HISTORICAL POPULATION GENETICS OF CALLORHINUS URSINUS (NORTHERN FUR SEALS) FROM THE ALEUTIAN ISLANDS
Ying Fang
A Thesis Submitted to the University of North Carolina Wilmington in Partial Fulfillment
of the Requirements for the Degree of Master of Science
Department of Biology and Marine Biology
University of North Carolina Wilmington
2009
Approved by
Advisory Committee
Michael A. McCartney Heather N. Koopman
Marcel van Tuinen Chair
Accepted by
_____________________________
Dean, Graduate School
TABLE OF CONTENTS
ABSTRACT....................................................................................................................... iii ACKNOWLEDGMENTS ................................................................................................. iv LIST OF TABLES .............................................................................................................. v LIST OF FIGURES ........................................................................................................... vi CHAPTER 1 ....................................................................................................................... 1 INTRODUCTION .............................................................................................................. 1 METHODS AND ANALYSES ........................................................................................ 12 RESULT............................................................................................................................ 18 DISCUSSION................................................................................................................... 24 CONCLUSION................................................................................................................. 43 LITERATURE CITED...................................................................................................... 44 CHAPTER 2 ..................................................................................................................... 52 INTRODUCTION ............................................................................................................ 52 METHODS AND ANALYSES ........................................................................................ 59 RESULT............................................................................................................................ 65 DISCUSSION................................................................................................................... 78 CONCLUSION................................................................................................................. 81 LITERATURE CITED...................................................................................................... 82
ii
ABSTRACT
Historically, the Northern Fur Seal (Callorhinus ursinus) has been a staple food
during human harvesting. While breeding localities exist primarily in Alaska today,
archaeological rookeries once were found along the entire Eastern Pacific. Given this
information, past genetic variation in C. ursinus may have been considerably higher than
today. In order to place modern population genetic variation in a proper historic context, I
obtained ancient DNA from the mitochondrial control region (157 base pairs) of 56 C.
ursinus sampled from three geographically isolated Aleutian Islands (Kodiak, Unalaska
and Shemya). Comparison of these data to unpublished data from extant individuals
indicated high haplotype diversity (h=0.962~0.988), high nucleotide diversity (π =
0.034~0.051) and low genetic structure (Fst=0.035) in C. ursinus before and after the
height of commercial sealing ~120 years ago. Lack of phylogeographic structure was
observed in a combined data set comprised of modern and ancient individuals, further
indicating that change in genetic diversity or structure has not occurred in the last 3000
years. Demographic analysis further indicated past growth in all populations with no
evidence of a bottleneck, but populations did vary in timing and extent of expansion.
Taken together, these results show that C. ursinus has not lost genetic variation or
experienced increased population structuring due to the historical commercial sealing.
iii
ACKNOWLEDGMENTS
I am especially grateful to my advisor Dr. Marcel van Tuinen who introduced me to
the exciting field of conservation biology and moves my improvement forward in my
career. My thanks go to my committee members Dr. Heather Koopman and Dr. Michael
McCartney who are always open to discuss my project and provide plenty of resources
and suggestions. I would like to thank Dr. Ann Pabst, Dr. Brian Arbogast and Dr. Ann
Stapleton. My thesis would never come out without their help and encouragement.
The Department of Biology and Marine Biology, the Graduate School of UNCW
and the National Science Foundation provided financial support for my research and
studies. Dr. Seth Newsome at Carnegie Institution of Washington provided
radiometrically dated samples for this project. People in Elizabeth Hadly’s laboratory at
Stanford University taught me statistic skills and introduced fresh ideas, specifically
Cheng Li, Malin Pinsky, and Dr. Elizabeth Hadly.
Special thanks go to my parents who are always supporting me.
iv
LIST OF TABLES
Table Page
1. Comparison of measures of genetic diversity and population changes in Alaskan
C. ursinus populations of different age. ............................................................. 19 2. Results from Analysis of Molecular Variance (AMOVA) on three ancient
populations......................................................................................................... 22 3. Studies showing high level of genetic variation in pinnipeds............................. 27 4. Studies showing reduced genetic variation in pinnipeds. ................................... 29 5. Studies showing pinnipeds with non-significant genetic structure..................... 34 6. Studies showing pinnipeds with significant genetic structure. ........................... 37 7. Comparison population changes in Alaskan C. ursinus populations of different
age using DnaSP ................................................................................................ 67 8. Comparisons of inter- and intra-specific clock rates of control region using C.
ursinus and outgroup species ............................................................................. 73
v
LIST OF FIGURES
Figure Page
1. The recent distribution of Northern Fur Seals. Black arrows indicate breeding localities. .............................................................................................................. 4
2. Map showing three archaeological sites where C. ursinus specimens were
collected. .............................................................................................................. 4 3. Fluctuations in population size of C. ursinus in the past 100 years on the Pribilof
Islands. ................................................................................................................. 7 4. Intraspecific Bayesian phylogeny of 58 ancient and 68 modern samples. ......... 22 5. Mismatch distributions of ancient and modern Aleutian populations of C.
ursinus................................................................................................................ 69 6. Rate comparisons between control region and cytochrome b region. ................ 71 7. Bayesian skyline plots of independent populations of Aleutian C. ursinus........ 76 8. Bayesian skyline plots showing the demographic history of Aleutian C. ursinus
............................................................................................................................ 77
vi
CHAPTER 1
PAST GENETIC CHANGES IN CALLORHINUS URSINUS USING ANCIENT DNA
INTRODUCTION
Overview
Callorhinus ursinus (Northern Fur Seal) is a widespread pinniped species found in
the North Pacific Ocean. It is considered as a “vulnerable” species under the U.S.
Endangered Species Act, with current populations of vital concern (Loughlin et al. 1994;
NMML 2007). C. ursinus belongs to the family Otariidae, and the class Pinnipedia (Rice
1998). It is the only species in the genus Callorhinus, having diverged from the sister
clades of the true fur seals (Arctocephalinae) and sea lions (Otariinae) at least 6 million
years ago (Wynen et al. 2001). Being a high trophic level predator, it plays an important
role in marine ecosystems. C. ursinus feeds opportunistically, primarily on a variety of
fish, cephalopods and crustaceans (COSEWIC 2006). Their predators are Killer Whales,
some sharks, foxes, and Steller Sea Lions (Gentry 1998).
C. ursinus has very thick fur that consists of with two layers--long guard hairs and a
dense waterproof underfur. Furthermore, this species has a good sense of hearing, keen
eyesight but lacks color vision. Males are much bigger than females (200-275 kg and
40-50 kg, respectively). They can live for more than 20 years (Gentry 1998; COSEWIC
2006). Growth rates of males and females are fairly similar until the fifth year of life,
when males develop more rapidly (York 1987).
Northern Fur Seals have a pelagic distribution that ranges across the North Pacific
including the east coast of Russia, Japan and the west coast of Canada and the United
States (Ream et al. 2005). The Pribilof Islands in the east Bering Sea are home to the
majority of the world’s breeding population today (Towell and Ream 2006). Additional
small rookeries exist on the Commander Islands in the west Bering Sea, Kuril Islands in
the north of Japan, Robben Island in the Sea of Okhotsk, and a few much smaller
rookeries on Bogoslof Island in the Aleutian Island chain, and San Miguel Island off the
coast of Southern California (Figure 1; Gentry 1998; Ream and Burkanov 2005).
Callorhinus ursinus is a colonial breeder and exhibits strong site fidelity. Outside the
breeding season, individuals spend nearly all their time in the open ocean to avoid
mainland predators. Certain offshore islands have been used for pupping and breeding,
where seals historically have been available to human exploitation (Gentry 1998; Ream
and Burkanov 2005).
Based on the abundance of females and unweaned pups in archaeological sites
(from ~5000 years ago to ~200 years ago), Newsome et al. (2007) confirmed that C.
ursinus was one of the most common and broadly distributed pinniped species along the
east Pacific Coast, stretching from the Pribilof Islands to southern California. Rookeries
distributed continuously along the eastern Pacific coastline (eastern Aleutian Islands to
California) indicated that C. ursinus once was more widespread than today (Figure 2).
The breeding and nursing behaviors of prehistoric and modern populations of C.
ursinus are also known to be different. Modern C. ursinus on the Pribilof rookeries
2
aggregate in May. Adult males remain on the rookeries throughout the breeding season
(Gentry 1998). Adult females give birth in late June. This pattern follows the “sub-polar”
maternal strategy, where females wean their pups and move south with them (Gentry
1998). Being faithful to their natal site, most females will return to the Pribilof Islands the
following March (Gentry 1998). Ream et al. (2005) monitored physical locations of 13
female C. ursinus from the Pribilofs during 2002-2003, and recorded their migration
routes southward and back. Pups may remain at sea for 22 months before returning to
their rookery of birth (COSEWIC 2006). Modern C. ursinus is weaned at about 4 months
of age. This is known as a “short term” strategy, while almost all other otariids nurse their
young for one to two years (Gentry and Kooyman 1986; Perrin et al. 2002; Newsome et
al. 2007). However, a similar “long term” maternal strategy is thought to have been used
prehistorically by Alaskan C. ursinus based on 5- to 12-month-old age profiles and
isotope data (Newsome et al. 2007). The δ15N isotope values that indicate nursing
persisted longer (until 9-12 months) in Alaskan specimens than in extant specimens from
these localities (Newsome et al. 2007).
3
Commander Islands
Robben Island
Kuril Islands
Bogoslof Island
Aleutian Islands
Pribilof Islands
San Miguel Island
Figure 1. The recent distribution of Northern Fur Seals. Black arrows indicate breeding localities. Yellow arrow and dotted line indicate Aleutian Islands. Map modified from http://commons.wikimedia.org/wiki/File:Bering_Sea_Location.gif#file, a composite of two maps on Wikimedia Commons: [http://commons.wikimedia.org/wiki/Image: Topographic90deg_N0E90.png and :Topographic90deg_N0W]
Bogoslof Island Kodiak Island
Shemya Island
Figure 2. Map (courtesy of Seth Newsome, Carnegie Institution of Washington) showing three archaeological sites where C. ursinus specimens were collected (Shemya Island, Unalaska Island and Kodiak Island) and one modern site (Bogoslof Island) locates in the Aleutian Chain. Pribilof Islands are located north of Aleutian Islands. Other archaeological sites along the Pacific Northwest (Ozette, Hesquiat, Toquaht, Ts’ishaa, Seal Rock, Umpqua), and California (San Miguel, Point Mugu, Ano Nuevo, Moss Landing, Duncans Point) are also shown for completeness.
4
As demographic data reported, populations from the Pribilof Islands have
experienced dramatic fluctuations in size (Loughlin et al. 1994; Gentry 1998; Balsiger et
al. 2005; Newsome et al. 2007). The Aleut people who colonized Alaska about 8,000
years ago harvested C. ursinus as a staple food for millennia, but abundance did not
significantly decline (Balsiger et al. 2005). Subsequent human exploitation led to three
significant declines in C. ursinus numbers: the first one was from the late 1700s to the
early 1800s; the second one was from the late 1800s to the early 1900s, and the third
included the experimental harvest of females in 1956 (Ream 2002; Towell and Ream
2006). The fur trade started approximately about 220 years ago, and approached its
height in the late 19th century, when the populations in Alaska declined dramatically with
the arrival of Russian and European fur traders (Balsiger et al. 2005). Census size
reached a low of 216,000 animals in 1912, close to a 90% population decline of the
pre-exploitation size (NMML 2007; Macklin et al. 2008). As a result, the North Pacific
Fur Seal Convention was signed in 1911 to limit hunting. Following this protection, the
Alaska C. ursinus population grew again to 2 million individuals by 1950 (Macklin et al.
2008). Between 1956 and 1970, the population experienced a significant decline again,
which was attributed primarily to an experimental harvest of females (Towell and Ream
2006). The population then began to increase to an estimated maximum population size
of 1.25 million in 1974. By 1983 the population had declined to about 877,000 animals
(Macklin et al. 2008). Following this period of decline, C. ursinus abundance on the
Pribilof Islands was relatively stable until the mid-1990s (Towell and Ream 2006).
5
However, a new decline in the Pribilofs began in 2000. The total abundance on the
Pribilof Islands was 636,000 during 2002 to 2006 (DFO 2007). Figure 3 shows changes
in population size in the past 100 years. Reasons for the current decline are unknown but
may include a decrease in reproductive rates, disease, natural predation, interspecies
competition, climate changes, illegal killing, or pollutants (Adams et al. 2007). The
species currently is officially considered as “threatened” under the Marine Mammal
Protection Act of 1988 (Loughlin et al. 1994), and under the Species at Risk Act (SARA)
by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) (DFO
2007). Overall, pup production on the Pribilof Islands declined by 38% over the last 30
years (DFO 2007). In contrast, a small population at Bogoslof Island (in eastern Aleutian
Islands) has grown more than 12% per year between 1997 and 2005 (NMML 2007;
Ream & Burkanov 2005). The number at Bogoslof is about 4.5% of the global
populations (Ream & Burkanov 2005).
6
Figure 3. Fluctuations in population size of C. ursinus in the past 100 years on the Pribilof Islands.Data are cited from DFO 2007; NMML 2007; Macklin et al. 2008.
7
To better understand the cause and effect of the recent decline in C. ursinus
populations in the Pribilof and contiguous breeding sites, genetic response to external
stressors should be studied. For example, a comparison of genetic diversity before and
after a population contraction can evaluate whether the species experienced a genetic
bottleneck. Perhaps the most famous mammal bottleneck and corresponding inbreeding
effects are described in studies of the cheetah. O’ Brien et al. (1987) compared allozymes
from Acinonyx jubatus raineyi (east African cheetah) and Acinonyx jubatus jubatus
(south African cheetah), and found low genetic distance between subspecies, and very
low levels of heterozygosity in both of them, therefore affirming the cheetah as one of
the least genetically variable mammal species. Recent inbreeding that resulted from loss
of genetic variation after a Pleistocene bottleneck has inflicted considerable damage on
cheetah’s survival rates and health.
In the context of my thesis on Northern Fur Seals, inter-specific comparison among
pinniped species in genetic response to known population bottlenecks is especially
informative. Several genetic studies have shown a correlation between historic
population size reduction and genetic bottlenecks. For example, Weber et al. (2000)
analyzed sequences of the mitochondrial DNA control region from Mirounga
angustirostris (Northern elephant seal) that lived before, during and after a bottleneck in
1892, and found a sharp reduction in genetic variation after this bottleneck, which
confirmed a significant reduction in population size. Similarly, Arctocephalus townsendi
(Guadalupe fur seal) from the coast of California had experienced a loss of genetic
8
variability correlating with a bottleneck during the late 18th and early 19th centuries
(Weber et al. 2004). Wynen et al. (2000) studied postsealing mitochondrial DNA
variation in Arctocephalus gazella (Antarctic fur seal) and Arctocephalus tropicalis
(Subantarctic fur seal) and suggested that both species recovered in population size
through the recolonization of islands after the bottleneck in the late 18th century.
Little is known about variation among current C. ursinus populations except that the
degree of geographical isolation is higher today than that in the past (Newsome et al.
2007). Furthermore, paleo-ecological data indicate that there may have been distinct
dietary differences along the Aleutian chain from west to east (Newsome et al. 2007).
This information derives from a study using isotopic ratios of carbon and nitrogen to help
indicate the geographic distribution of the prehistoric populations (Newsome et al. 2007).
The isotopic ratio of consumers reflects the isotopic composition fixed at the base of
food webs (Burton et al. 2001). δ 13C values inform about foraging location since δ 13C
values are higher in nearshore than open waters (Burton and Koch 1999). Based on δ 13C
values, prehistoric C. ursinus foraged offshore across their entire range (Newsome et al.
2007). Nitrogen isotopes vary with both trophic level and latitude (Burton et al. 2001;
Newsome et al. 2007). δ 15N isotope values were higher in Southern C. ursinus remains
than in remains from Northern populations. Both δ 13C and δ 15N isotope values declined
steeply from east to west along the Aleutian Islands, indicating that Aleutian populations
on the west (Shemya Island) were ecologically distinct from eastern Aleutian populations
(Unalaska Island, Kodiak Island; Newsome et al. 2007).
9
Objectives
The objectives of the first chapter in my thesis are to (a) provide historical context
to population genetic variation in extant C. ursinus, (b) quantify genetic variation and
structure in three Aleutian populations of differing ages, (c) present an intra-specific
comparison of pre- and post-bottleneck genetic variation, and (d) present an inter-specific
comparison of genetic response to population contraction in commonly harvested
pinniped species. While modern genetic data can predict past genetic variation, ancient
genetic data can test such predictions directly and determine lineage coalescences more
exactly. Several studies on mammals show that ancient DNA tells more information than
modern DNA by observing the past genetic parameters directly. Barnes et al. (2002) used
30 ancient specimens ranging from 45,000 to 14,000 years ago of Ursus arctos (Brown
bear), and found no phylogeographical structure in the past compared with high structure
in present. Consuegra et al. (2002) amplified DNA from 6 ancient samples of Salmo
salar (Salmon) from 41,000 to 3,250 years ago, and suggested ancient salmon have a
totally different haplotype. A higher genetic variation from past was confirmed by 9
ancient specimens of Ovibos moschatus (Musk ox) (MacPhee et al. 2005). 21 sequences
of ancient Ursus arctos (Brown bear) collected from Spain confirmed no genetic
structure of past populations (Valdiosera et al. 2007). Thus, I applied ancient DNA
methodologies to a selectively neutral, highly variable genetic marker to document
genetic variation, and geographic subdivision of populations in the Aleutian Islands of
the past. According to other bottlenecked seals (Weber et al. 2000; Wynen 2001; Weber
10
et al. 2004), C. ursinus likely experienced one or more population bottleneck(s). If true,
genetic variation in populations predating these bottleneck events will exceed that of
today. Secondly, Shemya seals differ in foraging strategy from other Aleutian populations
(Newsome et al. 2008), hence I will test the null hypothesis of no interbreeding between
Shemya seals and other populations. This project is part of a collaborative effort between
UNCW, Stanford University, National Marine Mammal Laboratory (NMML), Simon
Fraser University and Carnegie Institution for Science. My contribution in this effort was
to investigate the Alaskan populations. Different teams targeted other ancient and modern
populations: ancient Pacific Northwest by Simon Fraser University, ancient California by
Stanford University, and modern populations by NMML.
11
METHODS AND ANALYSES
1. Sample selection
I examined sequence variation in the control region of the mitochondrial genome
from 82 seals sampled at 3 archaeological sites. Radiometrically dated samples of
individuals were selected and identified by colleagues from the Carnegie Institution for
Science, and taken into the Ancient DNA laboratory at the Center for Marine Science of
UNCW. I obtained 38 samples from Rolling Bay (Kodiak Island, ~1,000-500 years BP),
28 samples from Amaknak Bridge (Unalaska, ~3,300-2,700 years BP), and 18 samples
from Shemya (western Aleutians, ~2,000-1,000 years BP). I had access to unpublished
modern data (samples collected from 1993 to 1998; n=365) along the east Pacific sites
(Pinsky et al. unpublished; Dickerson et al. unpublished). These DNA sequences derived
from St. George Island (n=92), St. Paul Island (n=91) in the Pribilof Islands, Bogoslof
Island (n=96) in the Aleutians, and San Miguel Island (n=86) in California. I compared
my ancient DNA results to modern data from the Bogoslof population because of
geographic proximity.
2. Ancient DNA decontamination and extraction
The ancient DNA technique is modified from modern DNA protocols to be more
stringent. The major difficulty in working with ancient or degraded DNA samples lies in
proving the authenticity of ancient amplified DNA. The high risk of contamination from
modern sources is due to the degraded nature and low copy number of ancient DNA. The
12
Polymerase Chain Reaction (PCR) is extremely sensitive and can easily pick up
contaminant DNA from many sources (Pääbo 1989; Wayne et al.1999; Rawlence et al.
2009).
I performed all the molecular work in an isolated clean laboratory at the Center for
Marine Science of UNCW, where we have never done any genetic work on pinnipeds.
This laboratory consists of an outside sample preparation room and an inside ultra-clean
room. Both laboratories were bleached and UV-irradiated overnight every time before
use, and rooms were limited to few workers that were not allowed to re-enter the
ultra-clean room for a period of 24 hours after PCR set up. Before DNA extraction, all
surfaces were cleaned again with alconox detergent and a bleach solution. Labeled bones
were stored in the sterile aDNA preparation room. The bones were pulverized, and small
bits (< 500mg) were digested in extraction buffer (0.5M EDTA, 0.5% SDS, pH=8.0) with
proteinase K (100 ug/mL) at 55°C for 24 hours (Yang et al. 2003). I used a Qiaquick
PCR cleanup Kit (Qiagen) to perform the final DNA extraction on a maximum of five
individuals per experiment, which always included a negative control. The DNA
extraction was performed in a stand-alone workstation with positive airflow preventing
additional contamination between samples and possible spread to the entire clean room.
3. PCR amplification and quantification
Contamination controls and detection are very important in ancient DNA studies
(Yang et al. 2003; Newsome et al. 2007). Both negative and positive controls were set up
along with ancient DNA samples for PCR amplification (Yang et al. 2003). Positive
13
controls were used to indicate whether PCR conditions were set up correctly and negative
controls including blank extracts showed amplification products if contamination
occurred. Positive controls were also used to indicate the sensitivity of individual PCR
amplification and the level of contamination if it occurred, because the brightness of the
stained DNA displayed its concentration. No PCR products were ever used when
amplified along with a suspected contaminant. PCR was run under low-annealing
temperature with high-fidelity DNA polymerase. PCR amplification of target DNA was
performed in 50-ul reactions. I used the forward primer CalloCR1 (5’-
CTCCCCCTATGTACTTCGTGCA-3’) and the reverse primer CalloCR2
(5’-GTACACTTTTCACAAGGGTTGCTG-3’) to amplify a 157 bp region of the
mitochondrial control region (200 bp including primers). The PCR recipe with final
concentrations was 0.2 uL (5U/uL) FirePol DNA polymerase (Solis BioDyne), 5 uL (10×)
Reaction buffer B, 10 uL (25mM) MgCl2, 5 uL (10mM) dNTPs, 5uL (20uM) primer for
each, 5 uL (13mg/mL) bovine serum albumin (BSA) (1.3 mg/mL), 3 uL of DNA template
and 11.8 uL sterile water. I used the following PCR conditions: after denaturation at 94°C
for 10 minutes, DNA was amplified by 45 cycles consisting of denaturation at 94°C for
30 seconds, annealing at 45 °C for 45 seconds, and extension at 72°C for 1 minute. A
final extension period of 10 minutes at 72°C was followed by cooling of the PCR product
to 4°C. PCR products and visualized on 2% agarose gels stained with ethidium bromide.
postPCR cleanup was performed with EXO-SAP, which was made up of 1.7 uL
Exonuclease 13.4 uL Shrimp Alkaline Phosphatase and 3.4 uL sterile water for 50 uL
14
PCR product.
4. Sequencing and Analyses
DNA sequencing was outsourced to Macrogen, Inc. in Seoul, Korea. Sequences
were cleaned up and aligned with Sequencher 4.8. All unique haplotypes (distinct DNA
sequences that occur in only a single individual) were verified by re-extraction, re-PCR
and re-sequencing.
I performed several analyses of population genetic variation and structure using the
programs Arlequin (version 3.1) (Excoffier et al. 2006) and DnaSP (version 5.0) (Rozas
et al 2003; Librado &Rozas 2009). These programs assume that all samples are from
extant individuals (Excoffier et al. 2006). I determined S (number of segregating sites), η
(number of mutations), k (average number of nucleotide differences), h (haplotype
diversity), and π (nucleotide diversity, per site) (Watterson 1975; Nei 1987). I estimated
Tajima’s D, Fu & Li’s D, Fu & Li’s F, and Fu’s Fs in modern and ancient populations, as
a test of selective neutrality or possible population expansion or contraction (Tajima 1989,
1993; Fu & Li 1993; Fu 1997). Finally, I estimated statistical separation of haplotype
groupings through Analysis of Molecular Variance (AMOVA) (Weber et al. 2000; Wynen
2001; Excoffier et al. 2006). I used BEAUti (version 1.4.8) and BEAST (version 1.4.8) to
calculate the α value (reflecting amount of rate heterogeneity among nucleotide sites)
under the HKY+Gamma model, running with 2 independent MCMC simulations of
10,000,000 iterations in BEAST (version 1.4.8) (Drummond et al. 2005). The posterior
output of α (=0.116) was applied to AMOVA in Arlequin, with the application of the
15
Kimura–two parameter distance model.
A rooted, time-measured Bayesian intraspecific phylogenetic tree using Monte
Carlo Markov Chain (MCMC) simulation was built to show haplotype relationships
within C. ursinus from modern and ancient populations. BEAST (Bayesian Evolutionary
Analysis by Sampling Trees) accounts for serial sampling through time instead of
assuming that all tips are of same age (Drummond et al. 2003). BEAST uses a MCMC
Bayesian approach to find the best set of parameters and searches for the tree weighted
with the highest posterior probability, and at the same time producing a credible sample
of trees (Drummond et al. 2005). For the modern sample set, I used MEGA (version 3.0)
to build a neighbor-joining tree and randomly picked 68 sequences that almost entirely
represented all haplotype groups from the available 365 modern sequences. These
included sequences from populations from St. George Island, St. Paul Island, Bogoslof
Island and San Miguel Island. The modern samples plus the amplified ancient DNA
sequences were taken together to build a tree.
I used BEAST (v 1.4.8) to build the intraspecific phylogenetic tree (Drummond &
Rambaut 2007). To calculate the clock rate appropriately considering the problem of
time inequality above and below species level, I designed and compared different sets of
parameters, and picked up the most suitable rate for intraspecific clock (see Chapter 2).
The first step was to determine the clock rate. I used 8.2 mya as the divergence time
between C. ursinus and sea lions. The sea lions partially included Eumetopias jubatus
(Steller sea lion), Zalophus californianus (California sea lion) and Zalophus
16
californianus japonicus (Japanese sea lion) (Slade et al. 1994; Higdon et al. 2007; Pinsky
et al. unpublished). The data set included the 126 individuals of C. ursinus as the
majority and one individual from each outgroup species. The sequences of outgroups
were obtained from GeneBank and aligned with the 157 bp control region of C. ursinus.
A nexus file of all 129 samples was imported in BEAUti. Rate was estimated under the
HKY+Gamma+Invariant (HKY+G+I) substitution model with strict clock. Priors were
set up as coalescent constant size, with population size as a uniform distribution from 0
to 60 million, and tree model root height as a normal distribution using 8.2 mya as the
mean, and 2.1 mya as the standard deviation. Samples from the posterior were drawn
every 20,000 MCMC steps over a total of 20,000,000 steps, with the first 10% discarded
as burn-in for they were less trustable. The second step was plotting Bayesian skyline, a
method for estimating past population dynamics through time from molecular sequences
without a prespecified parametric model of demographic history (Drummond et al. 2005).
The posterior output of the first step (intraspecific clock rate with 2.13E-7 per site per
mya as the mean, and 1.36E-8 per site per mya as the standard deviation) was applied to
build the tree. The intraspecific tree did not count the outgroup species. Bayesian
analyses were performed using HKY+G+I model and a Bayesian skyline coalescence
prior. The output tree file was applied to tree building. The third step used programs of
TreeAnnotator and Figtree to make a maximum clade credibility tree, with the first 10%
of the set of trees discarded as burn-in.
17
RESULT
I obtained a total of 58 sequences from 3 ancient populations, with 21 sequences
from Rolling Bay on Kodiak Island, 18 sequences from Shemya Island, and 19 from
Amaknak Bridge on Unalaska Island. Fifteen of the 58 sequences were obtained from
Stanford University conducted by Cheng Li. Seventy-two of the 82 samples were
successfully extracted and amplified, as shown by gel visualization, but of these, 14 out
of the 72 individuals did not produce clean sequencing signals likely due to degraded
bones and impurity of samples. The haplotype diversity (h) is high in all populations,
ranging from 0.962 to 0.988. Similarly, nucleotide diversity (π) (the average number of
nucleotide differences per site between any two DNA sequences chosen randomly from
the sample population) has an overall high level (range from 0.034 to 0.051) but is
considerably smaller in Unalaska (π=0.034) than other islands (π=0.05) (Table 1).
18
a)
Population n S η No. of hap Kodiak 21 25 26 15 Unalaska 19 20 22 17 Shemya 18 27 27 15 Total ancient 58 33 36 42 BOG modern 96 42 44 65
b)
Population h ±SD π ±SD k Kodiak 0.962±0.026 0.048±0.003 7.486 Unalaska 0.988±0.021 0.034±0.004 5.328 Shemya 0.980±0.024 0.051±0.004 7.941 Total ancient 0.986±0.006 0.045±0.002 6.967 BOG modern 0.988±0.004 0.048±0.002 7.595
c)
Population Tajima’s D Fu & Li’s D Fu & Li’s F Fu’s Fs Kodiak 0.138 0.127 0.152 -3.530 Unalaska -0.596 -0.566 -0.667 -10.900 * Shemya 0.046 -0.135 -0.095 -5.080 * Total ancient -0.345 -0.750 -0.717 -25.000 * BOG modern -0.359 0.648 0.294 -24.800 *
Table 1 Comparison of measures of genetic diversity and population changes in Alaskan C. ursinus populations of different age.
a), b) measures of genetic diversity; c) measures of population changes; n: number of individuals in each population; S: number of segregating sites; η: number of mutations; No. of hap: number of haplotypes; h: haplotype diversity; π: nucleotide diversity; k: average number of nucleotide differences; *P<0.02.
19
Values of Tajima’s D, Fu & Li’s D, and Fu & Li’s F do not show significant
deviation from the null (stable size) model (Table 1c; P>0.10). The direction of Tajima's
D, Fu & Li’s D, and Fu & Li’s F is potentially informative about the evolutionary and
demographic forces that a population has experienced (Tajima 1989, 1993; Fu & Li 1993;
Fu 1997; Excoffier et al. 2006). For example, Tajima's test compares a the total number
of segregating sites of a set of sequences from a population versus the average number of
mutations between pair of sequences sampled from the same population. If these two
numbers are the same, then the null hypothesis of neutrality cannot be rejected. If the
average number of polymorphisms found in pairwise comparison minus the total number
of polymorphic sites gives a negative value, it reflect an excess of rare polymorphisms in
a population, which is consistent with either purifying selection or a recent increase in
population size. Positive values indicate an excess of intermediate-frequency alleles and
can result from a population decline (Tajima 1989, 1993). The tests are to identify
whether sequences fit the neutral theory model with a null hypothesis of selective
neutrality and population size constancy (Tajima 1989, 1993; Fu & Li 1993; Fu 1997;
Excoffier et al. 2006). Here the values of Tajima’s D, Fu & Li’s D, and Fu & Li’s F do
not indicate a signal of population expansion or contraction in any population.
Fu’s Fs statistic is very sensitive to population demographic expansion or
contraction, which generally leads to large negative (or positive) Fs values (Fu 1997;
Excoffier et al. 2006). Similar to Tajima’s D, the null hypothesis of Fu’s Fs test is
20
selective neutrality and constant population size. Fu’s Fs indeed appears more sensitive
to departure from population equilibrium than Tajima’s D (Table 1c). The Fs values of
populations from Unalaska, Shemya, “Total ancient” and Bogoslof all show significant
signs of population expansion (P<0.02; Fu 1997).
Analysis of Molecular Variance (AMOVA) is a method that evaluates the amount of
genetic structure in a data set, often indicative of multiple independent populations.
AMOVA results show that 96.54% of the variation in the total population is found within
populations, with only a 3.46% of the variation among populations (Table 2a). The
overall Fst value of 0.0346 (P=0.039±0.006) indicates low structure and high gene flow
among populations. The population-specific Fst indices that are low in the three
populations further denote lack of structure within single populations (Table 2b). The Fst
values are 0.029 (Kodiak), 0.028 (Shemya) and 0.048 (Unalaska). Population pairwise Fst
values range from 0.024 to 0.045, which are not significantly different from a total lack
of structure, while the P-values are not significant enough (P> 0.05 for all the
comparisons; Table 2c).
The Bayesian phylogenetic tree of C. ursinus control region sequences is shown in
Figure 4. The evolutionary relationships between individual haplotypes do not indicate
presence of significant structure between ancient and modern populations and highlight
the high level of genetic diversity before and after the fur trade, even in a single modern
population. Consistent with demographic expansion, the majority of nodes in the tree are
supported by low bootstrap values.
21
a) Source of variation d.f. Sum of squares Variance
components Percentage of variation
Among populations
2 13.630 0.145 3.463
Within populations
55 221.500 4.027 96.540
Total 57 235.100 4.172 Fixation Index Fst: 0.035 P-value=0.039±0.006
b) Population number Name Fst
1 Kodiak750 0.029 2 Shemya1500 0.028 3 Unalaska3000 0.048
c) Kodiak Shemya Unalaska Kodiak 0.099±0.025 0.135±0.039 Shemya 0.036 0.072±0.030 Unalaska 0.024 0.045
Table 2 Results from Analysis of Molecular Variance (AMOVA) on three ancient populations. a) AMOVA output and overall Fst; b) intra-population Fst indices; c) inter-population pairwise Fst: Fst values are below the diagonal, P-values are above.
Figure 4 (see next page). Intraspecific Bayesian phylogeny of 58 ancient and 68 modern samples. High bootstrap values (>0.50) are specified with error bars of 95% confident intervals. Time line shows age of each lineage. Individual samples are represented as ages, ybp stands for years before present. 0 (ybp) modern, 750 (ybp) Kodiak, 1500 (ybp) Shemya, 3000 (ybp) Unalaska.
22
40000.
0
300000. 250000 200000 150000 100000 50000 0
0
00
750
3000
0
1500
750
3000
0
3000
0
0
750750
1500
0
1500
0
750
0
0
1500
0
1500
0
750
0
3000
0
3000
0
3000
0
3000
750
0
1500
3000
750
3000
0
1500
750
0
0
3000
750
0
0
0
0
0
0
3000
0
0
750
750
0
1500
3000
1500
0
0
3000
1500
0
750
0
3000
0
0
0
1500
0
0
0
1500
0
3000
0
0
1500
0
750
00
3000
0
0
0
150
0
0
1500
0
1500
0
750
1500
0
750
0
750
0
3000
0
3000
0
0
750
0
0
3000
1500
0
0
750
0
750
0
0
750
0
0
0.62
0.82
0.
7
0.94
0.53
1
1
0.99
0.81
0.59
0.68
0.
6
0.54
1
1
0.83
23
DISCUSSION
Both the ancient and modern populations show high levels of haplotype diversity,
high nucleotide diversity and low genetic structure. The mtDNA analyses bring out
considerable genetic variation in C. ursinus, before and after historical commercial
hunting. None of these analyses indicate a signal of a genetic bottleneck. The high degree
of phylogenetic clustering of modern and ancient haplotypes together also indicates that
little genetic variation has been lost. In addition, most of the variation was partitioned
within rather than among populations, which demonstrates high gene flow among the
Aleutian Islands. Population from Unalaska, rather than that from Shemya, appears
somewhat genetically different from others. The demographic differences among the
three populations are discussed in more detail in chapter 2. The genetic homogeneity
among the three Aleutian Islands provides no clue of the foraging differences from the
east to west Aleutian chain (as shown by paleoisotopic data).
Bottleneck/Genetic variety
A population bottleneck can sometimes lead to a genetic homogeneity and
inbreeding, and gives rise to a high chance of extinction. Molecular markers are
commonly applied to detect bottleneck-induced genetic change. High haplotype diversity
in C. ursinus is not unusual compared to other historically abundant pinniped species, and
is consistent with two unpublished data sets. Ream’s Ph.D. thesis (2002) documented
extensive genetic variation within modern C. ursinus using 8 microsatellite loci. C.
24
ursinus also showed high haplotypic diversity (control region) in both prehistoric
(200-1500 ybp) and modern populations across the eastern Pacific range (Pinsky et al.
unpublished).
Several studies estimated genetic diversity of pinnipeds after sealing in late 19th
century. Most of the studies utilized the mitochondrial DNA control region or nuclear
DNA microsatellites because both molecular markers have high mutation rates, resulting
in high resolution in population genetics studies. Fewer studies used other markers such
as the slower evolving mitochondrial coding gene cytochrome b, the non-neutral major
histocompatibility complex (MHC) genomic region or allozymic variation. Other than C.
ursinus, in the family Otariidae, Eumetopias jubatus (Steller sea lion) (Bickham et al.
1996; Bickham et al. 1998; Trujilo et al. 2004; Harlin-Cognato et al. 2005 and Baker et al.
2005), Arctocephalus gazella (Antarctic fur seal) (Wynen et al. 2000), Arctocephalus
tropicalis (Subantarctic fur seal) (Wynen et al. 2000), Arctocephalus forsteri (New
Zealand Fur Seal) (Lento et al. 1997), Arctocephalus pusillus pusillus (Cape fur seal)
(Matthee et al. 2006), Otaria flavescens (South American sea lion) (Freilich et al.
unpublished), and Arctocephalus philippii (Juan Fernandez fur seal) (Goldsworthy et al.
2000) showed high levels of modern genetic diversity. The family Phocidae that includes
Halichoerus grypus (Gray seal) (Graves et al. 2009), Mirounga leonina (Southern
elephant seal) (Hoelzel et al. 1993; Slade 1997 & 1998), Phoca vitulina (Harbor seal)
(Goodman 1998; Westlake and Ocorry 2002), Phoca largha (Spotted seal) (Mizuno et al.
2003) showed extensive modern genetic variation too. Similar results have also been
25
documented in other marine mammals, including Balaena mysticetus (Bowhead whale), a
species that shows high post-bottleneck genetic variability (control region) with
haplotype diversity of 0.949 (Borge et al. 2007). Table 3 summarizes diversity measures
reported from some studies with high genetic variation in pinnipeds.
26
Species Marker N No. (h) H Π Reference Arctocephalus forsteri cytochrome b 56 NA 0.6812; Ht=0.7909(high) NA Lento et al. 19971
Arctocephalus gazella control region 145 26 NA 0.032 Wynen et al. 2000 Arctocephalus philippii control region 28 13 0.905 0.030 Goldsworthy et al.2000 Arctocephalus.pusillus.pusillus control region 106 57 0.975±0.006 0.011±0.006 Matthee et al. 2006 Arctocephalus townsendi control region (prebottleneck) 26 25 0.997 0.055 Weber et al. 2004 Arctocephalus tropicalis control region 103 33 NA 0.048 Wynen et al. 2000 Callorhinus ursinus control region (750-3000ya) 58 42 0.986 0.045 This study Callorhinus ursinus control region (200-1500ya) 43 39 0.997 0.050 Pinsky et al. unpublished Callorhinus ursinus control region 365 186 0.987 0.046 Pinsky et al. unpublished Callorhinus ursinus Microsatellites 578 NA ~1.000 NA Ream 2002 PhD thesis Eumetopias jubatus control region 1568 121 0.916±0.004 0.010±0.006 Baker et al. 2005 Eumetopias jubatus control region 194 55 0.690 ~ 1.000 0.004 ~ 0.016 Trujilo et al. 2004 Eumetopias jubatus control region 336 107 0.890 ±0.010 NA Harlin-Cognato et al. 2005 Eumetopias jubatus control region 224 52 0.927 NA Bickham et al. 1996 Eumetopias jubatus control region 36 13 0.850 NA Bickham et al. 1998 Halichoerus grypus Microsatellites, mtDNA 1312, 1143 632,5 463 0.943 ~ 0.9653 0.016±0.0083 Graves et al. 2009 Mirounga leonina control region 48 26 0.944, 0.662 NA Hoelzel et al. 1993 Mirounga leonina mtDNA, nDNA 60 304 NA 0.0293, 9*10-5, 4 Slade 1998 Otaria flavescens control region 39 18 0.880 0.008 ~ 0.012 Freilich et al. unpublished Phoca largha control region 66 57 NA NA Mizuno et al. 2003 Phoca vitulina Microsatellites 1029 685 2.6(±0.04) ~ 7.0 (±0.28)6 NA Goodman 1998 Phoca vitulina control region 778 225 0.975 0.015 Westlake and Ocorry 2002
Table 3. Studies showing high level of genetic variation in pinnipeds. N: number of animals, No. (h): Number of haplotypes, H: haplotype diversity, π: nucleotide diversity. 1 This rate is considered as high for survival of two divergent lineages found in interspecific comparisons of other species. 2. microsatellite; 3. control region; 4 nDNA; 5 number of alleles; 6 allelic diversity.
27
Despite many pinnipeds not showing any obvious reduction in genetic variation due
to commercial sealing, several other species that were commercially hunted do indicate
low genetic variability. Weber et al. (2004) used the mitochondrial control region to
compare genetic diversity in Arctocephalus townsendi (Guadelupe fur seal) before and
after a population bottleneck and documented a reduction in genetic variability. Lento et
al. (2003) showed low diversity of a MHC Class II gene in Phocarctos hookeri (New
Zealand sea lion). In family Phocidae, Mirounga angustirostris (Northern elephant seal)
(Bonnell and Selander 1974; Hoelzel et al. 1993; Weber et al. 2000; Weber et al. 2004),
Monachus schauinslandi (Hawaiian monk seal) and Monachus monachus (Mediterranean
monk seal) exemplify well documented reductions in genetic diversity (Kretzmann et al.
1997; Aldridge et al. 2006; Harwood et al. 1996; Stanley and Harwood 1997; Pastor et al.
2004). Table 4 shows some of studies of pinnipeds with low genetic variation. It is
important to note that low diversity in one marker need not imply loss of genetic
variation, particularly if fast evolving markers have not been exploited yet. For example,
in Otariidae, the two subspecies of Brown fur seal, Arctocephalus pusillus doriferus
(Australian fur seal) and Arctocephalus pusillus pusillus (Cape fur seal) showed low
diversity in the mitochondrial cytochrome b gene (Lento et al. 1997). While suggestive of
a loss of genetic variation, another study (Table 3) using control region indicated high
haplotype diversity after the recent sealing in Arctocephalus pusillus pusillus (Matthee et
al. 2006). A similar trend exists in Zalophus californianus (California sea lion) with no
cytochrome b diversity but high control region variation (Maldonado et al. 1994&1995).
28
Species Marker N No. (h) H Π Reference Arctocephalus. pusillus doriferus cytochrome b 17 3 0.157 0.004 Lento et al. 1997 Arctocephalus. pusillus.pusillus cytochrome b 25 8 0.728 0.006 Lento et al. 1997
Arctocephalus townsendi control region (postbottleneck) 32 7 0.798 0.025 Weber et al. 2004
Mirounga angustirostris control region 149 2 0.410 0.007 Weber et al. 2000
Mirounga angustirostris control region(ca. 1000-1800) 5 4 0.900 0.007 Weber et al. 2000
Mirounga angustirostris control region (1914-1980) 8 2 0.530 0.009 Weber et al. 2000 Mirounga angustirostris control region 40 2 NA 0.004 Hoelzel et al. 1993 Mirounga angustirostris allozymes 67 NA 0* NA Hoelzel et al. 1993 Monachus monachus microsatellites 52 3 NA NA Pastor et al. 2004
Harwood et al. 1996; Stanley and Harwood. 1997 Monachus monachus control region 18 1 NA NA
Monachus schauinslandi control region 50 3 NA 0.007 Kretzmann et al. 1997 Phocarctos hookeri SSCP3 39 2 NA NA Lento et al. 2003 Zalophus californianus control region 40 11 NA NA Maldonado et al. 1995 Zalophus californianus cytochrome b 40 1 NA NA Maldonado et al. 1995
Table 4 Studies showing reduced genetic variation in pinnipeds.
N: number of animals, No. (h): Number of haplotypes, H: haplotype diversity, π: nucleotide diversity. * No detectable variation
29
Most studies listed above showed the same tendency of higher genetic diversity
being maintained in populations with larger post-bottleneck abundance. Nearly all
pinniped species were hunted during the sealing from the 17th to the 19th century. Species
with high modern genetic diversity had survived from sealing, and thereafter maintained
a comparatively large population size, ranging from hundreds of thousands to millions of
individuals. For example, demographic data show the size of C. ursinus on the Pribilofs
reached a low of 216,000 animals in 1912, which corresponds to a 90% population
decline of the pre-exloitation size (NMML 2007; Macklin et al. 2008), but the species
recovered to a stock size of 636,000 individuals during 2002 to 2006 (DFO 2007). On the
contrary, species that proved to experience a significant bottleneck and failed to recover
in abundance most often showed loss of genetic diversity. Thus, low levels of current
population size corresponds with decreased genetic diversity: Arctocephalus townsendi
(Guadelupe fur seal) had a recovered population of 10,000 by the late 1990s; Phocarctos
hookeri (New Zealand sea lion) had a population of 15,000 in the mid 1990s; Mirounga
angustirostris (Northern elephant seal) had a recovered size of 175,000 animals with a
lower genetic diversity of today; Monachus schauinslandi (Hawaiian monk seal) and
Monachus monachus (Mediterranean monk seal) are two species that are critically
endangered with 1200 individuals and 350-450 individuals remaining respectively
(Kretzmann et al. 1997; Pastor et al. 2004). Thus, it may be that despite major reduction
in abundance during the fur trade, some species did not reach low enough numbers for
genetic drift to act. Secondly, some species naturally vary in abundance, reflecting
30
variation in reproductive rates. Thus, species are expected to vary in overall abundance
and population recovery rate.
A relationship between extent of geographic range and population size of pinnipeds
is unclear, however. There is a trend that widely dispersed species are resilient to external
stressors. Broadly dispersed pelagic species such as Callorhinus ursinus, Eumetopias
jubatus (Steller sea lion) and Phoca vitulina (Harbor seal) have decent population size
today (Goodman 1998; Westlake and Ocorry 2002; Trujilo et al. 2004). Almost all
critically endangered or extinct pinnipeds are/were limited regionally and thus vulnerable:
critically endangered Monachus schauinslandi (Hawaiian monk seal) and Monachus
monachus (Mediterranean monk seal) are limited to Hawaiian waters and Mediterranean
Sea/Eastern Atlantic waters respectively (Kretzmann et al. 1997; Pastor et al. 2004).
Extinct Zalophus japonicus (Japanese sea lion) spread in the Sea of Japan (Sakahira and
Niimi 2007), while Monachus tropicalis (Caribbean monk seal) lived in the Caribbean
Sea and the Gulf of Mexico (Leboeuf et al. 1986). Though these correlations may be due
to many causes other than geographic distribution, habitat management and species
competition need careful consideration for conservation.
To evaluate genetic variability, it is also important to determine the strategy to
choose appropriate molecular markers. Commonly used markers (e.g. mitochondrial
DNA control region and nuclear DNA microsatellite) have high mutation rates that result
in considerable resolution in genetic analyses. However, it is necessary to interpret data
differently by taking into account the nature of every marker and specific species. For
31
example, the mtDNA control region is maternally inherited, so it is limited to the female
component of a population to identify the level of genetic diversity. Suppose the sex ratio
is balanced within C. ursinus, the maternal information can be indirectly applied to the
total population. Important reasons for choosing the mitochondrial control region are the
abundance of female fossils spread in archaeological sites, and ancient DNA techniques
being more amenable to high copy number mitochondrial genes than to single copy
nuclear markers. To modify this analyses, amplifying a longer length of sequence will
improve resolution.
Structure/natal fidelity/gene flow
The low Fst values among the three ancient populations of C. ursinus suggest a lack
of genetic structure (in time and space) and high gene flow. Genetic structure is affected
by habitat continuity, geographic distribution, population densities, level and direction of
drift, migration etc. (Davis et al. 2008). To properly interpret the level of genetic structure
found in C. ursinus, I compared previous genetic studies of pinnipeds that mentioned
(lack of) global or regional genetic subdivision.
The low structure among populations of C. ursinus confirms two previous studies
that used microsatellites and mtDNA control region data respectively (Ream 2002;
Pinsky et al. unpublished). Neither of the two studies found significant structure in C.
ursinus among breeding sites. Further studies might apply mitochondrial cytochrome b as
a marker to test the level of structure. Similar to C. ursinus, some pinnipeds show high
level of genetic diversity and low genetic structure. Mizono et al. (2003) found high level
32
of genetic diversity and low level of structure in Phoca largha (Spotted seal) using the
mitochondrial control region. Matthee et al. (2006) used the same marker and suggested
no geographic structure in Arctocephalus pusillus pusillus (Cape fur seal). The same
pattern exists in Cystophora cristata (Hooded seal) (Coltman et al. 2007). Pusa hispida
(Ringed seal), a species that occurs throughout the Arctic Ocean and can be found in the
Bering Sea, had little genetic differentiation between main breeding areas although
populations are clearly geographically subdivided (Palo et al. 2001; Palo et al. 2003). In a
comprehensive study on six ice-breeding seals, Pusa hispida (Ringed seal), Hydrurga
leptonyx (Leopard seal), Ommatophoca rossii (Ross seal) and Lobodon carcinophagus
(Crabeater seal) exhibited low structure while two other ice seals (Leptonychotes
weddellii Weddell seal and Erignathus barbatus Bearded seal) showed significant
structure (Davis et al. 2008). Critically endangered Monachus monachus (Mediterranean
monk seal) is a species with low genetic variability and no structure (Pastor et al. 2004).
Table 5 lists some pinnipeds with low genetic structure.
33
Species Marker N Φst Fst Rst Reference
Arctocephalus pusillus. pusillus control region 106 NS NS NA Matthee et al. 2006
Callorhinus ursinus microsatellites 578 NA 0.0004 (P=0.273) -0.004 (P=0.321) Ream 2002 PhD thesis
Cystophora cristata control region, microsatellites 123, 300 -0.001 (P=0.49) NA NA
Coltman et al. 2007
Hydrurga leptonyx microsatellites 150 NA 0.001 (P=0.001) NA
Davis et al. 2008
Lobodon carcinophagus microsatellites 303 NA 0.003 (P=0.045) NA Davis et al. 2008 Ommatophoca rossii microsatellites 90 NA 0.006 (P=0.221) NA Davis et al. 2008
Phoca largha control region 66 -0.003-0.006 (P>0.05) NA NA
Mizuno et al. 2003
Pusa hispida microsatellites 149 NA 0.017 0.002 Palo et al. 2001 Pusa hispida microsatellites 78 NA 0.023 (P=0.007) NA Palo et al. 2003 Pusa hispida microsatellites 303 NA 0.005 (P=0) NA Davis et al. 2008
Table 5 Studies showing pinnipeds with non-significant genetic structure.
NS. Not significant
34
Comparatively more pinnipeds exhibited genetic structure than not. Eumetopias
jubatus (Steller sea lion) shares a similar geographic range with C. ursinus (Ream 2002)
but, compared to C. ursinus, it maintains high haplotype diversity after sealing and high
genetic structure. It has been debated whether 2 or 3 populations (subspecies) of
Eumetopias jubatus exist globally (Bickham et al. 1996; Baker et al. 2005), and a recent
study suggested a phylogenetic break between eastern stock and Asian/western stock
(Hoffman et al. 2006). Phoca vitulina (Harbor seal), as the most wide-ranging pinniped in
the North Atlantic and Pacific Oceans, showed extensive subdivision among geographic
populations by several studies (Stanley et al. 1993 & 1996; Kappe et al. 1997; Goodman
1998; Burg et al. 1999; Westlake and OCorry 2002). Wynen et al. (2000) inferred the
population genetic history of Arctocephalus gazella (Antarctic fur seal) and
Arctocephalus tropicalis (Subantarctic fur seal), and found extensive genetic structure in
Arctocephalus tropicalis but less structure in Arctocephalus gazella. Other species that
showed population structure include: Arctocephalus forsteri (New Zealand fur seal)
(Lento et al. 1997), Otaria flavescens (South American sea lion), Arctocephalus australis
(South American fur seal) (Tunez et al. 2006), Halichoerus grypus (Gray seal) (Allen et
al. 1995; Boskovic et al. 1996; Graves et al. 2009), Mirounga leonina (Southern elephant
seal)(Gales et al. 1989; Hoelzel et al. 1993), Leptonychotes weddellii (Weddell Seal),
Erignathus barbatus (Bearded Seal) (Davis et al. 2008), and Odobenus rosmarus (Walrus)
(Goodman 1998; Born et al. 2001). In other studies, the existence of genetic structure was
dependent on which marker was chosen. For example, Arctocephalus pusillus pusillus
35
(Cape fur seal) showed no structure based on the control region (Matthee et al. 2006), but
did so with cytochrome b (Lento et al. 1997). Graves et al. (2009) suggested geographic
separation based on microsatellites, but no structure using the control region in
Halichoerus grypus (Gray seal). Monachus schauinslandi (Hawaiian monk seal)
presented structure by using multilocus fingerprinting but no structure in the control
region (Kretzmann et al. 1997). Table 6 lists some of pinnipeds with high genetic
structure.
The majority of phocids have a reproductive strategy of slight polygyny where
breeding females are sparsely distributed, and a considerable proportion breed on ice
(Cassini 1999; Coltman et al. 2007; Davis et al. 2008). All otariids breed on land and
females aggregate to territories for mating (Cassini 1999). Land-breeding otariids
commonly present significant population structure as a result of natal site fidelity
(Coltman et al. 2007; Crockford & Frederick 2007). However, C. ursinus provides an
exception. In contrast to C. ursinus, the land-breeding pinnipeds Eumetopias jubatus
(Steller sea lion), Arctocephalus gazella (Antarctic fur seal), Arctocephalus tropicalis
(Subantarctic fur seal), Arctocephalus forsteri (New Zealand fur seal), Otaria flavescens
(South American sea lion), Arctocephalus australis (South American fur seal) all exhibit
significant genetic structure. Some phocids breeding on land also display high level of
structure consistent with strong natal site fidelity. They are Phoca vitulina (Harbor seal),
Mirounga leonina (Southern elephant seal) and Halichoerus grypus (Gray seal).
36
Species Marker N Hst Φst Fst Rst Gst Reference Arctocephalus forsteri cytochrome b 56 0.191 NA NA NA NA Lento et al. 1997 Arctocephalus gazella control region 145 NA 0.074 NA NA NA Wynen et al. 2000 Arctocephalus tropicalis control region 103 NA 0.190 NA NA NA Wynen et al. 2000 Erignathus barbatus microsatellites 119 NA NA 0.064 NA NA Davis et al. 2008 Eumetopias jubatus control region 1568 NA 0.081- 0.339 0.046-0.084 NA NA Baker et al. 2005 Eumetopias jubatus control region 224 NA NA 0.050± 0.030 NA NA Bickham et al. 1996
Halichoerus grypus control region, microsatellites 131 NA NA 0.0089-0.0967 0.0082-0.226 NA Graves et al. 2009
Leptonychotes weddellii microsatellites 893 NA NA 0.030 NA NA Davis et al. 2008 Mirounga leonina control region 48 NA 0.570 NA NA NA Hoelzel et al. 1993 Monachus schauinslandi fingerprinting 50,22 NA NA 0.130-0.200 NA NA Kretzmann et al. 1997 Otaria flavescens cytochrome b 70 NA -0.210~0.610 NA NA Tunez et al. 2006 Phoca vitulina microsatellite 1029 NA NA NA 0.181 0.187 Goodman 1998 Phoca vitulina control region 227 NA 0.772 NA NA NA Stanley et al. 1996 Phoca vitulina control region 778 NA 0.004-0.405 0.004-0.168 NA NA Westlake and O' corry 2002
Table 6 Studies showing pinnipeds with significant genetic structure.
37
A large proportion of ice-breeding phocids exhibit little genetic structure. Ice-packed
seals most frequently show low genetic structure, e.g. Cystophora cristata (Hooded seal),
Phoca largha (Spotted seal), Hydrurga leptonyx (Leopard seal), Ommatophoca rossii
(Ross seal) and Lobodon carcinophagus (Crabeater seal). The reason for high level of
gene flow in ice-packed seals may be the dynamic and unstable nature of the breeding
habitat that does not assist natal site fidelity or polygyny (Coltman et al. 2007; Davis et al.
2008). Though expected to show elevated genetic differentiation because of ecological
stable territories, the fast-ice-breeding seal Pusa hispida (Ringed seal) displays little
structure, where fast-ice means stable ice attaches to land (Davis et al. 2008). Unlike
most ice-breeding phocids, Leptonychotes weddellii (Weddell seal) and Erignathus
barbatus (Bearded seal) displayed high genetic structure (Davis et al. 2008), and
Odobenus rosmarus (Walrus) breeding on packed ice represents large population
differentiation (Goodman 1998; Born et al. 2001).
The pattern of higher population differentiation in land-breeding pinnipeds and
lower structure in ice-breeding seals is consistent, albeit with few exceptions. In contrast
to ice-breeding seals whose breeding sites are dynamic, land-breeding seals mate on
stable continent or islands where they develop sexual dimorphism, polygyny, and natal
site fidelity for both males and females (Davis et al. 2008). The limit of gene flow among
isolated rookeries leads to population partition in land-breeding seals. The inability to
detect population structure in C. ursinus may seem astonishing considering the structure
displayed in other land-breeding pinnipeds with high polygyny. However, several
possible reasons could explain this phenomenon. First, despite the high level of
philopatry, even small amounts of gene flow can sometimes lead to genetic homogeneity
(Ream 2002). Second, the large abundance of C. ursinus compared to other endangered
species (e.g. Mediterranean monk seal) provides enough resilience to environmental
disturbance and sustains high levels of genetic exchange. Third, the degree of migration
is high in both males and females in C. ursinus. Males live in the open ocean year-round
except for the breeding season, while females travel a long distance south to return the
next year. These migratory behaviors may result in more extensive movement throughout
rookeries, settling away from natal rookeries, and flexibility in mating. Lastly, the
inequality in age-specific natal site fidelity improves the chance of gene flow. It is not
well documented if pups return to their original natal areas with their mothers. One study
suggested that young seals locate their natal areas better with age (Baker et al. 1995), so
the larger span of rookeries settlement makes for a higher level of gene exchange.
42
CONCLUSION
In summary, mtDNA analyses of ancient and modern seals indicate that C. ursinus
has not experienced a loss or reduction of genetic diversity. The high haplotype diversity
in C. ursinus is not unusual compared to other pinnipeds that were historically abundant.
There are other seals that do show a decrease in population size and genetic variability
after a major population bottleneck. As a land-breeding pinniped with natal site fidelity
and polygyny, C. ursinus however exhibits low level of genetic structure between the
three studied Alaskan populations. This result indicates extensive gene flow among the
geographically distant populations to prevent genetic differentiation.
43
LITERATURE CITED
1. Adams, G. P., Ward, T. J., Goertz, C. E. C., Ream, R. R., and Sterling. 2007. Ultrasonographic characterization of reproductive anatomy and early embryonic detection in the northern fur seal (Callorhinus ursinus) in the field. Marine Mammal Science. 23 (2): 445-452.
2. Aldridge, B. M., Bowen, L., Smith, B. R., Antonelis, G. A., Gulland, F., and Stott, J. L. 2006. Paucity of class I MHC gene heterogeneity between individuals in the endangered Hawaiian monk seal population. Immunogenetics. 58: 203-215.
3. Allen, P. J., Amos, W., Pomeroy, P. P., and Twiss, S. D. 1995. Microsatellite variation in grey seals (Halichoerus grypus) shows evidence of genetic differentiation between two British breeding colonies. Molecular Ecology. 4: 653-662.
4. Anderson, C. N. K., Ramakrishnan, U., Chan, Y. L., and Hadly, E. A. (2005) Serial SimCoal: A population genetic model for data from multiple populations and points in time. Bioinformatics. 21: 1733-1734.
5. Baker, A. R., Loughlin, T. R., and Burkanov, V. et al. 2005. Variation of mitochondrial control region sequences of Steller’s sea lions: the three-stock hypothesis. Journal of Mammalogy, 86, 1075–1084.
6. Baker, J. D., Antonelis, G. A., Fowler, C. W., and York, A. E. 1995. Natal site fidelity in northern fur seals, Callorhinus ursinus. Animal Behavior. 50: 237-347.
7. Balsiger, J. A., et al. 2005. Setting the annual substance harvest of northern fur seals on the Pribilof Islands. Final environmental impact statement. National Oceanic and Atmosphere Administration.
8. Barnes, I., Metheus, P., Shapiro, B., Jensen, D., and Cooper, A. 2002. Dynamcis of Pleistocene population extinctions in Beringian brown bears. Science. 5562: 2267-2269.
9. Bickham, J. W., Loughlin, T. R., Calkins, D. G., Wickliffe, J.K., and Pattern J. C. 1998. Genetic variability and population decline in Steller sea lions from the Gulf of Alaska. Journal of Mammalogy. 79: 1930-1935.
10. Bickham, J. W., Patton, J.C., and Loughlin T.R. 1996. High variability for control-region sequences in a marine mammal: implications for conservation and biogeography of Steller sea lions (Eumetopias jubatus). Journal of Mammalogy 77: 95–108.
11. Bonnell M. L., and Selander, R. K. 1974. Elephant Seals: Genetic Variation and Near Extinction. Science. 184: 908-909.
12. Borge, T., Bachmann, L., BjØrnstad, G., and Wiig, Ø. 2007. Genetic variation in Holocene bowhead whales from Svalbard. Molecular Ecology. 16: 2223-2235.
13. Born, E. W., Anderson, L. W., Gjertz, I. and Wiig, Ø. 2001. A review of the genetic relationships of Atlantic walrus (Odobenus rosmarus rosmarus) east and
44
west of Greenland. Polar Biology. 24: 713-718. 14. Boskovic, R., Kovacs, K. M., Hammill, M. O., and White, B. N. 1996.
Geographic distribution of mitochondrial DNA haplotypes in grey seals (Halichoerus grypus) Canadian Journal of Zoology. 74: 1787-1796.
15. Burg, T. M., Trites, A. W., and Smith, M. J. 1999. Mitochondrial and microsatellite DNA analyses of harbour seal population structure in the northeast Pacific Ocean. Canadian Journal of Zoology. 77: 930–943.
16. Burton, R., Snodgrass, J. J., Gifford-Gonzalez, D., Guilderson, T., Brown, T., and Koch, P. L. 2001. Holocene changes in the ecology of northern fur seals: insights from stable isotopes and archaeofauna. Oecologia. 128: 107-115.
17. Burton, R., Koch, P. 1999. Isotopic tracking of foraging and long-distance migration in northeastern Pacific pinnipeds. Oecologia. 119:578–585.
18. Cassini. M. H. 1999. The evolution of reproductive systems in pinnipeds. Behavioral Ecology. 10: 612-616.
19. Coltman, D. W., Stenson, G., Hammill, M. O., Haug, T., Davis, C. S., and Fulton, T. L. 2007. Panmictic population structure in the hooded seal (Cystophora cristata). Molecular Ecology.16: 1639-1648.
20. Consuegra, S., García De Leániz, C., Serdio, A., González Morales. M., and Straus, L. G. 2002. Mitochondrial DNA variation in Pleistocene and modern Atlantic salmon from the Iberian glacial refugium. Molecular Ecology. 11: 2037-2048.
21. COSEWIC. 2006. COSEWIC assessment and update status report on the northern fur seal Callorhinus ursinus in Canada. Committee on the status of endangered wildlife in Canada, Ottawa.
22. Crockford, S. J., and Frederick, S. G. 2007. Sea ice expansion in the Bering Sea during the Neoglacial: evidence from archaeozoology. The Holocene. 17 (6): 699-706.
23. Davis, C. S., Stirling, I., Strobeck, C., and Coltman, D. W. 2008. Population structure of ice-breeding seals. Molecular Ecology. 17: 3078–3094.
24. DFO. 2007. Recovery potential assessment for Northern Fur Seals (Callorhinus ursinus). DFO. Can. Sci. Advis. Sec. Sci. Advis. Rep. 2007/052.
25. Drummond, A. J, Rambaut, A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology. 7: 214.
26. Drummond, A. J., Rambaut, A., Shapiro, B., and Pybus, O. G. 2005. Bayesian Coalescent Inference of Past Population Dynamics from Molecular Sequences. Molecular Biology and Evolution 2005 22(5):1185-1192.
27. Drummond, A. J., Pybus, O.G., Rambaut, A., Forsberg, R., and Rodrigo, A.G. 2003. Measurably evolving populations. Trends in Ecology and Evolution. 18: 481-488.
28. Excoffier, L., Laval, G., and Schneider, S. 2006. Arlequin 3.1: An Integrated Software Package for Population Genetics Data Analysis.
29. Freilich, S., Hoelzel, A. R., and Choudhury. S. R. Poster. Genetic diversity and
45
population genetic structure in the South American sea lion (Otaria flavescens). 30. Fu, Y. X. 1997. Statistical tests of neutrality of mutations against population
growth, hitchhiking and backgroud selection. Genetics. 147:915-925. 31. Fu, Y. X., and Li, W. H. 1993. Statistical tests of neutrality of mutations. Genetics.
133: 693-709. 32. Gales, N. J., Adams, M., and Burton, H. R. 1989. Genetic relatedness of two
populations of the Southern elephant seal, Mirounga leonina. Marine Mammal Science. 5: 57-67.
33. Gentry, R. L. 1998. Behavior and ecology of the northern fur seal. Princeton University Press, New Jersey, pp. 1-132.
34. Gentry, R., and Kooyman G. eds. 1986. Fur seals: maternal strategies on land and at sea. Princeton University Press, Princeton.
35. Goldsworthy, S., Francis, J., Boness, D., and Fleischer, R. 2000. Variation in the mitochondrial control region in the Juan Fernández fur seal (Arctocephalus philippii). The Journal of Heredity. 91: 371-377.
36. Goodman, S. J. 1998. Patterns of extensive genetic differentiation and variation among European harbor seals (Phoca vitulina vitulina) revealed using microsatellite DNA polymorphisms. Molecular Biology and Evolution. 15: 104-118.
37. Graves, J. A., Helyar, A., Biuw, M., Jüssi, M., Jüssi, I., and Karlsson, O. 2009. Microsatellite and mtDNA analysis of the population structure of grey seals (Halichoerus grypus) from three breeding areas in the Baltic Sea. Conservation Genetics. 10: 59-68.
38. Hadly, E.A. 1996. Influence of late-holocene climate on northern rocky mountain mammals. Quaternary Research. 46: 298-310.
39. Harlin-Cognato, A., Bickham, J. W., Loughlin, T. R., and Honeycutt, R. L. 2005. Glacial refugia and the phylogeography of Steller’s sea lion (Eumetopias jubatus) in the North Pacific. Journal of Evolutionary Biology.
40. Harwood, J., Stanley, H., Beudels, M-O., and Vanderlinden, C. 1996. Metapopulation dynamics of the Mediterranean monk seal. In: Metapopulations and wildlife conservation and management (McCullough, D. R. ed). Covelo, CA: Island Press; 241–256.
41. Higdon, J., Bininda-Emonds, O., Beck, R., and Ferguson, S. 2007. Phylogeny and divergence of the pinnipeds (Carnivora: Mammalia) assessed using a multigene dataset. BMC Evolutionary Biology 7: 216.
42. Hoelzel, A. R., Halley, H., O’Brien, S. J., Campagna, C., Arnbom, T., Boeuf, B. L., Ralls, K., and Dover, G. A. 1993. Elephant seal genetic variation and use of simulation models to investigate historical population genetics. Journal of Heredity. 84: 443-449.
43. Hoffman, J. I., Matson, C. W., Amos, W., Loughlin, T. R., and Bickham, J. W. 2006. Deep genetic subdivision within a continuously distributed and highly
46
vagile marine mammal, the Steller’s sea lion (Eumetopias jubatus) Molecular Ecology. 15: 2821-2832.
44. Hunt, G.L., P. Stabeno, G. Walters, E. Sinclair, R.D. Brodeur, J.M. Napp, and N.A. Bond. 2002. Climate change and control of the southeastern Bering Sea pelagic ecosystem. Deep-Sea Res. II 49:5821-5853.
45. Kappe, A. L., Bijlsma, R., Osterhaus, A. D. M. E., van Delden, W., and van de Zande, L.1997. Structure and amount of genetic variation at minisatellite loci within the subspecies complex of Phoca vitulina (the harbour seal). Heredity. 78: 457–463.
46. Kretzmann, M. B., Gilmartin, W. G., Meyer, A., Zegers, G. P., Fain, S. R., Taylor, B. F., and Costa, D. P. 1997. Low Genetic Variability in the Hawaiian Monk Seal. Conservation Biology. 11: 482-490.
47. Leboeuf, B. J., Kenyon, K. W., and Villa-Ramirez, B. 1986. The Caribbean Monk Seak is extinct. Marine Mammal Science. 2: 70-72.
48. Lento, G. M., Baker, C. S., David, V., Yuhki, N., Gales, N. J., and O’Brien, S. J. 2003. Automated single-strand conformation polymorphism reveals low diversity of a Major Histocompatibility Complex Class II gene in the threatened New Zealand sea lion. Molecular Ecology Notes. 3: 346-349.
49. Lento, G. M., Haddon, M., Chambers, G. K., and Baker, C. S. 1997. Genetic variation of southern hemisphere fur seals (Arctocephalus spp.): investigation of population structure and species identity. The Journal of Heredity. 88(3):202-208.
50. Librado, P., and Pozas J. 2009. DnaSP v5: a software for comprehensive analysis of DNA. Bioinformatics. 25: 1451-1452.
51. Loughlin, T. R., Antonelis, G.A., Baker, J.D., York, A.E., Fowler, C.W., DeLong, R.L., and Braham, H.W. 1994. Status of the northern fur seal population in the United States during 1992. E. H. Sinclair, ed. Fur seal investigations, 1992. U. S. Department of Commerce, Seattle, WA. pp 9-78.
52. Macklin, S. A., Schumacher, J., Moore, S. E., and Smith S. 2008. Sustaining the marine ecosystem of the Pribilof Domain. Deep-Sea Research. II 55: 1698-1700.
53. MacPhee, R. D., Tikhonov, A. N., Mol, D., Greenwood, A. D. 2005. Late Quaternary loss of genetic diversity in muskox (Ovibos). BMC Evolutionary Biology. 5, doi: 10.1186/1471-2148-5-49.
54. Maldonado, J. E., Davila, F. O., Stewart, B. S., Geffen, E., and Wayne, R. K. 1995. Intraspecific genetic differentiation in California sea lions (Zalophus californianus) from southern California and the Gulf of California. Marine Mammal Sciences. 11: 46-58.
55. Matthee, C. A., Fourie, F., Oosthuizen, W. H., Meyër, M. A., and Tolley, K. A. 2006. Mitochondrial DNA sequence data of the Cape fur seal (Arctocephalus pusillus pusillus) suggest that population numbers may be affected by climatic shifts. Marine Biology. 148: 899–905.
56. Mizuno, A. W., Onuma, M., Takahashi, M., and Ohtaishi, N. 2003. Population
47
genetic structure of the Spotted seal Phoca largha along the Coast of Hokkaido, based on mitochondrial DNA sequences. Zoological Science. 20: 783–788.
57. Moss, M. L., Yang, D. Y., Newsome, S. D., Speller, C. F., McKechnie, I., McMillan, A. D., Losey R. J., and Koch P. L. 2006. Historical ecology and biogeography of North Pacific pinnipeds: isotopes and ancient DNA from three archaeological assemblages. Journal of Island & Coastal Archaeology. 1: 165-190.
58. Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University Press, New York.
59. Newsome, S. D., Etnier, M. A., Gifford-Gonzalez, D., Phillips, D.L., van Tuinen, M., Hadly, E. A., Costa, D. P., Kennett, D. J., Guilderson, T. P., and Koch, P. L. 2007. The shifting baseline of northern fur seal ecology in the northeast Pacific Ocean. PNAS. 104:9709-9714.
60. NMML (National Marine Mammal Laboratory), 2007 NMML (National Marine Mammal Laboratory), 2007. Northern fur seals: history and exploitation. WWW Page, http://www.afsc.noaa.gov/nmml/alaska/nfs/exploitation.php, (retrieved December 2007).
61. O’ Brien, S. J., Wildt, D. E., Bush, M., Caro, T. M., FitzGibbon C., Aggundey, I., and Leakey, R. E. 1987. East African cheetahs: evidence for two population bottlenecks? PNAS. 84: 508-511.
62. Palo, J. U., Hyvärinen, H., Helle, E., Mäkinen, H. S., and Väinölä. R. 2003. Postglacial loss of microsatellite variation in the landlocked Lake Saimaa ringed seal. Conservation Genetics 4: 117–128.
63. Palo, J. U., Mäkinen, H. S., Helle, E., Stenman, O., and Väinölä, R. 2001. Microsatellite variation in ringed seals (Phoca hispida): genetic structure and history of the Baltic Sea population. Heredity 86: 609-617.
64. Pastor, T., Garza, J. C., Allen, P., Amos, W., and Aguilar, A. 2004. Low Genetic Variability in the Highly Endangered Mediterranean Monk Seal. Journal of Heredity. 95(4): 291-300.
65. Pääbo, S. 1989. Ancient DNA: extraction, characterization, molecular cloning, and enzymatic amplification. PNAS. 86: 1939-1943.
66. Perrin, W., Wursig B., and Thewissen J. eds. 2002. Encyclopedia of marine mammals. Academic, San Diego. pp 1414.
67. Pinsky, M.L., Newsome, S., Dickerson, B., van Tuinen, M., Kennett, D., Ream, R., and Hadly E.A. unpublished. Refugia and high migration rates buffered Callorhinus ursinus from genetic impacts of range contraction and population decline.
68. Rawlence, N. J., Wood, J. R., Armstrong, K. N., and Cooper A. 2009. DNA content and distribution in ancient feathers and potential to reconstruct the plumage of extinct avian taxa. Proceedings of the Royal Society. 7: 1-8.
69. Razjigaeva, N. G., Grebennikova, T. A., Ganzey, L. A., Mokhova, L. M., Bazarova, V. B. 2004. The role of global and local factors in determining the
48
middle to late Holocene environmental history of the South Kurile and Komandar islands, northwestern Pacific. Palaeogeography, Palaeoclimatology, Palaeoecology. 209: 313– 333.
70. Ream, R., 2002. Molecular ecology of North Pacific Otariids: genetic assessment of Northern fur seal and Steller sea lion distributions. Ph. D. thesis.
71. Ream, R. R., Sterling J. T., and Loughlin T. R. 2005. Oceanographic features related to northern fur seal migratory movements. Deep-sea research part II- Topical studies in oceanography. 52 (5-6): 823-843.
72. Ream, R., and Burkanov, V. 2005. Trends in abundance of Steller sea lions and Northern fur seals across the North Pacific Ocean. PICES XIV Annual Meeting, Vladivostock, Russia.
73. Rice, D. W. 1998. Marine mammals of the world – Systematics and distribution (special publications of the society of marine mammalogy speical publications 4). Allen Press, Inc., Lawrence, Kansas. pp 231.
74. Rozas, J., Sanchez-DelBarrio, J.C., Messeguer, X., and Rozas, R. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19: 2496-2497.
75. Sakahira, F., and Niimi, M. 2007. Ancient DNA Analysis of the Japanese Sea Lion (Zalophus japonicus japonicus Peters, 1866): Preliminary results using mitochondrial control-region sequences. Zoological Society of Japan. 24: 81-85.
76. Slade, R. W., Moritz, C., Hoelzel, A. R., and Burton. H. R. 1998. Molecular population genetics of the Southern elephant seal Mirounga leonina. Genetics. 149: 1945-1957.
77. Slade, R. W., Moritz, C., and Heideman, A. 1994. Multiple nuclear-gene phylogenies: application to pinnipeds and comparison with a mitochondrial DNA gene phylogeny. Molecular Biology and Evolution. 11: 341-356.
78. Stanley, H. F., and Harwood, J. 1997. Genetic differentiation among subpopulations of the highly endangered Mediterranean monk seal. In: The role of genetics in conserving small populations (Tew, T. E., Usher, M. B., Crawford, T. J., Stevens, D., Warren, J., and Spencer, J., eds). Peterborough: Joint Nature Conservation Committee. pp 97–101.
79. Stanley, H. F., Casey, S., Carnahan, J. M., Goodman, S., Harwood, J., and Wayne, R. K. 1996. Worldwide patterns of mitochondrial DNA differentiation in the harbor seal (Phoca vitulina). Molecular Biology and Evolution. 13: 368-382.
80. Tajima, F. 1993. Measurement of DNA polymorphism. pp. 37-59. In Takahata, N. and Clark, A. G. (eds), Mechanisms of Molecular Evolution, Sinauer Associates. Inc., Sunderland, Massachusetts.
81. Tajima, F. 1989. Statistical method for testing the neutral mutationhypothesis by DNA polymorphism. Genetics. 123: 585-595.
82. Towell, R. G. and Ream, R. R. 2006. Decline in Northern fur seal (Callorhinus ursinus) pup production on the Pribilof Islands. Marine Mammal Science. 22 (2):
49
486-491. 83. Trujillo, R. G., Loughlin, T. R., Gemmell, N. J., Patton, J. C., and John W.
Bickham, J. W. 2004. Variation in microsatellites and mtDNA across the range of the Steller sea lion, Eumetopias jubatus. Journal of Mammalogy 85(2):338-346.
84. Túnez, J. I., Centrón, D., Cappozzo H. L., and Cassini, M. H. 2007. Geographic distribution and diversity of mitochondrial DNA haplotypes in South American sea lions (Otaria flavescens) and fur seals (Arctocephalus australis). Mammalian Biology. 72: 193-203.
85. Valdiosera, C. E., Garcia-Garitagoitia, J. L., Garcia, N. et al. 2008. Surprising migration and population size dynamics in anicent Iberian Brown bears (Ursus arctos). Proceeding of National Academy of Sciences, USA. 105: 5123-5128.
86. Valentine, K., Duffield, D. A., Patrick, L. E., Hatch, D. R., Butler, V. L., Hall R. L., and Lehman, N. 2007. Ancient DNA reveals genotypic relationships among Oregon populations of the sea otter (Enhydra lutris). Conservation Genetics.
87. Vermeij, G. J. 1993. Biogeography of Recently Extinct Marine Species: Implications for Conservation. Conservation Biology. 7: 391-397.
88. Watterson, G. A. 1975. On the number of segregating sites in geneticalmodels without recombination. Theor.Pop. Biol. 7: 256-276.
89. Wayne, R. K., Leonard, J. A., and Cooper, A. 1999. Full of sound and fury: the recent history of ancient DNA. Annu. Rev. Ecol. Syst. 30: 457-477.
90. Weber, D. S., Stewart, B. S., and Lehman, N. 2004. Genetic consequences of a severe population bottleneck in the Guadalupe fur seal (Arctocephalus townsendi). Journal of Heredity. 95 (2): 144-153.
91. Weber, D. S., Stewart, D., Garza, J., Lehman, N. 2000. An empirical genetic assessment of the severity of the Northern elephant seal population bottleneck. Current Biology. 10: 1287-1290.
92. Westlake, R. L., and O'Corry-Crowe, G. M. 2002. Macrogeographic structure and patterns of genetic diversity in Harbor seals (Phoca vitulina) from Alaska to Japan. Journal of Mammalogy. 83: 1111-1126.
93. Wynen L. P., Goldsworthy, S. D., Insley, S. J., Adams, M., Bickham, J. W., Francis, J., Gallo, J. P., Hoelzel, A. R., Majluf, P., White, R. W. G., and Slade, R. 2001. Phylogenetic relationships within the eared seals (Otariidae: Carnivora): implications for the historical biogeography of the family. Molecular Phylogenetics and Evolution. Vol. 21, No. 2, November, pp. 270-284.
94. Wynen, L. P., Goldsworthy, S. D., Guinet, C., Bester, M. N., Boyd, I. L., Gjertz, I., Hofmeyr, G. J. G., White, R. W. G., and Slade, R. 2000. Postsealing genetic variation and population structure of two species of fur seal (Arctocephalus gazella and A. tropicalis). Molecular Ecology. 9: 299-314.
95. Yang, D. Y., Eng, B., and Saunders, S. R. 2003. Hypersensitive PCR, Ancient Human mtDNA, and Contamination. Human Biology 75: 355-364.
96. York, A. E. 1987. Northern fur seal, Callorhinus ursinus, Eastern Pacific
50
population (Pribilof Islands, Alaska, and San Miguel Island, California). In J. P. Croxall and R. L. Gentry, eds., Status, Biology, and Ecology of Fur Seals. NOAA Tech. Rep. NMFS 51: 133-140, British Antarctic Survey, Cambridge, England.
51
CHAPTER 2
DEMOGRAPHIC HISTORY OF CALLORHINUS URSINUS FROM ALEUTIAN
ISLANDS
INTRODUCTION
Callorhinus ursinus is the only species in the genus Callorhinus (Wynen et al. 2001),
and it belongs to the family Otariidae, and the class Pinnipedia (Rice 1998). Callorhinus
ursinus has a pelagic distribution that ranges across the North Pacific (Ream et al. 2005).
The Pribilof Islands in the east Bering Sea have served as primary breeding sites for
millennia, and accommodate the largest population worldwide (Towell and Ream 2006).
As one of the classic examples in pinnipeds showing strong polygyny, C. ursinus also
presents natal site fidelity, where females from the Pribilofs wean their pups and move
south with them, and return to the same islands the following March (Gentry 1998).
For the past hundreds of years with an apex during the late 1700s, C. ursinus has
experienced several population collapses and recoveries. The Pribilofs population, in
particular, practiced dramatic fluctuations in size (Loughlin et al. 1994; Gentry 1998;
Balsiger et al. 2005; Newsome et al. 2007). Human exploitation in the last centuries led
to further declines in C. ursinus abundance. The most serious one was in the late 19th
century, resulting in a low of 216,000 animals in 1912, and corresponding to a 90%
population decline of the pre-exploitation size (Balsiger et al. 2005; National Marine
Mammal Laboratory 2007; Macklin et al. 2008). After this decline a law was signed to
52
limit hunting, and the population recovered in abundance. However, the population
declined again by 50-60% from 1974-2004 (COSEWIC 2006; Adams et al. 2007) for
unknown reasons. The species currently is officially considered as “threatened” under the
Marine Mammal Protection Act of 1988 (Loughlin et al. 1994), and under Species at Risk
Act (SARA) by the Committee on the Status of Endangered Wildlife in Canada
(COSEWIC) (DFO 2007).
Because of the drastic population contraction in the late 19th Century, C. ursinus
may have suffered from a genetic bottleneck, similar to other pinnipeds. In Chapter 1, I
used ancient DNA techniques to compare ancient and modern populations from four
Aleutian Islands adjacent to the Pribilofs (see Figure 2 in Chapter 1). The archaeological
sub-fossils were collected from Kodiak (samples were dated as old as 750 ybp; ybp
stands for years before present), Shemya (1,500 ybp), and Unalaska (3,000 ybp) while the
modern samples were collected from Bogoslof Island. Both ancient and modern
populations showed similar high level of haplotype diversity, high nucleotide diversity
and low genetic structure. The high genetic variation within C. ursinus before and after
the serious historical commercial hunting thus implied no genetic impact of the most
recent population bottleneck (see also Ream 2002; Pinsky et al. unpublished). However,
unlike other land-breeding pinnipeds with natal fidelity and polygyny, C. ursinus
exhibited genetic homogeneity that demonstrated a high extent of gene flow among the
geographically distant islands.
No populations showed significant difference in estimates of genetic diversity,
53
except Unalaska, which contained lower nucleotide diversity and fewer nucleotide
differences (see Table 1 in Chapter 1). Understanding of past demographic responses will
allow us to better predict for future changes. Thus, taking a closer look at each island
independently, I wanted to evaluate demographic history of individual Aleutian
populations by testing the size and timing of population expansion/contraction.
The direction and degree of population changes can be inferred from mismatch
distribution tests (Rogers and Harpending 1992) and statistics such as Tajima’s D and
Fu’s Fs tests (Tajima 1989; Fu 1997; Excoffier et al. 2006). To approach particular time
depth of demographic expansion or contraction, the Bayesian Skyline Plot recently was
developed to infer past changes in population size (Drummond et al. 2003; Finlay et al.
2009). Under coalescent theory, which demonstrates the relationship of demographic
history and root probabilities in a genealogical tree, the probability of lineages coalescing
at a particular time is inversely proportional to the population size at the same time
(Drummond et al. 2003). A Bayesian skyline therefore estimates effective population size
during the intervals between coalescences based on interval lengths (Pybus et al. 2000;
Drummond et al. 2003). BEAST (Bayesian Evolutionary Analysis by Sampling Trees) is
the program that constructs Bayesian skyline plots (Drummond et al. 2003), and it uses
Markov chain Monte Carlo (MCMC) sampling to explore the likelihood of parameter
estimates and genealogies fitting the data. BEAST provides a posterior distribution of
most likely genealogies and parameters, as well as population size (Ne) at different time
points (Drummond et al. 2003; Drummond et al. 2005).
54
Molecular sequences either sampled simultaneously (synchronously) or serially
(heterochronously) through time, can be used to reconstruct demographic histories
(Drummond et al. 2003). BEAST accounts for serial sampling through time instead of
assuming that all tips are of the same age so ancient DNA can be applied
heterochronously, with sufficiently measurably sampling ages (Drummond et al. 2003).
Modern genetic data can predict past molecular evolutionary processes, while ancient
genetic sequence can observe such predictions and determine lineage coalescences in
action with rate and magnitude estimation (Drummond et al. 2003; Chan et al. 2006).
Ancient DNA has the ability to estimate the timing and severity of population declines
(Chan et al. 2006), formation or cessation of gene flow (Hadly et al. 2004; Hofreiter et al.
2004; Depaulis 2009), and even has been utilized to estimate substitution rate (Lambert
et al. 2002; Ho et al. 2007).
In order to obtain a Bayesian skyline, BEAST requires reasonable priors on
population size and genetic parameters because the Bayesian posterior is a function of
these priors (Drummond et al. 2005). For example, the mutation rate of the molecular
marker serves as a critical prior for interpreting the relationship between coalescence
event and population size. Ho et al. (2005, 2007, 2008) hypothesized the time
dependency of substitution rate, which suggested interspecific calibrations versus
intraspecific calibrations may lead to disparate evolutionary scenarios. Deeper calibration
points result in slower observed rates and older coalescences (Ho et al. 2008). If an
evolutionary rate is calibrated from an external (deeper) calibration point, the
55
intraspecific rate will be underestimated (Ho et al. 2008). Therefore, employing deep
fossil calibrations or canonical substitution rates (e.g., broadly applied rate of 1%
substitution per site per million years for mammals) are not proper to apply to
population-level data (Ho et al. 2008). The idea of time-dependent substitution rate has
not been unanimously accepted (Bandelt et al. 2006; Emerson 2007; Bandelt 2008). For
example, Bandelt (2008) argued that Ho used inadequate data in his study, and not every
data set supports the time dependency theory. Therefore, it is best to verify on a
case-by-case basis, whether disparity exists in intraspecific vs. interspecific substitution
rate. In my case, strategy of selecting proper parameters for rate calculation should be
carefully considered. Therefore, distinguishing inter- from intra- specific substitution
rates is desirable for demographic studies when rate differences are suspected.
Rates of mutation and substitution show variation among lineages and genes.
Nucleotide changes within a population or species are often transient and can be removed
by genetic drift or selection (Penny 2005; Woodhams 2006), while in the long term
segregating sites of sequences taken among species are fixed with higher probability
(Penny 2005; Woodhams 2006). Because not all mutations can be fixed, calculating a
rate from only fixed substitutions will result in lower long-term rates compared with that
estimated from all short-term mutations incorporated. Analyses of the rapidly evolving
non-coding mitochondrial control region have yielded great uncertainty in the rates of
nucleotide substitution across both nucleotide positions and genealogies (Alter &
Palumbi 2009), because of such features as rate heterogeneity across positions
56
(mutational hotspots), bias in nucleotide composition, and substitutional saturation (Ho et
al. 2005, 2007; Woodhams 2006; Alter and Palumbi 2009). In modern DNA studies,
mutation rates of control region are often higher than those of mitochondrial protein
coding regions such as cytochrome b gene. The mutation rate of control region varies
considerably among species with documented rates ranging from as high as 5-10%
substitutions/site/my in pinnipeds (Slade et al. 1994), 43% substitutions/site/my in brown
bears (Drummond et al. 2003), 93% substitutions/site/my in Adélie penguins (Lambert et
al. 2002), 83%-234% substitutions/site/my in Tuatara (Hay et al. 2008) and 32%-260%
substitutions/site/my in humans (Parsons et al. 1997; Sigurdardóttir et al. 2000; Howell et
al. 2003). In contrast, rate variation of the slower cytochrome b gene is less drastic
among species. For example, Hawaiian drepanidines show 0.016 sequence divergence
per million years (Fleischer et al. 1998), which is similar to the traditionally recognized
substitution rate of 1% substitutions per site per million years for protein-coding
mitochondrial DNA (Ho et al. 2008; Hay et al.2008). Because of a slower substitution
rate, back substitutions accumulate more gradually, which makes the cytochrome b gene
appropriate especially for interspecific rate calibration. Therefore, a combination of
markers appropriate below (control region) and above (cytochrome b) the species level is
a valid strategy to better understand the interplay of mutation rate and population size
changes.
Effective population size is another important parameter in the Bayesian skyline
plot. Under the neutral theory (e.g. no selection, no migration, no mutation), genetic
57
diversity within a population is a formula of θ= 2Nef *µ in mitochondrial DNA, where
Nef is the female genetic effective population size, and µ is the mutation rate per site per
generation (Slade et al. 1998; Pavlova et al. 2005), applied from θ= 4Ne *µ for an
autosomal gene of a diploid organism (Fu and Li 1993). Total Ne is composed of male
and female, where Nem is the number of male parents and Nef is the number of female
parents, and effective population size is given by Ne=4NemNef/(Nem+Nef) when the sex
ratio varies from the 1:1 ratio (Wright 1931; Crow & Denniston 1988; Caballero 1994).
Although this Ne does not exactly equal to the total observed population count, the
direction of change in Ne through time (e.g. population expansion or contraction) will
likely reflect actual population size changes.
Thus, I propose in this article to analyze the past demographic history of Aleutian C.
ursinus using DnaSP, Arlequin and BEAST to compare the direction, degree and time
depth of population changes among different islands. Specifically, I used DnaSP and
Arlequin to access general population estimates such as Tajima’s D, Fu & Li’s and Fu’s
Fs tests and mismatch distributions. To best appreciate possible variation in mutation rate
with time depth, I jointly compared the inter- and intraspecific rates by grouping different
outgroup species with C. ursinus, and confirmed these rates with pairwise measures of
cytochrome b and control region distances.
58
METHODS AND ANALYSES
1. Sample selection
I used a total of 58 sequences of 157 bp mitochondrial control region from 3 ancient
populations, with 21 sequences from Rolling Bay on Kodiak Island (750 ybp; ybp stands
for years before present), 18 sequences from Shemya Island (1500 ybp), and 19 from
Amaknak Bridge on Unalaska Island (3000 ybp). I had access to unpublished modern
data (samples collected from 1993 to 1998; n=365) along the east Pacific sites (Pinsky et
al. unpublished; Dickerson et al. unpublished). I compared my ancient DNA results to
modern data from the Bogoslof population (96 sequences) because of its geographic
proximity to other Aleutian islands. I used MEGA (version 3.0) to build a neighbor
joining tree and randomly picked 68 sequences that represent the haplotype network from
the available 365 modern sequences, which included populations from St. George Island,
St. Paul Island, Bogoslof Island and San Miguel Island. The 68 modern DNA samples
plus the 58 ancient DNA samples were combined together to calculate the intraspecific
clock rate.
2. General population estimates using DnaSP and Arlequin
I performed several analyses of population genetic variation and structure using the
programs DnaSP (version 5.0) (Librado &Rozas 2009) and Arlequin (version 3.1)
(Excoffier et al. 2006). I estimated Tajima’s D, Fu & Li’s D, Fu & Li’s F, and Fu’s Fs in
modern and ancient populations, as tests of selective neutrality or possible population
59
expansion or contraction by the program DnaSP (Tajima 1989, 1993; Fu & Li 1993; Fu
1997). I used Arlequin to plot mismatch distributions of populations from the four islands
independently, and an additional plot with 3 ancient populations combined together.
3. Rate comparison between control region and cytochrome b region
I used Arlequin to plot mismatch distributions to compare the ratio of rates between
the cytochrome b region and control region. I obtained 10 sequences of 157 bp
cytochrome b region, corresponding with sequences of 157 bp control region from the
same animals. I also obtained another 5 sequences of 121 bp cytochrome b region,
coupled with 5 sequences of control region from the same animals. I obtained the means
of all the pairwise distributions to compare the rate differences of the two molecular
markers.
The fossil record was used to estimate the interspecific rate of cytochrome b region
using BEAST. Calculations were performed using each prior independently and
combining all priors (Arnason et al. 2006; Higdon et al. 2008; Koepfli et al. 2008). The
coalescent prior used here was the Yule process coalescent (speciation model). Rate was
estimated under the HKY+Gamma+Invariant (HKY+G+I) substitution model, with
partition into codon positions as 1+2, 3. The molecular clock model was uncorrelated
lognormal relaxed clock. Priors of root height were (mya stands for million years ago):
Pinniped-Mustelid, uniform distribution (23-52 mya)
Otariidae, uniform distribution (6-52 mya)
Odobenidae-Otariidae, uniform distribution (18.2-52 mya)
60
Phocidae, uniform distribution (14.6-52 mya)
Miroungini, normal distribution (7 MYA mean; 0.5 mya standard deviation)
Monachini, normal distribution (8 MYA mean, 0.5 mya standard deviation)
Mustela, normal distribution (5.3 MYA mean; 0.5 mya standard deviation)
Mustelidae, uniform distribution (10-52 mya)
Pinnipedia, uniform distribution (23-52 mya)
4. Calculation of intraspecific clock rate
I used BEAST (v 1.4.8) to calculate the intraspecific clock rate within C. ursinus
(Drummond & Rambaut 2007). The programs distributed as part of the BEAST package
included BEAUti, BEAST, and LogCombiner. Specifically, BEAUti was used for
specifying priors to run BEAST; LogCombiner was a program used to combine log and
tree files from multiple runs of BEAST. Tracer (v 1.4.1) was used for analysing results
from Bayesian MCMC programs. To calculate the clock rate appropriately considering
the problem of time inequality above and below species level, I designed different sets of
parameters, and picked up the most suitable rate of intraspecific clock. I used 8.2 mya as
the divergence time between C. ursinus and sea lions, where the sea lions included
Eumetopias jubatus (Steller sea lion), Zalophus californianus (California sea lion) and
Zalophus californianus japonicus (Japanese sea lion), which was applied in many studies
(Slade et al. 1994; Higdon et al. 2007; Pinsky et al. unpublished). The data set included
the 126 individuals of C. ursinus as the majority plus one individual from each outgroup
species. The sequences of outgroup species were obtained from GenBank and aligned
61
with the 157 bp control region of C. ursinus. Nexus file of the total 129 samples was
imported in BEAUti, while the ages of samples were not input because they were not
considered as sufficiently long compared with the long internal nodes of the genealogy
(8.2 mya) (Drummond et al 2003). Rate was estimated under the
HKY+Gamma+Invariant (HKY+G+I) substitution model with strict clock. Priors were
set up as coalescent constant size, with population size as a uniform distribution from 0
to 60 million, and tree model root height as a normal distribution using 8.2 mya as the
mean, and 2.1 mya as the standard deviation. Samples from the posterior were drawn
every 20,000 MCMC steps over a total of 20,000,000 steps, with the first 10% discarded
as burn-in. I assessed convergence of each MCMC run in Tracer and by checking for
sufficient ESS values (>200).
5. Comparing inter- and intra- specific rates
I compared the inter- and intra- specific rates by grouping the 126 individuals of C.
ursinus with different numbers from outgroup sea lions: Eumetopias jubatus (Steller sea
lion), Zalophus californianus (California sea lion) and Zalophus californianus japonicus
(Japanese sea lion). Rate was estimated under the HKY+G+I substitution model with
strict clock in BEAST. Priors were set up as coalescent constant size, and tree model root
height as a normal distribution using 8.2 mya as the mean, and 2.1 mya as the standard
deviation. I grouped species in different ways such as: a) 126 C. ursinus plus one
Eumetopias jubatus; b) 126 C. ursinus plus 30 Eumetopias jubatus; or c) 126 C. ursinus
62
plus 57 outgroup species which included 30 Eumetopias jubatus, 23 Zalophus
californianus and 4 Zalophus californianus japonicus. Numbers of sea lion comparisons
were based on all available Genbank sequences for these species. In general, the
posterior output of a) represented intraspecific rate, and c) represented interspecific rate.
As the number of outgroup species increased, the rate decreased and approached the
interspecific rate. Without the basic knowledge of the root height within Callorhinus, I
had to add in the interspecific node (8.2 mya) to simulate the intraspecific rate, but the
data contained mostly Callorhinus sequences.
6. Bayesian skyline plot
I built skyline plots of Kodiak, Shemya, Unalaska and Bogoslof independently,
without any outgroup species by BEAST. Here I added in the sample ages to directly
measure population dynamics in the recent thousands of years. Bayesian analyses were
performed using HKY+G+I model with strict clock and a Bayesian skyline coalescence
prior, and prior of population size as a uniform distribution from 0 to 60 mya. The
posterior output intraspecific clock rate from the first BEAST run (see step 4) was
applied to plot the Bayesian skyline. Samples from the posterior were drawn every
20,000 MCMC steps over a total of 20,000,000 steps, with the first 10% discarded as
burn-in. I used the program Tracer (v 1.4.1) to examine the posterior output of each
population. The medians of skyline population sizes were plotted through time using
posterior tree model root height, respectively, and the four Aleutian populations were
63
plotted together. The female effective population size can be obtained from the function
of Ne*Tau, where Tau is the female generation time of 10 years (Lander 1981; Wickens
1997).
64
RESULT
1. Estimates of population changes using the program DnaSP
Values of Tajima’s D, Fu & Li’s D, and Fu & Li’s F do not show significant
deviation from the null (stable size) model (P>0.10). The direction of Tajima's D, Fu &
Li’s D, and Fu & Li’s F is potentially informative about the evolutionary and
demographic forces that a population has experienced (Tajima 1989, 1993; Fu & Li 1993;
Fu 1997; Excoffier et al. 2006). Negative values reflect an excess of rare polymorphisms
in a population, which is consistent with either positive selection or an increase in
population size. Positive values indicate an excess of intermediate-frequency alleles and
can result from a population decline. The tests identify whether sequences fit the neutral
theory model with a null hypothesis of selective neutrality and population equilibrium
(Tajima 1989, 1993; Fu & Li 1993; Fu 1997; Excoffier et al. 2006). Here the values of
Tajima’s D, Fu & Li’s D, and Fu & Li’s F are not informative enough to indicate a signal
of population expansion or contraction in any population (Table 7).
Fu’s Fs statistic is very sensitive to population demographic expansion or
contraction, which generally leads to large negative (or positive) Fs values (Fu 1997;
Excoffier et al. 2006). Similar to Tajima’s D, the null hypothesis of Fu’s Fs test is
selective neutrality and population equilibrium. Fu’s Fs indeed appears more sensitive to
departure from population equilibrium than Tajima’s D (Table 7). The Fs values of
populations from Unalaska, Shemya, “Total ancient” and Bogoslof all show significant
65
signs of population expansion (P<0.02).
2. Demographical history generated from mismatch distributions
Distributions of pairwise differences (mismatch distributions) have been extensively
used to estimate the demographic parameters of past population expansions (Schneider &
Excoffier 1999; Atarhouch et al. 2006; Curtis et al. 2009). Mismatch indices give the
direction, time and rate of demographic changes. Opposed demographic changes (e.g.
population expansion versus bottleneck) display recognizable disparate signatures in
mismatch distributions (Harpending et al. 1998; Schneider & Excoffier 1999). I used
Arlequin (version 3.1) to simulate the “expansion” model and DnaSP (version 5.0) to
simulate the “constant” model. Figure 5 (a-e) shows mismatch distributions of 3 Aleutian
ancient populations (Kodiak, Shemya and Unalaska), the combined ancient populations,
and one Aleutian modern population (Bogoslof). The real data (observed) of all the
populations fit “expansion” model better than “constant” model. Unalaska is the one
showing most recent expansion while the other two ancient and the modern samples
show earlier expansion, because the peak of the distribution (tau) for Unalaska is 4.125
(Figure 5 c), which is almost half of the other three Aleutian populations (8.662, 8.559,
8.475 respectively, see Figure 5 a, b, e).
66
Population Tajima’s D Fu & Li’s D Fu & Li’s F Fu’s Fs Kodiak 0.138 0.127 0.152 -3.530 Unalaska -0.596 -0.566 -0.667 -10.900 * Shemya 0.046 -0.135 -0.095 -5.080 * Total ancient -0.345 -0.750 -0.717 -25.000 * Bogoslof -0.359 0.648 0.294 -24.800 *
Table 7 Comparison population changes in Alaskan C. ursinus populations of different age using DnaSP
. *P<0.02. Significance of Fu’s Fs: Kodiak (P=0.069); Unalaska (P=0); Shemya (P=0.009); Total ancient (P=0); Bogoslof (P=0).
67
a)
b)
c)
68
d)
e)
Figure 5 Mismatch distributions of ancient and modern Aleutian populations of C. ursinus.
a) Kodiak; b) Shemya; c) Unalaska; d) Ancient samples together (58 individuals); e) Bogoslof (modern). X-axis shows pairwise differences between each two sequences those are randomly picked, y-axis is the frequency of the differences. The “observed” are distribution of real data; “constant population” is a model of stable population size simulated by DnaSP (version 5.0); “simulated expansion” is a simulated model of population expansion by Arlequin (version 3.1), with the up bound and low bound 95% confidence interval. Means of pairwise differences: a) 7.729, b) 7.941, c) 5.327, d) 7.718, e) 7.595; taus (location of crest) of pairwise differences: a) 8.662, b) 8.559, c) 4.125, d) 7.145, e) 8.475.
69
3. Rate comparison between different mitochondrial genes
Mismatch distributions of cytochrome b region and control region are plotted using
Arlequin (version 3.1). In figure 6, the mean ratio of control region to cytochrome b rate
for 15 samples of 121 bp of control region (pmean=6.59) and f cytochrome b region
(pmean=0.61) equals 10.8. Similar estimates are obtained from the 15 samples of 157 bp
control region (pmean=7.66) against 15 samples of 121 cytochrome b(pmean=0.61); and
from 10 samples of 157 bp control region (pmean=6.44) against15 samples of 121
cytochrome b (pmean=0.87): 12.6 and 7.6 respectively. The estimated rate of control
region is roughly ten-fold that of the cytochrome b rate (range=7.6-12.6). A ten-fold
difference in rate is supported by the fossil calibration data that shows an interspecific
rate of cytochrome b gene at 1% substitution/site/my, which is about 1/10th that of the
interspecific rate estimate for the control region.
70
a)
b)
Figure 6 Rate comparisons between control region and cytochrome b region.
a) 15 sequences of 121 bp cytochrome b region paired with 15 sequences of 121 bp and 157 bp control region; b) 10 sequences of 157 bp cytochrome b region paired with 10 sequences of 157 bp control region. Paired sequences are from same individuals of C. ursinus. X-axis shows pairwise differences between each two sequences those are randomly picked, y-axis is the frequency of the differences.
71
4. Heterogeneity of inter- and intra-specific rates
The result of comparisons of inter- and intra-specific rates of control region is
demonstrated in Table 8. The prior parameters are identical except the strategy to group
C. ursinus with different outgroup species. Number 1-5 represent intraspecific rates of
C. ursinus with samples being primarily from C. ursinus. Number 9-11 stand for
intraspecific rates within outgroup species using C. ursinus as the minority. Number 12
and 13 represent interspecific rates of C. ursinus with several outgroup species and
individuals. The intraspecific clock rate of C. ursinus is estimated as 0.192-0.357
substitutions per site per million years, and the interspecific rate is computed as
0.097-0.105 substitutions per site per million years, about a two-fold difference. To
investigate the effect of adding in priors on the sample ages, I performed the same
analyses with the same priors and ages, and obtained a rate of 2.348 substitutions per
site per million years, which I considered too high for control region.
72
C. ursinus + outgroup species Output clock rate (substitutions
per site per mya) Number
1 126 C. ursinus + 3 outgroups 0.213 2 126 C. ursinus + 1 Z. californianus japonicus 0.198 3 126 C. ursinus + 1 Z. californianus 0.230 4 126 C. ursinus + 1 Z. californianus 0.357 5 126 C. ursinus + 1 E. jubatus 0.192 6 126 C. ursinus + 30 E. jubatus 0.152 7 126 C. ursinus + 23 Z. californianus 0.569 8 126 C. ursinus + 4 Z. californianus japonicus 0.183 9 1 C. ursinus + 30 E. jubatus 0.073 10 1 C. ursinus + 23 Z. californianus 0.249 11 1 C. ursinus + 4 Z. californianus japonicus 0.185 12 126 C. ursinus + 57 outgroups 0.097 13 68 C. ursinus + 57 outgroups 0.105
Table 8 Comparisons of inter- and intra-specific clock rates of control region using C. ursinus and outgroup species
Number 1-8, 12: the 126 C. ursinus are made up of 58 ancient samples and 68 modern samples. Number 9-11: the one C. ursinus is randomly picked up from the 126 individuals (same individual in different runs). Number 13: the 68 C. ursinus are all modern samples Number 1: the 3 outgroups include one individual picked randomly from Z. californianus japonicus, Z. californianus and E. jubatus respectively. Number 2-5: one outgroup species is randomly picked in each run. Two samples of Z. californianus are picked separately and calculated independently (Number 3, 4). Number 6-11: Grouped outgroup species are combined with 126 C. ursinus Number 12, 13: the 57 outgroups include 30 E. jubatus, 23 Z. californianus, and 4 Z. californianus japonicus
73
5. Demographical history generated by Bayesian skyline plot
Bayesian skyline plots summarize estimates of timing of population changes and
effective population size through time (Figure 7, 8). The medians of skyline population
sizes were plotted through time (Figure 8a), the clock rate was 0.213
substitutions/site/mya. All four populations show signals of population growth, though
starting at different times. Of the three ancient populations, the Kodiak population
expanded 250,000 ybp and practiced a population contraction after the size reached an
apex at 110,000 ybp. The Shemya population expanded later at 158,000 ybp and
remained stable from 67,800 ybp onwards. The Unalaska population started to grow most
recently (113,000 ybp) and consistently kept expanding. The modern population from
Bogoslof started a rapid growth 208,000 ybp until it stayed at constant size at 34,700 ybp.
From the skyline plot, the current female effective population sizes are 28,200 (Kodiak),
69,200 (Shemya), 188,000 (Unalaska) and 219,000 (Bogoslof) seals respectively,
assuming the generation time for females (Tau) is 10 years (Lander 1981; COSEWIC
2006). Using the breeding ratio of adult females to males is 9:1 (Gentry 1998), the male
and female effective population sizes applied to the function of Ne=4NemNef/(Nem+Nef)
give rise to Ne of 11,280 (Kodiak), 27,680 (Shemya), 75,200 (Unalaska) and 87,600
(Bogoslof) animals from each island. To compare the output from inter- and
intra-specific rates, I also used the interspecific rate (0.097 substitutions/site/mya), and
obtained approximately double the numbers of effective population sizes and twice the
time span (Figure 8b).
74
a)
b)
75
c)
d)
Figure 7 Bayesian skyline plots of independent populations of Aleutian C. ursinus
Vertical axis is Ne*Tau. Ne is effective population size of females, and Tau is the generation time (10 years). Horizontal axis stands for years before present. The medians, 95% confidence intervals of population sizes are shown. a) Kodiak; b) Shemya; c) Unalaska; d) Bogoslof.
76
a)
b)
Figure 8 Bayesian skyline plots showing the demographic history of Aleutian C. ursinus Vertical axis is Ne*Tau. Ne is effective population size of females, and Tau is the generation time (10 years). Horizontal axis stands for years before present. a) uses the high rate (intraspecific rate) within C. ursinus and b) uses the low rate (interspecific rate) among C. ursinus and outgroup species. The rates are 0.213 substitutions/site/mya and 0.097 substitutions/site/mya respectively.
77
DISCUSSION
This study jointly analyzed the inter- and intra- specific clock rates of mitochondrial
control region to estimate the past population dynamics of Aleutian C. ursinus. To certify
the control region rate, fossil record and phylogenetic computations were combined to
examine the clock rate of mitochondrial cytochrome b region, with approval of rate
comparison using sequence mismatch distribution from the two genes. The root height
within C. ursinus was difficult to estimate for the lack of prior fossil information, so
adding in external calibrations placed the root of the tree deeper. Adding in closely
related species produced more prior knowledge for the data set but, if not analyzed
properly, would yield interspecific rates that are two-fold lower than population rates.
BEAST assessed the intraspecific rate of control region ranging 0.192 to 0.357
substitutions per site per million years, and the rate above the species level was computed
as 0.097-0.105 substitutions per site per million years under the same prior parameters.
The interspecific rate was confirmed by mismatch distribution tests and phylogenetic
computation, where the substitution rate of cytochrome b gene was ~0.01 per million
years, and the control region rate was ~10 times higher. The results confirmed that rates
of substitution vary among lineages and genes but in C. ursinus also support the time
dependency theory (Ho et al. 2005, 2007, 2008).
Bayesian analysis proved especially suitable for ancient DNA study. BEAST can be
applied to large datasets, is faster than the regular evolutionary sampling methods using
78
maximum likelihood, counts samples with different ages, especially proper for ancient
DNA sequences where aging serves as additional prior knowledge, and gives a set of
parameters coalescent phylogenies and distribution for final comparisons. A Bayesian
skyline plot assesses the particular time depth of population changes, which reconstructs
demographic history. Bayesian approaches require strict prior parameters to input
accurate information. It can be difficult to draw strong conclusions if the associated
estimation error is taken into account. Considering the idea of measurably evolving
populations (Drummond et al. 2003), I used different strategies: for the first step of
calculating the clock rate, I did not add ages of samples as prior because they were not
considered as old enough to be informative about mutation rate, so the sequences can
effectively be treated as isochronous (Drummond et al. 2003); for the second step of
plotting Bayesian skyline which did not require the root height as the prior knowledge, I
added in sampling ages that provided an opportunity to see population dynamics
information that was otherwise unattainable from modern sequences.
All analyses in my thesis are consistent in showing no genetic impacts of a recent
population bottleneck. In Chapter 1 I proposed that both ancient and modern populations
had similar high level of genetic diversity and low genetic structure. In this chapter, Fu’s
Fs test, mismatch distributions and Bayesian skyline plots indicated a population
expansion in Shemya, Unalaska and Bogoslof populations that started tens of thousands
of years ago and continued to recent times. The Kodiak population experienced a
contraction (though not drastic) after initial population growth. Correspondingly, Fu’s Fs
79
test showed no significant expansion in Kodiak population, and mismatch index of this
population presented as “jagged” as probably multiple expansions instead of a single one.
The other potential explanation for the “jagged” shape could be lack of sufficient
sampling. Unalaska is the island showing most recent expansion by both mismatch
distribution and Bayesian skyline plot, but still long before the arrival of humans.
It is worthwhile to take into account recent climatic events to evaluate the
demographic changes, especially with the availability of documentation of climate events.
Glacial cycles probably influenced the distribution of C. ursinus (Gentry 1998).
Crockford and Frederick (2007) presented that sea ice extended further south to Unalaska
Island about 3,500 to 2,500 years ago (the height of the Neoglacial period), and persisted
longer into the summer. The expansion of sea ice in the Bering Sea would have prevented
fur seals from using the Pribilof Islands as a summer breeding rookery (Crockford &
Frederick 2007). In that case, fur seals’ ecological reaction may have differed at this time.
Interestingly, my samples from Unalaska were collected from Amaknak Bridge, and were
dated as 3,000 years old. The rapid population expansion (especially during the most
recent time) in Unalaska could potentially be an evolutional effect to the climatic events,
but is not strongly supported by the Bayesian skyline plots, since the expansion started at
an earlier time of hundreds of thousands years ago. The ecological responses of C.
ursinus need to be further studied. At this point, it is unclear what type of ecological
response could result in reduced genetic diversity and expansion time, while low
structure is being maintained.
80
CONCLUSION
Reconstruction of demographical history of Callorhinus ursinus from Aleutian
Islands does not indicate a genetic bottleneck. Instead, population expansions were
observed in Unalaska, Shemya and Bogoslof islands. Fur seals from Unalaska started to
increase in abundance most recently but still long before human arrival. More data are
needed to address a possible impact of the Neoglacial on Northern Fur seal’s
demographic history.
81
LITERATURE CITED
1. Adams, G. P., Ward, T. J., Goertz, C. E. C., Ream, R. R., and Sterling. 2007.
Ultrasonographic characterization of reproductive anatomy and early embryonic detection in the northern fur seal (Callorhinus ursinus) in the field. Marine Mammal Science. 23 (2): 445-452.
2. Alter, S. E., and Palumbi, S. R. 2009. Comparing evolutionary patterns and variability in the mitochondrial control region and cytochrome b in three species of Baleen whales. Journal of Molecular Evolution. 68: 97-111.
3. Arnason et al. 2006. Pinniped phylogeny and a new hypothesis for their origin and dispersal. Molecular phylogenetics and evolution. 41: 345-354.
4. Atarhouch, T., Rüber, L., Gonzalez, E. G., Albert, E. M., Rami, M., Dakkak, A., and Zardoya, R. 2006. Signature of an early genetic bottleneck in a population of Moroccan sardines (Sardina pilchardus). Molecular Phylogenetics and Evolution. 39: 373-383.
5. Balsiger, J. A., et al. 2005. Setting the annual substance harvest of northern fur seals on the Pribilof Islands. Final environmental impact statement. National Oceanic and Atmosphere Administration.
6. Bandelt, H. J. 2008. Clock debate: when times are a-changin': Time dependency of molecular rate estimates: tempest in a teacup. Heredity. 100: 1-2.
7. Bandelt, H. J., Kong, Q. P., Richards, M., and Macaulay, V. 2006. Estimation of mutation rates and coalescence times: some caveats. In: Bandelt H-J, Macaulay V, Richards M (eds). Human Mitochondrial DNA and the Evolution of Homo Sapiens. Springer-Verlag: Berlin Heidelberg. pp 47–90.
8. Caballero, A. 1994. Developments in the prediction of effective population size. Heredity. 73: 657-679.
9. Chan, Y. L., Anderson, C. N. K., and Hadly, E. A. 2006. Bayesian estimation of the timing and severity of a population bottleneck from ancient DNA. PloS Genetics. 2(4): e59.
10. COSEWIC. 2006. COSEWIC assessment and update status report on the northern fur seal Callorhinus ursinus in Canada. Committee on the status of endangered wildlife in Canada, Ottawa.
11. Crockford, S. J., and Frederick, S. G. 2007. Sea ice expansion in the Bering Sea during the Neoglacial: evidence from archaeozoology. The Holocene. 17 (6): 699-706.
12. Crow, J. F., and Denniston, C. 1988. Inbreeding and variance effective population numbers. Evolution. 42: 482-495.
13. Curtis, C., Stewart, B. S., and Karl, S. A. 2009. Pleistocene population expansions of Antarctic seals. Molecular Ecology. 18: 2112–2121.
14. Depaulis, F., Orlando, L., and Hänni, C. 2009. Using classical population genetics
82
tools with heterochroneous data: time matters! PLoS ONE. 4(5): e5541. 15. DFO. 2007. Recovery potential assessment for Northern Fur Seals (Callorhinus
ursinus). DFO. Can. Sci. Advis. Sec. Sci. Advis. Rep. 2007/052. 16. Drummond, A. J, Rambaut, A. 2007. BEAST: Bayesian evolutionary analysis by
sampling trees. BMC Evolutionary Biology. 7: 214. 17. Drummond, A. J., Rambaut, A., Shapiro, B. and Pybus, O. G. 2005. Bayesian
coalescent inference of past population dynamics from molecular sequences. Molecular Biology and Evolution. 22: 1185–1192.
18. Drummond, A. J., Pybus, O. G., Rambaut, A., Forsberg. R., and Rodrigo, A. G. 2003. Measurably evolving populations. Trends in Ecology and Evolution. 18: 481-488.
19. Emerson, B. C. 2007. Alarm bells for the molecular clock? No support for Ho et al.'s model of time-dependent molecular rate estimates. Systems Biology. 56: 337–345.
20. Excoffier, L., Laval, G., and Schneider, S. 2006. Arlequin 3.1: An Integrated Software Package for Population Genetics Data Analysis.
21. Finlay, E. K., Gaillard, C., Vahidi, S. M. F., Mirhoseini, S. Z., Jianlin, H., Qi, X. B., El-Barody, M. A. A ., Baird, J. F., Healy, B. C., and Bradley, D. G. 2007. Bayesian inference of population expansions in domestic bovines. Biology Letters. 3: 449-452.
22. Fleischer, R. C., McIntosh, C. E., and Tarr, C. L. 1998. Evolution on a volcanic conveyor belt: using phylogeographic reconstructions and K-Ar-based ages of the Hawaiian Islands to estimate molecular evolutionary rates. Molecular Ecology. 7: 533–545.
23. Fu, Y. X. 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and backgroud selection. Genetics. 147:915-925.
24. Fu, Y. X., and Li, W. H. 1993. Statistical tests of neutrality of mutations. Genetics. 133: 693-709.
25. Gentry, R. L. 1998. Behavior and ecology of the northern fur seal. Princeton University Press, New Jersey, pp. 1-132.
26. Hadly, E. A., Ramakrishnan, U., Chan, Y. L., van Tuinen, M., O’Keefe, K. et al. 2004. Genetic response to climatic change: insights from ancient DNA and phylochronology. PLoS Biol. 2(10): e290.
27. Harpending, H. C., Batzer, M. A., Gurven, M., Jorde, L. B. A. R. Rogers, A. R. et al. 1998. Genetic traces of ancient demography. PNAS. 95: 1961–1967.
28. Hay, J. M., Subramanian, S., Millar, C. D., Mohandesan, E., and Lambert, D. M. 2008. Rapid molecular evolution in a living fossil. Trends in Genetics. 24: 106-109.
29. Higdon, J., Bininda-Emonds, O., Beck, R., and Ferguson, S. 2008. Phylogeny and divergence of the pinnipeds (Carnivora: Mammalia) assessed using a multigene dataset. BMC Evolutionary Biology 8: 216.
83
30. Ho, S. Y. W., Saarma, U., Barnett, R., Haile, J., Shapiro, B. 2008. The effect of inappropriate calibration: three case studies in molecular ecology. PloS One. 3(2): e1615.
31. Ho, S. Y. W., Kolokotronis, S-O., and Allaby, R. G. 2007. Elevated substitution rates estimated from ancient DNA sequences. Biology Letters. 3: 702-705.
32. Ho, S. Y. W., Phillips, M. J., Cooper, A., and Drummond, A. J. 2005. Time dependency of molecular rate estimates and systematic overestimation of recent divergence times. Molecular Biology and Evolution. 22(7): 1561–1568.
33. Hofreiter, M., Serre, D., Rohland, N., et al. 2004. Lack of phylogeography in European mammals before the last glaciations. PNAS. 101: 12963-12968.
34. Howell, N., Smejkal, C. B., Mackey, D. A., Chinnery, P. F., Turnbull, D. M., and Herrnstadt, C. 2003. The pedigree rate of sequence divergence in the human mitochondrial genome: there is a difference between phylogenetic and pedigree rates. The American Journal of Human Genetics. 72: 659–670.
35. Koepfli, K. P., Deere, K. A., Slater, G. J., Begg, C., Begg, K., Grassman, L., Lucherini, M., Veron, G., and Wayne, R. K. 2008. Multigene phylogeny of the Mustelidae: Resolving relationships, tempo and biogeographic history of a mammalian adaptive radiation. BMC Biology. 14: 6-10.
36. Lambert, D. M., Ritchie, P. A., Millar, C. D., Holland, B., Drummond, A. J., and Baroni, C. 2002. Rates of evolution in ancient DNA from Ade´lie penguins. Science. 295: 2270–2273.
37. Lander, R. H. 1981. A life table and biomass estimate for Alaskan fur seals. Fisheries Research (Amsterdam) 1: 55-70.
38. Librado, P., and Rozas J. 2009. DnaSP v5: a software for comprehensive analysis of DNA. Bioinformatics. 25: 1451-1452.
39. Lloyd, D. S., Mcroy, C. P., and Day, R. H. 1981. Discovery of Northern Fur Seals (Callorhinus ursinus) breeding on Bogoslof Island, Southern Bering Sea. 34: 316-320.
40. Loughlin, T. R., Antonelis, G.A., Baker, J.D., York, A.E., Fowler, C.W., DeLong, R.L., and Braham, H.W. 1994. Status of the northern fur seal population in the United States during 1992. E. H. Sinclair, ed. Fur seal investigations, 1992. U. S. Department of Commerce, Seattle, WA. pp 9-78.
41. Macklin, S. A., Schumacher, J., Moore, S. E., and Smith S. 2008. Sustaining the marine ecosystem of the Pribilof Domain. Deep-Sea Research. II 55: 1698-1700.
42. Newsome, S. D., Etnier, M. A., Gifford-Gonzalez, D., Phillips, D.L., van Tuinen, M., Hadly, E. A., Costa, D. P., Kennett, D. J., Guilderson, T. P., and Koch, P. L. 2007. The shifting baseline of northern fur seal ecology in the northeast Pacific Ocean. PNAS. 104:9709-9714.
43. NMML (National Marine Mammal Laboratory), 2007 NMML (National Marine Mammal Laboratory), 2007. Northern fur seals: history and exploitation. WWW Page, http://www.afsc.noaa.gov/nmml/alaska/nfs/exploitation.php, (retrieved
84
December 2007). 44. Parsons, T. J., Muniec, D. S. Sullivan, K. et al. 1997. A high observed
substitution rate in the human mitochondrial DNA control region. Nature Genetics. 15: 363–368.
45. Pavlova, A., Zink, R. M., and Rohwer, S. 2005. Evolutionary history, population genetics, and gene flow in the common rosefinch (Carpodacus erythrinus). Molecular Phylogenetics and Evolution. 36: 669-681.
46. Penny, D. 2005. Relativity for molecular clocks. Nature. 426: 183–184 47. Pinsky, M.L., Newsome, S., Dickerson, B., van Tuinen, M., Kennett, D., Ream,
R., and Hadly E.A. unpublished. Refugia and high migration rates buffered Callorhinus ursinus from genetic impacts of range contraction and population decline.
48. Pybus, O. G., Rambaut, A., and Harvey, P. H. 2000. An integrated framework for the inference of viral population history from reconstructed genealogies. Genetics. 155: 1429–1437.
49. Ream, R., and Burkanov, V. 2005. Trends in abundance of Steller sea lions and Northern fur seals across the North Pacific Ocean. PICES XIV Annual Meeting, Vladivostock, Russia.
50. Ream, R., 2002. Molecular ecology of North Pacific Otariids: genetic assessment of Northern fur seal and Steller sea lion distributions. Ph. D. thesis.
51. Rice, D. W. 1998. Marine mammals of the world – Systematics and distribution (special publications of the society of marine mammalogy speical publications 4). Allen Press, Inc., Lawrence, Kansas. pp 231.
52. Rogers, A. R., and Harpending, H. 1992. Population growth makes waves in the distribution of pairwise genetic differences. Mol. Biol. Evol. 9:552-569.
53. Schneider, S., and Excoffier, L. 1999. Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates vary among sites: application to human mitochondrial DNA. Genetics. 152: 1079–1089.
54. Sigurdardo´ ttir, S., Helgason, A., Gulcher, J. R., Stefa´nsson, K., and Donnelly, P. 2000. The mutation rate in the human mtDNA control region. The American Journal of Human Genetics. 66: 1599–1609.
55. Slade, R. W., Moritz, C., Hoelzel, A. R., and Burton. H. R. 1998. Molecular population genetics of the Southern elephant seal Mirounga leonina. Genetics. 149: 1945-1957.
56. Slade, R. W., Moritz, C., and Heideman, A. 1994. Multiple nuclear-gene phylogenies: application to pinnipeds and comparison with a mitochondrial DNA gene phylogeny. Molecular Biology and Evolution. 11: 341-356.
57. Tajima, F. 1993. Measurement of DNA polymorphism. pp. 37-59. In Takahata, N. and Clark, A. G. (eds), Mechanisms of Molecular Evolution, Sinauer Associates. Inc., Sunderland, Massachusetts.
85
58. Tajima, F. 1989. Statistical method for testing the neutral mutationhypothesis by DNA polymorphism. Genetics. 123: 585-595.
59. Towell, R. G. and Ream, R. R. 2006. Decline in Northern fur seal (Callorhinus ursinus) pup production on the Pribilof Islands. Marine Mammal Science. 22 (2): 486-491.
60. Wickens, P. A., and York, A. E. 1997. Comparative population dynamics of fur seals. Marine Mammal Science. 13: 241-292.
61. Woodhams, M. 2006. Can deleterious mutations explain the time dependency of molecular rate estimates? Molecular Biology and Evolution. 23: 2271–2273
62. Wright, S., 1931. Evolution in mendelian populations. Genetics. 16: 97-159. 63. Wynen L. P., Goldsworthy, S. D., Insley, S. J., Adams, M., Bickham, J. W.,
Francis, J., Gallo, J. P., Hoelzel, A. R., Majluf, P., White, R. W. G., and Slade, R. 2001. Phylogenetic relationships within the eared seals (Otariidae: Carnivora): implications for the historical biogeography of the family. Molecular Phylogenetics and Evolution. Vol. 21, No. 2, November, pp. 270-284.
86