Article

Recent colonization and radiation of North American Lycaeides (Plebejus) inferred from mtDNA.

Zachariah Gompert, James A Fordyce, Matthew L Forister, Chris C Nice

Department of Biology, Population and Conservation Biology Program, Texas State University, San Marcos, TX 78666, USA.

Molecular Phylogenetics and Evolution (impact factor: 3.61). 06/2008; 48(2):481-90. DOI:10.1016/j.ympev.2008.04.035 pp.481-90

Source: PubMed

ABSTRACT North American Lycaeides populations exhibit remarkable variation in ecological, morphological, and behavioral characters, as well as an established history of introgressive hybridization. We examined mitochondrial DNA variation from 55 Eurasian and North American Lycaeides populations using molecular phylogenetics and coalescent-based methods in order to clarify the evolutionary and demographic history of this polytypic group. Specifically we addressed the following questions: (1) Do mitochondrial alleles sampled from North America form a monophyletic group, which would be expected if North American Lycaeides were descended from a single Eurasian ancestor? (2) When did Lycaeides colonize North America? and (3) What is the demographic history of North American Lycaeides since their colonization? Bayesian maximum likelihood methods identified three major mitochondrial lineages for Lycaeides; each lineage contained haplotypes sampled from both Eurasia and North America. This suggests a complex colonization history for Lycaeides, which likely involved multiple founding lineages. Coalescent-based analyses placed the colonization of North America by Eurasian Lycaeides sometime during or after the late Pliocene. This was followed by a sudden increase in population size of more than an order of magnitude for the North American population of Lycaeides approximately 100,000-150,000 years before the present. These mitochondrial data, in conjunction with data from previous ecological, morphological, and behavioral studies, suggest that the diversity observed in Lycaeides in North America is the result of a recent evolutionary radiation, which may have been facilitated, in part, by hybridization.

0 0

0 Bookmarks

35 Views

Article: Miocene and Pliocene dominated diversification of the lichen-forming fungal genus Melanohalea (Parmeliaceae, Ascomycota) and Pleistocene population expansions.

Steven D Leavitt, Theodore L Esslinger, Pradeep K Divakar, H Thorsten Lumbsch

[show abstract] [hide abstract]
ABSTRACT: BACKGROUND: Factors promoting diversification in lichen symbioses remain largely unexplored. While Pleistocene events have been important for driving diversification and affecting distributions in many groups, recent estimates suggest that major radiations within some genera in the largest clade of macrolichens (Parmeliaceae, Ascomycota) vastly predate the Pleistocene. To better understand the temporal placement and sequence of diversification events in lichens, we estimated divergence times in a common lichen-forming fungal genus, Melanohalea, in the Northern Hemisphere. Divergence times were estimated using both concatenated gene tree and coalescence-based multilocus species tree approaches to assess the temporal context of major radiation events within Melanohalea. In order to complement our understanding of processes impacting genetic differentiation, we also evaluated the effects of Pleistocene glacial cycles on population demographics of distinct Melanohalea lineages, differing in reproductive strategies. RESULTS: We found that divergence estimates, from both concatenated gene tree and coalescence-based multilocus species tree approaches, suggest that diversification within Melanohalea occurred predominantly during the Miocene and Pliocene, although estimated of divergence times differed by up to 8.3 million years between the two methods. These results indicate that, in some cases, taxonomically diagnostic characters may be maintained among divergent lineages for millions of years. In other cases, similar phenotypic characters among non-sister taxa, including reproductive strategies, suggest the potential for convergent evolution due to similar selective pressures among distinct lineages. Our analyses provide evidence of population expansions predating the last glacial maximum in the sampled lineages. These results suggest that Pleistocene glaciations were not inherently unfavorable or restrictive for some Melanohalea species, albeit with apparently different demographic histories between sexually and vegetatively reproducing lineages. CONCLUSIONS: Our results contribute to the understanding of how major changes during the Miocene and Pliocene have been important in promoting diversification within common lichen-forming fungi in the northern Hemisphere. Additionally, we provide evidence that glacial oscillations have influenced current population structure of broadly distributed lichenized fungal species throughout the Holarctic.

BMC Evolutionary Biology 09/2012; 12(1):176. · 3.52 Impact Factor
Source
Available from: PubMed Central

Article: A hierarchical bayesian approach to ecological count data: a flexible tool for ecologists.

James A Fordyce, Zachariah Gompert, Matthew L Forister, Chris C Nice

[show abstract] [hide abstract]
ABSTRACT: Many ecological studies use the analysis of count data to arrive at biologically meaningful inferences. Here, we introduce a hierarchical bayesian approach to count data. This approach has the advantage over traditional approaches in that it directly estimates the parameters of interest at both the individual-level and population-level, appropriately models uncertainty, and allows for comparisons among models, including those that exceed the complexity of many traditional approaches, such as ANOVA or non-parametric analogs. As an example, we apply this method to oviposition preference data for butterflies in the genus Lycaeides. Using this method, we estimate the parameters that describe preference for each population, compare the preference hierarchies among populations, and explore various models that group populations that share the same preference hierarchy.

PLoS ONE 01/2011; 6(11):e26785. · 4.09 Impact Factor

Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.

Page 1

Recent colonization and radiation of North American Lycaeides (Plebejus)
inferred from mtDNA
Zachariah Gomperta,b,*, James A. Fordycec, Matthew L. Foristerd, Chris C. Nicea
aDepartment of Biology, Population and Conservation Biology Program, Texas State University, San Marcos, TX 78666, USA
bDepartment of Botany, University of Wyoming, 1000 E. University Avenue, Laramie, WY 82071, USA
cDepartment of Ecology and Evolutionary Biology, University of Tennessee, Knoxville, TN 37996, USA
dDepartment of Natural Resources and Environmental Science, University of Nevada, Reno, NV 89512, USA
a r t i c l ei n f o
Article history:
Received 5 September 2007
Revised 24 April 2008
Accepted 24 April 2008
Available online 1 May 2008
Keywords:
Coalescent
Mitochondrial DNA
Lycaeides
Bayesian skyline plot
Pleistocene
Holarctic biogeography
a b s t r a c t
North American Lycaeides populations exhibit remarkable variation in ecological, morphological, and
behavioral characters, as well as an established history of introgressive hybridization. We examined
mitochondrial DNA variation from 55 Eurasian and North American Lycaeides populations using molec-
ular phylogenetics and coalescent-based methods in order to clarify the evolutionary and demographic
history of this polytypic group. Specifically we addressed the following questions: (1) Do mitochondrial
alleles sampled from North America form a monophyletic group, which would be expected if North
American Lycaeides were descended from a single Eurasian ancestor? (2) When did Lycaeides colonize
North America? and (3) What is the demographic history of North American Lycaeides since their coloni-
zation? Bayesian maximum likelihood methods identified three major mitochondrial lineages for Lycae-
ides; each lineage contained haplotypes sampled from both Eurasia and North America. This suggests a
complex colonization history for Lycaeides, which likely involved multiple founding lineages. Coales-
cent-based analyses placed the colonization of North America by Eurasian Lycaeides sometime during
or after the late Pliocene. This was followed by a sudden increase in population size of more than an order
of magnitude for the North American population of Lycaeides approximately 100,000–150,000 years
before the present. These mitochondrial data, in conjunction with data from previous ecological, morpho-
logical, and behavioral studies, suggest that the diversity observed in Lycaeides in North America is the
result of a recent evolutionary radiation, which may have been facilitated, in part, by hybridization.
? 2008 Elsevier Inc. All rights reserved.
1. Introduction
Gene genealogies can be used to infer historical evolutionary
and demographic processes operating within species or groups of
closely related species (Hein et al., 2005). For example, the effects
of past geological events (e.g. Pleistocene climatic cycles) on the
distribution and abundance of species have been inferred increas-
ingly through the analysis of molecular gene genealogies (e.g. Avise
and Walker, 1998; Knowles, 2001; Griswold and Baker, 2002; Pin-
ho et al., 2007; Spellman et al., 2007). Coalescent-based analyses
have been especially useful for the inference of historical pro-
cesses, as these analyses utilize the information present in DNA
sequence variation (Hein et al., 2005). Data on historical processes
are necessary for understanding current evolutionary and ecologi-
cal patterns and processes. Here we report the results of a large-
scale survey of mitochondrial DNA (mtDNA) variation in the
polytypic Lycaeides butterfly species group. This is a diverse genus
with a complex evolutionary history, including one of the few doc-
umented cases of homoploid hybrid speciation in animals (Nice
and Shapiro, 1999; Nice et al., 2005; Gompert et al., 2006a,b).
Butterflies of the genus Lycaeides are members of the family
Lycaenidae, a family notable for its species richness, diversity of
natural histories, and extreme local specialization (Downey and
Dunn, 1964; Grimaldi and Engel, 2005). Lycaeides is a circumpolar,
holarctic genus (Nice and Shapiro, 1999). Nabokov (1943, 1949) re-
vised the classification of North American Lycaeides based on qual-
itative differences in wing pattern and quantitative differences in
the size and shape of the male genitalia. He recognized two
species: L. idas (Linnaeus) and L. melissa (W.H. Edwards). The for-
mer has a holarctic distribution, while that latter is a North Amer-
ican endemic. Gompert et al. (2006b) identified a third, yet
unnamed, North American Lycaeides species, which is a homoploid
hybrid species distributed in the alpine habitat of the Sierra Neva-
da. This species originated via hybridization between L. idas and L.
melissa (Gompert et al., 2006b). Lycaeides idas and L. melissa are
broadly sympatric in several regions of North America (Fig. 1).
1055-7903/$ - see front matter ? 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2008.04.035
* Corresponding author. Address: Department of Botany, University of Wyoming,
1000 E. University Avenue, Laramie, WY 82071, USA. Fax: +1 307 766 2851.
E-mail address: zgompert@uwyo.edu (Z. Gompert).
Molecular Phylogenetics and Evolution 48 (2008) 481–490
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev

Page 2

Approximately 17 subspecies have been described based on differ-
ences in wing pattern, male genitalia, habitat specificity, and larval
host plant use (Nabokov, 1949; Scott, 1986; Opler, 1992; Lane and
Weller, 1994). Two of these subspecies have attracted the attention
of conservation biologists. The Karner blue butterfly (L. m. samuelis)
is listed in the United States as an endangered species (US Fish and
Wildlife Service, 1992, 2003); the Lotis blue (L. i. lotis) has the same
status, but is likely extinct (Arnold, 1993). Lycaeides was previously
placed within the genus Plebejus, and is still referred to as Plebejus
by some authors (e.g. Scott, 1986).
North American Lycaeides are diverse and exhibit remarkable
ecological specialization among taxa and populations. This diver-
sity suggests that North American Lycaeides could be thought of
as an evolutionary radiation. While Lycaeides at a specific locality
are generally monophagous, there is considerable variation in the
specific host-plant species used by geographically separated popu-
lations (Scott, 1986; Brock and Kaufman, 2003). This variation in
host-plant use is coupled with inter-population variation in female
host-plant fidelity and host-plant preference (Nice et al., 2002;
Gompert et al., 2006b). Populations also differ in the size and
placement of wing pattern elements (Fordyce et al., 2002). Some
of these population-level differences in wing pattern elements
are discernable by male Lycaeides and play a role in mate recogni-
tion (Fordyce et al., 2002). Population-level variation in egg mor-
phology (Forister et al., 2006) and male genital morphology (Nice
and Shapiro, 1999; Nice et al., 2005; Lucas et al., 2008) has also
been documented throughout North America; at present, an adap-
tive role forthis morphological
demonstrated.
Nice et al. (2005) examined geographic patterns of mtDNA (AT-
rich region) variation in Lycaeides across parts of North America in
order to elucidate the evolutionary history of this ecologically,
behaviorally, and morphologically diverse group. They found that
variationhasnotbeen
mitochondrial variation was partitioned geographically and only
loosely followed taxonomic boundaries. Specifically, three mito-
chondrial clades were detected (Nice et al., 2005). Several popula-
tions at the geographical boundaries of identified clades displayed
discordant patterns of mtDNA and morphological variation. For
example, populations of L. m. samuelis (the endangered Karner blue
butterfly) in the western portion of their range were morphologi-
cally indistinguishable from those in the eastern portion of their
range, but were assigned to a different mtDNA clade than the east-
ern L. m. samuelis populations. Nice et al. (2005) suggested that
these patterns were indicative of isolation of Lycaeides in at least
three refugia during Pleistocene glacial maxima followed by
post-Pleistocene range expansion, secondary contact, and intro-
gressive hybridization.
The hypothesis of introgressive hybridization at two of these
contact zones has been tested using nuclear genetic markers. Gom-
pert et al. (2006a) determined that L. m. samuelis populations in the
western portion of their range possess mtDNA haplotypes from a
different mtDNA clade than the eastern populations because of
mitochondrial introgression. Examination of genomic divergence
between the eastern and western L. m. samuelis populations using
AFLP markers revealed that mitochondrial introgression had oc-
curred from L. m. melissa to the western L. m. samuelis populations
with minimal to no nuclear introgression (Gompert et al., 2006a). A
second contact zone occurs in the Sierra Nevada. Lycaeides popula-
tions in the alpine (i.e. above tree-line) in the Sierra Nevada pos-
sess wing pattern elements indicative of L. melissa, L. idas-like
mtDNA haplotypes, and male genitalic morphology intermediate
to these two taxa (Nice et al., 2005; Gompert et al., 2006b). Based
on extensive molecular data (i.e. AFLP markers, nuclear sequence
data, mtDNA sequence data, and microsatellite markers), Gompert
et al. (2006b) concluded this discordant pattern resulted from
hybridization between L. idas and L. melissa and that the alpine-
associated populations represent a species of hybrid origin. Thus,
North American Lycaeides are highly specialized and have a history
of hybridization, which, in at least one case, resulted in the origin
of a species.
To better understand the potentially complex and dynamic pro-
cesses that have influenced and continue to act on North American
Lycaeides, data regarding the evolutionary and demographic his-
tory of this genus are required. Particularly we require information
concerning the colonization and subsequent diversification of
Lycaeides in North America. We do not know whether the special-
ized forms that exist in North America began to differentiate in
Eurasia, nor do we know if the hybridizing L. idas and L. melissa
of North America are each others’ closest relatives (L. idas also oc-
curs in Eurasia, and the Eurasian species L. argyrognomon is
thought to be closely related to L. idas and L. melissa (Nice et al.,
2005)). Here we examine mtDNA data using molecular phyloge-
netics and coalescent-based models to address the following ques-
tions regarding the evolutionary and demographic history of North
American Lycaeides: (1) Are North American Lycaeides the result of
colonization from Eurasia by a single ancestor, or did multiple dif-
ferentiated forms colonize North America? (2) When did coloniza-
tion of North America occur? (3) What is the demographic history
of North American Lycaeides since their colonization?
2. Methods
2.1. Collection and DNA extraction
Adult Lycaeides were obtained from 55 populations from across
North America, Europe, and Asia (Table 1). North American taxa
sampled included Lycaeides idas, L. melissa, the hybrid species from
the Sierra Nevada, and the Warners Mountain entity. Eurasian taxa
from which samples were obtained include L. idas and L. argyrogn-
Fig. 1. Range map for North American Lycaeides. The approximate ranges of L. idas,
L. melissa, and the unnamed hybrid species are shown in dark gray, light gray, and
black, respectively. The range maps for L. idas and L. melissa follow Nabokov
(1949),and Stanford and Opler (1996). The range map of the hybrid species is based
on field-work conducted by Z.G., J.A.F., and C.C.N. during the summer of 2006.
482
Z. Gompert et al./Molecular Phylogenetics and Evolution 48 (2008) 481–490

Page 3

omon. Both males and females were collected from most Lycaeides
populations; males were collected from L. m. samuelis populations
in accord with USFWS permit PRT842392. Additional specimens
were collected from several species in the genus Plebejus, which
are closely related to Lycaeides (Kandul et al., 2004). This includes
P. saepiolus, P. shasta, and P. icarioides from North America and
P. argus and P. sephirus from Eurasia. Taxonomic identification of
collected specimens was made based on morphological (wing
patterns and male genitalic variation), ecological, and behavioral
characters as well as geographical data. DNA was isolated follow-
ing the methods of Hillis et al. (1996) and Brookes et al. (1997).
2.2. Molecular methods
We sequenced portions of the mitochondrial genes cytochrome
oxidase c subunit I (COI) and cytochrome oxidase c subunit II (COII)
Table 1
Population data
Locality #TaxaLocality LatitudeLongitude mtDNA (n)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
L. i. alaskensis
L. i. anna
Koyukuk River, AK
Donner Pass, CA
Leek Springs, CA
Trap Creek, CA
Yuba Gap, CA
Indian Valley, CA
Lerida, Spain
S. Urals, Russia
Galich Mt., Macedonia
Cave Lake, CA
Deadfall Mdw., CA
Soda Mt., OR
Josephine Co., OR
Marble Mts., CA
Mt. Ashland, OR
Alberta, Canada
Mineral Creek, CO
San Juan Co., CO
Alta, UT
Teton Mts., WY
Big Pine, CA
Manzanar Airfield, CA
Beckwourth Pass, CA
Brandon, SD
Cedarville, CA
Elmore Co., ID
Garderville, CA
Granite Mts., NV
Gowanda, CO
Montague, CA
Pitkin Co., CO
Sierravalley, CA
Spring Creek, SD
Verdi, NV
Fish Lake, WI
Fort McCoy, WI
Indiana Dunes, IN
Necedah, WI
Saratoga, NY
Bodie Hills, CA
County Line Hill, CA
Carson Pass, CA
Jeff Davis Pk., CA
Lake Emma, CA
Mt. Rose, NV
Piute Pass, CA
Sonora Pass, CA
South Fork, CA
Sweetwater Mts., CA
Tioga Crest, CA
Wassuk Mts., NV
Eagle Pk., CA
Slovakia
Magadan, Russia
Atali, Russia
Russia
Plastovek, Slovakia
Raec, Macedonia
Sweetwater Mts., CA
Tioga Crest, CA
Tioga Crest, CA
Babuna mt., Macedonia
67? 2701000N
39? 1805300N
38? 3705900N
39? 2204300N
39? 2902400N
40? 4503300N
Not available
Not available
Not available
41? 5804600N
41? 1401100N
42? 0501900N
42? 2503500N
41? 4904000N
42? 0405200N
50? 4203700N
37? 4100000N
37? 5600000N
40? 3503300N
43? 5101300N
37? 1002300N
36? 4401900N
39? 4703500N
43? 3602900N
41? 3104700N
42? 5304000N
38? 4805400N
40? 5301800N
40? 1800000N
41? 4602100N
39? 1000000N
39? 3704800N
44? 2500400N
39? 0300100N
45? 4401400N
43? 4705900N
41? 4000700N
44? 0400000N
43? 0302400N
38? 1604300N
37? 2705100N
38? 4204700N
38? 3802000N
38? 1605400N
39? 1902100N
37? 1401100N
38? 1905400N
37? 1203800N
38? 2700300N
37? 5705700N
38? 3805400N
41? 1503800N
Not available
Not available
Not available
Not available
Not available
Not available
38? 2700300N
37? 5705700N
37? 5705700N
Not available
150? 0303000W
120? 2005700W
120? 1402400W
120? 4002700W
120? 3503900W
122? 5802500W
Not available
Not available
Not available
120? 1202500W
122? 5703800W
122? 2805800W
123? 2704100W
122? 4405200W
122? 4301600W
114? 3704200W
107? 4100000W
107? 3401700W
111? 3702700W
110? 4505300W
118? 1604400W
118? 0800300W
120? 0603800W
96? 3403900W
120? 1001100W
116? 0603000W
119? 4604400W
119? 3306000W
104? 5000000W
122? 2803800W
106? 4700000W
120? 2104000W
100? 2405000W
119?5504800W
92? 4604300W
90? 4905900W
87? 0300400W
90? 1102000W
73? 4804700W
119? 0504000W
118? 1103500W
120? 0101700W
119? 5400500W
119? 2805900W
119? 5504800W
118? 4000800W
119? 3800200W
118? 3400700W
119? 2000400W
119? 1502500W
118? 4903000W
120? 1301100W
Not available
Not available
Not available
Not available
Not available
Not available
119? 2000400W
119? 1502500W
119? 1502500W
Not available
h63(3)
h18(1) h19(1) h20(5) h21(3)
h18(8) h29(2)
h20(5) h21(1) h43(1) h44(1) h45(2)
H20(8) h48(1) h49(1)
h23(4) h24(2) h25(1) h26(3)
h30(1)
h56(1)
h51(2) h76(1)
h01(10)
h16(6) h17(4)
h01(4) h27(1)
h28(4)
h31(2) h32(1) h33(7)
h01(10)
h01(5)
h68(2)
h01(5) h38(1) h64(1)
h01(2) h46(8)
h01(3) h06(2)
h01(10)
h01(7)
h01(1) h02(8)
h01(5) h04(2) h05(1) h06(1) h07(1)
h01(1) h08(8) h09(1)
h01(1) h02(1) h09(1)
h01(3) h02(7)
h67(3)
h01(1)
h01(9) h34(1)
h69(4)
h01(10)
h01(8) h04(1) h39(1)
h02(10) h47(1)
h01(5)
h01(5)
h22(5)
h01(5)
h22(5)
h03(2)
h02(3) h10(1) h11(2) h12(2) h13(2)
h03(8) h14(1) h15(1)
h03(4) h42(4) h50(2)
h03(10)
h03(1) h35(9)
h03(1)
h03(6) h40(3) h41(1)
h03(8), h42(1)
h03(9) h40(1)
h03(6) h42(4)
h03(1)
h01(10)
h02(1) h36(4) h56(3) h62(1)
h57(3)
h58(2)
h65(1)
h59(2) h60(1)
h59(4) h75(1)
h74(1)
h71(1) h72(1)
h73(3)
h53(1)
L. i. azureus
L.i. degenera
L. i. idasa
L. i. magnagraecaa
L. i. ricei
L. i. scudderii
L. i. sublivens
L. m. anneta
L. m. inyoensis
L. m. melissa
L. m. samuelis
Lycaeides Alpine sp.
Lycaeides Warner Mts. entity
L. a. argyrognomona
L. a. jauticaa
P. a. argusa
P. a. orientalisa
P. icarioides
P. saepiolus
P. shasta
P. sephirus magnificusa
Locality number, nominal taxa, sample locality, latitude, longitude, and mtDNA haplotypes are given for each population.
aEurasian.
Z. Gompert et al./Molecular Phylogenetics and Evolution 48 (2008) 481–490
483

Page 4

for 1–11 individuals from each sampled population (Table 1). PCR
was performed using the primer pair C1-J-1751/C1-N-2191 for COI
(Simon et al., 1994) and Pierre/Eva for COII (Caterino and Sperling,
1999). This yielded fragments of approximately 450 and 550 base
pairs for COI and COII, respectively. Fluorescently labeled dideoxy
terminators were used for single stranded sequencing reactions
for both COI and COII. Labeled amplicons were separated and visu-
alized using a Beckman 8800 automated sequencer (Beckman
Coulter Inc.). Sequences were aligned using Sequencher 4.2.2. or
by eye. Sixty-nine unique haplotypes were identified (GenBank
Accession Nos. EU409321–EU409362) (Table 1).
2.3. Phylogeny reconstruction and Bayesian hypothesis tests
To examine the relationship among the sampled haplotypes
from North American, European, and Asian Lycaeides populations,
gene genealogies for the combined data from COI and COII were
constructed using Bayesian maximum likelihood methods. Plebejus
sephirus was set as the outgroup. We evaluated three alternative
partition schemes for the sequence data: (1) a single partition,
(2) separate partitions for each gene region, and (3) separate parti-
tions for each gene region and codon position. Sequence evolution
models for each partition were selected using Modeltest 3.7 based
on Akaike Information Criterion (AIC) (Posada and Crandall, 1998).
When the data were placed in a single partition, the HKY+I+c mod-
el was selected, while the GTR+I+c model was selected for both COI
and COII when the data were partitioned by gene region. When the
data were partitioned by gene region and codon position the mod-
els selected were TrN+I, F81, TIM+c, HKY+I, TrN+I, and TIM for COI
1st codon position, COI 2nd codon position, COI 3rd codon position,
COII 1st codon position, COII 2nd codon position, and COII 3rd co-
don position, respectively.
Bayesian phylogenetic analyses were conducted using MrBayes
ver 3.1.2 (Huelsenbeck and Ronquist, 2001) for each of our three
data partition models. For each model, Markov Chain Monte Carlo
was performed with one cold chain and three hot chains run for
8 ? 106generations with a burnin of 5 ? 105generations. Two
independent runs were conducted for each model, and the stan-
dard deviation of the split frequencies was monitored to ensure
convergence upon the stationary distribution. We used Bayes fac-
tors (see Ronquist et al., 2005) to compare the posterior odds of
our preferred Bayesian gene genealogy with the data partitioned
by gene region and codon position to both our preferred Bayesian
gene genealogy with the data in a single partition and with the
data partitioned by gene region. Bayes factors were calculated by
taking the difference between the marginal log likelihood values
of the model with data partitioned by gene region and codon posi-
tion, M1, and each of the models with fewer data partitions, M0(see
Nylander et al., 2004).
To explicitly test the hypothesis that North American Lycaeides
are derived from a single colonization event from Eurasia, and thus
have a single old world ancestor, we used Bayes factors (Ronquist
et al., 2005) to compare the posterior odds of our preferred Bayes-
ian gene genealogy with no topological constraints to genealogies
with either North American Lycaeides constrained to be monophy-
letic or North American and Eurasian Lycaeides constrained to be
reciprocally monophyletic. In both cases we used a model with
the data partitioned by gene region and codon position (see Section
3). If North America was colonized by a single Eurasian species,
then North American Lycaeides should be monophyletic. If this oc-
curred sufficiently long ago, then North American and Eurasian
Lycaeides should be reciprocally monophyletic. A single haplotype
(h02) was shared between North America and Eurasian popula-
tions (see Section 3). Two copies of this haplotype were included
in these analyses, so the haplotype could be constrained to be in
both the North American and Eurasian clades. Gene genealogies
for the constrained and unconstrained trees were produced as de-
scribed in the preceding paragraph using MrBayes ver 3.1.2 (Huel-
senbeck and Ronquist, 2001). Bayes factors were calculated by
taking the difference between the marginal log likelihood values
of the unconstrained topology, M1, and the constrained topology,
M0(see Nylander et al., 2004; Brandley et al., 2005; Hedin and
Bond, 2006).
To determine if our preferred Bayesian gene genealogy, which
was obtained using a model that partitioned the sequence data
by gene region and codon position (see Section 3), conformed to
a molecular clock, we ran an additional Bayesian phylogenetic
analysis using MrBayes ver 3.1.2 (Huelsenbeck and Ronquist,
2001). This analysis was performed with a strict molecular clock
enforced and an exponential prior for branch lengths. We then
compared the posterior odds of our preferred Bayesian gene gene-
alogy with that of the constrained Bayesian gene geology using a
Bayes factor. We did this by taking the difference between the mar-
ginal log likelihood values of the model with the molecular clock
enforced, M1, and the unconstrained model, M0(Nylander et al.,
2004). The constrained model was selected as M1, because it re-
sulted in a greater marginal likelihood than the unconstrained
model.
2.4. Estimation of TMRCA
We used the software BEAST v1.4.2 (Drummond and Rambaut,
2006) to estimate the time to the most recent ancestor (TMRCA) for
the extant mtDNA variation for North American and Eurasian hapl-
otypes. As both North American and Eurasian haplotypes occur in
all three major Lycaeides mitochondrial lineages (see Section 3), the
TMRCA for North American mtDNA haplotypes should be the same
as that for both North American and Eurasian mtDNA haplotypes.
Further, as a single haplotype is shared between North American
and Eurasian populations, the TMRCA provides an estimate of the
upper bounds on the divergence between North American and Eur-
asian Lycaeides. TMRCA was estimated under Bayesian skyline plot
and constant population size demographic models (see below and
Section 3). All sampled Lycaeides haplotypes were included in this
analysis. The SRD06 model of sequence evolution was used; the
SRD06 model has fewer parameters than the GTR+I+c model, but
has been shown to provide a better fit for protein coding sequence
data (Shapiro et al., 2006). The analysis was implemented under a
relaxed clock with the rate for each branch drawn from a lognor-
mal distribution. The Markov Chain Monte Carlo algorithm was
iterated for 5 ? 106generations with a burn-in of 5 ? 105genera-
tions. Two independent runs were conducted to ensure conver-
gence upon the stationary distribution. Effective sample size
(ESS) was examined for all parameters to determine if the MCMC
chain was run for enough generations to obtain sufficient indepen-
dent samples for all parameters from the posterior distribution.
The program TRACER v1.3 (Rambaut and Drummond, 2004) was
used to compile and visualize the results from BEAST v1.4.2.
2.5. Estimation of population growth
Several mitochondrial haplotypes are shared among North
American Lycaeides and all three major mitochondrial lineages
contain mitochondrial haplotypes from multiple Lycaeides species
(see Section 3); these patterns are likely due to substantial mito-
chondrial gene flow among North American Lycaeides (Gompert
et al., 2006a, 2006b). Therefore, using molecular variation to make
demographic inferences for each species separately could be prob-
lematic. Instead, we have chosen to treat North American Lycaeides
as a single interbreeding unit when estimating demographic
parameters. In particular, we employ a Bayesian skyline plot to
estimate h (2 ? effective population size ? mutation rate) through
484
Z. Gompert et al./Molecular Phylogenetics and Evolution 48 (2008) 481–490

Page 5

time for North American Lycaeides. Bayesian skyline plots (see be-
low) are generally used to estimate the demographic parameter h
through time for a given interbreeding population or species
(Drummond et al., 2005; Crandall et al., 2008). We believe this is
justified given the pattern of variation for mtDNA observed in
North American Lycaeides. While our analysis may inflate our esti-
mate of h, it should have little influence on the rate of change in h
through time, which is the relevant parameter for this study.
If Lycaeides has expanded its range, diversified, and/or invaded
new habitats, following its colonization of North America, there
should be evidence of population growth in North American Lycae-
ides. To examine the demographic history of North American
Lycaeides we generated a Bayesian skyline plot (Drummond et al.,
2005). The Bayesian skyline plot provides a coalescent-based esti-
mate of the demographic parameter h (2 ? effective population
size ? mutation rate), through time (Strimmer and Pybus, 2001;
Drummond et al., 2005). This method does not require a pre-spec-
ified parametric demographic model. Furthermore, the Bayesian
skyline plot, unlike the generalized skyline plot, provides credibil-
ity intervals on h that take into account both phylogenetic and coa-
lescent uncertainty (Drummond et al., 2005). The software BEAST
v1.4.2 (Drummond and Rambaut, 2006) was used to generate a
Bayesian skyline plot. For this analysis only mtDNA haplotypes
that were sampled in North America were included; this includes
haplotypes found both in North America and Eurasia. The SRD06
model of sequence evolution was used (Shapiro et al., 2006). The
analysis was implemented under a relaxed clock with the rate
for each branch drawn from a lognormal distribution. The Markov
Chain Monte Carlo algorithm was iterated for 2 ? 107generations
with a burn-in of 2 ? 106generations. Two independent runs were
conducted to ensure convergence upon the stationary distribution.
Effective sample size (ESS) was examined for all parameters to
determine if the MCMC chain was run for enough generations to
obtain sufficient independent samples for all parameters from
the posterior distribution. The program TRACER v1.3 (Rambaut
and Drummond, 2004) was used to compile and visualize the re-
sults from BEAST v1.4.2. We re-ran this analysis with all Eurasian
Lycaeides haplotypes included. Including the Eurasian haplotypes
did not alter the results of the analysis (results not shown).
To further examine the demographic history of North American
Lycaeides, we compared the fit of five parametric demographic
models to our mtDNA data. These models were as follows: con-
stant population size, exponential growth, logistic growth, expan-
sion growth, and piecewise constant population size with a
single change in population size (see Pybus and Rambaut, 2002
for model details). Likelihoods for each model were estimated
using the software Genie v3.0 (Pybus and Rambaut, 2002). The
ultrametric maximum likelihood tree used for these analyses was
generated using Paup 4.0b10 (Swofford, 2002) under the TrN+I+c
sequence evolution model, which was selected based on AIC re-
sults from ModelTest 3.7 when the Lycaeides mtDNA data was
not divided into partitions (Posada and Crandall, 1998). A molecu-
lar clock was enforced for the maximum likelihood tree estimation.
Likelihoods for the five demographic models were estimated by
Genie v3.0 (Pybus and Rambaut, 2002) using the Powell algorithm.
Model likelihoods were compared using corrected Akaike Informa-
tion Criterion (AICC).
3. Results
3.1. Phylogeny reconstruction
The model with data partitioned by gene region and codon po-
sition resulted in an increase in 308.23 and 249.61 likelihood units
relative to the models with a single partition and with the data par-
titioned by gene region, respectively (Table 2). Following the crite-
ria of Kass and Raftery (1995), this magnitude of difference
indicates that we should favor the model with data partitioned
by gene region and codon position and that we have ‘‘decisive” evi-
dence against the alternative models with fewer data partitions.
‘‘Strong” to ‘‘decisive” evidence against a hypotheses based on
Bayes factors provides a similar level of support for rejecting a null
hypotheses as a p-value of 0.001 (Goodman, 1999). Thus, we pres-
ent our phylogenetic results based on the model with sequence
data partitioned by gene region and codon position.
Bayesian maximum likelihood analysis strongly supported the
monophyly of the sampled Lycaeides mitochondrial haplotypes
(posterior probability 1.00; Fig. 2). Haplotypes sampled from the
Eurasian species P. argus formed a monophyletic clade sister to
Lycaeides (Fig. 2). Thus, based on this analysis, the genus Plebejus
appears to be paraphyletic with respect to Lycaeides. Within Lycae-
ides three differentiated mitochondrial lineages were detected
(Fig. 2). However these lineages did not correspond to nominal
species (L. idas, L. melissa, the alpine hybrid species, and L. argy-
rognomon) or geographic regions (North America, Asia, and
Europe). Lineage I consisted of individuals sampled from L. i. anna,
L. i. azureus, L. i. ricei, and the alpine species in North America, as
well as an L. i. degener individual from Spain. Lineage II included
individuals collected from L. m. melissa populations in North Amer-
ica and both L. i. idas and L. argyrognomon populations from Asia.
Within this clade haplotype h02 was shared between North Amer-
ican L. m. melissa and Asian L. argyrognomon, and haplotype h56
was shared between Asian L. i. idas and L. argyrognomon. Lineage
III consisted of individuals sampled from L. i. alaskensis, L. i. ricei,
L. i. sublivens, L. m. annetta, L. m. inyoensis, L. m. melissa, and L. m.
samuelis populations from North America, as well as L. argyrogno-
mon, L. a. jautica, and L. i. magnagraeca populations from Asia.
These patterns suggest recent gene flow between Eurasian and
North American taxa, and thus, suggest a recent origin for the var-
iation observed in North America.
There were several North American populations with haplo-
types from two major evolutionary lineages. The L. m. melissa pop-
ulations at Beckwourth Pass, CA, Elmore Co., ID, and Gardnerville,
CA possessed haplotypes from both lineage II and lineage III. The
L. i. ricei population at Marble Mts., CA had haplotypes from lin-
eages I and III, and the alpine population at County Line Hill in
the White Mountains had haplotypes from lineages I and II (Table
1 and Fig. 2).
Bayes factors were used to compare the posterior odds of our
unconstrained Bayesian mitochondrial genealogy to genealogies
with: (1) North American Lycaeides constrained to be monophy-
letic and (2) North American and Eurasian Lycaeides constrained
to be reciprocally monophyletic. The hypothesized unconstrained
genealogy resulted in an increase of 105.87 (North American Lycae-
ides monophyletic) and 107.06 (North American and Eurasian
Lycaeides reciprocally monophyletic) likelihood units compared
to the constrained genealogies (Table 2). Based on the criteria of
Kass and Raftery (1995) this magnitude of difference suggests that
we should favor the unconstrained genealogy and consider the evi-
dence against the constrained topologies as ‘‘decisive” for both the
hypothesis of North American monophyly and the hypothesis of
reciprocal monophyly.
Enforcing a strict molecular clock resulted in an increase in
37.03 likelihood units relative to a model without a clock enforced
(Table 2). When calculating marginal likelihoods it is not uncom-
mon for a constrained model to have a higher likelihood than an
unconstrained model, as the unconstrained model may not sample
as tightly around the maximum likelihood parameter values and
thus result in a lower harmonic mean of the likelihood values than
the constrained model (Nylander et al., 2004). Based on this mag-
nitude of difference in the marginal likelihoods of the models we
Z. Gompert et al./Molecular Phylogenetics and Evolution 48 (2008) 481–490
485

Full-text

View

1 Download

Available from

txstate.edu

Keywords

Bayesian maximum likelihood methods

Coalescent-based analyses

complex colonization history

demographic history

established history

Lycaeides colonize North America

major mitochondrial lineages

mitochondrial data

mitochondrial DNA variation

monophyletic group

multiple founding lineages

North America form

North American Lycaeides

North American Lycaeides populations

North American Lycaeides populations exhibit remarkable variation

North American population

polytypic group

recent evolutionary radiation

single Eurasian ancestor

sudden increase

Recent colonization and radiation of North American Lycaeides (Plebejus) inferred from mtDNA.

Article: Miocene and Pliocene dominated diversification of the lichen-forming fungal genus Melanohalea (Parmeliaceae, Ascomycota) and Pleistocene population expansions.

Article: A hierarchical bayesian approach to ecological count data: a flexible tool for ecologists.

Full-text

View other sources

Hide other sources

View other sources

Hide other sources

Keywords

Similar Publications

Diagnostic testing for α-globin gene disorders in a heterogeneous North American population.

The Pheromone Frontalin and its Dual Function in the Invasive Bark Beetle Dendroctonus valens.

Geography of the circadian gene clock and photoperiodic response in western North American populations of the three-spined stickleback Gasterosteus aculeatus.

Zachariah Gompert

Recent colonization and radiation of North American Lycaeides (Plebejus) inferred from mtDNA.

Article: Miocene and Pliocene dominated diversification of the lichen-forming fungal genus Melanohalea (Parmeliaceae, Ascomycota) and Pleistocene population expansions.

Article: A hierarchical bayesian approach to ecological count data: a flexible tool for ecologists.

Full-text

View other sources

Hide other sources

View other sources

Hide other sources

Keywords

Similar Publications

Diagnostic testing for α-globin gene disorders in a heterogeneous North American population.

The Pheromone Frontalin and its Dual Function in the Invasive Bark Beetle Dendroctonus valens.

Geography of the circadian gene clock and photoperiodic response in western North American populations of the three-spined stickleback Gasterosteus aculeatus.

Zachariah Gompert

Log in