Journal of Heredity 2010:101(2):154–163
doi:10.1093/jhered/esp102
Advance Access publication November 27, 2009
Ó The American Genetic Association. 2009. All rights reserved.
For permissions, please email: [email protected].
Genealogical Relationship among
Members of Selection and Production
Populations of Yellow Cedar (Callitropsis
nootkatensis [D. Don] Oerst.) in the
Absence of Parental Information
NASIM MASSAH, JINLIANG WANG, JOHN H. RUSSELL, ANNETTE VAN NIEJENHUIS,
YOUSRY A. EL-KASSABY
AND
From the Department of Forest Sciences, Faculty of Forestry, The University of British Columbia, 2424 Main Mall,
Vancouver, BC V6T 1Z4, Canada (Massah and El-Kassaby); Institute of Zoology, Zoological Society of London, London,
UK (Wang); British Columbia Ministry of Forests and Range, Research Branch, Cowichan Lake Research Station, Box 335,
Mesachie Lake, BC, Canada (Russell); and Saanich Forestry Centre, Western Forest Products Limited, Saanichton, BC,
Canada (Van Niejenhuis).
Address correspondence to Y. A. El-Kassaby at the address above, or e-mail: [email protected].
Abstract
We used DNA fingerprinting and pedigree reconstruction to determine the genetic relationship among members of
3 yellow-cedar (Callitropsis nootkatensis [D. Don] Oerst.) selection populations in the absence of their parental genotypes.
Selection population members consisted of the tallest individuals within seedling crops originated from natural stand seed
collected from multiple seed donors covering wide areas within 3 distinct locations (phenotypic mass selection). Pairwise
relative kinship estimates indicated the presence of extensive coancestry among the selected seedlings, and pedigree
reconstruction grouped each selection members into multiple full-sib families of different sizes (1–10) nested within several
half-sib families (19–21). The ‘‘STRUCTURE’’ program (Pritchard JK, Stephens M, Donnelly P. 2000. ‘‘Inference of
population structure using multilocus genotype data.’’ Genetics. 155:945–959.) provided a pictorial classification of the
3 selection populations and grouped their individuals into multiple cohorts (9–10). The STRUCTURE program’s results
corresponded with that of the pedigree reconstruction, indicating that members of the selection populations originated from
a subset of the seed donors forming the natural stand seed collections. The species’ silvics, reproductive biology, methods of
natural stand seed collection and seedling production, and the high selection intensity applied to form the selection
populations contributed to limiting the selection to a subset of the original donor trees. The associated buildup of
coancestry in selection and production populations is expected to result in inaccurate estimation of genetic parameters and
an unintentional reduction in genetic diversity in reforestation stocks.
Key words: Callitropsis nootkatensis, clonal testing, DNA fingerprinting, pairwise relative kinship, pedigree reconstruction, selection
The recurrent selection scheme of tree improvement
programs follows 3 main steps: phenotypic selection of
candidate trees from natural stands or plantations; breeding,
wherein mating designs are commonly used to create
structured pedigreed material for testing, evaluation, and
ranking of selected parents based on their offspring’s
performance; and the identification of elite genotypes for
production population(s) establishment and/or starting
a new round of selection (Namkoong et al. 1988). The
154
robustness of this approach has been proven, and substantial genetic gain has been captured for many species for
a multitude of attributes (e.g., Namkoong 1979; Xiang et al.
2003). This classical approach with its predictable mechanics
(i.e., breeding, testing, and selection) remained unchanged;
however, species biology such as predisposition to cloning
through vegetative propagation or somatic embryogenesis
provided opportunities to increasing testing fidelity (Russell
and Cartwright 1991; Heilman 1999; Sutton et al. 2004) and
Massah et al. Genealogy of Yellow Cedar
the application of clonal multiplication for deployment
(Karlsson and Russell 1990; Russell et al. 1990; Heilman
1999; Cyr et al. 2001).
Yellow cedar (Callitropsis nootkatensis [D. Don] Oerst.),
formerly known as Chamaecyparis nootkatensis [D. Don]
Spach), is an important timber species in Western North
America covering wide geographical and ecological range
spanning up to 20° in latitude and elevation from sea level
up to tree line (Harris 1990) (Figure 1). The species
taxonomy has gained recent attention since the discovery of
new species (Xanthocyparis vietnamensis [Farjon et Hiep]) in
Northern Vietnam (Farjon et al. 2002; Little et al. 2004), and
its contemporary taxonomy is the subject of intense debate
(Gadek et al. 2000; Ritland et al. 2001; Wang et al. 2003).
The species has a 3-year reproductive cycle as adaptation to
its high elevation habitat (Owens and Molder 1974, 1975,
1977); however, this cycle is plastic and is reduced to 2 years
in its southern and at low elevation in its northern range
(El-Kassaby et al. 1991).
An innovative approach to reducing efforts as well as
shortening the length of the breeding cycle was attempted
for yellow cedar. This scheme involved substituting
structured mating designs with wind-pollination matings
among trees in natural stands and the application of 2
phases of selection. The first was a high intensity
phenotypic mass selection (Allard 1960, p. 252) among
natural crosses’ bulked offspring conducted at very early
age (1-year-old seedlings), whereas the second took place
after intense field testing over multiple sites and years using
clonal material produced by vegetative propagation of the
Figure 1.
first selection phase members. The intent of the second
selection phase is to identify elite genotypes for their
inclusion into production populations. This approach was
successful in capturing appreciable amount of genetic gain;
however, the extent of this gain cannot be ascertained
specifically when the genetic relationship among members
of the first and second selection phases is unknown.
Genetic gain calculation for clonal selection differs from
that of half-sib families. The former is similar to that for
monozygotic twins and assumes that the covariance among
clones equals 100% of the additive genetic variance,
whereas the latter assumes the covariance among half-sibs
equals 25% of the additive genetic variance (Falconer and
Mackay 1996). Thus, if genetic gain calculations are based
on clonal material, as in the case when the genealogy
among selections is unknown, the estimated gain will be
higher than that based on structured half- and full-sib
families.
The objectives of this study are to determine 1) the
genealogical relationships among the first and second phase
selection members and 2) effective population size (extent
of genetic diversity) for the testing and deployment
populations. If members of the first or second phase
selections are related, then the original genetic parameters
derived before considering the presence of coancestry
require adjustment, and care must be directed toward the
vegetative multiplication of the production population
members to avoid potential reduction in genetic diversity
in reforestation material through the amplification of the
related material.
The range of yellow cedar in the Pacific North West (Harris 1990) and the location of the 3 studied populations.
155
Journal of Heredity 2010:101(2)
Materials and Methods
Seedlings representing 3 first-phase selection populations
originated from natural stand seed collections made in
Vancouver Island, British Columbia (Table 1; Figure 1),
were used in this study. At each location, seed was collected
using a helicopter equipped with cone rakes to strip seed
cone–bearing branches from tree tops. Collections were
made from 29, 30, and 36 nonneighboring trees to form
seedlots #1, 2, and 5, respectively; thus, it was assumed that
seed donors acted as pollen recipients from their restricted
neighborhood. For each location, the collected seed cones
were bulked prior to seed extraction; therefore, the genetic
identity of the seedlings produced from these seedlots was
unknown. The seedlots were sown in a commercial nursery
to produce seedlings (several hundred thousands) for
reforestation purposes. At the end of the nursery growing
season, the healthiest and tallest 1-year-old seedlings within
each seedlot were selected to form the testing populations
(phase 1 selection). A total of 170, 271, and 133 individual
seedlings were selected from seedlots #1, 2, and 5,
respectively. The selection populations (seedlings) were
planted in an arboretum and were repeatedly topped to form
‘‘stool beds’’ for the mass production of small shoots that
were used to produce clonal material for field testing
through rooting. The resulting seedlings (ramets) were
planted in multiple replicated clonal trials over Vancouver
Island. Nine years after the trial establishment, several
growth and yield attributes were assessed for each individual
within all test sites, and quantitative genetics analyses were
conducted to identify the top performing individuals
(clones) within each selection population. The top 5, 17,
and 8 clones representing seedlots #1, 2, and 5, respectively,
were selected as members of the production population.
Fresh foliage samples from each member of the selection
populations were collected from 5 test sites near Port
McNeill (3 sites) and Jordan River (2 sites), Vancouver
Island, British Columbia. Collected samples were placed in
marked plastic bags, stored on ice, and shipped to UBC for
further processing. The tips of each branchlet sample were
placed in a vial tube (2 ml) and stored at 80 °C until DNA
extraction.
DNA extraction followed the Doyle and Doyle (1987)
protocol after optimizations to remove organic compounds
such as terpenes, proteins, and phenols (Barton 1976).
Extracted DNA was qualified and quantified, and each
sample was diluted to 20 ng/ll and stored at 20 °C for
polymerase chain reaction (PCR) amplification. Four yellow
cedar–specific microsatellite primers (Y1F10, Y1G09,
Y2C12, and Y2H01) (Bérubé et al. 2003), one (Cos2590)
of 19 primers tested on Japanese cypress (Chamaecyparis
obtusa), a closely related species (Nakao et al. 2001;
Matsumoto et al. 2006), and 1 of 6 tested primers obtained
from the C. obtusa available expressed sequence tag (EST)
data, accessed via genebank, were selected for genotyping.
We used the SSR finder program (Robinson et al. 2004) to
search through the C. obtusa available EST clones to detect
microsatellite sequences with the following search criteria:
1) primer minimum, optimum, and maximum temperature
were 50, 55, and 65 °C, respectively, and maximum
temperature difference was set at default, 2) primers# G-G
content % (GC%) was set at minimum, optimum, and
maximum of 40%, 50%, and 60%, respectively, and 3) the
default settings were used for the rest. Because GC has
3 binding sites, we set the percentage of their existence on
high in order to reach a better binding and amplification.
PCR reaction conditions were set as recommended by the
primer developers, and PCR reaction products were
electrophoresed on polyacrylamide gels using a LiCor
4200 automated sequencer (LiCor Inc., Lincoln, NE).
For more accurate genotyping and minimizing scoring
errors, a search for allelic variation was done using 2 subsets
of 30 individuals from seedlots #1 and 2, and then the
selected 6 loci were amplified and genotyped with the
SAGA GT (LiCor Biosciences, Lincoln, NE, United States)
software. In addition to the standard marker/DNA ladder,
samples that captured allelic variation within each locus
were selected, mixed, and used as a DNA reference of
known allele size. Each individual PCR plate contained
samples from 2 seedlots at a time to aid in correct allelic
detection, and all gels were scored twice.
Data analyses were conducted using 3 independent
approaches: the first to explore if genetic relationship
existed among the selected seedlings within each seedlot
using Ritland’s (1996) ‘‘MARK’’ program, the second to
construct the pedigree within each seedlot and classify the
selected seedlings into multiple full-sib families nested
within half-sib families using Wang’s (2004) ‘‘COLONY’’
program (it was considered as the primary analysis), and
finally, Pritchard et al.’s (2000, 2007) STRUCTURE program to provide a pictorial representation of cohorts within
each seedlot. The first 2 analyses were used to assist in
hypothesis development (MARK) and testing (COLONY),
whereas the third analysis (STRUCTURE) was used to
verify the results obtained from the COLONY program.
The ‘‘MA R K’’ program (Ritland 1996) was used to
calculate the genetic relationship among each possible dyad
within each seedlot by pairwise kinship analysis. The kinship
Table 1. Location of the 3 yellow-cedar seed collections, the number of seed donors, number of seedlings selected during the firstphase selection, and their seedlot registration numbers
Seedlot
Location
Latitude
Longitude
Elevation
# Of seed donors
# Of selected seedlings
1
2
5
Port McNeill
Shawnigan
Adams River
50°19#
48°45#
50°25#
127°08#
124°00#
126°05#
550
800
570
30
29
36
170
271
133
156
Massah et al. Genealogy of Yellow Cedar
coefficients were classified into 5 biologically meaningful
distinct groups (0.0, 0.125, 0.250, 0.375, and 0.5, representing unrelated, half-sib, full-sib, full-sib with inbred parents,
and inbred, respectively). The COLONY program ver. 1.3
(Wang 2004) was used to reconstruct all sample relationships as a cohort by applying their genotype data after
setting genotyping error rate at 5%. The COLONY program
uses group-likelihood methods, and individuals are studied
as a group after integrating the genotypic information of all
samples. However, it is restricted to distinguish first-degree
relationships such as full-sibs and half-sibs in a cohort. The
STRUCTURE program is a model-based method, which
utilizes the Bayesian algorithm to cluster individuals within
a population into genetically similar cohorts based on their
multilocus genotypes (Pritchard et al. 2000). The program
calculates every locus allele frequency separately for each
population and clusters individuals to cohorts based on their
posterior probabilities (Q) calculated from their genotype
data (Pritchard et al. 2000). This analysis is commonly used
in population structure–related studies to allocate each
entity to its genetically related cohort even in cases where
there is no prior parental or population structure information or in cases where entities are originated from
more than one population. STRUCTURE program ver. 2.2
(Pritchard et al. 2000) was used with a length of burn-in
period parameter set at 50 000 (burn-in length between
10 000 and 100 000 were recommended by the author), and
the same number was taken for Markov chain Monte Carlo
replications after burn-in. We set K (cluster) to range
between 1 and 20 with number of independent runs set at
40 (i.e., each K was simulated 40 times). The number of
most likely cohorts was determined using the ad hoc statistic
known as DK, which is based on the log probability change
across successive K values (Evanno et al. 2005).
Genetic diversity of a population (selection or production) can be expressed as the effective population size,
which describes the proportion of individuals from various
parents and their genetic relationships. Considering related
and/or inbred individuals, the concept of status number was
used (Lindgren et al. 1996) and denoted as Ne in this study
(Ne 5 ½H). Group coancestry, H, represents the average
coancestry of all pairs of population members including selfcoancestry and reciprocals (Cockerham 1967).
Results
The kinship analysis (Ritland 1996) estimated all possible
pairwise relative kinship coefficients among individuals
within each selection population (n(n 1)/2) (Figure 2).
The mean kinships over all individual pairs within the
studied 3 populations were 0.061(standard error 5 0.00067),
0.052 (0.00041), and 0.098 (0.00140), respectively. This
analysis provided insight of the genetic relationship among
individuals within each of the 3 selection populations. In
cases where no genetic relationship exists among the
selected seedlings (relative kinship (r) 0.0), it is expected
that all of the pairwise combinations will produce a value of
Figure 2. Distribution of pairwise relative kinship estimates
among members of 3 yellow-cedar selection populations (top:
Port McNeill [seedlot #1]; middle: Shawnigan [seedlot #2]; and
bottom: Adams River [seedlot #5]). Pairwise values were
classified under 5 biologically meaningful classes with
midpoints of 0.0, 0.125, 0.25, 0.375, and 0.5.
zero. However, the pattern obtained from the 3 selection
populations indicated that the selected seedlings belonged to
groups of half-sib families (r 5 0.125) with several smaller
full-sib families (r 5 0.250) nested within these half-sib
families (Figure 2). It should be pointed out that the
pairwise kinship estimate representation of Figure 2 is done
to portray biologically meaningful scenarios as the obtained
kinship estimates of every selection population covered
a continuum between 0 and 1 (for instance, 0.125 class
represents a range of relative kinship estimates between 0.064
and 0.188). Additionally, the 3 seedlots produced 2 very small
peaks of higher coancestry, indicating that some individuals
resulted from mating between inbred parents and/or were
a product of selfing (Figure 2). This pattern is consistent
with the grouping of each population’s individuals into sets of
157
Journal of Heredity 2010:101(2)
full-sib families with different sizes nested within their
respective half-sib families that were obtained from the
pedigree reconstruction analyses (see below).
The pedigree reconstruction analysis (Wang 2004) was
conducted based on the following assumptions: 1) seed
donors received pollen from multiple male parents present
in their vicinity (i.e., mother trees were polygamous) and 2)
self-fertilization would fail to produce viable seed due to the
high genetic load, or selfed seedlings would have a selective
disadvantage and thus would not be included in the
selection populations (see Discussion). The resulting
analyses grouped the selected seedlings into multiple fullsib families nested within 19, 21, and 17 half-sib families, for
seedlots # 1, 2, and 5, respectively (Figure 3). The full-sib
family size within these half-sib families varied and ranged
between 2 and 5, 1 and 10, and 1 and 4, for seedlots # 1, 2,
and, 5, respectively (Figure 3). The production population
members’ genetic status was determined from the pedigree
reconstruction, and half-sib membership was detected for
selection populations # 2 (6 HS families accounting for 14
of the 17 selections) and 5 (2 HS families accounting for 5
of the 8 selections) (Table 2).
Following the recommendations of Bradbury et al.
(2008), the number of iterations for running the STRUCTURE program was set at 40 to allow reaching stable
results. The pictorial classification of cohorts within each
selection population was presented graphically, where every
cohort is defined with a specific color, and individuals are
represented by vertical bars demonstrating their cohort
membership posterior probability. The STRUCTURE program grouped the selected population members into 9
(seedlot # 2) and 10 (seedlots # 1 and 5) cohorts (Figure 4).
For each selection population, some correspondence was
observed between the COLONY and STRUCTURE results
as presented by the size of the largest single half-sib family
within a specific genetic cohort (Table 3). Although half-sib
Figure 3. Pictorial presentations of pedigree reconstruction results for 3 yellow-cedar selection populations (top: Port McNeill
[seedlot #1]; middle: Shawnigan [seedlot #2]; and bottom: Adams River [seedlot #5]). Individuals with each half-sib family are
grouped (nested) within their respective full-sib families.
158
Massah et al. Genealogy of Yellow Cedar
Table 2. Number of genetic classes within the production
population (elite genotypes) and the production population
effective population size (Ne) estimates under genetic
independence (no genetic relationship) and the COLONY family
classification assuming equal contribution to the reforestation
planting stock
Elite genotypes
Ne
Seedlot
Half-sib Half-sib
#
Unrelated (size 5 2) (size 5 3) Independent COLONYa
1
2
5
a
5
3
3
—
4
1
—
2
1
5
17
8
5.0
13.1
6.4
Wang 2005.
family membership was not restricted to a single genetic
cohort, some family representation was high and ranged
between 18% and 61% of the cohort’s size, indicating that
seed donors contributed unequally to the resultant seedlots
and selection reflected their seedlings phenotypic superiority.
Discussion
The attempt to circumvent the arduous long-term classical
breeding program by substituting parent-tree selection and
structured mating designs with extensive early selection
from naturally produced population of seedlings is novel. It
has good potential to shorten the length and substantially
reduce both costs and efforts of breeding programs. This
approach is based on several assumptions that include 1) the
nursery seedling population genetic structure is not affected
by the species silvics and its peculiar seed germination and
dormancy, 2) members within each selection population are
genetically independent, 3) nursery truncation selection does
not limit the selected seedlings to few seed donors (families),
and 4) the ageage correlation between nursery and field
performance is reasonably high.
Several factors must be considered during the interpretation of results obtained from pairwise, pedigree
reconstruction, and STRUCTURE analyses. These are
related to the genetic structure within the 3 natural stand
seedlots as well as the expected level of coancestry present
among their respective seedlings. Although the number of
seed-cone donors forming each seedlot is known, their
proportionate contribution to the bulk seedlot and subsequently their successful input to the seedling crops are
unknown. The reproductive biology of yellow cedar is
characterized by a 3-year cycle, an adaptation to its high
elevation habitat (Owens and Molder 1974). When seed
cones are collected, trees often carry a mixture of 1- and
2-year cones that are difficult to differentiate. The collected
cone crops contain both developmentally mature and
immature seed, and therefore, the volume of cones collected
from a particular seed donor is not indicative of its seed
contribution. Even when a donor contributes substantially
Figure 4. Results from the STRUCTURE program for 3 yellow-cedar selection populations (top: Port McNeill [seedlot #1];
middle: Shawnigan [seedlot #2]; and bottom: Adams River [seedlot #5]). Scale represents individuals’ group membership
posterior probability.
159
Journal of Heredity 2010:101(2)
Table 3. Correspondence between the COLONY and
STRUCTURE programs represented by the size of the largest
half-sib family within a single genetic cohort (%)
Genetic cohorts K (cluster)
Seedlot #
1
2
3
4
5
6
7
8
9
10
1
2
5
20
22
38a
23
61
33
31
44
33
42
44
22a
55
38
43
23
50
18b
23
35
19b
47
29
22
33
27
21
29
—
36
a
Two half-sib families with the same size.
b
Three half-sib families with the same sizes.
to the seed crop, this does not guarantee increased
representation in the seedling crop, due to genetic variation
in seed germination and the deep dormancy of fully
developed seed (Pawuk 1993; El-Kassaby 1995).
The species is commonly present in small patches
(clusters) with many scattered individuals (Harris 1990);
thus, nonneighboring seed-cone donors are expected to
receive pollen from their adjacent locations (i.e., they are
spatially and may be temporally isolated). Seed-cone
collections originating from multiple nonneighboring trees
are expected to represent multiple ‘‘local’’ gene pools, thus
increasing the chance for greater differentiation among the
wind-pollinated maternal half-sib families originated from
these seedlots. Owens and Molder (1977) estimated that
mature cones averaged 7.2 seeds per cone with only 29%
being filled, thus yielding only about 2 filled seeds per cone.
This low yield may result from sparse pollination, increased
abortion of selfed seed (Bishir and Pepper 1977), and the
rarity of 2 consecutive years of good environmental
conditions, needed for pollination, fertilization, and embryo
and seed development (El-Kassaby 1995).
The collective genetic structure of seedling populations
originating from natural stand seed is expected to harbor
high level of variability, reflecting the variation among the
maternal parents (seed donors) as well as their multiple local
pollen pools. British Columbia’s 2003–2008 average number
of yellow-cedar seedlings produced from natural stand seed
sources is 900 K (SPAR: BC Ministry of Forests and Range
2009). Natural stand seedlots are used to produce seedling
crops of size varying between 50 and 100 K seedlings.
Under mass selection for outcrossing species (Allard 1960,
p. 252), factoring in the total number of seedlings selected
from the 3 studied seedlots (average: 191; range: 133–271), it
is safe to assume that the selection intensity applied was
close to 1:500–1:1000. Thus, the high selection intensity
applied in the nursery to form the testing populations is
expected to limit the selection to a subset of the original
number of parents forming these seedlots. Considering
these factors collectively, it is not surprising that the results
from either the COLONY or STRUCTURE programs
produced a smaller number of parents/cohorts as compared
with the original seed donors forming the seedlots.
The pattern obtained from the pairwise kinship analyses
(Ritland 1996) confirmed the presence of half- and full-sib
families (Figure 2) and allowed the formulation of the main
160
hypotheses that no genetic relationship exists among the
nursery stage selected seedlings (selection populations) or
among the selected elite genotypes forming the production
populations. Naturally, the presence of sib-ship relationship
among these individuals formed our alternate hypothesis.
We therefore employed the COLONY program (Wang
2004) to determine genetic membership among the selected
seedlings within each population. We assumed that the seed
donors receive pollen from multiple pollen donors
(polygamous) and that self-fertilization does not result in
viable seed, or self-seedlings are selected against at the
nursery stage during the removal (thinning) of excess
germinants, or seedling development during the growing
season resulted in inferior seedlings (El-Kassaby and
Thomson 1995; El-Kassaby 2000). Thus, self-seedlings
would not be included in the selection populations. This
assumption is supported by reports of inbreeding depression
causing embryo development failure or selective disadvantage of selfed germinants, seedlings, and saplings has been
reported for multiple conifers including yellow cedar
(Sorensen 1969, 1982; Bramlett and Popham 1971; Koski
1971, 1973; Bishir and Pepper 1977; Bishir and Namkoong
1987; Namkoong and Bishir 1987; Savolainen et al. 1992;
Williams and Savolainen 1996; Williams et al. 1999).
Forest trees commonly practice a mixed-mating system
with reported high levels of outcrossing rate (O’Connell
2003) and often show low levels of correlated matings
(O’Connell et al. 2001; Liewlaksaneeyanawin 2006). However, low selfing might be contradictory to the application of
Wang’s (2004) COLONY pedigree reconstruction program.
Like other members of the Cuperssaceae family, yellow
cedar can self-fertilize, and high inbreeding coefficients were
reported by Ritland et al. (2001) (mean rangewide F 5 0.18,
also see Thompson et al. 2008). Using Wright’s (1965)
approach (t 5 (1 FST)/(1 þ FST)), we estimated an
inbreeding coefficient of F 5 0.08 for the 3 studied
populations. The difference between Ritland et al. (2001) and
the present study’s inbreeding coefficient estimates may
reflect the high selection intensity applied in the greenhouse
to form the 3 test populations. The observed pattern from
the pairwise analyses (Figure 2) as well as reported inbreeding
coefficient estimates for the species and that of the present
study could support the notion that some members of the
classified full-sib families may be the result of selfing.
The pedigree reconstruction analyses (Wang 2004)
grouped the test population members into multiple full-sib
families nested within half-sib families (Figure 3). A total
of 19, 21, and 17 half-sib families were observed within
seedlots #1, 2, and 5, respectively, and multiple full-sib
families with variable sizes (2–5, 1–10, and 1–4, for seedlots #
1, 2, and 5, respectively) were nested within them (Figure 3).
The observed number of half-sib families with their
variable sizes confirmed that the strict truncation selection
applied in the nursery during the test population formation
successfully eliminated some as well as restricted the
selection to a subset of seed donors. The observed family
structure reduced the testing populations’ effective population size to 39.9, 37.6, and 28.0 for seedlots # 1, 2, and 5,
Massah et al. Genealogy of Yellow Cedar
respectively, as compared with their census numbers (average
191; range: 133–271).
The STRUCTURE analyses provided pictorial classification of each seedlot presented by the program’s bar plots
(Figure 4). Cohorts are defined by different colors where
individuals are represented by vertical bars demonstrating
the posterior probability (q) of each individual belonging to
a specific cohort (Figure 4). It should be noted that the
posterior probability of some individuals indicates that they
belong to more than one cohort (i.e., individuals are
represented by 2 or more colors), or it could be the result of
incomplete information caused by missing data (genotype)
(Bergel and Vigilant 2007). In general, the number of
cohorts within a test population was smaller than the halfsib family number obtained from the pedigree reconstruction and ranged between 9 and 10 (Figure 4; Table 2). This
reflects the difference between COLONY and STRUCTURE, where the former is based on the exact nature of
jointly assigning parentage and sib-ships based on their
multilocus genotypes and the latter on providing a group
membership posterior probability of individuals.
Combining pedigree reconstruction and STRUCTURE
program in analyzing natural populations has been attempted by Vähä et al. (2007) in studying the natal homing of
Atlantic salmon populations. They repeatedly used the
STRUCTURE program to stratify the populations in
sequential manner, and finally, they used pedigree reconstruction (the COLONY program) to identify the
genealogical relationship among the spawning individuals
within each stratum. In the present study, the test
populations were already stratified as they originated from
seedlots collected from different locations (Figure 1). Our
approach was similar and the use of the 2 analytical
approaches (i.e., pedigree reconstruction and STRUCTURE) assisted in unraveling the genetic similarity/differences among each seedlot’s selected seedlings. However, it
should be noted that the STRUCTURE and COLONY
program results did not completely match in spite of using
the same multilocus genotypic data. These discrepancies
could be caused by violating the STRUCTURE program’s
assumptions, specifically those pertaining to the presence of
family structure within the seedling populations (Pritchard
and Wen 2004; Camus-Kulandaivelu et al. 2007). Under
increasing number of groups, the STRUCTURE program
analysis starts by identifying the most genetically distinct
groups followed by the identification of the less distinct
groups. The presence of related individuals creates sufficient
correlation among these individuals, forcing them to be
identified as a genetic group. This situation was unintentionally encountered by Vähä et al. (2007) when they
repeatedly applied the STRUCTURE analysis to stratifying
their Atlantic salmon populations into subgroups and finally
realized that the final ‘‘Structure’’ represented full- and halfsib families. In essence, we followed Yu et al.’s (2006)
approach to identify possible family structure in the
analyzed populations.
Results from the COLONY and STRUCTURE programs were compared to investigate the presence of any
correspondence between the 2 approaches (Table 2).
Although the COLONY and STRUCTURE programs
produced different number of groups (half-sib families
and cohorts, respectively), some correspondence between
their results is apparent by the size of the largest single halfsib family within a specific genetic cohort (Table 2).
Although half-sib family membership was not restricted to
a single genetic cohort, some family representation was high
and ranged between 18% and 61% of the cohort’s size,
indicating that seed donors may have contributed differently
to the seedlots and that some members of different cohorts
shared common alleles. Additionally, some cohorts included
members of more than one half-sib family (cohorts #1 and
4 encompassed 2 half-sib families, each representing 38% of
the cohort size, whereas cohorts # 6 and 7 included 3 halfsib families representing 18% and 19% of the cohort size,
respectively; Table 2).
The ultimate goal of the implemented selection program
was to identify elite genotypes for their inclusion in
production populations. After field testing, the second
selection phase resulted in identifying 5, 17, and 8
individuals, representing seedlots #1, 2, and 5, respectively,
as determined by the COLONY analyses (Table 3). These
selected genotypes will be used to produce large numbers of
rooted cuttings for reforestation. If these genotypes are
related, then it is expected that the genetic diversity of the
deployed material will be lower than that assumed under no
relationship; in fact, the effective population sizes of the
selection populations were found to be lower than their
perceived census numbers (see Table 3). With the exception
of seedlot # 1 where no genetic relationship was detected,
seedlots # 2 and 5 contained 5 and 2 half-sib families with
different sizes, respectively, resulting in effective population
sizes of 13.1 and 6.4 as opposed to census numbers of 17
and 8 (Table 3).
The utilization of matings among parents in their natural
settings coupled with the implementation of 2 rounds of
selection present a novel approach to breeding and selection
of forest trees and provides an opportunity to reducing
breeding time, efforts, and resources. The present study
demonstrates the necessity of applying molecular biology
techniques such as DNA fingerprinting and pedigree
reconstruction as essential tools to advance these new
breeding methods. The pioneering work of Lambeth et al.
(2001) and Grattapaglia et al. (2004) opened new avenues to
molecular tree breeding, and this work has been followed by
the development of even bolder approaches such as
‘‘Breeding Without Breeding’’ (El-Kassaby and Lstibůrek
2009). Tree breeding must capitalize on ongoing scientific
developments in molecular biology, spatial statistics,
parentage assignment, and association genetics to advance
population improvement to new frontiers.
Conclusion
The implementation of an unconventional approach to tree
improvement that deviates from classical breeding schemes
161
Journal of Heredity 2010:101(2)
was attempted. This approach forfeits the phenotypic
selection of ‘‘superior genotypes’’ phase as well as the use
of conventional mating designs for creating structured
pedigree for testing. The method involves high selection
intensity at very early age (1 year) from offspring originated
from natural crosses among parent trees in their native
habitats, followed by exploitation of the species’ biology
(easy vegetative propagation) for the production of clonal
material for long-term extensive field testing. Using this
approach, the pedigree of the offspring selected for testing
is unknown, which could cause inappropriate estimation of
quantitative genetics parameters as well as the deployment
of genetically related material, thus reducing genetic
diversity. We applied DNA fingerprinting and pedigree
reconstruction to unravel the presence of extensive family
structure among the selected offspring. In light of the
findings, we recommend revision of 1) genetic parameter
calculations based on assumed independence among tested
individuals (i.e., clones) and 2) genetic diversity estimates of
deployment material, as member propagules of half-sib
families are expected to produce lower effective population
size (a clone’s effective population size is equal to census
number, whereas a half-sib’s effective population size is
lower than the census number).
Funding
Johnson’s Family Forest Biotechnology Endowment, British
Columbia Forest Investment Account Forest Genetic
Conservation and Management program, Natural Sciences
and Engineering Research Council of Canada—Discovery
and Industrial Research Chairs Program grants to Y.A.K.
and Natural Resources Canada—Science and Technology
Internship Program to N.M.
Camus-Kulandaivelu L, Veyrieras J-B, Gouesnard B, Charcosset A,
Manicacci D. 2007. Evaluating the reliability of STRUCTURE outputs in
case of relatedness between individuals. Crop Sci. 47:887–892.
Cockerham C. 1967. Group inbreeding and coancestry. Genetics.
56:89–104.
Cyr DR, Attree SM, El-Kassaby YA, Ellis DD, Polonenko DR, Sutton BCS.
2001. Application of somatic embryogenesis to tree improvement in
conifers. In: Morohoshi N, Komamine A, editors. Molecular breeding of
woody plants. Elsevier Science B.V. p. 305–312.
Doyle JJ, Doyle JL. 1987. A rapid DNA isolation procedure from small
quantities of fresh leaf tissues. Phytochem Bull. 19:11–15.
El-Kassaby Y, Lstibůrek M. 2009. Forthcoming. 2009. Breeding without
breeding. Genet Res. 91:111–120.
El-Kassaby YA. 1995. The fitness of reproductive plasticity in yellow-cedar
(Chaemacyparis nootkatensis). Silvae Genet. 44:217–218.
El-Kassaby YA. 2000. Representation of Douglas-fir and western hemlock
families in seedling crops as affected by seed biology and nursery crop
management practices. For Genet. 7:305–315.
El-Kassaby YA, Maze J, Macleod DA, Banerjee S. 1991. Reproductive cycle
plasticity in yellow-cedar (Chamaecyparis nootkatensis). Can J For Res. 21:
1360–1364.
El-Kassaby YA, Thomson AJ. 1995. Effect of seed biology and common
seedling nursery production practices on parental contribution rank and
relative performance. For Sci. 42:228–235.
Evanno G, Regnaut G, Goudet J. 2005. Detecting the number of clusters of
individuals using the software STRUCTUTRE: a simulation study. Mol
Ecol. 14:2611–2620.
Falconer DS, Mackay TFC. 1996. Introduction to quantitative genetics. 4th
ed. New York: Longman.
Farjon A, Nguyen TH, Harder DK, Phan KL, Averyanov L. 2002. A new
genus and species in Cupressaceae (Coniferales) from Northern Vietnam,
Xanthocyparis vietnamensis. Novon. 12:179–189.
Gadek PA, Alpers DL, Heslewood MM, Quinn CJ. 2000. Relationships
within Cupressaceae sensu lato: a combined morphological and molecular
approach. Am J Bot. 87:1044–1057.
References
Grattapaglia D, Ribeiro VJ, Rezende GDSP. 2004. Retrospective selection
of elite parent trees using paternity testing with microsatellite markers: an
alternative short term breeding tactic for Eucalyptus. Theor Appl Genet.
109:192–199.
Allard RW. 1960. Principles of plant breeding. New York: John Wiley and
Sons Inc.
Harris AS. 1990. Alaska-cedar. Silvics of North America: 1. Conifers.
Agriculture Handbook 654. Washington (DC): USDA Forest Service.
Barton GM. 1976. A review of yellow cedar (Chamaecyparis nootkatensis
[D. DON] SPACH) extractives and their importance to utilization. Wood
Fiber. 8:172–176.
Heilman PE. 1999. Planted forests: poplars. New For. 17:89–93.
Bergel RA, Vigilant L. 2007. Genetic analysis reveals population structure
and recent migration within the highly fragmented range of the cross river
gorilla (Gorilla gorilla). Mol Ecol. 16:501–516.
Bérubé Y, Ritland C, Ritland K. 2003. Isolation, characterization, and crossspecies utility of microsatellites in yellow cedar (Chamaecyparis nootkatensis).
Genome. 46:353–361.
Bishir J, Namkoong G. 1987. Unsound seed in conifers: estimation of
numbers of lethal allele and magnitude of effect associated with maternal
parent. Silvae Genet. 36:180–185.
Bishir J, Pepper WD. 1977. Estimation of the number of embryonic lethal
alleles in conifers: I. Self-pollinated seed. Silvae Genet. 26:50–54.
Bradbury IR, Campana SE, Bentzen P. 2008. Estimating contemporary
early life-history dispersal in an estuarine fish: integrating molecular and
otolith elemental approaches. Mol Ecol. 17:1438–1450.
Bramlett DL, Popham TW. 1971. Model relating unsound seed and
embryonic lethals in self-pollinated pines. Silvae Genet. 20:192–193.
162
Karlsson IK, Russell JH. 1990. Comparison of yellow cypress trees of
seedling and rooted cutting origins after 9 and 11 years in the field. Can J
For Res. 29:37–42.
Koski V. 1971. Embryonic lethals of Picea abies and Pinus sylvestris. Commun
Inst For Fenn. 75:1–30.
Koski V. 1973. On self-pollination, genetic load and subsequent inbreeding
in some conifers. Commun Inst For Fenn. 78:1–42.
Lambeth C, Lee BC, O’Malley D, Wheeler N. 2001. Polymix breeding with
parental analysis of progeny: an alternative to full-sib breeding and testing.
Theor Appl Genet. 103:930–943.
Liewlaksaneeyanawin C. 2006. Genetic evaluation of natural and domesticated lodgepole pine populations using molecular markers. PhD thesis.
Vancouver (BC): The University of British Columbia.
Lindgren D, Gea LD, Jefferson PA. 1996. Loss of genetic diversity
monitored by status number. Silvae Genet. 45:52–59.
Little DP, Schwarzbach AE, Adams RP, Hsieh C. 2004. The circumscription and phylogenetic relationships of Callitropsis and the newly described
genus Xanthocyparis (Cupressaceae). Am J Bot. 91:1872–1881.
Massah et al. Genealogy of Yellow Cedar
Matsumoto A, Tani N, Li X, Nakao Y, Tomaru N, Tsumura Y. 2006.
Development and polymorphisms of microsatellite markers for hinoki
(Chamaecyparis obtusa). Mol Ecol Notes. 6:310–312.
Nakao Y, Iwata H, Matsumoto A, Tsmura Y, Tomaru N. 2001. Highly
polymorphic microsatellite markers in Chamaecyparis obtusa. Can J For Res.
31:2248–2251.
Namkoong G. 1979. Introduction to quantitative genetics in forestry.
Washington (DC): US Department of Agriculture, Forest Service. Tech
Bulletin No 1588.
Namkoong G, Bishir J. 1987. Frequency of lethal alleles in forest tree
populations. Evolution. 41:1123–1127.
Namkoong G, Kang HC, Brouard JS. 1988. Tree breeding: principles and
strategies. New York: Springer-Verlag Monograph, Theo Appl Genet 11.
O’Connell LM. 2003. The evolution of inbreeding in western redcedar
(Thuja plicata: Cupressaceae). PhD thesis. Vancouver (BC): The University
of British Columbia.
O’Connell LM, Viard F, Russell J, Ritland K. 2001. The mating system in
natural populations of western redcedar (Thuja plicata). Can J Bot. 79:
753–756.
Forestry Canada and British Columbia Ministry of Forests. FRDA
Report 171. p. 34–35.
Russell JH, Grossnickle SC, Ferguson C, Carson DW. 1990. Yellow cedar
stecklings: nursery production and field performance. Victoria (BC):
Forestry Canada and British Columbia Ministry of Forests. FRDA Report
148.
Savolainen O, Karkkainen K, Kuttinen H. 1992. Estimating numbers of
embryonic lethals in conifers. Heredity. 69:308–314.
Sorensen F. 1969. Embryonic lethals in Douglas-fir, Pseudotsuga menziesii.
Am Nat. 103:389–398.
Sorensen FC. 1982. The roles of polyembryony and embryo viability in the
genetic system of conifers. Evolution. 36:725–733.
SPAR: BC Ministry of Forests and Range. 2009. Seed planning and registry
application. http://www.for.gov.bc.ca/hti/spar/index.htm
Sutton BCS, Attree SM, El-Kassaby YA, Grossnickle SC, Polonenko DR.
2004. Commercialisation of somatic embryogenesis for plantation forestry.
In: Walter C, Carson M, editors. Plantation forest biotechnology for the
21st century. Kerala (India): Research Signpost. p. 275–301.
Owens JN, Molder M. 1974. Cone initiation and development before
dormancy in yellow cedar (Chamaecyparis nootkatensis). Can J Bot.
52:2075–2084.
Thompson SL, Bérubé Y, Bruneau A, Ritland K. 2008. Three-gene identity
coefficients demonstrate that clonal reproduction promotes inbreeding and
spatial relatedness in yellow-cedar (Callitropsis nootkatensis). Evolution.
62:2570–2579.
Owens JN, Molder M. 1975. Pollination, female gametophyte, and embryo
and seed development in yellow cedar (Chamaecyparis nootkatensis). Can J Bot.
53:186–199.
Vähä J-P, Erkinaro J, Niemelä E, Primmer CR. 2007. Life-history and
habitat features influence the within-river genetic structure of Atlantic
salmon. Mol Ecol. 16:2638–2654.
Owens JN, Molder M. 1977. Cone induction in yellow cypress
(Chamaecyparis nootkatensis) by gibberellin A3 and the subsequent development of seeds within the induced cones. Can J For Res. 7:605–613.
Wang J. 2004. Sibship reconstruction from genetic data with typing errors.
Genetics. 166:1963–1979.
Pawuk WH. 1993. Germination of Alaska-cedar seed. Tree Planter’s Notes.
44:21–24.
Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population
structure using multilocus genotype data. Genetics. 155:945–959.
Pritchard JK, Wen W. 2004. Documentation for structure software.
Chicago (IL): The University of Chicago Press.
Pritchard JK, Wen X, Falush D. 2007. Documentation for structure software Version 2.2. http://pritch.bsd.uchicago.edu/software/structure2_2.
html.
Ritland C, Pape T, Ritland K. 2001. Genetic structure of yellow cedar
(Chamaecyparis nootkatensis). Can J Bot. 79:822–828.
Ritland K. 1996. Estimators for pairwise relatedness and individual
inbreeding coefficients. Genet Res. 67:175–185.
Wang WP, Hwang CY, Lin TP, Hwang SY. 2003. Historical biogeography
and phylogenetic relationships of the genus chamaecyparis (Cupressaceae)
inferred from chloroplast DNA polymorphism. Plant Syst Evol. 241:13–28.
Williams CG, Barnes RD, Nyoka I. 1999. Embryonic genetic load for
a neotropical conifer, Pinus patula Schiede et Deppe. J Hered. 90:394–398.
Williams CG, Savolainen O. 1996. Inbreeding depression in conifers:
implications for breeding strategy. For Sci. 42:102–117.
Wright S. 1965. The interpretation of population structure by F-statistics
with special regard to systems of mating. Evolution. 19:395–420.
Xiang B, Li B, McKeand S. 2003. Genetic gain and selection efficiency of
loblolly pine in three geographic regions. For Sci. 49:196–208.
Yu J, Pressoir G, Briggs H, William HB, Bi IV, Yamasaki M, Doebley JF,
McMullen MD, Gaut BS, Nielsen DM, et al. 2006. A unified mixed-model
method for association mapping that accounts for multiple levels of
relatedness. Nat Genet. 38:203–208.
Robinson AJ, Love CG, Batley J, Barker G, Edwards D. 2004. Simple
sequence repeat marker loci discovery using SSR primer. Bioinformatics.
20:1475–1476.
Received April 2, 2009; Revised October 3, 2009;
Accepted October 22, 2009
Russell JH, Cartwright C. 1991. The genetics of yellow cypress. In: Lousier JD,
editor. Yellow cypress: Can we grow it? Can we sell it? Victoria (BC):
Corresponding Editor: James L. Hamrick
163