ArticlePDF Available

Two Antarctic penguin genomes reveal insights into their evolutionary history and molecular changes related to the Antarctic environment

Authors:

Abstract and Figures

Penguins are flightless aquatic birds widely distributed in the Southern Hemisphere. The distinctive morphological and physiological features of penguins allow them to live an aquatic life, and some of them have successfully adapted to the hostile environments in Antarctica. To study the phylogenetic and population history of penguins and the molecular basis of their adaptations to Antarctica, we sequenced the genomes of the two Antarctic dwelling penguin species, the Adélie penguin [Pygoscelis adeliae] and emperor penguin [Aptenodytes forsteri]. Phylogenetic dating suggests that early penguins arose ~60 million years ago, coinciding with a period of global warming. Analysis of effective population sizes reveals that the two penguin species experienced population expansions from ~1 million years ago to ~100 thousand years ago, but responded differently to the climatic cooling of the last glacial period. Comparative genomic analyses with other available avian genomes identified molecular changes in genes related to epidermal structure, phototransduction, lipid metabolism, and forelimb morphology. Our sequencing and initial analyses of the first two penguin genomes provide insights into the timing of penguin origin, fluctuations in effective population sizes of the two penguin species over the past 10 million years, and the potential associations between these biological patterns and global climate change. The molecular changes compared with other avian genomes reflect both shared and diverse adaptations of the two penguin species to the Antarctic environment.
Content may be subject to copyright.
R E S E A R C H Open Access
Two Antarctic penguin genomes reveal insights
into their evolutionary history and molecular
changes related to the Antarctic environment
Cai Li
1,2
,YongZhang
1
,JianwenLi
1
, Lesheng Kong
3
,HaofuHu
1
,HailinPan
1
,LuohaoXu
1
,YuanDeng
1
,QiyeLi
1,2
,
Lijun Jin
1
, Hao Yu
1
, Yan Chen
1
,BinghangLiu
1
, Linfeng Yang
1
, Shiping Liu
1
, Yan Zhang
1
, Yongshan Lang
1
,
Jinquan Xia
1
,WeimingHe
1
, Qiong Shi
1
, Sankar Subramanian
4
, Craig D Millar
5
, Stephen Meader
3
, Chris M Rands
3
,
Matthew K Fujita
3,20
,MatthewJGreenwold
6
,ToddACastoe
7,8
,DavidDPollock
7
,WanjunGu
9
, Kiwoong Nam
10,21
,
Hans Ellegren
10
, Simon YW Ho
11
,DavidWBurt
12
, Chris P Ponting
3
,ErichDJarvis
13
, M Thomas P Gilbert
2,14
,
Huanming Yang
1,15
,JianWang
1
, David M Lambert
4*
, Jun Wang
1,15,16,17,18*
and Guojie Zhang
1,19*
Abstract
Background: Penguins are flightless aquatic birds widely distributed in the Southern Hemisphere. The distinctive
morphological and physiological features of penguins allow them to live an aquatic life, and some of them have
successfully adapted to the hostile environments in Antarctica. To study the phylogenetic and population history of
penguins and the molecular basis of their adaptations to Antarctica, we sequenced the genomes of the two Antarctic
dwelling penguin species, the Adélie penguin [Pygoscelis adeliae] and emperor penguin [Aptenodytes forsteri].
Results: Phylogenetic dating suggests that early penguins arose ~60 million years ago, coinciding with a period of
global warming. Analysis of effective population sizes reveals that the two penguin species experienced population
expansions from ~1 million years ago to ~100 thousand years ago, but responded differently to the climatic cooling of
the last glacial period. Comparative genomic analyses with other available avian genomes identified molecular changes
in genes related to epidermal structure, phototransduction, lipid metabolism, and forelimb morphology.
Conclusions: Our sequencing and initial analyses of the first two penguin genomes provide insights into the timing
of penguin origin, fluctuations in effective population sizes of the two penguin species over the past 10 million years,
and the potential associations between these biological patterns and global climate change. The molecular changes
compared with other avian genomes reflect both shared and diverse adaptations of the two penguin species to the
Antarctic environment.
Keywords: Penguins, Avian genomics, Evolution, Adaptation, Antarctica
Background
Sphenisciformes (penguins), an avian order comprising
six extant genera and 18 species [1], are flightless aquatic
birds widely distributed in the Southern Hemisphere.
Although all extant penguins have completely lost the
capacity for aerial flight, they employ modified flipper-
like wings in wing-propelled diving or underwater flight
[2]. To be competent for an underwater life, penguins
have undergone multiple morphological adaptations. For
instance, penguins have developed densely packed, scale-
like feathers which are good for waterproof and thermal
insulation [3,4]; their eye lens and visual sensitivity of
penguins are adapted for the efficiency of underwater
predation [5-7]; to overcome buoyancy force in water,
penguins have developed dense bones [8] and stiff wing
joints [9], and reduced the distal wing musculature
[9,10].
* Correspondence: d.lambert@griffith.edu.au; wangj@genomics.cn;
zhanggj@genomics.cn
Equal contributors
4
Environmental Futures Centre, Griffith University, Nathan QLD 4111, Australia
1
China National GeneBank, BGI-Shenzhen, Shenzhen 518083, China
Full list of author information is available at the end of the article
© 2014 Li et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain
Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article,
unless otherwise stated.
Li et al. GigaScience 2014, 3:27
http://www.gigasciencejournal.com/content/3/1/27
Penguins are the most common birds in Antarctica.
Among 18 extant penguin species, eight (emperor (Apte-
nodytes forsteri), king (A. patagonicus), Adélie (Pygoscelis
adeliae), chinstrap (P. antarctica), gentoo (P. papua),
macaroni (Eudyptes chrysolophus), royal (E. chrysolophus)
and rockhopper (E. chrysocome)) live in the Antarctic
and sub-Antarctic areas, and two of them (Adélie and
emperor) make the Antarctic continent as their major
habitats [11,12]. Antarctica is one of the most hostile
environments on earth. The penguins living in Antarctica
are subject to extremely low temperatures, high winds,
and profound seasonal changes in the length of daylight
[13]. To live in such a harsh environment, penguins have
developed a complicated system in the head, wing, and
legs for enhanced thermoregulation [14,15], and an effect-
ive management of energy storage for long-term fasting
[16-18]. Because of their important roles in the Antarctic
ecosystem and their sensitive responses to changes in
marine and Antarctic climate, penguins are also among
the widely studied organisms in climate change research
[19-22].
The unique morphology and remarkable life histories
of penguins have attracted broad interest from scientists
as well as the general public. However, most previous
studies focused on ecological, physiological, behavioral,
or phylogenetic aspects of their biology, whereas the
molecular genetic bases of penguin adaptations remain
largely unknown. As part of the avian phylogenomics
project [23,24], we sequenced the genomes of two Ant-
arctic dwelling penguins (Adélie and emperor penguins)
in order to understand the evolutionary history of pen-
guins as well as the genomic and molecular bases of
their adaptations to the Antarctic environment.
Data description
A male emperor penguin captured from Emperor Island
near Zhongshan Station and a male Adélie penguin from
Inexpressible Island in the Ross Sea were used for DNA
collection and sequencing. Using the Illumina Genome
Analyzer platform, we generated more than 60× coverage
of usable reads for each of the two penguins (Additional
file 1: Table S1) [25]. The assembled draft genomes of
Adélie and emperor penguins resulted in contig N50 sizes
of 19.1 kb and 30.5 kb, and scaffold N50 sizes of 5.0 Mb
and 5.1 Mb, respectively (Table 1). The assemblies of 1.17
Gb (54 Mb of gaps) for Adélie penguin and 1.19 Gb
(72 Mb of gaps) for emperor penguin cover >85% of the
estimated genome sizes of 1.25 Gb and 1.39 Gb based
k-mer analysis (Additional file 2: Figure S1), respectively.
The GC content of the two genomes is 41.7% and 41.8%
(Table 1), towards the lower end of the range of the 48
birds sequenced (40.5-44.8%) [23].
A total of 15,270 and 16,070 protein-coding genes
were annotated in Adélie and emperor penguin genomes
(Table 1), respectively. We annotated 598 and 627 non-
coding RNAs (ncRNA) in Adélie and emperor genomes, of
which 172 (Adélie) and 180 (emperor) were microRNAs
(miRNAs). 6.47% (Adélie) and 7.38% (emperor) in the two
assemblies are predicted to be repetitive elements (Table 1;
Additional file 3: Table S2). However, the proportions of
repetitive elements in the penguin genomes should be lar-
ger, because the unassembled regions tend to contain many
repetitive elements. By mapping the reads of short insert
size libraries to the genome assembly of each penguin,
we identified 2,559,440 and 3,410,305 heterozygous sites in
emperor and Adélie, respectively. We obtained 8295 1:1
orthologs of 48 birds (including Adélie and emperor pen-
guins) from the avian phylogenomics project [23,24], and
the CDS alignments of these orthologs were used to analyze
the ratio of nonsynonymous substitution rate to synonym-
ous substitution rate (dN/dS) in penguin lineages.
Analyses
Phylogenetic relationships of two penguins and closely
related aquatic species
In the avian phylogenomics study [24], we produced a
highly resolved phylogenetic tree of 48 avian species
representing almost all extant avian orders with whole-
genome phylogenetic signals. The two penguins are in a
relatively basal position within a clade of aquatic birds
(Figure 1A). Molecular dating of 48 birds was performed in
the main companion study [24] using 19 fossil calibrations,
including the earlier Waimanu penguin fossil. We esti-
mated that penguins diverged from their closest relatives,
the order Procellariiformes (represented by the genome of
northern fulmar Fulmarus glacialis), ~60.0 million years
ago (MYA) with a 95% credibility interval (CI) of 56.8-62.7
MYA (Figure 1A). It is notable that global temperature
increased dramatically during the period from 60 MYA to
50 MYA [26] (Figure 1A), and at approximately 55.0 MYA,
global temperature rose by ~6°C within ~20,000 years [27].
This event, known as the PaleoceneEocene Thermal
Maximum (PETM), was associated with sea level rise
and a massive benthic extinction event [28]. The global
warming and massive extinction might have provided an
opportunityfortheoffshorebirdstoexpandtheirhabitatto
the sea, leading to the emergence of ancient penguins. The
estimated divergence time between the two penguin species
[24] is 23.0 MYA (95% CI: 6.9-42.8 MYA) (Figure 1A),
which is slightly more ancient than that reported in a
recent study using a few genes (11.715.4 MYA) [29].
Analysis of effective population sizes
The population dynamics of Adélie and emperor penguins
arestronglyinfluencedbytheAntarcticenvironment
and climatic variation. Based on the heterozygous sites
identified in the penguin genomes, we used the pairwise
sequentially Markovian coalescent (PSMC) method [30]
Li et al. GigaScience 2014, 3:27 Page 2 of 15
http://www.gigasciencejournal.com/content/3/1/27
to infer fluctuations in the effective population sizes of
the two penguins from 10 MYA to 10 thousand years
ago (KYA). From 10 MYA to 1 MYA, both species had
relatively small and stable effective population sizes
of <100,000, and the populations expanded gradually
from ~1 MYA (Figure 1B). The effective population size
of the Adélie penguin appears to have increased rapidly
after ~150 KYA, at a time when the penultimate glaciation
period ended and the climate became warmer. This
expansion is consistent with the prediction in a previous
study based on mitochondrial data from two Adélie
penguin lineages [31] and with the recent observations
that Adélie populations expanded when more ice-free
locations for nesting became available [32]. Notably,
at ~60 KYA, within a relatively cold and dry period
called Marine Isotope Stage 4 (MIS4) [33] in the last
glacial period, the effective population size of Adélie
penguins declined by ~40% (Figure 1B and C). By contrast,
the effective population size of emperor penguin remained
at a stable level during the same period.
It has been suggested that environmental conditions
during the last glacial period were favorable to emperor
penguins, because they do not require ice-free breeding
grounds and are able to incubate their eggs on their feet
and use an abdominal fold of skin to protect their eggs
from freezing temperatures [34]. It has also been hypothe-
sized that during the last glacial maximum (LGM, ~26.5-19
KYA, Figure 1C) [35], all penguin species except the
emperor penguin were displaced from Antarctica because
of the complete loss of nesting grounds and limited food
sources [34]. The contrasting patterns in PSMC-based
effective population sizes of the two penguins are consist-
ent with this hypothesis. However, because the estimated
divergence time between the two penguins has a large
Table 1 Basic statistics of assembly and annotation of the two penguin genomes
Species Contig N50 length (bp) Scaffold N50 length (bp) Assembly size (bp) (G + C)% Repeats (%) #Protein coding genes
Adélie 19,134 5,047,175 1,226,103,150 41.7% 6.47 15,270
Emperor 30,463 5,071,598 1,257,483,768 41.8% 7.38 16,070
Figure 1 Phylogenetic relationships and changes in effective population sizes of two penguin species. (A) Phylogeny of two penguins
and six closely related aquatic species (northern fulmar Fulmarus glacialis; great cormorant Phalacrocorax carbo; crested ibis Nipponia nippon; dalmatian
pelican Pelecanus crispus;littleegretEgretta garzetta; red-throated loon Gavia stellata) (blue names), and a land bird (zebra finch Taeniopygia guttata).
The estimates of topology and divergence times are from our avian phylogenomic study [24]. Horizontal bars at each node represent 95% credibility
intervals of estimated divergence times. Above the tree are the geological timescale and temperature changes over the past 65 million years, relative
to the present [26]. PETM, PaleoceneEocene Thermal Maximum. (B) Dynamic changes of effective population sizes (N
e
) of two penguins inferred by
the pairwise sequentially Markovian coalescent (PSMC) method. The thick curves depict the estimated N
e
values of the two penguins, and the thin
curves represent PSMC bootstrapping estimates. (C) Enlargement of the period from 100 KYA to 10 ka in panel (B). MIS 4, Marine Isotope Stage 4;
LGM, last glacial maximum. Temperature change data are from [33].
Li et al. GigaScience 2014, 3:27 Page 3 of 15
http://www.gigasciencejournal.com/content/3/1/27
95% CI, we cannot accurately date the population decline
in the Adélie penguin. Further studies will be needed to
resolve the exact timing of this decline.
Epidermis-related genes
Penguins exhibit many unique epidermal features (includ-
ing feathers) in comparison with other birds. The penguin
epidermis has a thick stratum corneum, consisting of
flattened heavily keratinized cells that lack basophilic
nuclear remnants [36]. The feathers of penguins are
short (30-40 mm), stiff, and evenly packed over the
body surface to help minimize heat loss, rather than
arranged in tracts as in other birds [37]. Beta(β)-keratins
make up 90% of the barbs and barbules of mature
feathers, and duplication and diversification of β-keratin
genes are known to play important roles in the diversi-
fication of avian feathers [38,39]. In the four subfamilies
of β-keratin genes [39], we found that the numbers of
keratinocyte β-keratin genes in the Adélie and emperor
penguins are among the highest of all avian species
(Figure 2A: emperor, 15; Adélie, 13) and only two other
birds have 13 keratinocyte β-keratin genes (Pekin duck
Anas platyrhynchos, 14; killdeer Charadrius vociferus,13).
The numbers of penguin keratinocyte β-keratin genes
are significantly larger than those of other aquatic birds
(7.1 ± 2.9 (mean ± S.D.); phylogenetic ANOVA, p <0.03)
and non-aquatic birds (6.8 ± 2.7; phylogenetic ANOVA,
p <0.01) (Figure 2A). Phylogenetic reconstruction of the
keratinocyte β-keratin genes indicates the two penguin
species have undergone several lineage-specific gene
duplications since their divergence from their closest
aquatic relatives (Figure 2B). And keratinocyte β-keratins
are expressed in both skin and feathers [38,40]. In
addition, the EVPL gene, which encodes the protein
envoplakin as a component of the cornified envelope of
keratinocytes [41], has undergone positive selection in
the ancestral lineage of the two penguins (branch-site
model in CODEML [42], likelihood-ratio test (LRT)
p=2.32×10
-14
). Another gene, DSG1, also predicted to
have evolved under positive selection in the ancestral
penguin lineage (CODEML branch-site model, LRT
p=1.33×10
-11
), is involved in a human dermatological
disorder characterized by thickening of the skin on the
palms and soles [43]. The expansion of keratinocyte β-
keratin genes and the positive selection on EVPL and
DSG1 may have contributed to generating the unique
skin and feathers in penguins.
Pseudogenes and positively selected genes involved in
phototransduction
The aquatic lifestyle and marked seasonal changes in the
length of daylight in Antarctica could affect the visual
abilities as well as non-visual phototransduction of pen-
guins. By comparing the genomes of 48 avian species,
we found that most birds had four (tetrachromatic) classes
of cone opsin genes, while Adélie and emperor penguins
Emperor penguin
Adélie penguin
Other aquatic birds
Non-aquatic birds
# of keratinocyte β-keratins
p < 0.03
p < 0.01
AB
0101214
16
PYGAD
APTFO
FULG
L
FULGL
APTFO
APTFO
APTFO
NIPNI
PELCR
APTF
O
FULGL
P
ELCR
EGRGA
NIPNI
EGRGA
PELCR
PELCR
FULGL
EGRGA
APTFO
PYGAD
PYGAD
PYGAD
APTF
O
APTFO
NIPNI
A
PTFO
EGRGA
PYGAD
PEL
C
R
PYGAD
PYGAD
PYGAD
APTFO
A
PTFO
PYGAD
PYGAD
NIPNI
PHACA
EGRGA
APTFO
PYGAD
PYGAD
FULGL
NIPNI
FULGL
PHACA
PHACA
NIPNI
PELCR
PYGAD
PELCR
FULGL
APTFO
APTFO
PHACA
APTFO
NIPNI
EGRGA
2468
Figure 2 Penguin-specific duplications of keratinocyte β-keratin genes. (A) Numbers of keratinocyte β-keratin genes in the two penguins and
other birds. Error bars indicate standard deviations. P-values were calculated by phylogenetic ANOVA. (B) Phylogenetic cladogram of keratinocyte
β-keratin genes of the Adélie penguin (PYGAD, in blue), emperor penguin (APTFO, in red), and five aquatic relatives (northern fulmar, FULGL; crested
ibis, NIPNI; great cormorant, PHACA; little egret, EGRGA; dalmatian pelican, PELCR). Shading in the tree indicates putative penguin-specific gene
duplication events.
Li et al. GigaScience 2014, 3:27 Page 4 of 15
http://www.gigasciencejournal.com/content/3/1/27
had only three (trichromatic) classes due to the pseudo-
genization of Rh2 [23]. In addition, we found that the
pinopsin gene OPSP, which is specifically expressed in
the pineal gland and involved in circadian rhythms
[44], has been pseudogenized in the two penguin species
(Additional file 4: Figure S2). This provides a potential
molecular explanation for a previous observation of the
absence of typical photoreceptor elements in the pineal
organ of gentoo penguins (Pygoscelis papua) [45], which is
a congener of the Adélie penguin. However, the mutations
leading to pseudogenization of OPSP in Adélie and em-
peror penguins were found in different codon positions
(Additional file 4: Figure S2), suggesting the pseudogeniza-
tion might have occurred independently in each lineage.
Intriguingly, we detected signals of positive selection
in either of the two penguin lineages on several genes
involved in phototransduction (CODEML branch-site
model, Additional file 5: Table S3; CNGB1,MYO3A, and
UACA in the emperor lineage and CRB1,CRY2 and
MYO3B in the Adélie lineage), suggesting different adap-
tations for light transduction in the two penguins. In
particular, CNGB1, which codes for a subunit of the cyc-
lic nucleotide-gated cation channel in retinal rods that is
important for visual perception [46], contains numerous
positively selected sites in the emperor penguin lineage
(Figure 3A). The different sets of positively selected
phototransduction genes in the two penguins might be
related to their different reproductive strategies Adélie
penguins reproduce in spring and summer with long
days [47], whereas emperor penguins reproduce in the
winter with short days [48].
Positively selected genes associated with lipid metabolism
The storage of fat is critical for penguins to withstand
cold and survive the long fasting periods (up to four
months in male emperor penguins) [18]. We found
eight, three, and four genes involved in lipid metabolism
exhibiting signals of positive selection in Adélie, emperor,
and their ancestral lineage of the two penguins, respect-
ively (CODEML branch-site model, Additional file 6:
Table S4). The gene FASN, which encodes fatty acid
synthase and plays essential roles in de novo lipogenesis,
exhibits significant positive selection in Adélie penguin
(LRT p = 8.78 × 10
-5
) and the ancestral penguin lineage
(LRT p=2.77×10
-4
), with some of the selected sites
located in its functional domains (Figure 3B, and
Additional file 6: Table S4). In contrast, no evidence for
positive selection on FASN was found in the emperor
penguin lineage. As with the genes involved in photo-
transduction, the two penguins seem to have exploited
different adaptive pathways for lipid metabolism in the
course of their evolution. Because the climate prior to
the divergence of the two penguins was warmer than
that after their divergence, this could potentially explain
why a large fraction of lipid-related positively selected
genes were found in the two individual penguin lineages,
rather than in the ancestral lineage.
Forelimb-related genes with non-neutral amino acid
substitutions
During their evolutionary history, the wings (or forelimbs)
of penguins changed profoundly for wing-propelled diving
in the water [9]. Although we did not find any dN/dS-
based positively selected genes and pseudogenes in the
ancestral penguin lineage that are linked to the forelimb
adaptation, we identified 17 forelimb-related genes (of 134
genes examined) harboring non-neutral penguin-specific
amino acid changes that might affect protein functions
(Additional file 7: Table S5), using Protein Variation Effect
Analyzer (PROVEAN) [49]. The EVC2 gene is of particu-
lar interest because it harbors five predicted non-neutral
substitutions (Figure 3C), the largest number of non-
neutral substitutions among the 17 candidate genes.
Mutations of EVC2 in human can cause Ellis-van Creveld
syndrome, patients of which manifest anomalies such
as short-limb dwarfism, short ribs, and postaxial poly-
dactyly [50], resembling some phenotypes in the wings
of penguins. Furthermore, another gene involved in Ellis-
van Creveld syndrome, EVC, also contains a predicted
non-neutral substitution. These genes may serve as candi-
dates for future functional studies.
Discussion
Genome sequencing for species living in extreme envi-
ronments has great potential to provide new insights
into the molecular basis of adaptation to the extreme
environments. For example, population genomics analysis
of polar bears revealed positively selected genes associated
with cardiomyopathy and vascular disease, implying im-
portant reorganization of the cardiovascular system in polar
bears to adapt to the Arctic environment [51]. The genome
of the Tibetan antelope exhibits signals of positive selection
and gene-family expansion in genes associated with energy
metabolism and oxygen transmission, suggesting high-
altitude adaptation in these genes [52]. Furthermore, the
recently published midge genome (Belgica antarctica)is
the first Antarctic eukaryote genome, and has a very com-
pact architecture which is thought to be constrained by
environmental extremes in Antarctica [53]. Given their
large populations and long history in Antarctica, the
Antarctic penguins are an excellent model for studying
how animals adapt to harsh environments, and how
climate changes affect the population dynamics.
Our sequencing and initial analyses of the two Antarctic
dwelling penguin species (Adélie and emperor) have shed
light on the timing of penguin origin and on the effective
population size changes of the two penguin species
over the past 10 million years. We found evidence of
Li et al. GigaScience 2014, 3:27 Page 5 of 15
http://www.gigasciencejournal.com/content/3/1/27
associations between these biological patterns and global
climate change. In particular, the contrasting patterns in
effective population sizes of the two penguins during the
last glacial period provide evidence for some previously
proposed hypotheses about how different penguin species
responded to climate change in the past. Morphological
changes in the epidermis and forelimbs are critical for
underwater flight in penguins, so the candidate genes that
0 500 1000 1500 2000 2500
0.5 0.6 0.7 0.8 0.9 1.0
Amino acid position
Posterior probability
KS MAT ER KR TE
NADP binding
NADP binding
Acyl carrier
EVC2_like(PF12297)
A
B
C
CNGB1
FASN
EVC2
0 200 400 600 800 1000 1200
0.5 0.6 0.7 0.8 0.9 1.0
Amino acid position
Posterior probability
Cytoplasmic Transmembrane
cAMP binding
Cytoplasmic
Figure 3 Cases of positively selected sites and non-neutral penguin-specific amino acid changes. (A) Positively selected sites in emperor
penguin CNGB1 protein sequence. Cytoplasmic and transmembrane regions are separated by the dashed lines. Blue shading represents the
membrane-spanning helix, and the cAMP binding domain is shown in grey. The posterior probabilities were calculated using BEB method in
CODEML. (B) Positively selected sites in the FASN protein in the Adélie lineage (green dots) and the ancestral lineage (blue dots). The molecular
binding domains of FASN are shown in light red, whereas the major catalytic domains are shown in grey. From left to right, beta-ketoacyl synthase
(KS), acyl and malonyl transferases (MAT), enoyl reductase (ER), beta-ketoacyl reductase (KR), and thioesterase (TE). The posterior probabilitieswere
calculated using BEB method in CODEML. (C) Non-neutral penguin-specific amino acid changes in the EVC2 protein. One substitution site is located in
the Pfam domain EVC_like (PF12297).
Li et al. GigaScience 2014, 3:27 Page 6 of 15
http://www.gigasciencejournal.com/content/3/1/27
we discovered in this study are highly valuable for future
functional studies. The genes involved in light trans-
duction and lipid metabolism exhibit signals of positive
selection or pseudogenization in penguins, suggesting
their evolutionary responses to the extreme conditions of
light and temperature in Antarctica. The pseudogeniza-
tion events also show examples of relaxed constraints in
the two penguins. Interestingly, we not only found shared
patterns in the molecular evolution of the two penguin
species, but also found distinct patterns between them,
such as the genes involved in phototransduction and lipid
metabolism. This implies that the diversity of molecular
evolution in different penguin species deserves further
investigation.
The genomic resources and the results presented here
lay the foundation for further genomic and molecular
studies of penguins. Given the genome sequences of two
penguin species, conducting other omicsstudies, such
as trascriptomics and population genomics, becomes
achievable in the near future. Based on the candidate
genes identified in this study, future work can involve
more in-depth experiments to investigate the functional
roles of target genes. The divergence times of modern
penguins remain somewhat unclear, and whole-genome
sequencing of other penguin species will help more
precise estimates to be obtained. Overall, we believe
that the two penguin genomes will likely facilitate related
research, such as penguin biology, avian evolution, polar
biology and climate changes.
Methods
Genome sequencing and assembly
DNA preparation
The male emperor penguin was captured from Emperor
Island near Zhongshan Station, and the male Adélie
Penguin was collected from Inexpressible Island in the
Ross Sea, Antarctica. High molecular weight genomic DNA
(>100 kbp) was extracted from the peripheral venous blood
of the two penguins. All the work done in this project
followed guidelines and protocols for research on animals,
and had the necessary permits and authorization.
DNA library construction and sequencing
Sequencing and assembling of the two penguin genomes
both followed the whole-genome shotgun approach. All
of the raw reads were generated using the Illumina
Genome Analyzer platform.
1) Short insert paired-end DNA libraries (size range
from approximately 200 bp to 500 bp)
Seven DNA paired-end (PE) libraries (insert size
of ~200 bp, ~350 bp and ~500 bp) (Additional
file 1: Table S1) were constructed for the emperor
penguin, and four DNA PE libraries (insert size
of ~200 bp and ~500 bp) (Additional file 1: Table
S1) were prepared for the Adélie penguin using the
following steps: 1) We fragmented 5 μg of the
genomic DNA by Adaptive Focused Acoustic
(Covaris); 2) The DNA end was polished and a
dATP was added to the 3end of the fragment;
3) The DNA adaptors were ligated with a dTTP
overhang at the 3end of the fragment; 4) The
ligated fragments were size-selected at 200 bp,
350 bp, 500 bp on agarose gels to yield the
corresponding short insert libraries. After 10 cycles
of PCR, the DNA fragments of the appropriate size
were excised and purified for sequencing.
2) Long insert mate-pair DNA libraries (size range from
ca. 2 kbp to 20 kbp)
Eight mate-pair libraries were generated for the
emperor penguin, and five mate-pair libraries were
constructed for the Adélie penguin (Additional file
1: Table S1). Long-insert mate-pair libraries were
generated based on circularization and random
fragmentation. Then, 20-50 μg of genomic DNA
was fragmented using the hydroshear apparatus to
obtain the concentrated DNA smear. The products
were end-polished and biotin-labeled with the
biotinylated dNTP. After different insert-size
libraries were size-selected at 2 kb, 5 kb, 10 kb, and
20 kb, 1 μg of biotinylated DNA was circularized to
join the two ends, and the linear DNA was digested.
The circularized DNA was fragmented randomly
using the Covaris apparatus to about 400-600 bp
smear, and the biotinylated DNA fragments were
enriched by biotin/streptomycin on the surface of
magnetic beads. End polishing, A-tailing, adaptor
ligation, and PCR of 18 cycles were all performed on
M-280 beads.
All DNA libraries were sequenced on Illumina GAII
or Hiseq-2000 platforms in PE 50 cycles, PE 91 cycles or
PE101 cycles.
Read filtering
For the de novo data we avoid mistakes from man-made
or technology-system errors by a series of checking and
filtering measures on reads generated by the Illumina-
Pipeline.
1) Discard reads in which N or polyA sequence
constitutes more than 10% of bases.
2) Discard low quality reads. Reads that have 40 bases
with Q20 less than or equal to 7 for the large
insert-size library data and 50 bases for the short
ones were filtered.
3) Discard reads with adapter contamination. Reads
with more than 10 bp aligned to the adapter
Li et al. GigaScience 2014, 3:27 Page 7 of 15
http://www.gigasciencejournal.com/content/3/1/27
sequence (allowing less than or equal to 3 bp
mismatches) were removed.
4) Discard small insert-size reads in which read1 and
read2 overlapped more than or equal to 10 bp
allowing a 10% mismatch. Read1 and read2 are ends
of one paired-end reads.
5) Discard PCR duplicates. When read1 and read2 of
two paired end reads are identical, these reads were
considered to be duplicates and were discarded.
In total, we obtained approximately 78.45 Gb of reads
for the Adélie Penguin and approximately 80.69 Gb for
the emperor penguin (Additional file 1: Table S1).
Estimating the genome size using k-mer frequencies
We used a method described in the panda genome paper
[54] to estimate the genome sizes of the two penguins.
First, we used corrected data of short insert libraries
to calculate the 17-mer distribution (Additional file 2:
FigureS1).Wecanestimatethegenomesizeas
G = K_num/K_depth, where K_num is the total num-
ber of kmers, and K_depth is the peak frequency (i.e.
the mode) which occurs more often than all others
(approximate sequencing depth). We otained G
Adélie
=
K_num/K_depth =38,874,560,013/31 = 1254,018,064
(bp), and G
emperor
= K_num/K_depth =31,909,923,872/
23 = 1,387,387,994 (bp).
Genome assembly
We used SOApdenovo2 [54] to construct contigs (kmer
size =19) and scaffolds, and filled gaps in the intra-scaffolds
using GapCloser (a companion program released with
SOAPdenovo). Total scaffold lengths of 1,257,483,768 bp
and 1,226,103,150 bp; N50 values of 5,071,598 bp and
5,047,175 bp; and, contig N50 values of 31,902 bp and
19,134 bp were obtained for the emperor and Adélie
penguins, respectively (Additional file 8: Table S6). There
were 80,973 and 108,883 gaps in the emperor and Adélie
assemblies, covering 71,985,685 bp (5.72% of assembly size)
and 54,269,351 bp (4.43% of assembly size), respectively.
Core Eukaryotic Genes Mapping Approach (CEGMA)
[55] is a pipeline that can be used to evaluate the com-
pleteness of a genome assembly by annotating the 248
ultra-conserved Core Eukaryotic Genes (CEGs). The
CEGMA results (Additional file 9: Table S7) revealed that
the completeness of the emperor assembly is close to that
of chicken and zebra finch assemblies [56,57]. Although
the Adélie assembly is not as good as the emperor, its
completeness is close to the published turkey assembly
[58]. One main reason for the relatively worse assembly
of Adélie is that Adélie has a higher heterozygosity rate,
which was also shown in the k-mer frequecy curve
(Additional file 2: Figure S1). The high heterozygosity
rate resulted in shorter contigs during the assembling.
We also used the whole-genome alignment between
the penguin and zebra finch to assess the rearrangement
events. We first generated a raw whole-genome alignment
with LASTZ [59], and then used the chainNet package [60]
to generate the reciprocal best net alignment. The chainNet
results predicted two kinds of rearrangement events (trans-
location events with the nonSynlabel and inversion event
with invlabel in netSyntenicsoutput).Tohaveacontrol,
we also generated the reciprocal best net alignment be-
tween the previously published peregrine falcon assembly
[61] and zebra finch assembly. We obtained very similar
numbers of rearrangement events (relative to zebra finch)
in the Adélie and emperor penguins, and the peregrine
falcon (Additional file 10: Table S8).
Genome annotation
Repeat annotation
We identified known transposable elements (TEs) in the
two penguin genomes using RepeatMasker (version 3.2.6)
[62] against the Repbase [63] library (version 15.01). We
also used RepeatModeler (version 1.44) [64] to construct
the de novo TE library for each penguin genome. Repeat-
Masker was used again with the de novo libraries to identify
new TEs in both genomes. We predicted tandem repeats
using Tandem Repeat Finder (TRF) [65] (version 4.00),
with parameters set to Match =2, Mismatch =7, Delta =7,
PM =80, PI =10, Minscore =50, and MaxPeriod =12.The
statistics of repeat sequences of two penguins are listed in
Additional file 3: Table S2.
Protein-coding gene annotation
The protein-coding gene annotation of the two penguins
were from the avian phylogenomic project (APP here-
after) [23,24]. Methodological details of gene annota-
tion can be found in the avian comparative genomics
study [23].
Non-coding RNA (ncRNA) annotation
We used tRNAscan-SE [66] to identify transfer RNA
genes. The snRNA genes were predicted by INFERNAL
[67] software against the Rfam database [68]. First we
ran BlastN against the Rfam sequence with an E-value
acceptance threshold of 1. All hits were then extended
and collected as input for INFERNAL [67]. The micro-
RNA (miRNA) genes of two penguins were predicted
using two independent methods, and then combined to
make a non-redundant set.
1) Method A
First, hairpin sequences from miRBase [69] (V.16)
were used as query sequences and then aligned to
the penguin genome assemblies using WU-BLAST
[70]. The outputs that matched longer than 20 bp
were extended to the length of the query sequences
Li et al. GigaScience 2014, 3:27 Page 8 of 15
http://www.gigasciencejournal.com/content/3/1/27
as putative miRNA. To determine how likely the
putative fragment resembled miRNA, Randfold [71]
was run and sequences were retained with minimum
free folding energy smaller than -15 kcal/mol and
p-value smaller than 0.05. An RNAshapes [72]
analysis was performed with the single-stem
probabilityofthesequencelargerthan0.9to
identify the putative miRNA, which folded into a
simple stem-loop structure more than 99% of
miRNAs from miRBase had a simple stem-loop
structure. Subsequently, miRNAs were retained
when its seed region was 100% identical to the
querysequenceandthematurepartwasbetter
than 90% conserved. Finally, we reserved the hit
with the highest overall percent alignment identity
for each locus as the miRNA sequence.
2) Method B
We performed a WU-BLAST search, as in method
A, to obtain putative miRNAs. To acquire precursor
miRNAs, three filters were used in parallel for each
hit: 1) Randfold with 1000 iterations per sequence,
the outputs were filtered with Minimum Free Energy
(MFE) smaller than -20 kcal/mol, and p-value
smaller than 0.015; 2) PRSS [73], which works by
constructing local alignments. A PRSS analysis was
run with 1000 iterations in order to confirm the
homology between the two sequences. The outputs
were filtered with the following cutoff: E-value
smaller than 10
-5
and similarity greater than 0.65; 3)
We performed a global alignment between the
subjectsequence(theextendedhit)andquery
sequence (hairpin sequence from miRBase) with
T-COFFEE [74], and the outputs were filtered with
the similarity of >0.95. The three outputs were
merged together as Set B.
3) Combined miRNA predictions from Set A and Set B
The two miRNA prediction sets were combined:
predicted miRNAs were considered to represent a
single locus if genome coordinates overlapped on
the same DNA strand, reserving the highest global
identity sequence as the final precursor miRNA.
Putative microRNAs were checked to remove
repetitive sequence and transposable elements.
Finally, 172 and 180 miRNAs were identified in
Adélie penguin and emperor penguin, respectively.
The statistics of annotated ncRNAs are provided in
Additional file 11: Table S9.
Phylogenetics and effective population size analysis
Phylogeny of two penguins and other birds
The phylogeny and divergence times for the two penguins
with six closely related aquatic species (northern fulmar,
great cormorant, crested ibis, dalmatian pelican, little
egret, and red-throated loon) and a land bird (zebra finch)
were derived from the ExaML TENT tree as described in
the avian phylogenomics study [24].
Analysis of effective population sizes
We first identified heterozygous sites in each of the two
penguin genomes. For each penguin, we used SOAPa-
ligner [75] to map all the reads of short insert size
libraries to the genome assembly, not allowing indels in
the alignments. Based on the short-read alignments, we
used SOAPsnp [76] to identify the heterozygous single
nucleotide polymorphisms (SNPs). We performed add-
itional screening to reduce false positives, keeping only can-
didates with: 1) quality score 20; 2) sequence depth >20;
3) the approximate copy number of flanking sequences <2;
4) at least 1 uniquely mapped read for each allele; and 5) a
minimum distance between SNPs 5bp.
We utilized PSMC [30] to infer the population histories
of the two penguins. In order to evaluate the substitution
rate of the two penguins, which is required for PSMC ana-
lysis, we aligned the two penguin genomes using LASTZ
(v1.01.50) [59] with parameters T = 2 C = 2 H = 2000
Y = 3400 L = 6000 K = 2200. We calculated their substitu-
tion rate as: μ= (C/L)/2(T/g) =79551994/1066586108/2
(22998300/5) =8.11×10
-9
substitutions per site per gen-
eration; where C is the number of the mismatch loci
between the two penguin genomes, without insertions
or deletions; L is the length of the aligned sequences,
without insertions or deletions; T is the divergence time
between the two penguins; g = 5 is the generation time,
according to [48].
After filtering SNPs located within repeat elements and
putative scaffolds of the Z chromosome (based on align-
ment against zebra finch chromosomes), we obtained
2,559,440 and 3,410,305 SNPs in emperor and Adélie
assemblies, respectively. SNPs in the assemblies of the
two penguins were replaced by degenerate bases and
converted to PSMC fasta-style sequence. We then ran
PSMC with parameters -N30 -t15 -r5 -p 4 + 25*2 + 4 + 6.
We also performed 100 bootstraps to estimate uncertainty
in the estimates. We identified a population shrinking of
ancestral Adélie penguins during the last glaciation. We
used the δ18O data in [26,33] to measure the temperature
changes relative to the present; a δ18O increase of 0.22
is considered to be equivalent to a 1°C (1.8°F) cooling [77].
dN/dS analysis
Ortholog identification
We obtained 8295 1:1 orthologs of 48 birds (including
Adélie and emperor penguins) and the corresponding
alignments from APP [23,24], then applied the methods
of ortholog assignment and alignment as described in
the avian phylogenomics study [24]. CDS alignments of
these orthologs were used to analyze the ratio of
Li et al. GigaScience 2014, 3:27 Page 9 of 15
http://www.gigasciencejournal.com/content/3/1/27
nonsynonymous substitution rate to synonymous substi-
tution rate (dN/dS).
Branch model
We sought to identify genes with accelerated dN/dS values
in the penguin lineages. We investigated three different
scenarios: accelerated dN/dS in the Adélie lineage, in the
emperor lineage, and in the ancestral lineage of the two
penguins. We ran the two-ratio branch model (one dN/dS
for the investigated branch, another dN/dS for other
branches; set parameters model =2, NSsites =0, fix_o-
mega =0) and one-ratio model (one dN/dS estimate for
all branches, as null model; set parameters model =0,
NSsites =0, fix_omega =0) using CODEML within the
PAML package [42]. After obtaining the results of two-
ratio and one-ratio models, we performed LRTs to obtain
the p-values for quantifying the significance of accelerated
evolution. False discovery rates (FDR) were computed
using the Benjamini-Hochberg procedure to adjust for
multiple testing. With a FDR cut-off of 0.05, we obtained
245, 123, and 72 genes that had accelerated dN/dS (fast
evolving genes) in the Adélie lineage, emperor lineage and
the ancestral penguin lineage, respectively. We performed
Gene Ontology (GO) enrichment analysis on the three
lists of genes. Only the list of fast-evolving genes in
the emperor penguin exhibits enriched GO categories
(Additional file 12: Table S10).
Because the results of this dN/dS analysis could have
been affected by incomplete gene sequences or incorrect
alignments, we manually checked the genes of particular
interest (e.g., genes involved in vision, lipid; see related
sections below). We checked the CDS alignments and
removed very short gene sequences that could bias the
analysis, and re-ran PAML on these revised alignments. In
some cases, if the gene models appeared to be incorrectly
annotated, we also performed re-annotation to obtain
better gene models.
Branch-site model
We also ran branch-site models with CODEML in PAML
to identify the genes containing positively selected sites
in the penguin lineages. As with the branch models, we
considered three different scenarios: the Adélie lineage,
emperor lineage, and the ancestral lineage of the two
penguins. The parameters for the null model were set
as model =2, NSsites =2, fix_kappa =0, fix_omega =1,
omega =1, while the parameters for the alternative
model were set as model =2, NSsites =2, fix_kappa =0,
fix_omega =0. LRT and FDR were computed as for the
branch models. With an FDR cut-off of 0.05, we obtained
382, 225, and 107 positively selected genes in Adélie,
emperor, and the ancestral penguin lineage, respectively.
We performed GO enrichment analysis on the three lists
of genes but did not find any enriched GO category.
As with the branch models, we performed an additional
manual check for the genes of particular interest (see later
sections). We checked whether the surrounding align-
ments of the positively selected sites were reliable. For
suspicious alignments (which had low percent identities
or many gaps), we removed problematic sequences and
reran PAML on these revised alignments.
Penguin-specific amino acid changes
We extracted the sub-alignments for two penguins and six
closely related aquatic birds (Fulmarus glacialis, Pelecanus
crispus, Egretta garzetta, Nipponia nippon, Phalacrocorax
carbo, and Gavia stellate) from the protein alignments of
8295 orthologs [24]. Based on these alignments, we identi-
fied 14,751 penguin-specific amino acid changes (one geno-
type in both of the two penguins and another genotype in
their close relatives) in 4922 genes, including deletions and
insertions (Additional file 13: Table S11).
We used PROVEAN v1.1 [49] to predict whether a
single amino acid substitution or an indel has an impact
on the biological function of a protein. For each variation,
PROVEAN introduced a score indicating the functional
effects of this variation, and we used the default cutoff
of -2.5 to determine whether the effect was non-neutral
(affecting the protein function) or neutral. Under this cri-
terion, we detected 1887 genes that harbor non-neutral
amino acid changes in penguins (Additional file 13: Table
S11). We also performed GO enrichment analysis on
these genes. Most of the enriched GO terms were related
to basic cellular functions (Additional file 14: Table S12).
Gene family expansion and contraction
In order to investigate gene family evolution in penguins,
we performed gene clustering and detected gene family
expansion or contraction based on the clustering results.
We chose six closely related aquatic birds (Fulmarus
glacialis, Pelecanus crispus, Egretta garzetta, Nipponia
nippon, Phalacrocorax carbo, and Gavia stellate)and
zebra finch as the outgroup for gene clustering with the
two penguins.
First, all-vs-all BLASTp for all protein sequences of
nine species was performed to obtain alignments with
an E-value upper threshold of 10
-5
. Then BLASTp hits
were further filtered if the alignment length was smaller
than 25% of the query or target length. Based on filtered
BLASTp hits, hcluster_sg(v0.5.0) in Treefam [78] was
used to cluster the genes (parameters for hcluster_sg: -w
10 -s 0.34 m500 -b 0.1), and the resulting gene clusters
were considered to be gene families. The basic statistics of
gene clustering are listed in Additional file 15: Table S13.
We used CAFE [79] to identify potential gene families
under significant expansion or contraction but failed to
find any, probably because the difference in copy numbers
in most clusters was too small. We then used the
Li et al. GigaScience 2014, 3:27 Page 10 of 15
http://www.gigasciencejournal.com/content/3/1/27
Wilcoxon rank sum test to identify gene families for
which numbers in the two penguins are significantly
different from those of the other seven birds. We fur-
ther manually checked the significant families to ensure
that the gene annotation and clustering result were
correct. Finally, we obtained 10 potentially expanded
and three contracted gene families (Additional file 16:
Table S14). Note that the expansion of beta-keratin
genesdescribedlaterinthisstudywasnotpredicted
with this method, because the beta-keratin genes were
annotated and analyzed separately.
Analyses of genes of particular interest
Alpha and beta-keratins
The epidermis (including feathers) of penguins possesses
many unique features compared with those of other
birds. Two multigene families, alpha (α) and beta (β)kera-
tins, play important roles in the formation of the general
epidermis and epidermal appendages of birds (e.g., claws,
scales, beaks, and feathers) [39,80-83]. Therefore, we
decided to investigate αand β-keratins in the two pen-
guins and attempted to identify their differences between
these two penguins and their close relatives.
The annotation of the α-andβ-keratins of 48 birds
were obtained from [38], as were the copy numbers of
subfamilies of β-keratins in each bird. We did not find
significant differences in gene numbers of α-keratins of
the two penguins and the six other closely related
aquatic birds (Additional file 17: Table S15). The eight
investigated aquatic birds have very similar numbers of
α-keratins, ranging from 32 to 36.
For the four subfamilies of β-keratins, the copy numbers
of claw, scale, and feather β-keratin subfamilies tended
to be affected by sequencing depth (Additional file 18:
Figure S3) and the copy numbers of two penguins did not
show clear differences with two high-depth sequenced
close relatives (little egret and crested ibis) (Additional
file 19: Table S16). However, we found significantly higher
numbers of keratinocyte β-keratins in the two penguins
(13 for Adélie, 15 for emperor) compared with other
closely related aquatic birds (Additional file 19: Table S16).
The difference in copy numbers was probably not due to
sequencing depth (Additional file 18: Figure S3). The two
high-depth genomes of birds that are close relatives of
penguins, little egret and crested ibis, only contain six and
seven keratinocyte β-keratin genes, respectively. Therefore
the keratinocyte β-keratins of penguins might have under-
gone an independent expansion and contributed to the
unique features of the epidermis of penguins. We also
generated the protein sequence alignments of keratinocyte
β-keratins of seven species (Aptenodytes forsteri, Pygoscelis
adeliae, Fulmarus glacialis, Pelecanus crispus, Egretta gar-
zetta, Nipponia nippon, and Phalacrocorax carbo)using
Prank (v.140110) [84], and inferred the phylogenetic tree
using RAxML (v8.0.14) (parameters: -m GTRGAMMA -f
a -# 1000). The phylogeny inferred by RAxML is shown in
Additional file 20: Figure S4.
Genes involved in phototransduction
In order to understand the light-sensing ability of
penguins, we investigated several opsin genes in the
penguin genomes, including RH1 (rhodopsin), RH2 (green
light-sensitive), SWS1 (violet light-sensitive), SWS2 (blue
light-sensitive), LWS (red light-sensitive), OPN3 (encepha-
lopsin), OPN4 (melanopsin), OPN5 (neuropsin), and OPSP
(pinopsin). We used protein sequences for these genes
from the chicken genome as our reference and used
TBLASTN to find the gene locations in the genome with
an E-value cutoff of 10
-5
. GeneWise was used to predict
the gene structure when the alignment length was more
than 50% of the query sequence. The protein sequences of
predicted genes were extracted and the reciprocal best hits
method was used to determine the orthology relationships
with chicken opsin genes. The gene structures were then
manually checked to investigate whether frameshift sites
and premature stop codons were caused by errors in
assembly or annotation. Among these genes, OPSP was
found to be pseudogenized. We found two frameshifts
in OPSP of emperor penguin, and one frameshift and
one premature stop codon in OPSP of Adélie penguin
(Additional file 21: Table S17; Additional file 4: Figure S2).
We did not identify LWS,SWS1, or SWS2 in either of the
two penguin genomes. Upon searching for these three
genes in other avian genomes, only a few birds were found
to have these genes. The absence of these genes might be
due to incomplete genome assemblies.
Based on the CODEML branch-site model test, we
also identified some phototransduction-related genes that
exhibit positive selection in Adélie penguin and emperor
penguin lineages respectively. We further manually
checked this result and excluded genes or sites with
suspicious alignments. We finally found three genes with
high confidence in each of the two penguins (Additional
file 5: Table S3).
Lipid-related genes
Based on the results of CODEML branch models
(FDR <0.05), we found seven, two, and four candidate
lipid-related genes that had accelerated dN/dS in the
Adélie lineage, emperor lineage, and ancestral penguin
lineage. We further manually checked the results of
CODEML, and redid the alignment and CODEML ana-
lysis for those with suspicious alignments. Following this
process, we found six, two, and three candidate genes with
high confidence (Additional file 22: Table S18).
Based on the results of CODEML branch-site models
(FDR <0.05), we found eight, five, and four candidate
lipid-related genes that contained positively selected sites
Li et al. GigaScience 2014, 3:27 Page 11 of 15
http://www.gigasciencejournal.com/content/3/1/27
in the Adélie, emperor, and ancestral penguin lineage.
We further manually checked the results of CODEML,
and redid the alignment and CODEML analysis for
those with suspicious alignments. Following this process,
we found eight, three, and four candidate genes with
high confidence (Additional file 6: Table S4). Among
these candidate genes, the most interesting is the FASN
gene, which encodes fatty acid synthase and is essential
for fatty acid synthesis. We observed positive selection
signals in FASN in the Adélie and ancestral lineages.
Forelimb-related genes
The wings (or forelimbs) of penguins have been heavily
modified for wing-propelled diving over the course of
evolution [2]. We downloaded a list of forelimb-related
genes (250 genes) from MGI (MGI id: MP:0000550
abnormal forelimb morphology) [85]. 59 of the 250
genes were found in 8295 orthologs mentioned in previ-
ous sections. We did not find any overlap between the
59 genes and the gene loss list and the positively selected
gene list (ancestral lineage) described above. However,
11 of 59 forelimb-related genes contained non-neutral
penguin-specific amino acid changes predicted to affect
protein function (see the Penguin-specific amino acid
changessection).
In addition to these 59 genes in 8295 orthologs, we
generated orthologs for additional 75 genes using recipro-
cal best BLAST hits. Furthermore, we generated multiple
sequence alignments for these 75 genes using PRANK
[84]. Of the 75 orthologs, we found 17 genes harboring
penguin-specific amino acid changes and six genes were
predicted to harbor sites predicted to affect protein func-
tion with PROVEAN [49].
In total, we found 43 forelimb-related genes harbor-
ing penguin-specific amino acid changes in penguins,
with 17 predicted to harbor non-neutral amino acid
changes (Additional file 23: Table S19 and Additional
file 7: Table S5).
Availability of supporting data
The raw sequencing reads of the two penguins have been
deposited in NCBI under accession numbers of SRA129317
and SRA129318. The datasets (assembly and annotation
files) supporting the results of this article have been depos-
ited in the GigaScience database, GigaDB [86,87].
Additional files
Additional file 1: Table S1. Sequencing data generated for the Adélie
and emperor penguins.
Additional file 2: Figure S1. Distribution of 17-mer frequency in the
sequencing reads of short-insert libraries after correction. We used all
reads from the short insert-size libraries (<1000 bp). The peak depth for
Adélie and emperor are 31 and 23, respectively.
Additional file 3: Table S2. Statistics of repeat annotation in the Adélie
and emperor penguins. The predicted elements by TRF were merged with
the tandem repeats predicted by RepeatMasker. Othersrefers to the repeats
that can be classified by RepeatMasker, but not included by the classes above;
Unknownrefers to the repeats that cant be classified by RepeatMasker.
Additional file 4: Figure S2. The premature stop codon and frameshift
sites in OPSP.
Additional file 5: Table S3. Phototransduction-related genes that exhibit
positive selection in Adélie and emperor penguin lineages. #species
indicates the number of avian species used for analysis. P-values were
calculated by likelihood-ratio test based on the results of modified model A
(alternative model) and corresponding null model (fixed ω=1).
Additional file 6: Table S4. Lipid-related genes with positively selected
sites in Adélie, emperor, or the ancestral lineages, predicted by PAML
branch-site models. #speciesindicates the number of avian species used
for analysis. P-values were calculated by likelihood-ratio test based on the
results of modified model A (alternative model) and corresponding null
model (fixed ω=1).
Additional file 7: Table S5. Non-neutral penguin-specific amino acid
changes in forelimb-related genes predicted by PROVEAN.
Additional file 8: Table S6. Basic statistics of genome assemblies of
the two penguins.
Additional file 9: Table S7. CEGMA results. Prots = number of 248
ultra-conserved CEGs present in genome; %Completeness = percentage
of 248 ultra-conserved CEGs present.
Additional file 10: Table S8. Assessment of rearrangements based on
whole-genome alignments against zebra finch assembly.
Additional file 11: Table S9. Non-coding RNA genes in the genomes.
Additional file 12: Table S10. Enriched GO terms in fast evolving
genes in emperor penguin lineage. A cutoff of 0.05 for the FDR adjusted
p-values was used.
Additional file 13: Table S11. Non-neutral penguin-specific amino acid
changes found in the 8295 ortholog alignments, predicted by PROVEAN.
Additional file 14: Table S12. Enriched GO terms in genes harboring
non-neutral penguin-specific amino acid changes. A cutoff of 0.05 for the
FDR adjusted p-values was used.
Additional file 15: Table S13. Basic statistics of gene clustering with
Treefam.
Additional file 16: Table S14. Expanded and contracted gene families
in two penguins. The numbers indicate the family sizes in each species. The
abbreviation of each species indicates the gene number in the family.
APFTO: Aptenodytes forsteri,PYGAD:Pygoscelis adeliae, FULGL:Fulmarus
glacialis, PELCR: Pelecanus crispus, EGRGA: Egretta garzetta, NIPNI: Nipponia
nippon, PHACA: Phalacrocorax carbo, GAVST: Gavia stellate,TAEGU:
Taeniopygia guttata. P-values were calculated using Wilcoxon rank sum tests.
Additional file 17: Table S15. Type I and II α-keratins for two penguins
and six closely related aquatic birds.
Additional file 18: Figure S3. Correlation analysis between sequencing
depth and copy number of each beta-keratin subfamily. The copy numbers
of claw, scale, and feather β-keratin subfamilies are positively correlated with
sequencing depth (p <0.05, Pearsons test), but there is no significant
correlation between sequencing depth and copy number of keratinocyte
beta-keratins.
Additional file 19: Table S16. The numbers for four β-keratin subfamilies
for the Adélie and emperor penguins and six closely related aquatic birds.
Additional file 20: Figure S4. RAxML phylogeny of keratinocyte
β-keratins. Adélie penguin (PYGAD, in blue), emperor penguin (APTFO, in
red), and five aquatic relatives (northern fulmar, FULGL; crested ibis, NIPNI;
great cormorant, PHACA; little egret, EGRGA; dalmatian pelican, PELCR).
Additional file 21: Table S17. Number of frameshift and premature
stop codons in opsin genes.
Additional file 22: Table S18. Lipid-related genes with accelerated
dN/dS in Adélie, emperor, or the ancestral lineages, predicted by PAML
branch models. #speciesindicates the number of avian species used for
Li et al. GigaScience 2014, 3:27 Page 12 of 15
http://www.gigasciencejournal.com/content/3/1/27
analysis. P-values were calculated by likelihood-ratio test based on the
results of one-ratio model (null) and two-ratio model (alternative).
Additional file 23: Table S19. Penguin-specific amino acid changes in
forelimb-related genes.
Abbreviations
KYA: Thousand years ago; LRT: Likelihood-ratio test; MYA: Million years ago.
Competing interests
The authors declare that they have no competing interests.
Authorscontributions
JuW, GZ and DML designed the study; JL, YC, BL, LY, SL, YaZ, YL, JX, WH, QS,
TAC, DDP and WG performed genome assembly and annotation; CL, YoZ, JL,
LK, HH, HP, LX, YD, QL, LJ, HY, SS, CDM, SM, CMR, MKF, MJ, KN, HE, and
SYWH performed the comparative genomic analyses; CL, YoZ, JL, LK, JuW, GZ
wrote the manuscript; DWB, CPP, EDJ, MTPG, HuY, JiW helped improve the
manuscript. All authors read and approved the final manuscript.
Acknowledgements
The majority of this study was supported by internal funding from BGI. GZ was
supported by a Marie Curie International Incoming Fellowship grant (300837).
CL was partially supported by the Lundbeck Foundation grant R52-A5062. LK
has been funded by the European Research Council (Project Reference 249869,
DARCGENs). DML acknowledges the support from Australian Research Council
(LP110200229) and Australia India Strategic Research Fund. DDP acknowledges
the support of the National Institutes of Health (NIH; GM083127) . We thank
Jianzhi Zhang (University of Michigan) and Huabin Zhao (Wuhan University) for
commenting on early drafts of the manuscript.
Author details
1
China National GeneBank, BGI-Shenzhen, Shenzhen 518083, China.
2
Centre for
GeoGenetics, Natural History Museum of Denmark, University of Copenhagen,
Øster Voldgade 5-7, 1350 Copenhagen, Denmark.
3
MRC Functional Genomics
Unit, Department of Physiology, Anatomy and Genetics, University of Oxford,
South Parks Road, Oxford OX1 3QX, UK.
4
Environmental Futures Centre, Griffith
University, Nathan QLD 4111, Australia.
5
Allan Wilson Centre for Molecular Ecology
and Evolution, School of Biological Sciences, University of Auckland, Private Bag
92019, Auckland, New Zealand.
6
Department of Biological Sciences, University of
South Carolina, Columbia, SC, USA.
7
Department of Biochemistry and Molecular
Genetics, School of Medicine, University of Colorado, Aurora, CO 80045, USA.
8
Biology Department, University of Texas Arlington, Arlington, TX 76016, USA.
9
Research Centre of Learning Sciences, Southeast University, Nanjing 210096,
China.
10
Department of Evolutionary Biology, Uppsala University, Norbyvagen 18D,
SE-752 36 Uppsala, Sweden.
11
School of Biological Sciences, University of Sydney,
Sydney,NSW2006,Australia.
12
Department of Genomics and Genetics, The Roslin
Institute and Royal (Dick) School of Veterinary Studies, University of Edinburgh,
Easter Bush Campus Midlothian, Edinburgh EH25 9RG, UK.
13
Department of
Neurobiology, Howard Hughes Medical Institute, Duke University Medical Center,
Durham NC27710, USA.
14
Trace and Environmental DNA Laboratory, Department
of Environment and Agriculture, Curtin University, Perth, WA 6102, Australia.
15
Princess Al Jawhara Center of Excellence in the Research of Hereditary Disorders,
King Abdulaziz University, Jeddah 21589, Saudi Arabia.
16
Department of Biology,
University of Copenhagen, Ole Maaløes Vej 5, 2200 Copenhagen, Denmark.
17
Macau University of Science and Technology, Avenida Wai long, Taipa, Macau
999078, China.
18
Department of Medicine, University of Hong Kong, Hong Kong,
Hong Kong.
19
Centre for Social Evolution, Department of Biology,
Universitetsparken 15, University of Copenhagen, Copenhagen DK-2100, Denmark.
20
Current address: Department of Biology, University of Texas at Arlington,
Arlington, TX 76019, USA.
21
Current address: Bioinformatics Research Centre (BiRC),
Aarhus University, C.F.Møllers Allé 8, 8000 Aarhus C, Denmark.
Received: 1 September 2014 Accepted: 6 November 2014
Published: 12 December 2014
References
1. Gill F, Donsker D: IOC World Bird List (v 4.1). 2014.
2. Ksepka DT, Ando T: Penguins past, present, and future: trends in the
evolution of the Sphenisciformes. In Living Dinosaurs. Edited by Dyke G,
Kaiser G. Oxford: Wiley; 2011:155186.
3. Watson M: Report on the Anatomy of the Spheniscidae Collected by HMS
Challenger, During the Years 18731876. Edinburgh: Neill and Company; 1883.
4. Taylor JRE: Thermal insulation of the down and feathers of pygoscelid
penguin chicks and the unique properties of penguin feathers. Auk 1986,
103:160168.
5. Sivak JG: The role of a flat cornea in the amphibious behaviour of the
blackfoot penguin (Spheniscus demersus). Can J Zool 1976, 54:13411345.
6. Sivak JG, Millodot M: Optical performance of the penguin eye in air and
water. J Comp Physiol 1977, 119:241247.
7. Bowmaker JK, Martin GR: Visual pigments and oil droplets in the penguin,
Spheniscus humboldti.J Comp Physiol A 1985, 156:7177.
8. Meister W: Histological structure of the long bones of penguins. Anat Rec
1962, 143:377387.
9. Raikow RJ, Bicanovsky L, Bledsoe AH: Forelimb joint mobility and the
evolution of wing-propelled diving in birds. Auk 1988, 446451.
10. Schreiweis DO: A comparative study of the appendicular musculature of
penguins (Aves: Sphenisciformes). Smithsonian Contrib Zool 1982, 341:146.
11. Wilson GJ: Distribution and abundance of Antarctic and sub-Antarctic
penguins: a synthesis of current knowledge. Cambridge: SCAR; 1983.
12. Woehler EJ, Croxall JP: The status and trends of Antarctic and sub-Antarctic
seabirds. Marine Ornithology 1997, 25:4366.
13. Goldsmith R, Sladen WJ: Temperature regulation of some antarctic
penguins. J Physiol 1961, 157:251262.
14. Frost PGH, Siegfried WR, Greenwood PJ: Arterio-venous heat exchange
systems in the Jackass penguin Spheniscus demersus.J Zool 1975,
175:231241.
15. Thomas DB, Fordyce RE: The heterothermic loophole exploited by
penguins. Aust J Zool 2008, 55:317321.
16. Groscolas R: Metabolic adaptations to fasting in emperor and king
penguins. In Penguin Biology. Edited by Davis LS, Darby JT. San Diego:
Academic; 1990:269296.
17. Cherel Y, Gilles J, Handrich Y, Le Maho Y: Nutrient reserve dynamics and
energetics during long-term fasting in the king penguin (Aptenodytes
patagonicus). J Zool 1994, 234:112.
18. Groscolas R, Robin JP: Long-term fasting and re-feeding in penguins.
Comp Biochem Physiol A Mol Integr Physiol 2001, 128:645655.
19. Barbraud C, Weimerskirch H: Emperor penguins and climate change.
Nature 2001, 411:183186.
20. Ainley DG: The Adélie Penguin: Bellwether of Climate Change. New York:
Columbia University Press; 2002.
21. Forcada J, Trathan PN: Penguin responses to climate change in the
Southern Ocean. Glob Chang Biol 2009, 15:16181630.
22. Jenouvrier S: Impacts of climate change on avian populations. Glob
Chang Biol 2013, 19:20362057.
23. Zhang G, Li C, Li Q, Li B, Larkin DM, Lee C, Storz JF, Antunes A, Greenwold
MJ, Meredith RW, Ödeen A, Cui J, Zhou Q, Xu L, Pan H, Wang Z, Jin L,
Zhang P, Hu H, Yang W, Hu J, Xiao J, Yang Z, Liu Y, Xie Q, Yu H, Lian J, Wen
P, Zhang F, Li H, et al:Comparative genomics reveal insights into avian
genome evolution and adaptation. Science 2014, in press.
24. Jarvis ED, Mirarab S, Aberer AJ, Li B, Houde P, Li C, Ho SYW, Faircloth BC,
Nabholz B, Howard JT, Suh A, Weber CC, Fonseca RR, Li J, Zhang F, Li H,
Zhou L, Narula N, Liu L, Ganapathy G, Boussau B, Bayzid MS, Zavidovych V,
Subramanian S, Gabaldón T, Gutiérrez SC, Huerta-Cepas J, Rekepalli B,
Munch K, Schierup M, et al:Whole genome analyses resolve the early
branches to the tree of life of modern birds. Science 2014, in press.
25. Cai Q, Qian X, Lang Y, Luo Y, Xu J, Pan S, Hui Y, Gou C, Cai Y, Hao M,
Zhao J, Wang S, Wang Z, Zhang X, He R, Liu J, Luo L, Li Y, Wang J: Genome
sequence of ground tit Pseudopodoces humilis and its adaptation to high
altitude. Genome Biol 2013, 14:R29.
26. Zachos J, Pagani M, Sloan L, Thomas E, Billups K: Trends, rhythms, and
aberrations in global climate 65 Ma to present. Science 2001, 292:686693.
27. Katz ME, Pak DK, Dickens GR, Miller KG: The source and fate of massive
carbon input during the latest paleocene thermal maximum. Science
1999, 286:15311533.
28. Sluijs A, Schouten S, Pagani M, Woltering M, Brinkhuis H, Sinninghe
Damste JS, Dickens GR, Huber M, Reichart GJ, Stein R, Matthiessen J,
Lourens LJ, Pedentchouk N, Backman J, Moran K: Subtropical Arctic Ocean
temperatures during the Palaeocene/Eocene thermal maximum. Nature
2006, 441:610613.
29. Subramanian S, Beans-Picon G, Swaminathan SK, Millar CD, Lambert DM:
Evidence for a recent origin of penguins. Biol Lett 2013, 9:20130748.
Li et al. GigaScience 2014, 3:27 Page 13 of 15
http://www.gigasciencejournal.com/content/3/1/27
30. Li H, Durbin R: Inference of human population history from individual
whole-genome sequences. Nature 2011, 475:493496.
31. Ritchie PA, Millar CD, Gibb GC, Baroni C, Lambert DM: Ancient DNA
enables timing of the Pleistocene origin and Holocene expansion of two
Adélie penguin lineages in Antarctica. Mol Biol Evol 2004, 21:240248.
32. Jenouvrier S, Barbraud C, Weimerskirch H: Sea ice affects the population
dynamics of Adélie penguins in Terre Adélie. Polar Biology 2006,
29:413423.
33. Lisiecki LE, Raymo ME: A Pliocene-Pleistocene stack of 57 globally
distributed benthic δ18O records. Paleoceanography 2005, 20:PA1003.
34. Thatje S, Hillenbrand CD, Mackensen A, Larter R: Life hung by a thread:
endurance of Antarctic fauna in glacial periods. Ecology 2008, 89:682692.
35. Clark PU, Dyke AS, Shakun JD, Carlson AE, Clark J, Wohlfarth B, Mitrovica JX,
Hostetler SW, McCabe AM: The last glacial maximum. Science 2009,
325:710714.
36. Spearman RIC: The epidermis and feather follicles of the king penguin
(Aptenodytes patagonica) (Aves). Zeitschrift für Morphologie der Tiere 1969,
64:361372.
37. Dawson C, Vincent JFV, Jeronimidis G, Rice G, Forshaw P: Heat transfer
through penguin feathers. J Theor Biol 1999, 199:291295.
38. Greenwold MJ, Bao W, Jarvis ED, Hu H, Li C, Gilbert MTP, Zhang G,
Sawyer RH: Dynamic evolution of the alpha (α) and beta (β) keratins has
accompanied integument diversification and the adaptation of birds
into novel lifestyles. BMC Evol Biol 2014, in press.
39. Greenwold MJ, Sawyer RH: Molecular evolution and expression of
archosaurian beta-keratins: diversification and expansion of archosaurian
beta-keratins and the origin of feather beta-keratins. J Exp Zool B Mol Dev
Evol 2013, 320:393405.
40. Vanhoutteghem A, Londero T, Ghinea N, Djian P: Serial cultivation of
chicken keratinocytes, a composite cell type that accumulates lipids and
synthesizes a novel beta-keratin. Differentiation 2004, 72:123137.
41. Ruhrberg C, Hajibagheri MA, Simon M, Dooley TP, Watt FM: Envoplakin, a
novel precursor of the cornified envelope that has homology to
desmoplakin. J Cell Biol 1996, 134:715729.
42. Yang Z: PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol
Evol 2007, 24:15861591.
43. Rickman L, Simrak D, Stevens HP, Hunt DM, King IA, Bryant SP, Eady RA,
Leigh IM, Arnemann J, Magee AI, Kelsell DP, Buxton RS: N-terminal deletion
in a desmosomal cadherin causes the autosomal dominant skin disease
striate palmoplantar keratoderma. Hum Mol Genet 1999, 8:971976.
44. Okano T, Yoshizawa T, Fukada Y: Pinopsin is a chicken pineal
photoreceptive molecule. Nature 1994, 372:9497.
45. Piezzi RS, Gutierrez LS: Electron microscopic studies on the pineal organ of
the Antarctic penguin (Pygoscelis papua). Cell Tissue Res 1975, 164:559570.
46. Chen TY, Peng YW, Dhallan RS, Ahamed B, Reed RR, Yau KW: A new
subunit of the cyclic nucleotide-gated cation channel in retinal rods.
Nature 1993, 362:764767.
47. Alten M: Penguin parenting: Adelie penguins reunite for their annual
breeding rituals. Animals 1997, 130:2023.
48. Williams TD: The Penguins; Spheniscidae. Oxford: Oxford University Press; 1995.
49. Choi Y, Sims GE, Murphy S, Miller JR, Chan AP: Predicting the functional
effect of amino acid substitutions and indels. PLoS One 2012, 7:e46688.
50. Ruiz-Perez VL, Tompson SW, Blair HJ, Espinoza-Valdez C, Lapunzina P,
Silva EO, Hamel B, Gibbs JL, Young ID, Wright MJ, Goodship JA: Mutations
in two nonhomologous genes in a head-to-head configuration cause
Ellis-van Creveld syndrome. Am J Hum Genet 2003, 72:728732.
51. Liu S, Lorenzen ED, Fumagalli M, Li B, Harris K, Xiong Z, Zhou L,
Korneliussen TS, Somel M, Babbitt C, Wray G, Li J, He W, Wang Z, Fu W,
Xiang X, Morgan CC, Doherty A, O'Connell MJ, McInerney JO, Born EW,
Dalen L, Dietz R, Orlando L, Sonne C, Zhang G, Nielsen R, Willerslev E, Wang
J: Population genomics reveal recent speciation and rapid evolutionary
adaptation in polar bears. Cell 2014, 157:785794.
52. Ge RL, Cai Q, Shen YY, San A, Ma L, Zhang Y, Yi X, Chen Y, Yang L, Huang Y,
He R, Hui Y, Hao M, Li Y, Wang B, Ou X, Xu J, Zhang Y, Wu K, Geng C, Zhou W,
Zhou T, Irwin DM, Yang Y, Ying L, Bao H, Kim J, Larkin DM, Ma J, Lewin HA,
et al:Draft genome sequence of the Tibetan antelope. Nat Commun 2013,
4:1858.
53. Kelley JL, Peyton JT, Fiston-Lavier AS, Teets NM, Yee MC, Johnston JS,
Bustamante CD, Lee RE, Denlinger DL: Compact genome of the Antarctic
midge is likely an adaptation to an extreme environment. Nat Commun
2014, 5:4611.
54. Li R, Zhu H, Ruan J, Qian W, Fang X, Shi Z, Li Y, Li S, Shan G, Kristiansen K,
Yang H, Wang J: De novo assembly of human genomes with massively
parallel short read sequencing. Genome Res 2010, 20:265272.
55. Parra G, Bradnam K, Korf I: CEGMA: a pipeline to accurately annotate core
genes in eukaryotic genomes. Bioinformatics 2007, 23:10611067.
56. Hillier LW, Miller W, Birney E, Warren W, Hardison RC, Ponting CP, Bork P,
Burt DW, Groenen MAM, Delany ME, Dodgson JB, Map G, Assembly SA,
Chinwalla AT, Cliften PF, Clifton SW, Delehaunty KD, Fronick C, Fulton RS,
Graves TA, Kremitzki C, Layman D, Magrini V, McPherson JD, Miner TL, Minx
P, Nash WE, Nhan MN, Nelson JO, Oddy LG, et al:Sequence and
comparative analysis of the chicken genome provide unique
perspectives on vertebrate evolution. Nature 2004, 432:695716.
57. Warren WC, Clayton DF, Ellegren H, Arnold AP, Hillier LW, Kunstner A, Searle
S, White S, Vilella AJ, Fairley S, Heger A, Kong L, Ponting CP, Jarvis ED, Mello
CV, Minx P, Lovell P, Velho TA, Ferris M, Balakrishnan CN, Sinha S, Blatti C,
London SE, Li Y, Lin YC, George J, Sweedler J, Southey B, Gunaratne P,
Watson M, et al:The genome of a songbird. Nature 2010, 464:757762.
58. Dalloul RA, Long JA, Zimin AV, Aslam L, Beal K, Blomberg LA, Bouffard P,
Burt DW, Crasta O, Crooijmans RPMA, Cooper K, Coulombe RA, De S, Delany
ME, Dodgson JB, Dong JJ, Evans C, Frederickson KM, Flicek P, Florea L,
Folkerts O, Groenen MAM, Harkins TT, Herrero J, Hoffmann S, Megens H-J,
Jiang A, de Jong P, Kaiser P, Kim H, et al:Multi-platform next-generation
sequencing of the domestic turkey (Meleagris gallopavo): genome
assembly and analysis. PLoS Biol 2010, 8:e1000475.
59. Harris RS: Improved Pairwise Alignment of Genomic DNA. Pennsylvania State
University; 2007. PhD thesis.
60. Kent WJ, Baertsch R, Hinrichs A, Miller W, Haussler D: Evolution's cauldron:
Duplication, deletion, and rearrangement in the mouse and human
genomes. Proc Natl Acad Sci U S A 2003, 100:1148411489.
61. Zhan X, Pan S, Wang J, Dixon A, He J, Muller MG, Ni P, Hu L, Liu Y, Hou H,
Chen Y, Xia J, Luo Q, Xu P, Chen Y, Liao S, Cao C, Gao S, Wang Z, Yue Z, Li
G, Yin Y, Fox NC, Wang J, Bruford MW: Peregrine and saker falcon genome
sequences provide insights into evolution of a predatory lifestyle.
Nat Genet 2013, 45:563566.
62. Chen N: Using RepeatMasker to identify repetitive elements in genomic
sequences. Curr Protoc Bioinformatics 2004, 25:4.10:4.10.14.10.14.
63. Jurka J, Kapitonov VV, Pavlicek A, Klonowski P, Kohany O, Walichiewicz J:
Repbase Update, a database of eukaryotic repetitive elements. Cytogenet
Genome Res 2005, 110:462467.
64. Smit AFA, Hubley R: RepeatModeler Open-1.0. 2008, http://www.
repeatmasker.org.
65. Benson G: Tandem repeats finder: a program to analyze DNA sequences.
Nucleic Acids Res 1999, 27:573580.
66. Lowe TM, Eddy SR: tRNAscan-SE: a program for improved detection of
transfer RNA genes in genomic sequence. Nucleic Acids Res 1997, 25:955964.
67. Nawrocki EP, Kolbe DL, Eddy SR: Infernal 1.0: inference of RNA alignments.
Bioinformatics 2009, 25:13351337.
68. Gardner PP, Daub J, Tate JG, Nawrocki EP, Kolbe DL, Lindgreen S, Wilkinson
AC, Finn RD, Griffiths-Jones S, Eddy SR, Bateman A: Rfam: updates to the
RNA families database. Nucleic Acids Res 2009, 37:D136140.
69. Griffiths-Jones S, Saini HK, van Dongen S, Enright AJ: miRBase: tools for
microRNA genomics. Nucleic Acids Res 2008, 36:D154158.
70. Altschul SF, Gish W: Local alignment statistics. Methods Enzymol 1996,
266:460480.
71. Bonnet E, Wuyts J, Rouze P, Van de Peer Y: Evidence that microRNA
precursors, unlike other non-coding RNAs, have lower folding free
energies than random sequences. Bioinformatics 2004, 20:29112917.
72. Janssen S, Giegerich R: Faster computation of exact RNA shape
probabilities. Bioinformatics 2010, 26:632639.
73. Pearson WR: Flexible sequence similarity searching with the FASTA3
program package. Methods Mol Biol 2000, 132:185219.
74. Notredame C: Computing multiple sequence/structure alignments with
the T-coffee package. Curr Protoc Bioinformatics 2010, 4:3.8:3.8.13.8.28.
75. Li R, Yu C, Li Y, Lam TW, Yiu SM, Kristiansen K, Wang J: SOAP2: an
improved ultrafast tool for short read alignment. Bioinformatics 2009,
25:19661967.
76. Li R, Li Y, Fang X, Yang H, Wang J, Kristiansen K, Wang J: SNP detection for
massively parallel whole-genome resequencing. Genome Res 2009,
19:11241132.
77. Visser K, Thunell R, Stott L: Magnitude and timing of temperature change in
the Indo-Pacific warm pool during deglaciation. Nature 2003, 421:152155.
Li et al. GigaScience 2014, 3:27 Page 14 of 15
http://www.gigasciencejournal.com/content/3/1/27
78. Ruan J, Li H, Chen Z, Coghlan A, Coin LJ, Guo Y, Heriche JK, Hu Y,
Kristiansen K, Li R, Liu T, Moses A, Qin J, Vang S, Vilella AJ, Ureta-Vidal A,
Bolund L, Wang J, Durbin R: TreeFam: 2008 Update. Nucleic Acids Res 2008,
36:D735740.
79. De Bie T, Cristianini N, Demuth JP, Hahn MW: CAFE: a computational tool
for the study of gene family evolution. Bioinformatics 2006, 22:12691271.
80. Greenwold MJ, Sawyer RH: Genomic organization and molecular
phylogenies of the beta (beta) keratin multigene family in the chicken
(Gallus gallus) and zebra finch (Taeniopygia guttata): implications for
feather evolution. BMC Evol Biol 2010, 10:148.
81. Li YI, Kong L, Ponting CP, Haerty W: Rapid evolution of Beta-keratin genes
contribute to phenotypic differences that distinguish turtles and birds
from other reptiles. Genome Biol Evol 2013, 5:923933.
82. O'Guin WM, Sawyer RH: Avian scale development: VIII. Relationships
between morphogenetic and biosynthetic differentiation. Dev Biol 1982,
89:485492.
83. Sawyer RH, Knapp LW, OGuin WM: Epidermis, dermis and appendages.
In Biology of the Integument. Springer; 1986:194238.
84. Loytynoja A, Goldman N: Phylogeny-aware gap placement prevents
errors in sequence alignment and evolutionary analysis. Science 2008,
320:16321635.
85. Blake JA, Bult CJ, Kadin JA, Richardson JE, Eppig JT, Mouse Genome
Database G: The Mouse Genome Database (MGD): premier model
organism resource for mammalian genomics and genetics. Nucleic Acids
Res 2011, 39:D842848.
86. Zhang G, Lambert D, Wang J: Genomic data from Adelie penguin
(Pygoscelis adeliae). GigaScience 2011, http://dx.doi.org/10.5524/100006.
87. Zhang G, Lambert D, Wang J: Genomic data from the Emperor penguin
(Aptenodytes forsteri). GigaScience 2011, http://dx.doi.org/10.5524/100005.
doi:10.1186/2047-217X-3-27
Cite this article as: Li et al.:Two Antarctic penguin genomes reveal
insights into their evolutionary history and molecular changes related
to the Antarctic environment. GigaScience 2014 3:27.
Submit your next manuscript to BioMed Central
and take full advantage of:
Convenient online submission
Thorough peer review
No space constraints or color figure charges
Immediate publication on acceptance
Inclusion in PubMed, CAS, Scopus and Google Scholar
Research which is freely available for redistribution
Submit your manuscript at
www.biomedcentral.com/submit
Li et al. GigaScience 2014, 3:27 Page 15 of 15
http://www.gigasciencejournal.com/content/3/1/27
... TLR4 and TLR5 (Levy et al. 2020), we undertook a comprehensive analysis of penguin TLRs to consider processes that may have shaped the evolution of TLR-mediated immunity across this entire order of vertebrates. Using genomes derived from all extant species of penguin (Li et al. 2014;Pan et al. 2019), we investigate patterns of adaptive evolution in penguin TLRs. In addition, we examine a highly unusual case of cryptic pseudogenization of TLR15 in the Eudyptes (crested) penguins, which may have important implications for the susceptibility of penguins to fungal pathogens. ...
... BLAST (Altschul et al. 1990) was used to identify TLR sequences in assembled penguin genomes generated using Illumina short-read sequences (Li et al. 2014;Pan et al. 2019). The majority of penguin species possessed representatives of each of the ten avian TLRs (TLR1A, TLR1B, TLR2A, TLR2B, TLR3, TLR4, TLR5, TLR7, TLR15, and TLR21). ...
... Twenty-one assembled penguin genome sequences were downloaded from the GigaScience Database (doi: http://gigadb.org/dataset/100649) from two studies exploring penguin evolution (Li et al. 2014; A. forsteri and Pygoscelis adeliae; Pan et al. 2019; all other penguin species). TLR sequences from the annotated Emperor penguin (A. ...
Article
Full-text available
Penguins (Sphenisciformes) are an iconic order of flightless, diving seabirds distributed across a large latitudinal range in the Southern Hemisphere. The extensive area over which penguins are endemic is likely to have fostered variation in pathogen pressure, which in turn will have imposed differential selective pressures on the penguin immune system. At the front line of pathogen detection and response, the Toll-like receptors (TLRs) provide insight into host evolution in the face of microbial challenge. TLRs respond to conserved pathogen-associated molecular patterns and are frequently found to be under positive selection, despite retaining specificity for defined agonist classes. We undertook a comparative immunogenetics analysis of TLRs for all penguin species, and found evidence of adaptive evolution that was largely restricted to the cell surface expressed TLRs, with evidence of positive selection at, or near, key agonist-binding sites in TLR1B, TLR4 and TLR5. Intriguingly, TLR15, which is activated by fungal products, appeared to have been pseudogenized multiple times in the Eudyptes spp., but a full-length form was present as a rare haplotype at the population level. However, in vitro analysis revealed that even the full-length form of Eudyptes TLR15 was non-functional, indicating an ancestral cryptic pseudogenization prior to its eventual disruption multiple times in the Eudyptes lineage. This unusual pseudogenization event could provide an insight into immune adaptation to fungal pathogens such as Aspergillus, which is responsible for significant mortality in wild and captive bird populations.
... We analyzed 27 genomes comprising all extant and recently-extinct penguin species, subspecies, and major lineages. 21 of the high-coverage genomes have been published by members of our consortium for this project 8,9 . To supplement the dataset, we sequenced three highcoverage genomes from the remaining Pygoscelis papua lineages from Falkland Islands/Malvinas "FAL", Kerguelen Island "KER" and South Georgia "SG" (see 68 ), and partial genomes from the recently-extinct Eudyptes warhami, M. a. richdalei and M. a. waitaha (see ref. 5 and citations within). ...
... We do not include Eudyptes warhami, M. a. richdalei, and M. a. waitaha or additional P. papua lineages ("FAL", "SG", "KER") in these analyses. Our analyses expand on previous analyses that have only examined A. forsteri and P. adeliae (e.g., 8,49 ), or those that have relied on only on-site analysis for penguins (e.g., 7 ). ...
Article
Full-text available
Penguins lost the ability to fly more than 60 million years ago, subsequently evolving a hyper-specialized marine body plan. Within the framework of a genome-scale, fossil-inclusive phylogeny, we identify key geological events that shaped penguin diversification and genomic signatures consistent with widespread refugia/recolonization during major climate oscillations. We further identify a suite of genes potentially underpinning adaptations related to thermoregulation, oxygenation, diving, vision, diet, immunity and body size, which might have facilitated their remarkable secondary transition to an aquatic ecology. Our analyses indicate that penguins and their sister group (Procellariiformes) have the lowest evolutionary rates yet detected in birds. Together, these findings help improve our understanding of how penguins have transitioned to the marine environment, successfully colonizing some of the most extreme environments on Earth.
... On one hand, selection tests based on phylogenies are, of course, influenced by size and composition of the set of species included in the tree. In contrast with Li et al. (2014), where 48 bird species of which only two penguins were analyzed, we selected non-target species both in close (seven penguins) and in more distant (13 other birds) clades. On the other hand, the bioinformatic pipeline applied for identifying ortholog coding sequences across the species in the phylogeny may determine which genes are included or excluded. ...
... Our pipeline, successfully tested in Drosophila (Ometto et al., 2013), aligned ca. 30% more genes than in Vianna et al. (2020) and slightly less than in Li et al. (2014). We also note that in Vianna et al. (2020), the selection scan was performed for 18 penguin species, not only on the King and Emperor ones, and searched for candidate genes in all of the penguin lineages. ...
Preprint
Full-text available
The eco-evolutionary history of penguins is profoundly influenced by their shift from temperate to cold environments. Breeding only in Antarctica during the winter, the Emperor penguin appears as an extreme outcome of this process, with unique features related to insulation, heat production and energy management. However, whether this species actually diverged from a less cold-adapted ancestor, thus more similar in ecology to its sister species, the King penguin, is still an open question. As the Antarctic niche shift likely resulted in vast changes in selective pressure experienced by the Emperor penguin, the identification and relative quantification of the genomic signatures of selection, unique to each of these sister species, could answer this question. Applying a suite of phylogeny-based methods on 7,651 orthologous gene alignments of seven penguins and 13 other birds, we identified a set of candidate genes showing significantly different selection regimes either in the Emperor or in the King penguin lineage. Our comparative approach unveils a more pervasive selection shift in the Emperor penguin, supporting the hypothesis that its extreme cold adaptation is a derived state from a more King penguin-like ecology. Among the candidate genes under selection in the Emperor penguin, four genes (TRPM8, LEPR, CRB1, and SFI1) were identified before in other cold adapted vertebrates, while, on the other hand, 161 genes can be assigned to functional pathways relevant to cold adaptation (e.g., cardiovascular system, lipid, fatty acid and glucose metabolism, insulation, etc.). Our results show that extreme cold adaptation in the Emperor penguin largely involved unique genetic options which, however, affect metabolic and physiological traits common to other cold-adapted homeotherms.
... Over the past 12-15 million years, Adélie penguins have contended with a wide range of climates and consequent impacts to their habitat. Ice sheets have repeatedly expanded and retreated hundreds of kilometers, destroying or creating nesting habitat, and Adélie populations have grown and shrunk corresponding to interglacial and glacial periods, respectively (Li et al. 2014). The comings and goings of Adélies through geologic time, as determined by dating subfossil bones, have been used to validate the dates of ice sheet advances and retreats. ...
... To identify the demographic history on an evolutionary time scale (i.e., >10,000 years), ancient DNA (or aDNA) has been used. For example, genetic methods have been successfully employed to identify the population trends of the extinct passenger pigeon (Ectopistes migratorius) (Hung et al., 2014), the population recovery of the crested ibis (Nipponia nippon) (Feng et al., 2019), and population trends in Adélie penguins (Pygoscelis adeliae) and emperor penguins (Aptenodytes forsteri) in response to climate changes (Li et al., 2014). However, aDNA is less effective for more recent population trends (i.e., <10,000 years) because insufficient time has elapsed for the accumulation of evidence of evolutionary change. ...
Article
Full-text available
The lack of long-term monitoring data for many wildlife populations is a limiting factor in establishing meaningful and achievable conservation goals. Even for well-monitored species, time series are often very short relative to the timescales required to understand a population's baseline conditions before the contemporary period of increased human impacts. To fill in this critical information gap, techniques have been developed to use sedimentary archives to provide insights into long-term population dynamics over timescales of decades to millennia. Lake and pond sediments receiving animal inputs (e.g., feces, feathers) typically preserve a record of ecological and environmental information that reflects past changes in population size and dynamics. With a focus on bird-related studies, we review the development and use of several paleolimnological proxies to reconstruct past colony sizes, including trace metals, isotopes, lipid biomolecules, diatoms, pollen and non-pollen palynomorphs, invertebrate sub-fossils, pigments, and others. We summarize how animal-influenced sediments, cored from around the world, have been successfully used in addressing some of the most challenging questions in conservation biology, namely: How dynamic are populations on long-term timescales? How may populations respond to climate change? How have populations responded to human intrusion? Finally, we conclude with an assessment of the current state of the field, challenges to overcome, and future potential for research.
... As with other epidermal appendages, many of the genes involved in the development and structure of feathers are located within the EDC locus and originated from a single or small number of ancestral genes [9]. The physical diversity observed across feathers is accompanied by the genetic diversity displayed by several differentially expressed avian EDC genes [9,[23][24][25]. ...
Article
Full-text available
The transition of amniotes to a fully terrestrial lifestyle involved the adaptation of major molecular innovations to the epidermis, often in the form of epidermal appendages such as hair, scales and feathers. Feathers are diverse epidermal structures of birds, and their evolution has played a key role in the expansion of avian species to a wide range of lifestyles and habitats. As with other epidermal appendages, feather development is a complex process which involves many different genetic and protein elements. In mammals, many of the genetic elements involved in epidermal development are located at a specific genetic locus known as the epidermal differentiation complex (EDC). Studies have identified a homologous EDC locus in birds, which contains several genes expressed throughout epidermal and feather development. A family of avian EDC genes rich in aromatic amino acids that also contain MTF amino acid motifs (EDAAs/EDMTFs), that includes the previously reported histidine-rich or fast-protein (HRP/fp), an important marker in feather development, has expanded significantly in birds. Here, we characterize the EDAA gene family in birds and investigate the evolutionary history and possible functions of EDAA genes using phylogenetic and sequence analyses. We provide evidence that the EDAA gene family originated in an early archosaur ancestor, and has since expanded in birds, crocodiles and turtles, respectively. Furthermore, this study shows that the respective amino acid compositions of avian EDAAs are characteristic of structural functions associated with EDC genes and feather development. Finally, these results support the hypothesis that the genes of the EDC have evolved through tandem duplication and diversification, which has contributed to the evolution of the intricate avian epidermis and epidermal appendages.
... To assess capacity for adaptive potential, entry points include whole-genome sequencing, tests of functional responses to stress to evaluate adaptive plasticity, and estimation of genetic variability in populations to gauge survival potential. Only a handful of Antarctic marine vertebrate genomes have been sequenced and analysed to date: for two penguins (Li et al., 2014) and five notothenioid fishes (Shin et al., 2014;Bargelloni et al., 2019;Chen et al., 2019;Kim et al., 2019a). One of these studies compared the genomes (and transcriptomes) of the high-latitude cold-adapted Antarctic toothfish (Dissostichus mawsoni) and a basal non-Antarctic and closest sister species, which never experienced the same selective pressure, to identify Antarctic-specific adaptations . ...
Article
Full-text available
p>Important findings from the second decade of the 21st century on the impact of environmental change on biological processes in the Antarctic were synthesised by 26 international experts. Ten key messages emerged that have stakeholder-relevance and/or a high impact for the scientific community. They address (i) altered biogeochemical cycles, (ii) ocean acidification, (iii) climate change hotspots, (iv) unexpected dynamism in seabed-dwelling populations, (v) spatial range shifts, (vi) adaptation and thermal resilience, (vii) sea ice related biological fluctuations, (viii) pollution, (ix) endangered terrestrial endemism and (x) the discovery of unknown habitats. Most Antarctic biotas are exposed to multiple stresses and considered vulnerable to environmental change due to narrow tolerance ranges, rapid change, projected circumpolar impacts, low potential for timely genetic adaptation, and migration barriers. Important ecosystem functions, such as primary production and energy transfer between trophic levels, have already changed, and biodiversity patterns have shifted. A confidence assessment of the degree of ‘scientific understanding’ revealed an intermediate level for most of the more detailed sub-messages, indicating that process-oriented research has been successful in the past decade. Additional efforts are necessary, however, to achieve the level of robustness in scientific knowledge that is required to inform protection measures of the unique Antarctic terrestrial and marine ecosystems, and their contributions to global biodiversity and ecosystem services.</p
Article
Full-text available
Global warming is increasingly exacerbating biodiversity loss. Populations locally adapted to spatially heterogeneous environments may respond differentially to climate change, but this intraspecific variation has only recently been considered when modelling vulnerability under climate change. Here, we incorporate intraspecific variation in genomic offset and ecological niche modelling to estimate climate change-driven vulnerability in two bird species in the Sino-Himalayan Mountains. We found that the cold-tolerant populations show higher genomic offset but risk less challenge for niche suitability decline under future climate than the warm-tolerant populations. Based on a genome-niche index estimated by combining genomic offset and niche suitability change, we identified the populations with the least genome-niche interruption as potential donors for evolutionary rescue, i.e., the populations tolerant to climate change. We evaluated potential rescue routes via a landscape genetic analysis. Overall, we demonstrate that the integration of genomic offset, niche suitability modelling, and landscape connectivity can improve climate change-driven vulnerability assessments and facilitate effective conservation management.
Article
Anthropogenic climate change is causing observable changes in Antarctica and the Southern Ocean including increased air and ocean temperatures, glacial melt leading to sea‐level rise and a reduction in salinity, and changes to freshwater water availability on land. These changes impact local Antarctic ecosystems and the Earth’s climate system. The Antarctic has experienced significant past environmental change, including cycles of glaciation over the Quaternary Period (the past ~2.6 million years). Understanding Antarctica’s paleoecosystems, and the corresponding paleoenvironments and climates that have shaped them, provides insight into present day ecosystem change, and importantly, helps constrain model projections of future change. Biological archives such as extant moss beds and peat profiles, biological proxies in lake and marine sediments, vertebrate animal colonies, and extant terrestrial and benthic marine invertebrates, complement other Antarctic paleoclimate archives by recording the nature and rate of past ecological change, the paleoenvironmental drivers of that change, and constrain current ecosystem and climate models. These archives provide invaluable information about terrestrial ice‐free areas, a key location for Antarctic biodiversity, and the continental margin which is important for understanding ice sheet dynamics. Recent significant advances in analytical techniques (e.g., genomics, biogeochemical analyses) have led to new applications and greater power in elucidating the environmental records contained within biological archives. Paleoecological and paleoclimate discoveries derived from biological archives, and integration with existing data from other paleoclimate data sources, will significantly expand our understanding of past, present and future ecological change, alongside climate change, in a unique, globally significant region.
Article
Full-text available
Inferring the selective forces that orthologous genes underwent across different lineages can help us understand the evolutionary processes that have shaped their extant diversity and the phenotypes they underlie. The most widespread metric to estimate the selection regimes of coding genes—across sites and phylogenies—is the ratio of nonsynonymous to synonymous substitutions (dN/dS, also known as ω). Nowadays, modern sequencing technologies and the large amount of already available sequence data allow the retrieval of thousands of orthologous genes across large numbers of species. Nonetheless, the tools available to explore selection regimes are not designed to automatically process all genes, and their practical usage is often restricted to the single-copy ones which are found across all species considered (i.e., ubiquitous genes). This approach limits the scale of the analysis to a fraction of single-copy genes, which can be as low as an order of magnitude in respect to those which are not consistently found in all species considered (i.e., nonubiquitous genes). Here, we present a workflow named BASE that—leveraging the CodeML framework—eases the inference and interpretation of gene selection regimes in the context of comparative genomics. Although a number of bioinformatics tools have already been developed to facilitate this kind of analyses, BASE is the first to be specifically designed to allow the integration of nonubiquitous genes in a straightforward and reproducible manner. The workflow—along with all relevant documentation—is available at github.com/for-giobbe/BASE.
Article
Full-text available
The third review of the status and trends of Antarctic and sub-Antarctic seabird populations compiled by the Bird Biology Subcommittee of the Scientific Committee on Antarctic Research at the request of the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR) is presented.
Article
We describe a program, tRNAscan-SE, which identifies 99-100% of transfer RNA genes in DNA sequence while giving less than one false positive per 15 gigabases. Two previously described tRNA detection programs are used as fast, first-pass prefilters to identify candidate tRNAs, which are then analyzed by a highly selective tRNA covariance model. This work represents a practical application of RNA covariance models, which are general, probabilistic secondary structure profiles based on stochastic context-free grammars. tRNAscan-SE searches at approximately 30 000 bp/s. Additional extensions to tRNAscan-SE detect unusual tRNA homologues such as selenocysteine tRNAs, tRNA-derived repetitive elements and tRNA pseudogenes.
Article
We measured wing joint mobility in penguins, alcids, diving-petrels, and non-diving fliers. Great reduction in mobility of the intrinsic wing joints was found in penguins, but not in alcids or diving-petrels. This reduction is correlated with simplification of the intrinsic wing musculature. In contrast, alcids and diving-petrels, which use their wings in both air and water, retain the full functional capacities for flight. Movement through the air probably requires a capability for subtle and varied motions, forces, and shape changes that preclude stiffening and simplification of the wing. Hence, the conversion of an aerial wing to a flipper, as in penguins, must be possible only after the evolutionary loss of flight.
Chapter
The avian integument is uniquely characterized by the presence of feathers, whose distribution, structure and development have been reviewed recently on a comparative basis by Lucas and Stettenheim (1972). Brush (1974, 1975, 1980) and Brush and Wyld (1982) have used the biochemical composition of feathers in comparative and evolutionary studies of several bird species. In general, the body of most birds appears covered by feathers with obvious naked areas restricted to the feet, beak, comb and wattle. However, as discussed by Lucas and Stettenheim (1972) and Stettenheim (1972) feathers occur in tracts (pterylae) on most birds, which are separated by naked regions of skin known as apteria. The apteria may exist within individual feather tracts as the naked skin between feathers or between pterylae. In the developing chicken skin, Sengel (1976) distinguished two types of apteric regions; one which is entirely devoid of feathers (for example, the midventral apterium) and one which develops a few loosely distributed down feathers (for example, the apterium between the breast and ventral pterylae). This distinction is of importance in developmental studies where one is interested in the control of appendage formation. In a later section we will discuss some of the more recent findings on the development and structure of feathers, with special reference to the localization of keratins with indirect immunofluorescence, using non-cross-reacting antisera to both α -and ß-(feather) keratins.
Article
[1] We present a 5.3- Myr stack ( the " LR04'' stack) of benthic delta(18)O records from 57 globally distributed sites aligned by an automated graphic correlation algorithm. This is the first benthic delta(18)O stack composed of more than three records to extend beyond 850 ka, and we use its improved signal quality to identify 24 new marine isotope stages in the early Pliocene. We also present a new LR04 age model for the Pliocene- Pleistocene derived from tuning the delta(18)O stack to a simple ice model based on 21 June insolation at 65degreesN. Stacked sedimentation rates provide additional age model constraints to prevent overtuning. Despite a conservative tuning strategy, the LR04 benthic stack exhibits significant coherency with insolation in the obliquity band throughout the entire 5.3 Myr and in the precession band for more than half of the record. The LR04 stack contains significantly more variance in benthic delta(18) O than previously published stacks of the late Pleistocene as the result of higher-resolution records, a better alignment technique, and a greater percentage of records from the Atlantic. Finally, the relative phases of the stack's 41- and 23- kyr components suggest that the precession component of delta(18)O from 2.7 - 1.6 Ma is primarily a deep- water temperature signal and that the phase of delta(18)O precession response changed suddenly at 1.6 Ma.