grateful to V. Nolte and R. Tobler and two anonymous reviewers for
ture with immune-related genes in that they may be helpful feedback on earlier versions of the manuscript. The laboratory of
involved in an arms race caused by male–male and C.S. is supported by the ERC (ArchAdapt) and the Austrian Science Fund
male–female conflicts [39,40]. (FWF, P27630, W1225, P22834).

4
TIGS-1186; No. of Pages 5

Review Trends in Genetics xxx xxxx, Vol. xxx, No. x

References melanogaster. BMC Evol. Biol. 14, http://dx.doi.org/10.1186/s12862-

1 Jacob, F. (1977) Evolution and tinkering. Science 196, 1161–1166 014-0241-9 Published online December 4, 2014
2 Dujon, B. (1996) The yeast genome project: what did we learn? Trends 26 Chen, S. et al. (2010) New genes in Drosophila quickly become
Genet. 12, 263–270 essential. Science 330, 1682–1685
3 Fischer, D. and Eisenberg, D. (1999) Finding families for genomic 27 Mummery-Widmer, J.L. et al. (2009) Genome-wide analysis of Notch
ORFans. Bioinformatics 15, 759–762 signalling in Drosophila by transgenic RNAi. Nature 458, 987–992
4 Tautz, D. and Domazet-Loso, T. (2011) The evolutionary origin of 28 Ruiz-Orera, J. et al. (2014) Long non-coding RNAs as a source of new
orphan genes. Nat. Rev. Genet. 12, 692–702 peptides. Elife 3, http://dx.doi.org/10.7554/eLife.03523 Published
5 Khalturin, K. et al. (2009) More than just orphans: are taxonomically- online September 16, 2014
restricted genes important in evolution? Trends Genet. 25, 404–413 29 Clark, M.B. et al. (2011) The reality of pervasive transcription. PLoS
6 Wissler, L. et al. (2013) Mechanisms and dynamics of orphan gene Biol. 9, http://dx.doi.org/10.1371/journal.pbio.1000625 Published online
emergence in insect genomes. Genome Biol. Evol. 5, 439–455 July 12, 2011
7 Neme, R. and Tautz, D. (2013) Phylogenetic patterns of emergence of 30 Aspden, J.L. et al. (2014) Extensive translation of small open reading
new genes support a model of frequent de novo evolution. BMC frames revealed by Poly-Ribo-Seq. Elife 3, http://dx.doi.org/10.7554/
Genomics 14, http://dx.doi.org/10.1186/1471-2164-14-117 Published eLife.03528 Published online August 21, 2014
online February 21, 2013 31 Wilson, B.A. and Masel, J. (2011) Putatively noncoding transcripts
8 Sabath, N. et al. (2012) Evolution of viral proteins originated de novo by show extensive association with ribosomes. Genome Biol. Evol. 3,
overprinting. Mol. Biol. Evol. 29, 3767–3780 1245–1252
9 Donoghue, M.T. et al. (2011) Evolutionary origins of Brassicaceae specific 32 Cai, J. et al. (2008) De novo origination of a new protein-coding gene in
genes in Arabidopsis thaliana. BMC Evol. Biol. 11, 47 http://dx.doi.org/ Saccharomyces cerevisiae. Genetics 179, 487–496
10.1186/1471-2148-11-47 Published online February 18, 2011 33 Zhao, L. et al. (2014) Origin and spread of de novo genes in Drosophila
10 Capra, J.A. et al. (2010) Novel genes exhibit distinct patterns of melanogaster populations. Science 343, 769–772
function acquisition and network integration. Genome Biol. 11, http:// 34 Yang, H. et al. (2015) Expression profile and gene age jointly shaped the
dx.doi.org/10.1186/gb-2010-11-12-r127 Published online December 27, genome-wide distribution of premature termination codons in a
2010 Drosophila melanogaster population. Mol. Biol. Evol. 32, 216–228
11 Toll-Riera, M. et al. (2009) Origin of primate orphan genes: a 35 Levine, M.T. et al. (2006) Novel genes derived from noncoding DNA in
comparative genomics approach. Mol. Biol. Evol. 26, 603–612 Drosophila melanogaster are frequently X-linked and exhibit testis-
12 Palmieri, N. et al. (2014) The life cycle of Drosophila orphan genes. Elife biased expression. Proc. Natl. Acad. Sci. U.S.A. 103, 9935–9939
3, http://dx.doi.org/10.7554/eLife.01311 Published online February 19, 36 Begun, D.J. et al. (2007) Evidence for de novo evolution of testis-
2014 expressed genes in the Drosophila yakuba/Drosophila erecta clade.
13 Li, W-H. (1997) Molecular Evolution, Sinauer Genetics 176, 1131–1137
14 Carvunis, A.R. et al. (2012) Proto-genes and de novo gene birth. Nature 37 Metta, M. and Schlötterer, C. (2008) Male-biased genes are
487, 370–374 overrepresented among novel Drosophila pseudoobscura sex-biased
15 Li, C.Y. et al. (2010) A human-specific de novo protein-coding gene genes. BMC Evol. Biol. 8, http://dx.doi.org/10.1186/1471-2148-8-182
associated with human brain functions. PLoS Comput. Biol. 6, http:// Published online June 24, 2008
dx.doi.org/10.1371/journal.pcbi.1000734 Published online March 26, 38 Begun, D.J. et al. (2006) Recently evolved genes identified from
2010 Drosophila yakuba and D. erecta accessory gland expressed
16 Wallace, E.W. et al. (2013) Estimating selection on synonymous codon sequence tags. Genetics 172, 1675–1681
usage from noisy experimental data. Mol. Biol. Evol. 30, 1438–1453 39 Chapman, T. (2006) Evolutionary conflicts of interest between males
17 Hershberg, R. and Petrov, D.A. (2008) Selection on codon bias. Annu. and females. Curr. Biol. 16, R744–R754
Rev. Genet. 42, 287–299 40 Lessells, C.M. (2006) The evolutionary outcome of sexual conflict.
18 Powell, J.R. and Gleason, J.M. (1996) Codon usage and the origin of P Philos. Trans. R. Soc. Lond. B: Biol. Sci. 361, 301–317
elements. Mol. Biol. Evol. 13, 278–279 41 Abrusan, G. (2013) Integration of new genes into cellular networks, and
19 Berretta, J. and Morillon, A. (2009) Pervasive transcription constitutes their structural maturation. Genetics 195, 1407–1417
a new level of eukaryotic genome regulation. EMBO Rep. 10, 973–982 42 Kaessmann, H. et al. (2009) RNA-based gene duplication: mechanistic
20 Knowles, D.G. and McLysaght, A. (2009) Recent de novo origin of and evolutionary insights. Nat. Rev. Genet. 10, 19–31
human protein-coding genes. Genome Res. 19, 1752–1759 43 Wu, D.D. and Zhang, Y.P. (2013) Evolution and function of de novo
21 Wu, D.D. et al. (2011) De novo origin of human protein-coding genes. originated genes. Mol. Phylogenet. Evol. 67, 541–545
PLoS Genet. 7, http://dx.doi.org/10.1371/journal.pgen.1002379 Published 44 Arendsee, Z.W. et al. (2014) Coming of age: orphan genes in plants.
online November 10, 2011 Trends Plant Sci. 19, 698–708
22 Reinhardt, J.A. et al. (2013) De novo ORFs in Drosophila are important 45 Schlötterer, C. et al. (2014) Combining experimental evolution with
to organismal fitness and evolved rapidly from previously non-coding next-generation sequencing: a powerful tool to study adaptation from
sequences. PLoS Genet. 9, http://dx.doi.org/10.1371/journal.pgen. standing genetic variation. Heredity http://dx.doi.org/10.1038/
1003860 Published online October 17, 2013 hdy.2014.86 Published online October 1, 2014
23 Heinen, T.J. et al. (2009) Emergence of a new gene from an intergenic 46 Ekman, D. and Elofsson, A. (2010) Identifying and quantifying orphan
region. Curr. Biol. 19, 1527–1531 protein sequences in fungi. J. Mol. Biol. 396, 396–405
24 Colbourne, J.K. et al. (2011) The ecoresponsive genome of Daphnia 47 Guo, W.J. et al. (2007) Significant comparative characteristics between
pulex. Science 331, 555–561 orphan and nonorphan genes in the rice (Oryza sativa L.) genome.
25 Liu, H.Q. et al. (2014) Integrative analysis of young genes, positively Comp. Funct. Genomics http://dx.doi.org/10.1155/2007/21676 Published
selected genes and lncRNAs in the development of Drosophila online 2007