GENETIC IMPROVEMENT OF THE PACIFIC WHITE SHRIMP (PENAEUS VANNAMEI) IN CHINA Shengjie Ren M.Sc. (Hydrobiology) B.Sc. (Aquaculture) Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy School of Biological and Environmental Sciences Science and Engineering Faculty Queensland University of Technology 2020
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
GENETIC IMPROVEMENT OF THE PACIFICWHITESHRIMP (PENAEUS VANNAMEI) IN CHINA
Shengjie Ren
M.Sc. (Hydrobiology)
B.Sc. (Aquaculture)
Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy
School of Biological and Environmental Sciences
Science and Engineering Faculty
Queensland University of Technology
2020
Genetic Improvement of the Pacific White Shrimp (Penaeus vannamei) in China i
Keywords
Aquaculture, quantitative genetics, population genetics, microsatellite markers,
genetic diversity, population structure, prawn aquaculture, Bayesian assignment,
1.3.1 Geographic Distribution and Global Production.................................................. 4
1.3.2 Production Lifecycle of P. vannamei................................................................... 6
1.4 Genetic Breeding of Penaeid Shrimps............................................................................ 9
1.4.1 Basic Concepts in Genetic Breeding.................................................................... 9
1.4.2 Domestication History of Penaeid Shrimps........................................................12
1.4.3 Status of Stock Improvement of Penaeid Shrimp...............................................13
1.4.4 Genetic Parameters and Genetic Gains in Selected Traits in PenaeidShrimp.................................................................................................................16
1.5 Bridging the Gap between Population Genetics/Genomics and Quantitative Genetics28
1.6.3 Issues of Base Population................................................................................... 37
1.7 Aims of the Current Project.......................................................................................... 38
1.8 Objectives and Thesis Outline...................................................................................... 39
1.8.1 Chapter (1): General Introduction.......................................................................39
1.8.2 Chapter (2): Characterization for the Culture Resources of Pacific WhiteShrimp of Genetic Diversity, Genetic Structure, and Genetic Relatedness........39
1.8.3 Chapter (3): Genetic Parameters for Body Weight and Survival in theBase Population.................................................................................................. 39
1.8.4 Chapter (4): Comparison of Reproductive Performance of Female PacificWhite Shrimp Reared in Recirculating Tanks vs Earthen Ponds....................... 40
1.8.5 Chapter (5): Quantitative Genetic Analysis of Female ReproductiveTraits................................................................................................................... 40
1.8.6 Chapter (6): General Discussion.........................................................................41
Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeusvannamei Culture Resources in China: Implications for the Production of aBroad Synthetic Base Population for Genetic Improvement................................43
2.4.1 Genetic Variation Levels Within and Among Stocks.........................................67
2.4.2 Population Differentiation and Origins of Genetic Resources........................... 70
2.4.3 Implications for Forming a Genetic Foundation (Base) Population forGenetic Improvement......................................................................................... 74
vi Genetic Improvement of the Pacific White Shrimp (Penaeus vannamei) in China
Chapter 3: Genetic parameters for growth and survival traits in a basepopulation of Pacific white shrimp (Penaeus vannamei) developed fromdomesticated strains in China..................................................................................77
Chapter 4: Comparison of Reproductive Performance of Domesticated P.vannamei Females Reared in Recirculating Tanks and Earthen Ponds: AnEvaluation of Reproductive Quality of Spawns in Relation to Female Body Sizeand Spawning Order...............................................................................................107
4.4.1 Comparative Reproductive Performance of Broodstock in the RT and EPTreatments........................................................................................................ 129
4.4.2 Impacts of Female Body Size on Reproduction Performance..........................131
4.4.3 Quality of Reproductive Performance in Relation to Spawning Order............ 133
5.3.2 Relationships Between Body Weight and Number of Eggs/Nauplii perSpawn............................................................................................................... 150
5.3.3 Frequency Distribution of Number of Females Spawning............................... 151
5.3.4 Genetic (Co)variances Among Traits............................................................... 152
5.3.5 Genetic and Phenotypic Correlations among Reproductive Traits...................153
Chapter 6: GENERAL DISCUSSION............................................................166
viii Genetic Improvement of the Pacific White Shrimp (Penaeus vannamei) in China
6.1 Characterization of Pacific White shrimp genetic diversity and genetic structure inChina........................................................................................................................... 167
6.2 Genetic parameters for body weight and survival in the base population.................. 170
6.3 Comparison of reproductive performance of female Pacific white shrimp reared inrecirculating tanks vs earthen ponds........................................................................... 171
6.4 Quantitative genetic analysis of female reproductive traits........................................ 174
6.5 Future direction for Pacific white shrimp breeding programs.................................... 176
6.5.1 Breeding Strain for AHPND Disease Resistance............................................. 176
first in value among aquaculture species that contributed 24.40 billion USD to the
global market in 2016 alone (FAO, 2018a).
1.3.2 Production Lifecycle of P. vannamei
There are three basic stages in the lifecycle when farming P. vannamei: i)
broodstock maturation in hatchery tanks; ii) larval-culture in nursery tanks; and iii)
growout in commercial ponds (Figure 1.2). Broodstock used for seed supply in
commercial P. vannamei production come from three main sources. Wild
populations from sea-capture (size of broodstock usually >40g at approximately 1
year in age) were used as the original source for broodstock. Historically, wild
broodstock had been popular because they were thought to produce higher quality
nauplii than those from farmed stocks, and that they were also easier to manage in
larval-culture (Browdy 1998). However, up until the late 1990s, broodstock from
wild stocks were not used extensively by the commercial seed sector due to both
biosecurity concerns and development of technology for the maturation process for
domesticating the species (Briggs, 2006; Cock et al., 2017; Cock et al., 2009).
Alternatively, small entrepreneurs in Asia and Latin America prefer maturing
cultured shrimp taken from growout ponds.
In practice, shrimp from ponds at 4-6 months of age and 15-25 g in weight used to
be transferred to maturation tanks in the hatchery. After 2-3 months of pre-
maturation rearing, these shrimp were used for nauplii production (Briggs et al.,
2004). Currently, tank-reared Specific Pathogen Free (SPF) or Specific Pathogen
Resistant (SPR) broodstock are preferred; this type of broodstock are from
genetically improved individuals that are sourced from improvement programs
(Alday‐Sanz et al., 2018; Cock et al., 2017; Moss et al., 2012a). Broodstock at 8-10
Chapter 1: Introduction 7
months of age after purchase are usually sourced from international breeding
companies and supplied to local hatcheries by air.
Female Pacific white shrimp at 8-10 months of age can be induced to reproduce
by unilateral eyestalk ablation that can trigger repeated maturation and multiple
spawning events. The period for female nauplii production lasts approximately 3
months, with 5 to 15% of females spawning per night in commercial hatcheries.
In larval-culture, nauplii develop through a number of larval stages that include
six non-feeding nauplii stages (N1-N6), three feeding zoea stages (Z1-Z3), three
mysis stages (M1-M3) following which they reach the post-larval (PL) stage.
Generally, it takes ~20 days for nauplii in culture to reach the PL10-12 stage, which
is required by farmers for growout. Feeding during larval culture stages normally
includes mixed ingredients of live food (microalgae and Artemia) and artificial
micro-encapsulated feeds (~40% crude protein content).
Depending on the pond conditions and the management practices employed,
shrimp farming types can be divided into four main categories that include: extensive,
semi-intensive, intensive and super-intensive, with each approach applying different
stocking densities. Well-managed shrimp in a pond can reach market size within
three months, and are usually harvested at ~ 90 days.
8 Chapter 1: Introduction
Figure 1.2 Schematic diagram depicting stages in the general commercial productioncycle for Penaeus vannamei. From FAO:http://www.fao.org/fishery/culturedspecies/Penaeus_vannamei/en
Chapter 1: Introduction 9
1.4 Genetic Breeding of Penaeid Shrimps
1.4.1 Basic Concepts in Genetic Breeding
1.4.1.1 Genetic Variance
The fundamental tool for production improvement of penaeid shrimp is based on
quantitative genetics. Genetic breeding is an artificial breeding strategy that exploits
the additive genetic variance (VA) that is present in a breeding line (Gjedrem, 2005).
Variance in phenotypic values (VP) however, is influenced by both effects of genetic
variance (VG), environmental effects (VE), and the interaction between these factors
(VG×E) (Falconer, 1960), such that:
VP = VG + VE + VG×E.
VG is the most useful source of variation for breeders and can be further
partitioned into three components: additive genetic variance (VA), dominance genetic
variance (VD), and epistatic genetic variance (VI):
VG = VA + VD + VI.
VA can be transferred from parents to progeny via inheritance, but while other
components are also generally considered to be inherited, their inheritance patterns
cannot be predicted. (Falconer and Mackay, 1996). Therefore, the amount of
phenotypic variance that is due to VA in the breeding line primarily determines the
response of the population to selection. Hence, the focus of selection in animal
breeding is on the statistical partitioning of phenotypic variance in the breeding line
into additive genetic and environmental components (Hickey et al., 2017b).
1.4.1.2 Heritability
Heritability (h2) is defined as the percentage of VP that is inherited in a predictable
manner. In the narrow sense, it can be described as the ratio between VA and VP
(Falconer and Mackay, 1996):
h2= VA/VP.
10 Chapter 1: Introduction
Estimates of h2 on target phenotypic traits are essential when initiating a breeding
program, to calculate breeding values for candidate broodstock individuals and to
predict the genetic response to selection (Gjedrem, 2005). Estimates of heritability
range from 0 -1 and in general, traits with h2<0.15 are considered to show low
heritability which implies that these traits are likely to be difficult to change using a
selection approach (Cassell, 2009; Tave, 1986). Traits with h2 between 0.15 and 0.4
are considered to show moderate heritability, and traits with h2 greater than 0.4 show
high heritability; moderate and high heritability traits should respond well to
selection. When the h2 for a target trait is known and the selection intensity (S) is set,
the response to selection (R) can be predicted:
R = Sh2.
1.4.1.3 Methods of Selection
Methods for genetic selection can be divided into two major types, largely
depending on whether pedigree information is available or not: individual or mass
selection (without pedigree) and family selection (with pedigree). Family-based
selection schemes can also be split into three sub-types: between family selection,
within family selection, and combined selection. To determine the most appropriate
selection method however, for a specific genetic breeding program, the available
genetic variation in a target stock should be considered in association with levels of
heritability of the desired traits to be subjected to selection. Furthermore, recording
methods, the capacity of available facilities, as well as the reproductive
characteristics of the target species should also be considered (Fjalestad, 2005).
There are both advantages and disadvantages associated with each method of
selection.
Chapter 1: Introduction 11
Individual, or mass selection is based on individual performance (i.e. individual
phenotype) (Fjalestad, 2005). This method is popular because it does not incur high
costs associated with developing a pedigree record system. It simply relies on
selecting superior individuals from a population that are grown en mass. Therefore
the overall demands for facility capacity and the budget for mass selection are
relatively economical. Mass selection however, can only target traits that can be
measured on living candidate individuals and are unlikely to be effective when traits
with low heritability are targeted. For such traits, environmental and maternal effects
usually play dominant roles in determining phenotype and, as these factors are not
heritable, this results in poor genetic gain responses. Moreover, where pedigree
information is not available, mass selection is likely to lead to rapid accumulation of
high inbreeding risks in particular in highly fecund species as is the case with most
aquatic species.
Family-based methods in general, provide a better option than mass selection
where spawning of aquatic species can be controlled, when commercial traits with
low heritability are targeted, or when traits cannot be measured on live individuals
(e.g. flesh colour or disease resistance). This approach however, requires a pedigree
record system to be maintained for each family and extensive facilities to maintain
and rear progeny from each family, separately.
The between-family selection approach requires selecting breeding candidates
based on family ranks of mean trait values; those with the best means are selected for
mating to produce progeny for the next breeding cycle. This approach requires
retention of numerous families that are communally reared in a single large holding
environment marked with an efficient pedigree record system so that the breeding
value of each family can be estimated by the Best Linear Unbiased Prediction (BLUP)
12 Chapter 1: Introduction
approach or other similar algorithms e.g. Restricted Maximum Likelihood (REML)
or Bayesian inference via Markov Chain Monte Carlo (MCMC) methods (Gianola
and Rosa, 2015). These methodologies require substantial infrastructure resources to
be maintained for an appropriate number of families. Another potential negative is
that this approach may evolve biases when unequal numbers of individuals are
present among families.
Within-family selection assumes each family to be an independent sub-population,
with the best performing individuals in each sub-population selected as candidate
broodstock. This method is often used to minimise environmental effects when
families are reared separately. If sexual dimorphism is present in the target species,
selection on individuals should be undertaken separately for each sex.
A combined selection approach optimises the approaches of family and within
family selection and estimates family breeding values from both full and half sibs in
addition to that of each breeding candidate. Best performing individuals from the
best performing families are then selected to produce progeny for the new breeding
cycle. As a consequence, this method maximizes the genetic gain so it is generally
considered to be the best selection method. Inbreeding rates however, may
potentially accelerate if the contribution from each family is not equal; this is usually
dealt with actively in well-managed programs.
1.4.2 Domestication History of Penaeid Shrimps
Before family selection was initiated in the late 1990s, there had been a long
history of domestication of penaeid shrimps. In general however, this practice was
more ‘art’ rather than ‘science’ (Alday-Sanz, 2010). Even today, most commercial
Chapter 1: Introduction 13
shrimp larviculture management practices are similar to those used in the early 1930s
(Treece and Fox, 1993).
Dr. Motosaku Fujinaga (Hudinaga) is recognised as the pioneer of domesticated
penaeid shrimps. He was the first person to close the life cycle of wild-caught gravid
female P. japonicus and to rear larvae to the sub-adult stage under lab conditions in
1934 (Hudinaga, 1942). This method was the only way to for approximately 40 years
to domesticate penaeid species and to induce spawning in captivity until the
unilateral eyestalk ablation method was developed (Treece and Fox, 1993).
Knowledge gained by Dr. Fujinaga was instrumental in the development of a
global shrimp farming industry. When Japan’s domestication technology was
transferred to the USA, an important contribution towards domesticating penaeid
shrimps occurred at the National Marine Fisheries Laboratory in Galveston, Texas
(Fast and Lester, 2013). The technique was named the ‘Galveston Method’ and is
characterised by use of high larval culture densities, application of reliable water
management practices, and feeding both live algae and Artemia that together provide
high, predictable post larval survival in farms (Fast and Lester, 2013). Around the
same time, Dr. I Chiu Liao who studied under Dr. Fujinaga modified the
domestication technique developed for P. japonicus and applied it to P. monodon
production in Taiwan. This technique is now referred to as the ‘Taiwanese Method’
that shaped development of the modern shrimp farming industry in Asia.
1.4.3 Status of Stock Improvement of Penaeid Shrimp
The first genetic improvement project directed at penaeid shrimp culture was
initiated at the Oceanic Institute in Hawaii, USA. This institute developed the SFP
(specific pathogen-free) concept and in combination with genetic family selection,
promoted and shaped the modern shrimp farming industry by providing genetically
14 Chapter 1: Introduction
improved P. vannamei broodstock with fast growth performance and high health
status (Lightner et al., 2009b). Following this development over the last four decades,
penaeid shrimp genetic improvement programs have been developed in North and
South America, Asia, Australia and New Caledonia based on both family selection
and mass selection approaches. Specifically, these projects have focused primarily on
selecting for key commercial traits that include fast growth and improved disease
resistance (Gjedrem and Robinson, 2014).
While fully domesticated stocks are now available for all commercially important
penaeid species, in recent times the main focus has been on two species: P. vannamei
and P. stylirostris (Table 1.1) and genetically improved strains now provide seed to
some culture industries (Benzie, 2009). For the other five farmed penaeid species,
most breeding improvement programs are still at the research stage or have only
reached the early stages of commercial development and do not yet provide
improved broodstock to industry (Moss et al., 2009). Most commercial breeding
programs for penaeid species are based in the Western Hemisphere (Table 1.1). This
essentially matches geographical availability of the majority of the world’s
quantitative genetics expertise. The pattern is distorted however, because shrimp
culture in Asian countries accounts for 90% of global shrimp aquaculture but only a
very limited number of improved lines are available there (FAO, 2014). Currently,
there is an urgent need in the Asia-Pacific region, to establish stock improvement
programs for penaeid shrimps to meet growing regional and world demand for this
important commodity and to address critical problems that include; low relative
productivity in culture, poor biosecurity and poor product quality in an expanding,
competitive world market.
Chapter 1: Introduction 15
Table 1.1 Existing domestication programs for penaeid shrimps across the world
Species Regions of the breeding Generations Origin of the base Ownership Marketed References
Penaeus vannamei
Hawaii, USA (several) 27 Mexico and Ecuador Private International sales 1
Florida, USA 22+ Mexico and Ecuador Private International sales
Colombia 19+ Columbia, CostaRica, Ecuador,Hawaii, Panama,Peru, Salvador andVenezuela
Private/
government
Domestic sales 2
Mexico (several) Various upto 18+
Mexico, Ecuador,Venezuela,Colombia, andFlorida
Private Domestic sales 3,4
Venezuela 26 years Mexico, Panama,and Colombia
Private Private use 5
Brazil (several) Various upto 14+
Costa Rica,Ecuador, Panamaand Venezuela
Private Domestic sales 6
China (several) Unknown USA and SouthAmerica
Private/
government
Domestic sales 7
Thailand (several) Unknown USA and SouthAmerica
Private Domestic sales 7
Penaeus monodon
Hawaii, USA 12 Indo-Pacific Private International sales 8
Madagascar 24+ South-west IndianOcean
Private No 9
Australia 15+ East and north coastof Australia
Private/
Government
Domestic sales 10
Thailand 12+ Thailand waters Government Domestic sales 9
resistance), are all traits important for shrimp growout farms. Hatcheries and
nurseries in contrast, want strains that show high fecundity, more frequent spawning
by ablated females, a high ratio of mating success, high incubation rate of eggs, and
better survival rate from nauplii to post larval stage. These traits together can be
described as reproductive traits. To date, most breeding programs for shrimp have
paid more attention to fast growth, high survival rate and disease resistance traits
(Benzie, 2009; Gjedrem and Rye, 2016), with no published studies available on
Chapter 1: Introduction 25
reproductive traits as defined breeding goals for penaeid shrimps, although
reproductive characters are essential to the viability of both hatcheries and post-
larvae nurseries (Arcos et al., 2004).
Time of first spawning after ablation shows the highest heritability among
reproductive traits, with estimates ranging from 0.41 to 0.47 (Arcos et al., 2004;
Macbeth et al., 2007). Four studies have estimated the numbers of eggs (NE) per
spawning event as a trait, with the heritability estimate quite high, suggesting
promise if a family-based selection program was implemented. h2 estimates ranged
from 0.09 to 0.41, with an average of 0.2 (Arcos et al., 2004; Caballero-Zamora et al.,
2015a; Macbeth et al., 2007). Spawn frequency is also a crucial reproductive trait. In
commercial shrimp hatcheries, a large percentage of mature females never spawn or
spawn only a few times while only a relatively small percentage of females spawn
multiple times, a critical trait that if improved, will increase total nauplii production
(Ibarra et al., 2007b; Ibarra et al., 2005). Therefore, a focus on this trait could
significantly improve nauplii production and save a large proportion of the cost of
maintaining mature females that never spawn or spawn only a limited number of
times. Ibarra et al. 2005 showed that “number of spawns” showed moderate h2 (0.20)
which could potentially underpin improvement via family selection enhancing total
nauplii production per female. Conversely, the number of nauplii produced per
female showed very little additive genetic variation and heritability for this trait
ranged only from 0.03 to 0.07 (Caballero-Zamora et al., 2015a; Macbeth et al., 2007).
Finally, h2 estimates for other reproductive traits including egg chemical composition,
egg diameter, ovary maturity etc. are also of interest (Arcos et al., 2004; Ibarra et al.,
2007b). In the future, more studies of these novel traits will be required to allow a
26 Chapter 1: Introduction
more complete and comprehensive approach to stock improvement of the target
species to be undertaken.
1.4.4.6 Additional Traits Identified for Genetic Improvement in Penaeid Shrimp
In breeding programs for aquatic species, with the exception of selecting for
growth rate, survival rate, disease resistance, and reproductive traits, a number of
other commercially important traits have been identified as targets for breeding goals,
including: meat quality, external pigmentation, special environmental tolerance,
frequency of deformity, feed conversion efficiency, and uniformity in body weight
(Gjedrem and Baranski, 2010; Gjedrem and Rye, 2016; Sae‐Lim et al., 2016).
Among these, uniformity in body harvest weight is considered to be a major target
(Khaw et al., 2016; Sae‐Lim et al., 2016). Improving levels of uniformity could
potentially reduce competition among growout animals and increase their
commercial value at harvest that would benefit both retailers and food processors
(Khaw et al., 2016).
In Nile tilapia, heritability for uniformity at harvest and correlations with harvest
weight have been estimated at 0.23 and 0.17, respectively (Khaw et al., 2016). This
implies that it should be possible to improve uniformity at harvest by family
selection and that high growth rate and better uniformity can be targeted,
simultaneously. Genetic heterogeneity in harvest weight has been demonstrated
experimentally and it has been evaluated in some livestock species but has yet to be
evaluated in shrimp (Garreau et al., 2008; Gutiérrez et al., 2006; Rönnegård et al.,
2013; Vandenplas et al., 2013). It would be interesting therefore, to determine if it is
possible to combine selection for fast growth rate with better uniformity at harvest
simultaneously in farmed shrimp species.
Chapter 1: Introduction 27
1.4.5 Genotype-by-Environment (G×E) Interactions
Ideally, improved seed developed in breeding programs are also more productive
under a range of different commercial culture environments (Sae‐Lim et al., 2016).
Relative performance of a specific animal phenotype will depend however, on both
their individual genotype, the production environment to which they are raised and
the interaction between these factors. G×E interactions are a phenomenon where the
same genotypes can produce different phenotypic responses under different
environmental conditions (Falconer et al., 1996; Lynch and Walsh, 1998; Sae‐Lim et
al., 2016). It is essential therefore to assess G×E interactions to determine if an
improved animal will perform equally well in different production environments.
Studies of G×E interactions in penaeid shrimp species have focussed on correlations
between specific growth traits and different culture environments, in particular,
effects of stocking density, location and temperature (Caballero-Zamora et al., 2015b;
Campos-Montes et al., 2009; Castillo-Juárez et al., 2007; Coman et al., 2004;
Fjalestad et al., 1997; Gitterle et al., 2005a; Ibarra and Famula, 2008a; Jerry et al.,
2006b; Li et al., 2015; Pérez‐Rostro and Ibarra, 2003a; Suarez et al., 1999; Sui et al.,
2016c).
Perez-Rostro and Ibarra (2003) reported no differences in family breeding value
ranks in P. vannamei for body weight at 200 days testing when stocks were grown at
two densities (2.5 and 4.3 individuals/m2). Gitterle et al. (2005) also did not detect
any significant G×E interactions in P. vannamei for body weight at 160 days under
seven different commercial shrimp farming environments where stock density and
salinity level varied. Similarly, Castillo-Juarez et al. (2007) found high genetic
28 Chapter 1: Introduction
correlations (0.80-0.86) in P. vannamei for body weight after 130 days at two
different locations with stocking densities of 9 and 14 individuals/m2. Campos-
Montes et al. (2009) did not observe evidence for significant G×E interactions in P.
vannamei for body weight at 130 days under three culture densities (10, 30 and 85
shrimp/m2). In contrast, genetic parameters for P. vannamei can vary significantly
under some suboptimal environmental conditions including low temperature (Li et al.,
2015) and the presence of a natural WSSV outbreak (Caballero-Zamora, Montaldo et
al., 2015).
In Asia, including in China which is the largest P. vannamei producer, extremely
diverse farming conditions exist due to production across a wide geographical range
from the tropics to temperate climates, the use of different culture densities, pond
types, management practices, etc. Significantly, farming P. vannamei in freshwater
contributes equally with marine culture in China. Evaluating growth traits for G×E
interactions between freshwater and marine culture environments will be essential to
developing a sustainable shrimp industry in Asia in the future.
1.5 Bridging the Gap between Population Genetics/Genomics and
Quantitative Genetics
1.5.1 Molecular Markers
The fundamental tool for production improvement of penaeid shrimp is based on
quantitative genetics. Though most principles have been developed over the last one
hundred years (Hill, 2014), our understanding of the nature of quantitative traits is
still rudimentary (Hill, 2010). For example, how many genes affect a specific trait
and how they interact, how their levels affect layering, their individual relationships
to overall fitness, and why it is important to maintain variation (Hill, 2010). We can
achieve significant genetic gains, but this process is largely achieved via a ‘black
Chapter 1: Introduction 29
box’ approach that essentially ignores the molecular level. Development and
application of molecular marker technologies via population genetic and genomic
approaches have assisted stock improvement programs. The primary molecular
markers of interest in genetic improvement of penaeid shrimps have been PCR-based
techniques that include; microsatellites (SSRs), single nucleotide polymorphisms
(SNPs), and mitochondrial DNA markers (mtDNA) (Alfaro-Montoya et al., 2018;
Guppy et al., 2018). While application of these markers do not require highly trained
skilled staff, costly lab facilities and sophisticated data analysis and interpretation,
they have been applied successfully to characterise genetic diversity, and population
structure of wild genetic resources of P. vannamei (Mendoza-Cano et al., 2013;
Valles-Jimenez et al., 2004), P. monodon (Abdul‐Aziz et al., 2015; Alam et al., 2016;
Benzie et al., 2002; Walther et al., 2011; Waqairatu et al., 2012; You et al., 2008), P.
japonicus (Tsoi et al., 2005; Tsoi et al., 2007; Tsoi et al., 2014), P. merguiensis
(Wanna et al., 2004), P. indicus (Alam et al., 2014; De Croos and Pálsson, 2010),
and P. semisulcatus (Alam et al., 2017). In addition, molecular marker technologies
have also facilitated broodstock management of captive penaeid stocks (Goyard et al.,
2003; Knibb et al., 2014; Maggioni et al., 2013; Perez-Enriquez et al., 2009). The
addition of genotypic information and knowledge of levels of genetic diversity can
be used to assess stock differentiation, and genetic relatedness, so that the markers
can be implemented to broaden genetic variation in base lines for genetic
improvement programs using domesticated strains (Ren et al., 2018). Compared
however, with research in the poultry industry and on other domesticated terrestrial
farm animal strains, application of molecular markers to improving production of
peneaid shrimps still lags far behind. We still do not have public widely available
marker panels, therefore results achieved in different labs or in different countries
30 Chapter 1: Introduction
cannot be compared or shared easily. Moreover, for farmed penaeids at least,
knowledge about genetic variation generated from neutral molecular markers to date,
cannot be linked directly to genetic variation in quantitative traits using a
conventional selective breeding approach. However, pedigree information based on
molecular markers (SSRs or SNPs) via parentage assignment have recently offered
new ways to exploit genetic variation of quantitative traits in farmed shrimp.
1.5.2 Parentage Assignment
Parentage assignment is based on SSR or SNP markers for family pedigree
identification systems (Jones et al., 2010). Basically, there are two computational
methods for pedigree reconstruction: exclusion methods and likelihood methods
(Vandeputte et al., 2011). The exclusion approach makes no hypotheses other than
Mendelian segregation of alleles, is very simple and has been applied effectively in
genetic improvement of aquaculture species via software programs such as
PROBMAX (Danzmann, 1997), VITASSIGN (Vandeputte et al., 2006), and FAT
(Taggart, 2007). In contrast, likelihood methods are based on probability of allele
frequencies, and likelihood software programs are now available for aquaculture
species and include CERVUS (Kalinowski et al., 2007), PAPA (Duchesne et al.,
2002), and PARENTE (Cercueil et al., 2002). Generally, 8-15 microsatellite markers
can result in an assignment power >99% in aquatic species involving a few tens or
hundreds of parents (Vandeputte and Haffray, 2014). When SNPs are used, it is
estimated that ~6 SNPs can achieve the same assignment power equal to a single
microsatellite (Glaubitz et al., 2003). In experimental research, parentage assignment
panels have been developed for genetic improvement projects in P. vannamei (Harris
et al., 2016; Perez-Enriquez and Max-Aguilar, 2016), P. monodon (Jerry et al., 2006a;
Zhu et al., 2017), P. chinensis (Dong et al., 2006), P. japonicus (Jerry et al., 2004),
Chapter 1: Introduction 31
and P. merguiensis (Nguyen et al., 2014). Additionally, the potential of parentage
assignment has been demonstrated in a real commercial breeding program for P.
vannamei (Nolasco-Alzaga et al., 2018).
Implementation of parentage assignment in conventional selective breeding
programs can have many benefits for production improvement of penaeid shrimps.
The most important benefit is that it can allow more accurate estimation of genetic
parameters from communally reared families at very early developmental stages
which decreases common environmental effects for estimates of genetic parameters
(Vandeputte and Haffray, 2014). Moreover, it provides an efficient tool for
controlling inbreeding that allows the breeder to exploit a higher selection pressure in
commercial programs, while still controlling inbreeding effectively (Perez-Enriquez
and Max-Aguilar, 2016; Vandeputte and Haffray, 2014). It can also significantly
decrease infrastructure demands in breeding programs including the need for
maintaining families in separate tanks (Yue and Xia, 2014). Future directions for
study of parentage assignment in penaeid shrimps will however, require researchers
to establish widely accepted high quality public panels for target species that enable
comparisons and sharing of results among different labs and countries.
1.5.3 Quantitative Trait Loci (QTL) Mapping
The primary drivers behind integrating population genetics/genomics into
selective breeding is to understand the links between genetic variation and
phenotypic variation in economic traits (Abdelrahman et al., 2017). QTL mapping
using molecular markers can be used to detect the effects of individual functional
genes on a trait and to assist in selective breeding (MAS) to improve commercial
traits (Naish and Hard, 2008). Due to the limited number of markers used and the
relatively small numbers of individuals used in early QTL studies, this approach had
32 Chapter 1: Introduction
only limited power to accurately isolate QTL effects (Alcivar-Warren et al., 2007;
Du et al., 2010; Wang et al., 2012). Early findings about the inferred roles of specific
genes and genomic regions in these studies still require future validation
(Andriantahina et al., 2013a; Li et al., 2006b; Lyons et al., 2007). While recent QTL
studies of penaeid shrimp stocks have been remarkably improved in terms of marker
density (3, 959 – 4, 626 SNPs), it remains difficult to integrate identified QTLs (Lu
et al., 2016; Robinson et al., 2014; Yu et al., 2015) into industrial-scale genetic
improvement programs. Even after an extensive review of QTL studies conducted on
aquaculture species (more than 40 species examined), only two case studies were
identified that had applied MAS in a genetic improvement program effectively (Yue,
2014). Over the longer term, genomic selection (GS) which applies large numbers of
SNP markers in a panel is likely to be a better option than approaches used in the
past (Castillo-Juarez et al., 2015).
1.5.4 Genomic Selection
Genomic selection (GS) is a relatively new approach for selective improvement of
quantitative traits and is based on high density molecular marker panels distributed
over the whole genome. This method integrates marker data with phenotypic and
pedigree data that can improve selection outcomes compared with conventional
selective breeding. In this context, the link between population genetics/genomics
and quantitative genetics is tighter than ever before with respect to improving
domesticated animals (Hill, 2014). A landmark article on GS was published in 2001
(Meuwissen et al., 2001) that provided the first statistical model directed at using
high-density genomic data to increase accuracy of selection. Subsequent modelling
published on GS has had a large effect accelerating progress in this field (Schaeffer,
2006), and this approach has been applied rapidly to commercial improvement of
Chapter 1: Introduction 33
livestock species (Hayes et al., 2009). In Dairy Science, GS has almost replaced the
conventional breeding approach of progeny testing and produced significant
increases in rate of genetic gains achieved while shortening the generation interval
between selection events (Hickey et al., 2017a). The transition to GS for stock
improvement has enabled an almost doubling of genetic gains in milk yield in the
USA (Wiggans et al., 2017). Application of GS to genetic improvement of penaeid
shrimps however, lags behind developments in terrestrial livestock research, due to
both the shortage of genome information on aquatic species and the high cost of
genotype sequencing. The prerequisite for successful application of GS to a new
target species is the power density (>10k) of commercial SNPs chips that are
available. Most commercial genotyping arrays used for GS in livestock species
contain 50 k ~ 500 k SNP markers (Kranis et al., 2013; Matukumalli et al., 2009;
Wiggans et al., 2017). In penaeid shrimps, there have been two commercial SNPs
arrays developed, one for P. monodon (Baranski et al., 2014) and another for P.
vannamei (Jones et al., 2017), but both are less than 10 k in size and hence may not
be adequate for GS in a commercial project. Moreover, individual genotype
sequencing costs remain expensive, and this alone may restrict future application of
GS to genetic improvement of penaeid shrimps. Of equal importance is availability
of a high quality draft genome assembly that is fundamental for the development of
high density SNP arrays. The first draft of a penaeid shrimp genome for P. vannamei
has only just been released publicly (Zhang et al., 2019).
1.5.5 Whole Genome Sequencing of Aquatic Species
Information on whole genome architecture provides a fundamental tool for
assisting improvement of aquatic species in breeding programs. In particular, this
cutting-edge genetic technology can be applied in aquaculture to; develop a reference
34 Chapter 1: Introduction
genome, identify SNPs for genome-wide association studies (GWAS) and to apply
genomic selection (GS), to characterise the roles of functional genes linked to
commercially important traits, and to address specific biological questions (Guppy et
al., 2018; Yue and Wang, 2017). Since the emergence of NGS technology in 2005,
aquaculture genome sequencing projects have expanded rapidly in particular, in the
USA, China and the EU but also elsewhere. Different sequencing strategies and
different assembly pipelines have been used already to sequence the complete
genomes of a small number of farmed aquatic species. While a cost-effective fosmid-
pooling strategy was used for genome sequencing in the Pacific oyster (Crassostrea
gigas) (Zhang et al., 2012), a whole-genome shotgun strategy was used for whole
genome sequencing of common carp (Cyprinus carpio) (Xu et al., 2014). Currently,
short-read sequencing in combination with extremely long fragment sequencing (e.g.
SMRT sequencing) is the most widely-used method applied in aquaculture genome
projects (Ao et al., 2015; Roberts et al., 2013; Vij et al., 2016). To date, whole
genome sequences from 24 farmed aquatic species have been published (Yue and
Wang, 2017). For shrimp species however, the first assembled genome sequence for
P. vannamei was published recently based on a sequence assembly from a single
adult male and covered only ~1.66 Gb of the whole genome (Zhang et al., 2019).
Incomplete sequences are an issue with shrimp genomes because they are in general,
comparatively large and often contain high numbers of repeat sequence regions that
make genome sequencing problematic (Guppy et al., 2018; Huang et al., 2011).
Since accuracy problems with SMRT sequencing (PacBio) have now largely been
resolved, this major issue is likely to be addressed in the near future. A high quality
draft genome for P. vannamei would deliver many benefits to the shrimp farming
industry. It will not only unleash the genetic potential for new breed improvement
Chapter 1: Introduction 35
projects, but will also allow a better understanding of key functional traits to be
developed, including for osmoregulation and sex determination traits, while
improving our understanding of the genetic architecture of the complex shrimp
immune defence system. This is a key issue related to addressing the major disease
problems common in shrimp farming, worldwide.
1.6 Issues with P. vannamei Broodstock Quality in China
1.6.1 Limitation of the Imported SPF Broodstock
While the Asia-Pacific is the dominant region for shrimp production worldwide
and currently contributes more than 90% to total world production (FAO, 2016a; b),
almost all of the genetically improved broodstock in this region come from SPF lines
imported from the Western Hemisphere. Annual demand for SPF broodstock is
approximately 1 million P. vannamei pairs (Gitterle and Diener, 2014), of which,
nearly 400,000 pairs were marketed to China. While imported SPF seed show
improved growth performance, they also have low survival rates and high disease
susceptibility both serious problems that have developed with use of SPF seed in
Asia (Cock et al., 2017; Cock et al., 2009). Most SPF stocks purchased from the
USA are susceptible to many novel Asian pathogens and diseases, and this problem
has also become an issue in the South American shrimp industry (Cock et al., 2017;
Moss et al., 2012b). To develop SPF stock, larvae, juveniles, and pre-mature adults
are cultured under bio-secure conditions (Briggs, 2005; Lightner, 2003a) and genetic
parameters for selecting traits in experimental environments are likely to be very
different in regions where little attention is paid to biosecurity, in particular, in the
shrimp farming industry in the Asia-Pacific region. This would be particularly true
for traits related to pond survival and disease resistance that are significantly
36 Chapter 1: Introduction
impacted by G×E interactions (Cock et al., 2017; Li et al., 2015; Sae‐Lim et al.,
2016). This implies that in general, current SPF breeding goals are unlikely to be
appropriate for non-biosecure shrimp culture environments (Cock et al., 2017; Cock
et al., 2009). The main reason why classical SPF breeding does not test breeding data
based on real commercial pond systems is because of limitations on pedigree
recording methods that require elastomer (VIE) tagging or genetic markers (SNPs or
SSR) to be used for parentage assignment, and these technologies are both expensive
and labour intensive. In the future, new approaches to family selection in penaeid
shrimp that test full and half sibs under commercial production conditions may
address poor survival of SPF lines in many regional industries.
1.6.2 Inbreeding
There are several reasons why in general, inbreeding levels in P. vannamei
broodstock are relatively high in Asian shrimp culture. Major reasons include;
shortage of supply of genetically improved SPF broodstock resulting in small
entrepreneurs using second generation parents sourced directly from their own or
local culture ponds without applying any controls on inbreeding level (Doyle, 2016;
Thitamadee et al., 2016). For instance, broodstock demand in China is approximately
2.5-3 million pairs annually but imported SPF parents contribute only 400,000 pairs
(pers. comm.). This problem also affects Thailand, Ecuador and a number of other
developing countries (FAO, 2016b). Inbreeding problems are also impacted by the
actions of SPF line breeders because in order to protect their investment, SPF
suppliers have a tendency to export broodstock from only a single line and
unauthorized breeders in developing countries then use stock from the next
generation as parents. This practice can result in high inbreeding rates with
consequential potential for inbreeding depression to develop over time (Doyle, 2016;
Chapter 1: Introduction 37
Moss et al., 2007). But more importantly, specific reproductive traits, the high
fecundity of penaeid shrimp and a small proportion of females that spawn multiple
times all contribute to the majority of progeny produced (Ibarra et al., 2007b; Ibarra
et al., 2005). Together these factors can significantly increase inbreeding risk in a
captive (closed) population. The only real solution to this problem will be for
regional shrimp producers to initiate their own local breeding programs and provide
better support for both local hatcheries and shrimp farms.
1.6.3 Issues of Base Population
Decisions about different approaches for producing the best base population can
be critical for success of any stock improvement program (Fernandez et al., 2014;
Holtsmark et al., 2008a; b). Essentially, long-term success of a genetic breeding
program will be contingent on how much broad additive genetic variance is captured
in the foundation base population (Loughnan et al., 2016). A fundamental obstacle
for initiating a genetic breeding program for P. vannamei in Asia is how to collect an
appropriate level of genetic variation in the base population because Pacific white
shrimp is an exotic species to this region (FAO, 2016b).
To date, there have been very few studies that have applied genetic theory and
molecular data directly to capture broad genetic variation in founding broodstock
populations of target aquatic species (Hayes et al., 2006). In fact, the majority of
breeding programs for aquatic species have essentially built founding populations
based largely on the stock they had available at the time. A simulation study has
suggested that assuming individuals for a founding population are genetically
unrelated, and that similar levels of genetic variation are present among
subpopulations, optimal levels of genetic variation can be achieved by choosing four
or more discrete subpopulations and cross mating them to produce a diverse base
38 Chapter 1: Introduction
population (Holtsmark et al., 2006; Holtsmark et al., 2008a; b). A modelling case
study also showed that computer simulations that combine genome-wide DNA
marker information and phenotypic values from selected broodstock can maximize
genetic variation (6% higher) than sampling equal numbers from each contributing
strain (Fernandez et al., 2014). No published studies of real aquaculture breeding
programs up to now have applied or tested the above ideas.
Neutral molecular markers (e.g. microsatellites) can provide information about
relative levels of genetic diversity, population structure, stock relatedness and
kinship in both cost and time effective ways. This technology has been applied
successfully in stock conservation and management of breeding programs in a
number of terrestrial domesticated farm animals (Carvalho et al., 2015; Revidatti et
al., 2014; Wilkinson et al., 2011; Wilkinson et al., 2012). For aquatic farm species
however, to date there are only a relatively few examples where genotype
information on domesticated strains has been included in genetic improvement
programs (FAO, 2011).
1.7 Aims of the Current Project
Pacific white shrimp have become the pre-eminent farmed aquatic species and a
major food commodity in terms of trade value in world aquaculture. As this change
has occurred, China at the same time has become the world’s largest farmed shrimp
producer. Sustainability of the farmed shrimp industry in China however, faces
significant challenges; in particular, pond survival rates of farm strains are generally
very low. This problem offers an opportunity for industries to design better breeding
programs and to develop locally adapted strains that show high survival and
improved growth rates that will generate greater profits while targeting the specific
farm and market conditions present in China. Thus, the main objective of the current
Chapter 1: Introduction 39
project was to develop a high performing locally adapted culture strain for the shrimp
farming industry in China.
1.8 Objectives and Thesis Outline
1.8.1 Chapter (1): General Introduction
Chapter (1) is an introduction that illustrates the background and rationale for the
current project via a literature review approach. It also outlines the gaps in
knowledge that the project seeks to address.
1.8.2 Chapter (2): Characterization for the Culture Resources of Pacific White
Shrimp for Genetic Diversity, Genetic Structure, and Genetic Relatedness
Chapter (2) investigates the origins and relative levels of genetic variation in
domesticated strains in China, developing an argument for development of a broad
synthetic base population for a genetic improvement project there. This chapter
forms an article published recently in the journal Aquaculture (Aquaculture, (2018),
491, 221-231). Material presented here provides the most comprehensive analysis of
the status of genetic diversity levels and inferred genetic origins of farmed Pacific
white shrimp broodstock in China. In addition, the section combines information on
historical origins of Pacific white shrimp in China and extent of genetic
differentiation among 36 breeding lines sourced there for this study.
1.8.3 Chapter (3): Genetic Parameters for Body Weight and Survival in the Base
Population
Chapter (3) examines how much additive genetic variation is available for
production of a base population via applying genotypic information. Quantitative
data on 89 families were used in the analysis. This chapter is the first report of a
40 Chapter 1: Introduction
penaeid shrimp breeding program that employed domesticated strains based on an
experimental design that applies a “genotype approach” to initiate a genetic selection
program. This chapter forms an article currently ‘under review’ in the journal -
Aquaculture.
1.8.4 Chapter (4): Comparison of Reproductive Performance of Female Pacific
White Shrimp Reared in Recirculating Tanks vs Earthen Ponds
Chapter (4) examines the two candidate production environments (RT vs EP) used
for rearing Pacific white shrimp in China and assesses in which system cultured
females show better reproductive performance. This chapter presents the first
analysis of several key reproductive traits in mature females, in a scenario that
impacts commercial nauplii production based on a natural mating design system.
Data generated in this study will be essential for improving management of Pacific
white shrimp in a genetic improvement program to optimise reproductive
performance of commercial broodstock, to improve husbandry practices, and will
assist in developing an effective strategy for future seed dissemination of improved
broodstock. This chapter forms an article currently being prepared for the journal
‘Aquaculture’.
1.8.5 Chapter (5): Quantitative Genetic Analysis of Female Reproductive Traits
Chapter (5) investigates the levels of additive genetic variation for reproductive
traits in females that are considered among the top commercial factors that impact
successful nauplii production in hatcheries in China and addresses the specific
question ‘can we improve female reproductive traits via genetic selection. This
chapter also examines genetic correlations between body weight and certain
Chapter 1: Introduction 41
reproductive traits after spawning, to answer the scientific question: ‘does selecting
for improved body weight in females produce any potentially negative effects on
broodstock reproductive quality?’ This chapter forms an article currently being
prepared for the journal ‘Aquaculture’.
1.8.6 Chapter (6): General Discussion
Chapter (6) summarises and integrates results from each experimental chapter to
address the overall objective of how in combination, they can contribute to genetic
improvement of Pacific white shrimp in China. The section also provides final
conclusions and recommendations for future research to assist meeting the overall
project goals.
43 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
Chapter 2: Levels of Genetic Diversity and Inferred Origins of
Penaeus vannamei Culture Resources in China: Implications for the
Production of a Broad Synthetic Base Population for Genetic Improvement
This Chapter has been published in Aquaculture:
Ren, S., Mather, P. B., Tang, B., & Hurwood, D. A. (2018). Levels of genetic diversity and
inferred origins of Penaeus vannamei culture resources in China: Implications for the
production of a broad synthetic base population for genetic improvement. Aquaculture, 491,
221-231.
44 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
ABSTRACT
Knowledge about the origins and relative levels of genetic variation available in
domesticated strains can be used to develop robust synthetic populations for use in stock
improvement programs, in particular, where wild stocks are unavailable or the target species
is alien. Here, we undertook a case study of Pacific white shrimp (Penaeus vannamei) where
we utilised neutral molecular genetic variation data, population structure, and genetic
relatedness as an initial step towards producing a base line for a long-term family-selection
breeding program in China. The objective was to capture and maximise genetic variation in
the base population while at the same-time controlling inbreeding levels. Genetic diversity in
1162 individuals from 36 breeding lines sourced from 22 hatcheries with domesticated
broodstock available in China were assessed using seven microsatellite loci. Genetic diversity
levels in the sampled lines were considered to be healthy, with a mean number of alleles per
locus (A) 5.85, ranging from 4.00 in B4 to 12.43 in T1, and mean allelic richness (Ar) 4.46,
ranging from 3.35 in B4 to 6.69 in T1 breeds. Mean inbreeding coefficient (Fis) was 0.07
among the 36 lines indicating acceptable levels of inbreeding. Genetic differentiation
between lines from different hatcheries was moderate, with a mean FST estimate of 0.09
among all pairs of hatchery lines. Bayesian assignment and phylogenetic analyses indicated
that hatchery lines in China could be broadly divided into four groups: i) a line from North
America (NA1), originally sourced from Mexico (north Sinaloa) and Ecuador, ii) two
representatives of the Kona line developed in Hawaii (Oceanic Institute) (NA2,3), iii) a line
developed in Thailand (Southeast Asia 1 (SA1)), and iv) a mixed group containing stocks that
originated in Latin America (predominantly from Panama and Colombia) (LA; SA2; CN1,2;
NA4,5). From these data, we argue that a complete 4×4 diallel cross (16 crosses) of the four
subpopulations should be trialled in China to develop a base line for a family selection
45 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
breeding program with optimal levels of genetic variation and a relatively low inbreeding
coefficient.
Keywords: Microsatellite, genetic diversity, population structure, prawn aquaculture,
Bayesian assignment
46 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
2.1 INTRODUCTION
Capturing high levels of genetic variation in a base population in stock improvement
programs is crucial for long-term success (Holtsmark et al., 2006; Holtsmark et al., 2008a; b).
Achieving this however, can be problematical in places where the culture species is not
native and it is difficult to source wild germplasm. High levels of genetic variation is
important because genetic gains, when selecting for specific traits, will depend on the levels
of additive genetic variance present in the base population, the selection intensity applied and
the relative level of heritability of the target traits (Falconer and Mackay, 1996; Gjedrem and
Robinson, 2014). A number of well-organized breeding programs for aquatic species have
ultimately failed because exploitable levels of genetic variation in the base population were
low (Huang and Liao, 1990; Teichert-Coddington and Smitherman, 1988). In contrast,
lessons learned in two well-documented highly successful aquatic breeding programs that
provide models for other species i.e. Atlantic salmon in Norway (Gjedrem et al., 1991) and
the Genetic Improvement of Farmed Tilapia (GIFT) (Eknath et al., 1998) in Asia included
both wild and domesticated strains when base populations were established. The importance
of combining wild populations and ‘domestic’ strains contributing to a highly diverse base
population is clear in the GIFT case where change in growth rate was impressive with an
average genetic gain of 10% per generation when compared with a control line (Hamzah et
al., 2014; Nguyen, 2016). Estimates of heritability for body weight (h2=0.28) in the latest data
(year of 2012) indicate that exploitable genetic variation remains high in GIFT strains so
genetic gains from the family selection approach should be ongoing (Nguyen, 2016). In
general, results from incorporating domestic strains in aquatic animal stock improvement
programs suggest that this approach is productive and can address shortages of high quality
broodstock in non-native finfish aquaculture. Currently, farming of exotic aquatic species
accounts for a significant component of aquaculture across the world and has contributed
47 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
significant economic and great societal benefits, in particular in the Asia-Pacific region (Silva
et al., 2009; Singh and Lakra, 2011).
There can be a number of advantages associated with using domesticated strains to
produce a base population in selective breeding programs of alien species. For example,
much of the worldwide production of farmed Pacific white shrimp (Penaeus vannamei) and
tilapia occurs in areas outside the natural geographic ranges of both species that make
availability of wild strains challenging due to geographical distance, quarantine requirements
for introductions and costs associated with sampling and transportation. Local domestic
strains, where available, can provide a much easier and lower cost option for use in stock
improvement programs. Secondly, domesticated strains have often accumulated favorable
traits for artificial culture environments over many years that make them easier to handle and
to breed compared with wild strains (Olesen et al., 2015). When developing a base population,
use of domestic or genetically improved strains, rather than wild animals as the starting
resource, can provide the new breeding line with a competitive start, in particular for growth
performance (Fernández de Alaiza García Madrigal et al., 2018).
Farmed shrimp is a major commodity in the world seafood market, with annual production
of more than 4.3 million tonnes worth more than $USD 22 billion (FAO, 2016c; Fernández
de Alaiza García Madrigal et al., 2018). Pacific white shrimp (Penaeus vannamei) that is
native to southern North America, Central America and northern South America, now plays a
substantial role in shrimp farming worldwide. In the Asia-Pacific region (where
approximately 87% of global farmed shrimp is now produced), this exotic species contributed
78% to total production in 2014, with a biomass of 3.0 million tonnes. This represents the
largest relocation of a single species in the planet’s history (Saumena, 2015; Walker and
Mohan, 2009). Pacific shrimp farming in Asia is now threatened however, by the quality and
48 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
availability of seed for the industry. Almost all genetically improved broodstock available in
this region come from Specific Pathogen Free (SPF) lines imported from the western
hemisphere; current annual demand for P. vannamei SPF broodstock is approximately 1
million pairs (Giltterle and Diener, 2014). While SPF seed show improved growth rate
performance, relatively low survival rates and high disease susceptibility are serious
problems that have developed in recent times (Cock et al., 2017; Cock et al., 2009). Most
SPF stocks are susceptible to many local pathogens and diseases (e.g. white spot disease
(WSD), early mortality syndrome (EMS)) (Thitamadee et al., 2016), and this problem has
become a significant issue for the industry, particularly for the Asia-Pacific shrimp culture
industry (Cock et al., 2017; Moss et al., 2012a). One way to address these issues is to develop
locally adapted breeding lines that can be subjected to selection for disease resistance,
improved growth rate, etc. The first step in setting up these lines is to survey available genetic
variation (GV).
Molecular markers (e.g. microsatellites) can provide effective tools for evaluating neutral
genetic diversity levels, population structure and relatedness and kinship in domestic animal
stocks. These characteristics have implications for both domesticated stock conservation and
management of breeding programs (Carvalho et al., 2015; Glowatzki‐Mullis et al., 2009;
Loughnan et al., 2016; Revidatti et al., 2014; Wilkinson et al., 2011; Wilkinson et al., 2012).
Applying information about relatedness among broodstock in artificial mating designs for
stock improvement can provide an empirically based solution to reducing inbreeding levels
when the initial generation is developed (Porta et al., 2006; Rodzen et al., 2004; Sekino et al.,
2004). This approach is highly relevant for genetic improvement in aquaculture, particularly
when alien aquatic species are targeted. This is because there can be critical issues associated
with high inbreeding levels in alien species due to difficulties associated with sourcing new
and genetically diverse broodstock. But more importantly, the high fecundity inherent in
49 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
many cultured aquatic animals means that only a few broodstock are required to produce
sufficient offspring to meet demand. The issues of inbreeding are twofold: i) inbreeding can
lead to a decrease in fitness (inbreeding depression) through a reduction in heterozygosity
and an associated increase in the expression of recessive deleterious alleles, and ii) low levels
of genetic diversity can result in a poor response to selection (Li et al., 2004; Schwartz and
Beheregaray, 2008).
Despite the important role of P. vannamei in the shrimp farming industry, studies on
genetic background of both global domesticated broodstock and wild populations are lacking
or have only been limited in scope (Maggioni et al., 2013; Mendoza-Cano et al., 2013; Perez-
Enriquez et al., 2009; Valles-Jimenez et al., 2004; Zhang et al., 2014). The aim of the current
study therefore, was to characterise genetic diversity levels and relationships among
domesticated lines of the exotic farmed penaeid (P. vannamei) in China based on
microsatellite DNA markers. The data generated will be used to inform decisions about the
best approach to establishing a founding stock to capture the highest levels of genetic
variation possible while controlling for inbreeding level from domesticated lines currently
available in country.
2.2 MATERIALS AND METHODS
2.2.1 Sampling
Twenty-two commercial P. vannamei hatcheries from three provinces in southern China
(Fujian, Guangzhou, and Hainan) were sampled to assess levels of genetic diversity,
population structure and relatedness (Table 2.1). Standards for inclusion of specific
hatcheries were as follows: (1) sampling from at least 10 hatcheries that were ranked among
the top 20 performing P. vannamei hatcheries determined by the China Aquatic Products
50 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
Processing and Marketing Association (CAPPMA) and where possible, (2) to sample
broodstock with known different origins. The 22 hatcheries are identified but referred to
anonymously as A to V (Table 2.1) to ensure that agreed confidentiality was addressed
appropriately. Provenance of culture lines used to generate the original broodstock in each
hatchery can be attributed to four geographical regions: i) North America (NA), ii) Latin
America (LA), iii) Southern Asia (SA), and iv) China (CN). If multiple breeding lines have
come from the same region, a unique number (1 to 5) was added after the name of the source
region (Table 2.1). Detailed information on line sources are provided on Table 2.1. Up to 50
individuals were sampled per line, with equal contribution by sex. In total, 1162 individuals
from 36 breeding lines from 22 hatcheries were sampled. Pleopods were taken from each
individual and the tissue preserved in 95% ethanol prior to DNA extraction and genotyping.
51 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
Table 2.1 Penaeus vannamei sample information.
Yeara Location(Province)
Hatchery PopulationID
Samplesizeb
Source
2013 Hainan A A1 50 NA1c
2013 Fujian B B1 50 NA1c
2013 Guangdong C C1 51 NA1c
2013 Hainan D D1 43 NA1c
2013 Hainan D D2 50 NA1c
2013 Guangdong E E1 50 NA1c
2014 Fujian B B2 20 NA1c
2014 Guangdong E E2 20 NA1c
2014 Hainan F F1 20 NA1c
2014 Fujian B B3 15 NA1c
2014 Fujian G G1 20 NA1c
2015 Hainan B B4 25 NA1c
2015 Hainan H H1 19 NA1c
2013 Guangdong I I1 50 NA2d
2013 Guangdong E E3 50 NA2d
2013 Guangdong J J1 50 NA2d
2014 Fujian K K1 15 NA2d
2014 Fujian L L1 15 NA2d
2014 Guangdong I I2 15 NA2d
2015 Fujian K K2 25 NA2d
2015 Hainan M M1 20 NA2d
2013 Hainan N N1 50 SA1e
2014 Hainan N N2 15 SA1e
2014 Hainan N N3 20 SA1e
2014 Hainan O O1 20 SA1e
2014 Hainan O O2 30 SA1e
2015 Hainan N N4 20 SA1e
2014 Guangdong P P1 30 NA3f
2014 Guangdong P P2 30 NA3f
2014 Fujian Q Q1 30 NA4g
2014 Fujian Q Q2 23 NA4g
2013 Hainan R R1 50 LAh
2013 Hainan S S1 50 CN1i
2013 Fujian T T1 50 CN2j
2015 Fujian U U1 50 SA2k
2015 Hainan V V1 21 NA5l
(a Year of sampling; b Number used for microsatellite markers genotyping; NA1, North America 1; NA2, North America 2; SA1, South Asia
1; NA3, North America 3; NA4, North America 4; LA, Latin America; CN1, China 1; CN2, China 2; SA2, South Asia 2; NA5, North
America 5; c Nucleus line located in USA, source of stock from Mexico and Ecuador; d Nucleus line located in USA, originated form The
Kona line; e Nucleus line located in Thailand, source of stock from USA and South America; fNucleus line located in USA, originated from
The Kona line; g Nucleus line located in USA, originated from a population in Hawaii and a SPR line; h From Colombia; i Developed in
China, source of stock from USA and South America; jDeveloped in China, source of stock from several local populations in China; k
Nucleus line located in Thailand, source of stock from two populations of USA, lNucleus line located in USA, source of stock from two
population from USA).
52 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
2.2.2 DNA Extraction and Genotyping
Total genomic DNA was extracted from pleopod samples using a Marine Animal DNA
Extraction Kit (TIANGEN BIOTECH CO. LTD.) following the manufacturers specified
protocols. Seven microsatellite loci developed for the target species: Lvan05 (Perez-Enriquez
et al., 2009), Pvan1758 (Cruz et al., 2002), and TUMXLv5.45, TUMXLv7.121, TUMXLv8.256,
TUMXLv9.103, TUMXLv10.312 (Meehan et al., 2003) were used in the genetic analysis.
Amplification volume was 15 µl with the following conditions: 1×Buffer I (Invitrogen), 0.2
mM dNTP (Invitrogen), 2.5 µM of each primer (forward primers of seven loci were labelled
fluorescently with HEX or FAM), 0.75 U Taq DNA polymerase (Invitrogen) and milliQ
water to a final volume of 15 µl. PCR cycling parameters included initial denaturing step of
95℃, then 35 cycles of 94℃ for 30 s, specific annealing temperature for 45 s, 72℃ for 1 min,
and final extension at 72℃ for 10 min. Genotyping of PCR products was undertaken on an
ABI 3730xl genetic analysis system (Applied Biosystems) and allele size was generated
using GeneMapper 4.0 software (Applied Biosystems) with a GenScan ROX-500 (Applied
Biosystems) internal size standard.
2.2.3 Data Analysis
2.2.3.1 Genetic Diversity Estimates
Number of alleles (A), expected heterozygosity (He), and observed heterozygosity (Ho)
were estimated in GENALEX 6.502 (Peakall and Smouse, 2012). Allelic richness (Ar) and
private allele richness (PAr) were calculated in HP-RARE 1.1 (Kalinowski, 2005), with
parameters set for a minimum of 14 alleles per sample (7 diploid individuals). Ar is a
weighted estimate of the number of alleles at a locus, independent of sample size, by
applying a rarefaction process. PAr is a related estimator of the number of unique or rare
alleles within a stock. Significant differences among genetic diversity measures for each
53 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
breeding line were determined using a random complete block ANOVA that recognizes the
inherent lack of independence of each microsatellite locus (block) between strains. Analyses
were performed for mean number of alleles across loci (Ā), and mean allelic richness (Ār –
the number of alleles detected at a locus standardized on the number of individuals genotyped
across all samples). PAr was not tested due to a large zero count. Where significant
differences were detected, a Tukey’s HSD posthoc test was used to identify specific
differences. Inbreeding coefficient (F) was estimated via three methods. Firstly, the widely
used Fis index of the degree of random mating within populations was estimated with
FSTAT 2.9.3.2 (Goudet, 2001) using the Weir and Cockerham method (Weir and Cockerham,
1984). A recent simulation has shown however, that Fis can underestimate real inbreeding
levels for domestic breeding animals compared with a more recent Wang’s TrioML (Ft)
method (Doyle, 2016; Wang, 2014). Thus, a likelihood method that uses genotypes from a
triad of individuals was used to estimate the inbreeding coefficient (Ft) (Wang, 2007; 2011).
Finally, a third inbreeding coefficient index (DyadML (Fd) (Milligan, 2003)) was estimated
to assess the correlation of Fis and Ft values. Deviations from Hardy-Weinberg (HWE) and
Linkage Equilibrium (LE) were assessed using GENEPOP 4.1 (Rousset, 2008), applying a
Markov Chain method with 10,000 dememorization steps followed by 20 batches (100
batches for LE) of 5000 iterations per batch. Sequential Bonferroni was used to adjust
analysis-wide tests for the significance of HWE and LE (Rice, 1989).
2.2.3.2 Population Genetic Differentiation
Genetic differences among 36 stocks from the 22 sampled hatcheries were estimated using
several methods. Firstly, pairwise FST (Weir and Cockerham, 1984), a statistic that partitions
genetic diversity within and among each stock was calculated in ARLEQUIN 3.5.2.2
(Excoffier and Lischer, 2010). Also, a pairwise matrix assessing allele frequency
54 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
heterogeneity among samples was constructed using Nei’s DA (Nei et al., 1983), calculated
with Populations 1.2.31 software (Langella, 1999). The PHYLIP program (Felsenstein, 1993)
was used to construct an unrooted neighbour-joining cladogram based on the DA matrix.
Results were entered into FigTree 1.4.3 (Rambaut, 2012) to produce a high-quality
phylogenetic tree representation. Grouping of individuals was performed using a Bayesian
clustering method executed in STRUCTURE 2.3.4 software (Pritchard et al., 2000).
Individuals were assigned to the most probable cluster out of k putative clusters with and
without the use of sample location as a prior reference (‘locprior’). Parameter settings
included admixture and correlated allele frequencies (Falush et al., 2003). For each k (1-35),
20 replicates were run, applying set parameters of 1,000,000 iterations of a Markov Chain
Monte Carlo (MCMC) process with a burn-in length of 100,000 iterations (Gilbert et al.,
2012). Threshold q-value set for stock cluster was >0.9 for a single cluster and <0.9 for the
detection of admixture. The most likely number of genetic populations (k) was estimated with
STRUCTURE HARVESTER (Earl, 2012; Evanno et al., 2005). Admixture proportions of
each stock applying 20 replicates were averaged via CLUMPP 1.1.2 for the best k (Jakobsson
and Rosenberg, 2007) and final barplots were generated with DISTRUCT 1.1 (Rosenberg,
2004).
2.2.3.3 Relatedness Estimates
Estimated relatedness (rxy) among stocks (36 stocks) and average rxy of randomised mating
between populations were estimated using the estimator, rQG (Queller and Goodnight, 1989).
rQG is a widely used index of relatedness used for kinship studies of both captive and natural
populations (Blouin, 2003). In the current study, 1139 individuals (genotyped with a
minimum of four microsatellite loci) were used to generate a relatedness estimates among
stocks. In addition, a moment estimator of pairwise relatedness from LynchRd (rLR) (Lynch
and Ritland, 1999) was calculated for comparison with the rQG method. All relatedness
55 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
estimators were generated with COANCESTRY 1.0.1.7 software (Wang, 2011). This
software incorporates seven relatedness and four inbreeding estimators, to generate the most
accurate estimates from multi-locus genotype data. Irrespective of the most appropriate
relatedness estimator, accuracy depends on the dataset in each study; i.e. the number of
microsatellite loci screened and the relative polymorphism information content (PIC) of the
markers employed (Van de Casteele et al., 2001; Wang, 2011).
2.2.3.4 Effective Population Size
Effective population size (Ne1) was estimated using a linkage mating model by assuming
monogamous mating type as reported for both P. vannamei (Harris et al., 2016) and other
closely related penaeid species: P. merguiensis (Knibb et al., 2014) and P. monodon
(Marsden et al., 2013). A linkage mating model Ne1 from 36 hatchery populations was
estimated in LDNE 1.31, applying a minimum allele frequency of 0.05 (Waples and Do,
2008). When considering the impacts of artificial mating design of populations for estimating
Ne, a molecular co-ancestry method (Ne2) was used (Nomura, 2008). Ne2was estimated with
NEESTIMATOR 2.01 that is considered to be a powerful method when there is missing data
as it screens out rare alleles, and it is designed for large data sets (Do et al., 2014).
56 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
2.3 RESULTS
2.3.1 Genetic Diversity and HWE Estimates
Mean number of alleles (Ā) per locus in the sampled stocks ranged from 4.00 in B4 to
12.43 in T1, with an overall average across all sampled 36 stocks of 5.85 (Table 2.2). A
similar ranking pattern was observed for allele richness estimates (Ār) compared with Ā, with
a highest value of 6.69 in T1 to the lowest value of 3.35 in B4, and an overall average of 4.46
(Table 2.2). Highest private allele richness (PAr) (0.48) was observed in the V1 stocks,
followed by T1 (0.39) and Q2 (0.28) and stocks from NA1, NA2 and SA1 possessed
significantly lower values of PAr except for the K2 stocks. Expected heterozygosity (He)
across all sampled loci ranged from 0.58 in B4 to 0.81 in T1, while observed heterozygosity
(Ho) ranged from 0.47 in B4 to 0.77 in T1 (Table 2.2). Mean estimates of He and Ho across
all loci and stocks were 0.69 and 0.66, respectively.
57 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
Table 2.2 Genetic diversity measures for 36 batches of P. vannamei broodstock (N=1162) from 22 hatcheries inChina based on 7 microsatellite loci: North America (NA), South Asia (SA), Latin America (LA), and China(CN). Sample size (N), average number of alleles (A), allele richness (Ar), private allele richness (PAr),observed heterozygosity (Ho), expected heterozygosity (He), inbreeding coefficient of the Weir and Cockerhammethod (Fis), inbreeding coefficient of Wang’s TrioML method (Ft), inbreeding coefficient of DyadML method(Fd), effective population size of the linkage mating model (Ne1), and effective population size of the molecularco-ancestry method (Ne2).
58 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
While the inbreeding coefficient estimate (Fis) was low at 0.07 in all stocks, this value
was 3-fold greater where the other two calculation methods were applied, with 0.21 (Ft) and
0.22 (Fd), respectively. The RCB ANOVA identified significant differences among genetic
diversity estimates for the 36 breeding lines (for A, F35, 210 = 7.448, p<0.001; Ar, F35, 210 =
3.738, p<0.001,). In general, the Tukey’s posthoc test demonstrated that breeding lines
derived from CN2, NA4 and SA2 possessed the highest diversity (see Figure 2.1 for A, Ar
(not presented) showed a similar pattern).
Figure 2.1: A bar plot showing mean number of alleles (Ā) for 36 breeding lines. Error bars
represent ±1 SD. Letters above bars indicate where significant differences among breeding
lines exist determined from Tukey’s HSD test (P < 0.05). A1-V1 are 36 breeding lines ID
from 22 hatcheries in China; and NA1-NA5, SA1, SA2, LA, CN1, CN2 are 10 group ID of
the origins history of the above 36 breeding lines (details information see Table 2.1).
59 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
Among 252 HWE tests for 7 loci among 36 stocks, only 3 tests showed significant
deviation (P<0.05) from HWE after standard Bonferroni correction. Among these three
significant deviations, there was no obvious pattern at specific loci or population. After
Bonferroni correction, 6 out 756 pairs of loci were found to deviate significantly from LE
(P<0.05), but again, no specific patterns were detected. As such, all loci were considered to
be reliable, independent estimators of neutral diversity in each of the 36 breeding lines.
2.3.2 Population Genetic Differentiation
Genetic differentiation among stocks (pairwise FST estimates) from lowest to highest,
ranged from 0.00 to 0.29. Mean FST estimates among the 36 stocks were moderate and ranged
from 0.06 (D1 and I1 stocks) to 0.14 (B4 stocks). Overall mean genetic differentiation (FST)
among stocks was 0.09. Genetic differentiation (FST) of stocks from different hatcheries that
had originated from the same geographical region historically was lower compared with the
overall FST, with a mean value of 0.05 within NA1 , 0.01 within NA2, 0.03 within SA1, 0.00
within NA3 and NA4 (Table 2.3). Among 78 pairwise FST comparisons of 13 breeding line
within NA1 group, 53 tests were statistically significant, which indicated most breeding lines
tested within the NA1 group were diverged. While all 28 pairwise FST tests within NA2 group
were not significant, 13 pairwise FST estimates among 15 tests within SA1 group were also
not significant. In addition, no significant differences were observed for pairwise FST
comparisons within NA3, or NA4 respectively. It is interesting to note that 37 of 40 pairwise
FST tests were not significant between breeding lines in China that originated from NA2, NA3
and NA4, suggesting a close genetic relationship among NA2, NA3, and NA4. Pairwise
genetic distance estimates (DA) ranged from 0.03 (I1 vs. J1) to 0.56 (B4 vs. R1). Average
pairwise genetic distance estimates ranged from 0.17 for I1 stocks to 0.41 for R1 stocks
(Table 2.3). Pairwise genetic distance estimates between hatcheries that had originated from
60 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
the same geographical region historically also were lower compared with the overall average
DA (0.23), with average DA 0.12 between NA1, 0.11 between NA2, 0.15 between SA1, 0.06
between NA3 and 0.09 between NA4.
61 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources in China: Implications for the Production of a Broad Synthetic BasePopulation for Genetic Improvement
Table 2.3 Population genetic differentiation among 36 P. vannamei stocks (ID: name of 36 breeding lines; GP: The 10 places of origin for thesebreeding lines). The pairwise genetic differentiation amongst 36 breeds estimated using FST (Weir and Cockerham, 1984) (below diagonal). Thepairwise genetic distance DA (Nei et al., 1983) between 36 breeds (above diagonal). Values in shaded areas represent comparison amongbreeding lines within same groups of origin. The last two rows on the table represent mean FST and DA values of 36 breeding lines. Values of FSTin bold are statistically significance at α=0.05 after Bonferroni correction.
62 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
A phylogenetic reconstruction of stock (Hatchery) relationships is represented in Figure
2.2 Stocks with a NA1 origin clustered to form a monophyletic clade, as did all stocks in the
SA1 group. NA2 and NA3 stocks together formed a third clade, consistent with pairwise FST
comparisons showing that NA2 and NA3 shared a close genetic relationship. Seven breeding
lines originating from NA4, NA5, LA, SA2, CN1 and CN2 loosely clustered together to form
a fourth, polyphyletic clade. However, due to a lack of multiple lines being available from the
various source populations represented in this clade, we cannot discard the possibility that
each of these lineages is genetically distinct and could potentially be readily identified
through multi-locus analysis.
63 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
Figure 2.2: An unrooted neighbour joining tree for 36 P. vannamei breeding lines based on
seven microsatellite loci using Nei’s DA genetic distance method. Each breeding line label
includes the abbreviations of origins and hatchery ID as per Table 2.1.
64 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
The most appropriate number of clusters applying the Bayesian method in STRUCTURE
was, k=2 even though k=4 was also strongly supported (see Supplementary Figure S2.1).
With k=2, all stocks that originated from NA1 clustered into a single group with stocks from
LA forming a separate cluster. The other 22 stocks showed an admixture pattern after
Bayesian clustering (Figure 2.3_a). When k=4, all stocks from NA1 and LA clustered into
two discrete genetic populations, (Figure 2.3_b) but stocks from NA2 separated to form a
separate cluster. Despite a pattern of minor admixture for some stocks from SA1, all SA1
stocks formed a single genetic cluster (Figure 2.3_b). Remaining stocks showed an
admixture pattern. Some individuals from D2 stocks showed a different Bayesian cluster
assignment within populations that suggested that they may constitute a mix of different
genetic resources, historically. Overall, the Bayesian cluster results were consistent with the
population structure analysis based on FST and the reconstructed phylogenetic tree.
Figure 2.3: Individual assignment based on Bayesian analysis of 36 breeding lines at: a)
Structure plot for K=2; b) Structure plot for K=4.
a
b
65 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
2.3.3 Relatedness Estimates
There was a high correlation between results from different relatedness estimate methods
(r=0.85). Average relatedness estimates using RQGwithin stocks ranged from 0.02 (R1 and T1)
to 0.32 (B4), with an average value of 0.16. In general, average relatedness values using RDML
were larger than for RQG, and ranged from 0.12 (T1) to 0.4 (B4) with an average estimate of
0.26. Significantly higher values were found for relatedness comparisons within groups than
between different groups (Table 2.4).
66 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources in
China: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
Table 2.4 Average relatedness estimates amongst 36 P. vannamei stocks. The pairwise relatedness estimates amongst 36 breeds calculated using
RQG (Queller and Goodnight, 1989) (above diagonal). The pairwise relatedness estimates of RDML (Milligan, 2003) between 36 breeds (below
diagonal). Tables in shadow areas were the comparison between same groups of origin. The last two rows on the table represent average RQG and
67 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
2.3.4 Effective Population Size (Ne)
Based on the linkage model, effective population size (Ne1) estimates ranged from
19.9 to 749.6, with four populations recorded as infinite. Applying the molecular co-
ancestry method (Ne2), most stocks showed estimates (Ne2) that were much lower
than for Ne1 and ranged from 5.0 to 68.6. Nine populations for Ne2 were infinite,
indicating that their Ne size was very large (Table 2.2). Infinite Ne estimates
generated with both Ne1 or Ne2 estimators may result from either mixing of stocks
from different genetic stocks in history (Hartl and Clark, 1997) or analysis of small
sample sizes per line. A parallel result has been reported for Ne estimates in captive
barramundi stocks where lines with small sample sizes were estimated to be infinite
(Loughnan et al., 2016).
2.4 DISCUSSION
2.4.1 Genetic Variation Levels Within and Among Stocks
The current project provides the most comprehensive analysis available for the
status of genetic diversity levels and inferred genetic origins of farmed P. vannamei
broodstocks in China. The results presented here can also be used as a reference and
foundation for assessing genetic resources of P. vannamei stocks in other countries in
Asia where this species is becoming the major farmed penaeid.
Overall, levels of genetic diversity in 36 farm stocks assessed here were similar
compared with another earlier, but less comprehensive study on genetic diversity in
P. vannamei stocks in China (Zhang et al., 2014). The small differences observed
68 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources in
China: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
between results obtained in the two studies most likely reflect the different panel of
microsatellite loci used to assess diversity in the stocks examined. In general, genetic
diversity levels in farm stocks of P. vannamei in China were slightly higher when
compared with stocks farmed in Mexico (Perez-Enriquez et al., 2009; Vela Avitúa et
al., 2013), consistent with the fact that genetic resources in Mexican hatcheries are
believed to have come from a single origin (i.e. the ‘Los Malago’ breeding line that
forms the basis of the industry there (Perez-Enriquez et al., 2009)). In contrast in
China, farm stocks have been developed from a number of different sources that
originated from different geographical locations (both from the native and introduced
ranges) before they were introduced to China.
Among the breeding lines examined here, estimates of genetic diversity in stocks
that originated from NA1, NA2 and SA1 sources in general, were lower compared
with others (Table 2.2). Potentially, this difference may reflect the longer time span
they have spent managed as closed breeding lines (Benzie, 2009) or differences in
relative levels of genetic variation present in the initial founders in each breeding
program. It is interesting to note that genetic diversity levels in the CN2 stocks were
much higher than in other Chinese stocks tested here (Table 2.2). Broodstock
management practices and strategies employed in the CN2 hatchery are different
from many others in China. First, the number of broodstock maintained in the CN2
hatchery is significantly higher than in other hatcheries, with approximately 15,000
to 25,000 pairs maintained each year for PL production. Assuming that mating
success approaches 10% per night during the nauplii supply season, this implies that
1,500 to 2,500 pairs likely contribute to each larval batch. Assuming that breeders
are unrelated and contributions from pairs are approximately equal, this number of
breeders can maintain a high effective population size and avoid genetic erosion of
69 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
the population due to genetic drift effects. Each year the CN2 hatchery also obtains a
small proportion of new stock from other hatcheries and thoroughly mixes their
diversity into their own stocks, introducing new alleles to their breeding program.
Inbreeding levels (Fis) were similar to those reported in an earlier study of P.
vannamei in China (Zhang et al., 2014). These estimates suggest that most
broodstocks currently used at hatcheries in China likely apply effective inbreeding
management and follow the recommendations proposed by (Moss et al., 2007) who
stated that inbreeding levels should not exceed 10%. Chinese commercial
broodstocks show much lower inbreeding levels compared with those reported in
farmed P. vannamei stocks in Mexico (Fis 0.25) (Perez-Enriquez et al., 2009) and
Brazil (Fis 0.30) (Maggioni et al., 2013). It should be noted however, that while high
homozygosity in two studies mentioned above may reflect significant inbreeding,
deviations from HWE more likely suggests that homozygote excess was a function
of null alleles (a common problem with microsatellite studies) and that inbreeding in
these studies may have been overestimated. Inbreeding coefficient Fis estimates in a
number of the Chinese stocks tested here however, were much higher than that
reported in the stocks imported originally. A possible explanation for this
observation is that while they probably received a representative gene pool initially,
poor management of broodstock genetic resources subsequently over time may
account for observed high Fis estimates in these stocks. Notably, Ft (mean Ft = 0.21)
and Fd (mean Fd = 0.22) estimates for inbreeding level were three times higher than
the Fis estimates (mean Fis= 0.07). Doyle (2016) argued that Fis can be a poor
indicator of inbreeding level in stocks. In our study, Fd and Ft estimates were much
higher than for the more widely used Fis estimates, a difference that could suggest a
70 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources in
China: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
more serious inbreeding issue in Chinese stocks. We must be cautious however, in
concluding that stocks are inbred based solely on Fd or Ft estimates, as maximum
likelihood methods can overestimate the real value, particularly for stocks with low
relationship estimates (Milligan, 2003). The essential point about estimating
inbreeding level is that it will always be a relative estimate and not an absolute value,
so there needs to be some comparison point rather than a discussion based solely on
the inbreeding estimation method that is employed. For example, in order to
acknowledge any real decline in genetic variation in captive populations of P.
vannamei, ideally we would need a comparison point with a reference wild
population (Vela Avitúa et al., 2013). While this was not possible in the current study,
for the purpose of genetic monitoring of broodstocks of P. vannamei in closed
rearing systems in China, we can obtain a relative assessment of the success or
otherwise of management practices employed to manage genetic diversity over time
by sampling successive broodstock cohorts over years in the same hatchery using the
same marker loci (De Lima et al., 2008). Based on Fis comparisons with that in
Mexico (Perez-Enriquez et al., 2009) and Brazil (De Lima et al., 2008), there is an
indication that management of genetic diversity levels in China may have been better
than in hatcheries elsewhere.
2.4.2 Population Differentiation and Origins of Genetic Resources
The estimate of mean level of genetic differentiation (FST) among Chinese P.
vannamei culture stocks was 0.09 (Table 2.3) indicating a moderate degree of allelic
heterogeneity among stocks. This result is consistent with those reported in previous
studies in China (FST, 0.08) (Zhang et al., 2014) and Brazil (FST, 0.06) (Maggioni et
al., 2013), but were higher than estimates reported for stocks in Mexico (FST, 0.02)
(Perez-Enriquez et al., 2009). Differences reflect that Mexican P. vannamei culture
71 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
stocks have had a common ancestry, while stocks in Brazil and China most likely
have resulted from mixing of multiple lineages following the importation of exotic
germplasm because no native wild stocks were available in China.
Genetic differentiation estimates among the 36 Chinese stocks showed pairwise
FST estimates ranging from 0.00 to 0.29 (Table 2.3). Stocks with the same inferred
wild origin as expected showed the lowest pairwise FST estimates, (average of 0.05
among NA1, 0.01 among NA2, 0.03 among SA1, 0.00 between NA3 and 0.00
between NA4). Most pairwise FST tests within NA2 or SA1 groups (Table 2.3,
shadowed areas) were low and not significant, which indicates that breeding lines
from different hatcheries in China with the same origins (either from NA2 or SA1)
show limited genetic differentiation. In contrast, breeding lines in China with NA1
origins show moderate levels of genetic heterogeneity. This potentially results from
several independent nucleus breeding centres utilising NA1 stocks. Of interest,
pairwise FST estimates between NA2, NA3 and NA4 were low, indicating a common
ancestry for these groups. When the founding information for NA2 and NA3 was
traced back, we established that both had been developed independently from the
Kona line in Hawaii. Population NA4 was sourced from a Hawaian population but
information was not available as to whether this was from the Kona line or other
lines maintained in Hawaii. While results of the STRUCTURE analysis here provide
little confidence for a significant contribution from the Kona line in NA4, FST
analysis shows relatively low levels of differentiation among NA2, 3 and 4. This
suggests that while NA2 and NA3 had a common Kona line ancestry, NA4 resulted
from mixing of Kona line genes with those from other lines.
72 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources in
China: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
In the present study, the unrooted neighbour-joining dendrogram is consistent
with the historical records of P. vannamei culture stocks origins in China (Figure
2.2). Four main clades were distinguished: i) NA1; ii) NA2 and NA3; iii) SA1; and
iv) a clade representing all other stocks. It is reasonable to suggest that NA2 and
NA3 grouped together in a single clade because they were both derived from a
common founding population (the Kona line developed in Hawaii). China has been
the major P. vannamei producer for decades, and there have been many records of
importations since 1987 (Li et al., 2006a). When we combine our results from the
Bayesian cluster analysis with those from a review paper of global exchange of
penaeid shrimps by Benzie (2009) and information from shrimp companies and
research institutes in China regarding the origins of their founding populations, we
can synthesize a more detailed account of founding genetic resources for each group.
Founding populations of NA1 came from Mexico (north Sinaloa) and from
Ecuador, while the NA2 and NA3 lines originated from genetic resources sourced
from the Kona line in Hawaii. A recent study on seven culture populations of P.
vannamei in China (Zhang et al., 2014) indicated the two clusters in their
phylogenetic reconstruction of genetic relationships align with our clades i and ii.
Origins of SA1 are more obscure, but potentially may have had contributions from
stocks in the USA and South America. According to the records, the LA stock was
developed from multiple founding populations that were sourced from wild
populations in South America. The official records for the founding population for
CN1 suggest that they are the result of the mixing of several cultured genetic lines
from the USA and a wild population from Ecuador. SA2 was developed from
germplasm sourced from Hawaii and a Kentucky culture population, confirmed by
its mixed ancestry in the STRUCTURE analysis. The founding population for NA5
73 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
came from Texas but originated from South America. CN2, the stock with the
highest diversity, was developed from a mixture of Chinese culture populations that
includes genetic resources possibly from many stocks introduced to China, but
predominantly displaying a Latin American ancestry in the STRUCTURE analysis.
While no official information was available about the origins of the NA4 population,
the STRUCTURE analysis suggests that it likely had a base population developed
from several lines from the USA and South America. The only available information
about the origin of the NA4 line was that it was constructed from a combination of
Hawaiian resources and a special specific pathogen resistant (SPR) line. According
to a history of genetic improvement of P. vannamei culture stocks (Cock et al., 2009),
SPR resources originated from survivors of stocks in commercial ponds in South
America sourced from Ecuador and Colombia in the mid-1990s when Taura
syndrome virus (TSV) was prevalent in the shrimp industry. The broodstock
generated from the survivors apparently possessed useful genetic variation for TSV
resistance and this was applied in a family selection program in Hawaii.
In our study, all of the stocks that have South American origins (i.e. NA4, LA,
CN2, SA2, & NA5), either as a uniform or admixed population following Bayesian
Given the very limited information available on the genetic background of virtually
all stocks sampled here and the broad geographical distribution of natural
populations of the target species, the markers appear to have traced the ancestry of
the lines quite successfully. Of interest, assignment of individuals from the D2 stock
(for either k = 2 or 4), suggested that it consisted of two distinct breeding lines that
have been kept isolated from each other (i.e. there is no pattern of admixture between
74 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources in
China: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
groups). While all individuals sampled for this study were collected at the same time
from breeding tanks from this hatchery, broodstock used for nauplii production can
however, be between 12 to 16 weeks old after eyestalk ablation and sometimes
hatchery managers keep healthy females from previous hatchery runs and mix them
with broodstock that are younger. Theoretically, a pattern of two different lines
without admixture could occur if the sets of genotypes in the two age groups differ
significantly by chance. In the current situation however, there are records of
importation of two different lines that clearly assign to the NA1 and LA groups. It is
obvious that broodstock from these two lines have been kept separate within the
hatchery.
2.4.3 Implications for Forming a Genetic Foundation (Base) Population for
Genetic Improvement
Our results provide a strong foundation for making decisions about how to form a
synthetic base population of an exotic aquaculture species, P. vannamei with high
levels of genetic variation for a stock improvement program in China. Assuming that
a classical complete diallel cross (Gjedrem and Robinson, 2014; Hung et al., 2014;
Nguyen, 2016) will be conducted to establish the base population, the first question
will be how to identify discrete subpopulations of the available stocks. Results from
the genetic differentiation analyses (pairwise FST, Da, phylogenetic tree and Bayesian
clustering) of the available domesticated stocks of P. vannamei indicate four discrete
subpopulations, NA1, Kona line, SA1 and Latin America (LA) closed group. While
at present there is no available information on additive genetic variance of
commercially important traits for these groups in China, such an experiment tested
that the Kona line is known to have been selected to improve growth performance
but performed poorly for resistance against WSSV (Cuéllar-Anjel et al., 2012). All of
75 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources inChina: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
the four subpopulations described here however, have been used for shrimp farming
in China, widely. We suggest that the best approach will be to undertake a complete
4×4 diallel cross with equal offspring contributed from each of the four
subpopulations. This approach will capture not only high levels of genetic variation,
as the stocks are successful domesticated lines they will also carry copies of different
QTL alleles that have been selected indirectly, potentially also including alleles
influencing growth and survival traits (Sae‐Lim et al., 2016) as an outcome of past
domestication and farming processes. Stochastic simulation has suggested that using
four subpopulations to form the base population will maximise genetic gains in
future generations of selection while saving sampling costs associated with using a
greater number of lines in the base population (Holtsmark et al., 2006).
Relatedness estimates generated in the current study can provide a guideline for
avoiding inbreeding risk across the proposed generations of selection. Based on RQG
relatedness index results (Table 2.4), the proposed diallel cross mating design among
stocks should theoretically maintain RQG near zero because none of the four
subpopulations identified for inclusion in the base population have shared a recent
common ancestry as a result of many generations of isolation during their
domestication process. While initially there had been potential for higher rates of
inbreeding for mating pairs within subpopulations, this can be addressed by sourcing
stocks of the same groups from different hatcheries with high genetic diversity levels
for mating. By incorporating relatedness information, and clear pedigree records, any
risk of significant inbreeding potentially leading to inbreeding depression and any
negative impacts of genetic drift can be minimised.
76 Chapter 2: Levels of Genetic Diversity and Inferred Origins of Penaeus vannamei Culture Resources in
China: Implications for the Production of a Broad Synthetic Base Population for Genetic Improvement
2.5 CONCLUSIONS
Here we have documented genetic diversity levels, genetic differentiation among,
and relatedness patterns of, P. vannamei culture lines currently available in China.
Our results confirm that modern P. vannamei hatchery lines have come from
multiple genetic sources and the genetic differentiation patterns described in hatchery
lines in China are consistent with the historical records of introductions. The data
described here provide foundation information for both hatchery stock management
and for developing a base population with high levels of genetic diversity for a future
breed improvement program.
77 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
s
Chapter 3: Genetic parameters for growth and survival traits in
a base population of Pacific white shrimp (Penaeus vannamei)
developed from domesticated strains in China
ABSTRACT
Historically, breeding programs directed at genetic improvement of penaeid
shrimp farm lines have differed remarkably in the eastern and western hemispheres
with respect to the source of their base stock. In the East, broodstocks were
commonly sourced from wild populations, while the majority of culture industries in
the West are based on domesticated strains with genotype/pedigree information
applied where available. While the majority of genetically improved broodstock used
in shrimp farming around the world have been supplied from these programs in the
West, it is essential to consider how much genetic variation correlated with important
commercial phenotypic traits was captured when the breeding lines were developed
from the available domesticated strains because this provides the fundamental
resource on which the program will depend long-term. In an earlier study, we applied
a complete diallel cross approach to produce a base population for genetic selection
from a number of domesticated P. vannamei strains in China based on their relative
levels of microsatellite variation. Here, we assess quantitative data on growth traits in
families generated in the full 4 x 4 diallel cross. In total, body weight was measured
at two ages (BW1 and BW2) from 2,752 and 2,452 individuals respectively from 89
full-sib families and analysed using a univariate animal model following REML
methodology. Estimated heritabilities (h2 ± SE) for BW1, BW2 and survival (S) were
0.52 ± 0.09, 0.44 ± 0.07, and 0.01 ± 0.02, respectively. The genetic correlation
78 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
between growth traits (BW1 and BW2) was 0.95 ± 0.03, a result significantly
different from zero. Genetic correlations between survival and body weight were low
however, 0.26 (S vs BW1) and 0.18 (S vs BW2) respectively, and not significantly
different from zero. High heritability estimates for growth traits confirm that a
substantial component of additive genetic variance is available for growth in our P.
vannamei culture line families in China prior to a family selection breeding program
to improve relative productivity.
Keywords: Penaeus vannamei, Strain, Heritability, Genetic parameters, Body weight
79 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
3.1 INTRODUCTION
Opinions on how to source foundation stocks for genetic improvement of penaeid
shrimp culture lines differ remarkably in the ‘West’ and the ‘East’ (Boyd et al.,
2006). The majority of penaeid genetic improvement programs in the West have
developed breeding lines that have been based on available farm strains of
domesticated shrimp often guided by genotypic information where it was available.
In contrast, in the East, animal breeders have preferred to source foundation stocks
directly from wild populations (Boyd et al., 2006), often where there was only
limited or even no knowledge of the genetic attributes of the source populations.
While 87% of annual global farmed shrimp production (estimated at > 3.0 million
tonnes) is produced in the East (FAO, 2016c), specifically in the Asia-Pacific region,
there continues to be a significant shortage of available genetically improved farm
broodstock for this important industry with current annual demand for P. vannamei
broodstock estimated at approximately 1 million pairs (Giltterle and Diener, 2014).
Most broodstock in Asia are currently imported from commercial companies in the
West.
Despite there having been considerable progress made in the Asia-Pacific region
with selective breeding of native penaeid species e.g. P. chinensis (Sui et al., 2016a),
P. indicus (Benzie, 2009) and P. monodon (Krishna et al., 2011)), it will take some
time for current breeding programs to be able to provide sufficient numbers of
genetically improved stock in terms of both quality and quantity to meet demand
from the shrimp farming industry across this region. For ongoing sustainable
development of shrimp farming in the Asia-Pacific region, it is important to
acknowledge and apply lessons learned in earlier genetic improvement programs
80 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
conducted on penaeid farm stocks in the West. Essentially, genetic selection is an
artificial breeding strategy that exploits the additive genetic variance that is present
in a breeding line. Therefore, to a significant extent the total amount of genetic
variation that is captured in a founding population will largely determine the relative
success or failure of a breeding program over extended time frames (Falconer and
Mackay, 1996; Gjedrem and Robinson, 2014; Nguyen, 2016). As a consequence,
when existing unimproved domesticated strains are used as founding stocks for a
breed improvement program, an important first step can be to apply a “genotypic
approach” to determine the relationship between levels of genetic diversity in neutral
markers (SNPs/SSR) and genetic variance of commercially-important phenotypic
traits, in particular how much genetic variance that influences growth and survival
traits was captured in the base population.
Applying genetic theory and protocols to capture broad genetic variation in
founding stocks of cultured aquatic species is a relatively recent development (Hayes
et al., 2006). In addition, much of the available literature on this subject has been
based primarily on simulation studies (Fernández et al., 2014; Gianola and Rosa,
2015; Holtsmark et al., 2006; Holtsmark et al., 2008a; b), with very few studies
completed on actual aquaculture breeding programs that apply or test the above
approaches. While several well planned breeding programs undertaken on aquatic
species have been completed that have involved sophisticated strain comparison
experiments prior to starting a selection program (e.g. GIFT tilapia - (Eknath et al.,
1998); Atlantic salmon - (Gjedrem et al., 1991); European sea bass - (Vandeputte et
al., 2014)), significant financial outlays and extended timeframes were required in
each case, illustrating that there are still major issues when pursuing genetic
improvement of aquatic animals more widely. Neutral molecular markers (e.g.
81 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
microsatellites) can provide information about relative levels of genetic diversity,
population structure, stock relatedness and kinship in both cost and time effective
ways. This technology has been applied successfully in stock conservation and
management of breeding programs in a number of terrestrial domesticated farm
animals (Carvalho et al., 2015; Revidatti et al., 2014; Wilkinson et al., 2011;
Wilkinson et al., 2012). For aquatic farm species however, to date there are only
relatively few examples where genotype information on domesticated strains has
been included in genetic improvement programs (FAO, 2011).
Different opinions have been offered about the relative merits of using
domesticated strains vs wild stocks as the starting point for producing a base
population for genetic improvement programs. The main advantage of choosing
domesticated strains over wild populations is that locally-sourced farm strains in
general are often much easier to source and much less expensive to collect compared
with wild populations particularly if the target species is an exotic species that can
involve issues with quarantine requirements and large costs associated with sampling
and transportation. Another advantage of domesticated strains over wild strains is
that in most cases they have already accumulated some favourable traits suitable for
culture as a result of the effects of indirect selection in artificial culture environments
over many generations that often make them easier to handle and breed compared
with wild stocks (Olesen et al., 2015). Thirdly, using domesticated strains as the
starting resource can provide a new breeding line with a competitive start, in
particular for commercially important traits including growth performance
(Fernández et al., 2014).
82 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
Genetic selection is widely recognised as an efficient tool for improving the
sustainability of fish farming because it can be used to enhance biological production,
feed conversion, and increase survival of aquatic animals (De Verdal et al., 2018;
Gjedrem, 2012; Gjedrem and Rye, 2018; Gjedrem et al., 2012; Nguyen, 2016). Breed
improvement programs undertaken on Pacific white shrimp (P. vannamei) have now
been trialled for decades in a number of countries (Argue et al., 2002; Benzie, 2009;
Castillo-Juárez et al., 2007; Gitterle et al., 2005c; Li et al., 2015) and some have
achieved significant genetic gains for some commercially important traits including
growth rate (Andriantahina et al., 2013b; Gjedrem and Rye, 2018) and disease
resistance (Cock et al., 2017; Cock et al., 2009; Moss et al., 2012a).
Evaluating the relative additive genetic component (heritability, h2) of target traits
is a crucial step in any breed improvement program because these data allow the
selection strategies employed to be optimised. Most published estimates of
heritability for important morphological traits (or stature traits, including body size
and body weight) for Pacific white shrimp to date have been moderate to high
(h2 >0.15) indicating that significant genetic gains are possible via a genetic selection
approach (Argue et al., 2002; Castillo-Juárez et al., 2007; Li et al., 2015). In contrast,
most estimates for fitness traits have been relatively low (h2 <0.15), including for
pond survival (Caballero-Zamora et al., 2015b; Gitterle et al., 2005a; Li et al., 2015;
Zhang et al., 2017), cold temperature tolerance (Li et al., 2015), White Spot
Syndrome Virus (WSSV) disease resistance (Gitterle et al., 2005a), and growth in the
presence of WSSV (Caballero-Zamora et al., 2015b); however TSV resistance (Moss
et al., 2013) and ammonia tolerance (Lu et al., 2017) are exceptions in this regard.
Furthermore, estimates for fatty acid composition suggest only limited additive
genetic variance exists for this trait in the target species (Nolasco-Alzaga et al., 2018).
83 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
Pacific white shrimp is currently, one of the top ranking species in terms of value
among traded seafood commodities worldwide (Kumar and Engle, 2016), and it
constitutes an extremely large and growing industry in China, with annual production
there exceeding 1 million tonnes in recent decades (FAO, 2016c). The seed
production component of Pacific white shrimp farming in China however, is
vulnerable as most seed producers depend on imported Specific Pathogen Free (SPF)
lines that often show susceptibility to a number of local pathogens and diseases
(Cock et al., 2017; Cock et al., 2009; Moss et al., 2012a; Thitamadee et al., 2016).
An alternative way to address this problem is to develop broodstock of improved
lines that are naturally adapted to local farm conditions across the target region. In a
previous study, we evaluated Pacific white shrimp domesticated strains from
hatcheries in China for their relative levels of genetic variation, extent of stock
differentiation and genetic relatedness using microsatellite markers and produced a
complete diallel cross approach to produce a local synthetic base population for
selection, with goals of maximising genetic variation and maintaining inbreeding
rates at appropriate levels (Ren et al., 2018). Here, we examine quantitative data on
genetic variance for growth in the same domesticated lines in China that potentially
could be captured using the above defined “genotypic approach”. In parallel to this
first goal, our objective was to develop a base population of locally adapted stock to
be used in a future breed improvement program.
3.2 MATERIALS AND METHODS
3.2.1 Animal Material and Crossing Design
Four strains with three replicated breeding lines of each strain (referred to as:
NA_1; SA_1; Kona; and LA) were sourced from 10 different P. vannamei nauplii
84 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
suppliers from three Provinces in China (Fujian, Guangdong and Hainan). The initial
target was to produce 96 full-sib families via a complete 4 × 4 diallel of 16 crosses
with 6 replicates. Strains were defined by their relative degree of genetic
differentiation from each other and their documented breeding history (Ren et al.,
2018). Criteria used for inclusion of an individual breeding line into the program
combined the quality of the stock management practices applied at each respective
hatchery and their rankings for genetic diversity levels identified from microsatellite
markers (Ren et al., 2018). A complete diallel cross design was then applied among
the four P. vannamei ‘best’ strains to form the synthetic base population (Table 3.1).
The mean inbreeding estimate among the 16 crosses used in the diallel cross was
approximately 0 based on a RQG relatedness estimation (Ren et al., 2018) conducted
earlier, resulting from the 98 full-sib families generated via monogamous pairing
(Table 3.1).
Table 3.1 Number of families produced from 16 complete diallel crosses betweenfour Penaeus vannamei strains (NA_1, SA_1, KONA, and LA).
Dam
Sire
NA_1 SA_1 KONA LA
NA_1 7 6 6 4
SA_1 6 8 5 6
KONA 9 6 8 5
LA 3 6 4 9
85 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
3.2.2 Broodstock Management
Candidate broodstock of 6 to 8 months in age were collected from the four strains
(12 breeding lines with three breeding lines replicated for each strain) and stock
transported to maturation tanks located in Wanning, Hainan Province, China. Each
breeding line was stocked separately into two maturation tanks per line (male and
females kept separately) at a density of 10-15 shrimp/m2. Maturation tanks were
circular polypropylene fibre tanks (3.5 m diameter, 0.9 m depth), with the water
column depth maintained at 50 cm (Figure 3.1a and Figure 3.1b). A biological
recirculating system was used to maintain water quality at an exchange rate of 600%
to 800% per day. The biological filter was constructed of four layers consisting of;
filter biological cotton, silica sand, crushed coral stone, and volcanic rock, with a
total approximate volume of 30% marine water in each tank. During the 3-4 months
pre-maturation stage after arrival, broodstock were fed with a combined diet (2:1) of
commercial pellets (containing 35% to 40% crude protein) and fresh squid, with
water temperature maintained at 22 °C to 27 °C. When all broodstock had reached 10
months of age, females were subjected to unilateral eyestalk ablation. Daily feed
composition consisted of a mixture of fresh meal diet (50% polychaetes, 30% squid
and 20% mussels) at approximately 5% of total biomass per tank. Tank water was
maintained at 28 ± 2 °C and 31-35 ppt salinity.
86 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
Figure 3.1 a, b) Maturation tanks system used in the experiment; c, d) selecting
candidate females with ovarian development at IV ~ V stage for artificial
insemination.
3.2.3 Synthesis of Families
Females with ovarian maturation at late IV and V stage (with dark yellow-green
colour ovarian lobes) (Figure 3.1c and Figure 3.1d) were inseminated artificially
using a single male to produce full-sib families. Inseminated females were
transferred to 500 L tanks for spawning. At 1:00 am, berried females were returned
to maturation tanks, and eggs collected using a 25L bucket. Eggs were then
disinfected with iodine (50 ppm) and placed into another 500 L tank (Figure 3.2a)
filled with clean seawater. Gentle aeration was provided to all tanks in the hatchery.
Nauplii (Figure 3.2b) from each family with more than 4,000 individuals were
collected for the next larviculture step (Figure 3.2c and Figure 3.2d). In total, 98
full-sib families were produced over a 21 day period.
87 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
Figure 3.2 a) 500L tanks used for families reared separately (with capacity to
produce 245 families each breeding cycle); b) collecting nauplii (cloudy white area)
for the next larviculture step; c) larviculture for families reaching Z2~Z3 stage; d)
family successful reaching PL stage.
3.2.4 Larviculture
Nauplii-5 density in each family was adjusted to approximately 120 individuals/L.
Live microalgae (Chaetoceros sp. and Tetraselmis sp.) were provided to incubation
tanks according to the standard schedule outlined by Treece and Yates (1990).
Freshly hatched Artemia nauplii were prepared for the later protozoeal-3 stage,
following the feeding regime used by Treece and Yates (1990). Commercial
88 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
microbound diets (INVE-Frippak, 1CAR, 2CD, 3CD, INVE, Belgium) also served as
a supplement following INVE recommendations. When post-larvae had reached the
PL7-8 stage, 450 individuals from each family were removed and maintained in a
single tank until their body weights had reached 1 to 2 grams prior to family pedigree
tagging.
3.2.5 VIE Tagging
Visible Implant Elastomer (VIE, Northwest Marine Technology) tagging was
used for pedigree identification of full-sib families. A combination of six different
colours (red, green, yellow, purple, blue, and orange) and three different tagging
positions on the body (fifth left and right abdominal segments, and sixth dorsal
abdominal segment) were employed so that each individual could be injected with a
unique family code. A total of 120 individuals from each family were used for VIE
tagging.
3.2.6 Growth Rate and Survival Experiment
Tagged individuals from each family were stocked randomly into 12 recirculating
system tanks (10 m2, the same dimensions as maturation tanks) at a rate of 100
individuals/m2. The feed used in all tanks was a combination of commercial pellets
(35%-40% crude protein) and adult Artemia, with the daily feeding rate fluctuating
between 3%-5% of total biomass per tank. Physical water parameters were
maintained as per the maturation stage for broodstock, except that water temperature
ranged from 22 to 29 °C, depending on both local weather conditions and ocean
water temperature in the hatchery. Stocking density in experimental tanks was
adjusted to 50 individuals/m2 after VIE tagging (body weight (BW1)), following
which this was adjusted to 20 individuals/m2 (sampled at random) when shrimp had
reached 8 months (pre-maturation stage; body weight (BW2)).
89 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
On 18th October 2016, the hatchery was hit by typhoon Sarika, interrupting the
electricity supply for 36 hours, resulting in 60% mortality of all experimental shrimp
(after VIE tagging, ~90 days post hatching). Surviving shrimp however, recovered
well (death ratio ~0.01% per day, feeding ratio ~4.5% per day of biomass,
observation of swimming actively) and we estimated that on average ~30 individuals
had survived in all families that could be used to continue the experiment. We
therefore adjusted density by redistributing shrimps among the 12 experimental tanks
equally. The survival trait (S) was defined as the proportion of shrimp alive at BW2
relative to the number alive at BW1. In total, 89 full-sib families data were available
for further data analysis.
3.2.7 Statistical Analysis
The three variables were tested for normality using SPSS. Both BW1 and BW2
were normally distributed, whereas survival was not. As such, we chose to examine
the significance of fixed effects by applying a linear model (GLM) using
UNIVARIATE in SPSS. The variance components and heritability values for
individual traits were estimated using an animal model in WOMBAT (Meyer, 2007)
via restricted maximum likelihood (REML) methodology. The animal model was as
follows:
y = Xβ + Zα + e,
Where:
y is a vector of observations for a trait (body weight: BW1 or BW2; survival:
S (where ( 0=dead, 1=alive));
X refers to a design matrix to fixed effects,
90 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
β is the vector of fixed effects including sex, strain effect of dam, strain effect
of sire, strain effect of different combinations, age and tank;
Z is an incidence matrix to animal effects,
α is the vector of random additive genetic effects of the animals, and e is the
random residual. Both a and e follow a normal distribution with mean zero
and variance Aσa2 and Iσe2, respectively. Here, σa2 and σe2 are additive genetic
and error variances and A is the numerator relationship matrix based on
pedigree information.
Total phenotypic variance (σp2) was the sum of additive genetic variance (σα2) and
random residual components (σe2). Heritability was calculated as the ratio of additive
genetic variance to the total phenotype variance (h2= σα2/ σp2).
A multivariate mixed linear animal model was fitted to estimate the genetic
correlation between three traits, expressed in matrix notation as:
yBW1yBW2yS
� �� � �� � � �
(2)
Where yBW1 and yBW2 are body weight at 150 days post hatching and 8 months post
hatching, respectively. yS refers to survival trait between stage between yBW1 and
yBW2. Fixed effects in Model 2 are the same with Model 1. The genetic correlation
between estimated traits (BW1, BW2 and S) were calculated as the covariance of the
standard deviation between two traits as: � � �12
�12 �2
2where σ12 was the calculated
additive genetic covariance between two traits, and σ12 and σ22 were the additive
genetic variances of traits 1 and 2, respectively. Presence of heterosis was assessed
as the percentage change in traits compared with the mean values observed from the
91 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
two respective parental purebred strains, significant differences were assessed via
Tukey’s HSD posthoc test in SPSS.
3.3 RESULTS
3.3.1 Survival in Experimental Tanks
Mean survival over 72 days from BW1 (Feb. 9th, 2017) to BW2 (Apr. 12th, 2017)
was 89.1% (Table 3.2). Survival in experimental tanks systems was consistently
high and could be managed, with a cumulative mortality rate per day of 0.15% where
biomass of shrimp per tank ranged from 427.7 to 582.8 g/m2.
Table 3.2 Descriptive statistics for body weight at two different stages (BW1 andBW2) and survival (S)
TraitStructure of data Statistics of data
No. offamilies
No. ofshrimps
Mean no. ofshrimps/family
Mean Minimum Maximum Standarddeviation
BW1 89 2752 30.92 18.65g 1.39g 43.59g 6.74g
BW2 89 2452 27.55 28.52g 4.77g 63.02g 8.69g
S 89 --- --- 89.10% 0 100% 39%
3.3.2 Descriptive Statistics
Details of ANOVA results for individual fixed effects are provided in Table 3.3.
It is clear that all fixed effects impacted the growth and survival traits.
92 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
Table 3.3 Summary of analysis of variance for fixed effects (F-statistic value andsignificant level).
Fixed factors DF BW1 BW2 S
Tank 11 91.53 *** 122.52 *** 21.45 ***
Sex 1 N. E. 45.61 *** N. E.
Batch age 10 56.71 *** 83.62 *** 20.66 ***
Strain of dam 3 130.73 *** 172.04 *** 11.26 **
Strain of sire 3 107.20 *** 184.18 *** 2.86 *
Strain
combination
9 60.12 *** 81.32 *** 29.31 ***
The number of recorded families, shrimp and details of the three study traits are
presented in Table 3.2. Mean weight at BW1 across tanks was 18.56 ± 6.74g with
individual weights ranging from 1.39g to 43.59g, while mean weight at BW2 was
28.52 ± 8.69g and individual weights ranged from 4.77g to 63.02g (Table 3.2).
Coefficient of variation for growth was high, but decreased with shrimp age from
36.13% at BW1 to 30.56% at BW2. Coefficient of variation for survival was also
high (28.59%) indicating large differences in mean survival among families.
3.3.3 Effects of strain on growth and survival
Table 3.4 presents differences for three estimated traits among the four purebred
strains and the six crosses. There were no significant differences found between
reciprocal strain crosses (P>0.05) for the studied traits, the data from each bi-
directional cross between pairs of strains were pooled together.
93 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
Table 3.4 Estimated means for four purebred strains and six crosses for body weight(g) at two stages (BW1 and BW2) and survival (S %) .
Mating types BW1 BW2 S (%)
N1 means ± SD N2 means ± SD means ± SD
NA_1 × NA_1 111 20.71 ± 7.17 de 62 32.13 ± 7.87 de 0.56 ± 0.05 a
SA_1 × SA_1 219 24.97 ± 7.76 f 212 36.70 ± 9.82 f 0.97 ± 0.01 d
KONA × KONA 382 19.56 ± 5.62 bcd 321 33.86 ± 7.31 ef 0.84 ± 0.02 bc
LA × LA 505 15.03 ± 4.72 a 465 23.27 ± 6.00 a 0.92 ± 0.01 cd
NA_1 × KONA 210 20.01 ± 6.33 cd 168 29.87 ± 8.96 cd 0.80 ± 0.07 b
NA_1 × LA 547 17.43 ± 6.97 b 492 26.33 ± 8.43 b 0.90 ± 0.01 bcd
SA_1 × KONA 152 21.17 ± 4.74 de 147 29.68 ± 6.28 cd 0.97 ± 0.02 cd
SA_1 × LA 173 20.70 ± 5.48 de 149 28.98 ± 6.42 bc 0.86 ± 0.03 bcd
KONA × LA 350 18.41 ± 4.19 bc 343 26.08 ± 5.87 ab 0.98 ± 0.01 d
VIE tag loss 14 --- 25 --- ---
All 2752
18.65 ± 6.74 2452
28.52 ± 8.69 0.89
Specific differences for studied traits among four purebred strains and the sixcrosses were detected via a Tukey’s HSD posthoc test.
N = Number of individuals for data analysis.
94 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
From the Table 3.4, it is clear that there were significant mean body weight
differences among strains, with SA_1 (SA_1 x SA_1) always showing the highest
mean body weight at both BW1 and BW2. The LA strain (LA x LA) always showed
the lowest mean body weight at both BW1 and BW2. No significant differences were
evident between the NA_1 and KONA strains at BW1 and BW2. The largest
difference in mean body weight were 66.13% at BW1 (SA_1 vs LA) and 57.71% at
BW2 (SA_1 vs LA). Overall heterosis for BW1 was -0.4%, ranging from -4.9%
(SA_1 × KONA) to 6.5% (KONA × LA). A significant decline (P<0.01) in heterotic
effects was evident in BW2 with overall mean heterosis at -0.08, ranging from -0.16
(SA_1 × KONA) to -0.03 (SA_1 × LA).
Strain SA_1 also showed the highest survival rate, with a mean survival rate of
97%. In contrast, strain NA_1 was ranked lowest for survival rate, with a mean of
56%. The LA and KONA strains were intermediate for mean survival rate, with
means of 92% and 84%, respectively. Thus, the largest difference in overall mean
survival rate was as much as 73% (SA_1 vs NA_1). Overall mean heterosis of S was
0.04 ranging from -9.0% (SA_1 × LA) to 21.6% (NA_1 × LA), a significant
difference compared with that observed in purebred strains (P<0.01).
3.3.3 Genetic Analysis of Growth and Survival Traits
Estimates of variance components, heritability and genetic correlations for growth
traits (at BW1 and BW2) and survival (S) are presented in Table 3.5. Heritability for
growth rate was high (h2>0.4), with 0.52 ± 0.09 at BW1 and 0.44 ± 0.07 at BW2,
respectively. In contrast, h2 for survival (S) was low (0.01 ± 0.02) and was not
significantly different from zero indicating only very limited additive genetic
variance for this trait. The genetic correlation between BW1 and BW2 was moderate
95 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
(0.95 ± 0.03) and was also significantly different from zero. Genetic correlations
between survival (S) and growth (BW1 and BW2) however, were low at 0.26 ± 0.51
(S and BW1) and 0.18 ± 0.01 (S and BW2), respectively and both estimates were not
significantly different from zero (p>0.05).
Table 3.5 Estimates of variance components (σ2p, the phenotypic variance; σ2a, theadditive genetic variance; σ2e, the random residual error variance), heritabilities (h2,ratio of additive genetic variance; e2, ratio of random residual error variance), andgenetic correlations for the body weight (BW1 and BW2) and survival (S) traitsbased on univariate animal model analysis.
*Estimate is highly significantly different from zero (P<0.01).NS Estimate is not significantly different from zero (P>0.05).
3.4 DISCUSSION
As far as we are aware, the current study is the first report of a penaeid shrimp
breeding program employing domesticated strains based on an experimental design
that applies a “genotype approach” to initiate the genetic selection program.
Heritability for body weight at two ages (BW1, h2=0.52; BW2, h2=0.44) were quite
high with quantitative data available from 89 full families and approximately 30
individuals per family providing strong evidence that broad genetic variation for
96 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
growth rate was available in the base population developed for the project. This is
evident even though we employed a monogamous mating design where we could
only estimate genetic variance of family differences, an approach that has potential to
overestimate heritability as a result of maternal or random effects common to full-
sibs. From this base and with appropriate stock management and applying the
breeding strategies rigorously to preserve maximum levels of genetic variation
within families and by controlling inbreeding, we are confident that future genetic
gains can be optimised and that the planned selection to produce a fast growth stock
will be successful.
3.4.1 Experimental Tank System
A major constraint on penaeid shrimp breeding in the past has been designing and
building reliable, high quality environmental closed water systems for domesticated
stock in captivity (Coman et al., 2005; Duy et al., 2012; Yano, 2000). Traditional
maturation or indoor culture systems that have employed flow-through water systems
have regularly experienced average mortality rates of ~0.5% per day. This rate is in
general, too high to support efficient shrimp genetic improvement programs when we
consider mean age at maturation (commonly >10 months) vs mortality rate (~0.5%
per day) and the bio-secure environmental conditions required to produce SPF
broodstock. In our experimental tanks, except for the unpredictable impacts of a
typhoon, water conditions were highly suitable for an improvement program for
domesticated shrimp as is evident in the survival data where average mortality rate
per day was less than 0.1% from VIE tagging age through to BW1 data collection
(data not provided), followed by an average mortality rate of only 0.15% per day
during the pre-maturation stage from BW1 to BW2. Biomass reared in our
experimental tanks ranged between 427.71 to 582.76 g/m2 and this already meets
97 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
basic requirements for broodstcok maturation to produce nauplii successfully in most
commercial hatcheries in China maintaining a stocking density of Pacific white
shrimp of size 40-60g at approximately 6-8 individuals/m2.
3.4.2 Genetic Parameters for Growth and Survival Traits
In the present study, heritability estimates for growth rate (BW1 and BW2) were
high indicating that large genetic variation was present at both developmental stages
tested for this trait in our base population. In general, our estimates were consistent
with results of reviews of heritability estimates for body size (or stature) across a
wide range of species (from model species like Drosophila, wild animals,
domesticated species to humans) that is generally moderate to high with h2 estimates
ranging from 0.15 to 0.85 (Visscher et al., 2008). Estimation of heritability in a
specific population however, will depend on partitioning observed phenotypic
variation into unobserved additive genetic variance and environmental factors, thus
heritability will vary under different environmental (culture) conditions (Hill, 2014;
Visscher et al., 2008). Studies on genetic breeding programs for penaeid shrimps that
have examined environmental effects on heritability estimates for growth rate have
identified three factors that can influence the estimate (a) sub-optimal environmental
conditions (including suboptimal low water temperature, the presence of disease,
pathogens, etc.); (b) normal commercial farming conditions; and (c) the ability to
maintain optimum environmental conditions in recirculating water tank systems or
clear water systems. Our results are consistent with the last defined group (c), where
h2 ranged from 0.23 to 0.84 with a mean >0.4 (Argue et al., 2002; Coman et al., 2010;
Kenway et al., 2006; Macbeth et al., 2007). Heritability estimates in the current study
were higher than reported for group b (most estimates for h2 were between 0.15 to
98 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
0.4) (Campos-Montes et al., 2013; Campos-Montes et al., 2017; Castillo-Juárez et al.,
2007; Gitterle et al., 2005c; Ibarra and Famula, 2008b; Krishna et al., 2011; Nolasco-
Alzaga et al., 2018; Pérez‐Rostro and Ibarra, 2003a; Sui et al., 2016a; Sui et al.,
2016b; Sun et al., 2015a; Zhang et al., 2017). Heritability estimates for growth were
also much lower in sub-optimal culture conditions with h2 = 0.09 in the presence of
WSSV disease vs 0.15 in a culture pond without WSSV present (Caballero-Zamora
et al., 2015b), and 0.30 in under low and suboptimal water temperature conditions vs
0.48 in controlled water temperature culture ponds (Li et al., 2015).
It is widely recognised that heritability estimates for growth rate can also change
with age class. Our estimates for h2 at the sub-adult BW1 stage (~24 weeks) were
marginally higher than that at BW2 the maturation stage. In general, it is expected
that heritability for growth should increase across the growth period. In Giant
freshwater prawn (M. rosenbergii), heritability h2 estimates for body weight
increased from pre-market to market size, from 0.11 to 0.15 (Hung et al., 2014).
Similar patterns have also been reported in Pacific white shrimp (Campos-Montes et
al., 2013; Pérez‐Rostro and Ibarra, 2003b; Zhang et al., 2017). In the current study in
contrast, we observed higher heritability estimates for growth at the earlier life stage
(BW1) examined. This pattern has also been reported in some other penaeid shrimp
breeding programs in particular ones applied to P. monodon stock improvement
(Coman et al., 2010; Kenway et al., 2006; Macbeth et al., 2007).
Knowledge about genetic parameters for traits at different developmental ages and
associated genetic correlations among traits can allow earlier application of selection
to achieve genetic gains later in the growth cycle or selection of a mixed age cohort
to maximise genetic gains (Campos-Montes et al., 2013; Hung et al., 2014).
Moreover, applying selection early during the growth phase can reduce costs
99 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
associated with maintenance of facilities in the hatchery and/or grow out systems. In
penaeid shrimp breeding however, it is often necessary to delay initiating selection
because the larval nursery phase is extended delaying the stage at which physical
tagging of families can take place (60-80 days to reach body size ~1.5g) and
estimating heritability during the nursery phase potentially may be confounded by
maternal and/or common environmental effects in family tanks (Montaldo et al.,
2013). Moreover, where heritability of growth traits has been estimated at early
stages during development in shrimp taxa, estimates have been quite low (i.e.
h2<0.15) (Campos-Montes et al., 2013; Hung et al., 2014), which if applied in
models would suggest only limited genetic gains were possible.
h2 estimates for survival were low in our study, a result that is consistent with
many previous studies of cultured shrimp species, including for P. vannamei
(Gitterle et al., 2005c; Li et al., 2015; Zhang et al., 2017) and M. rosenbergii (Luan et
al., 2015). Estimates of heritability for survival in aquatic species in general however,
vary widely. Tan et al., (2017a) reported back-transformed h2 estimates for survival
in P. vannamei were 0.36 for low and 0.22 for high density treatments, respectively.
In red tilapia, h2 estimates for survival ranged from 0.02 to 0.68 and were affected by
both test environments and the data analysis methods employed (Nguyen et al.,
2017). In general, survival is a fitness trait that can be affected by many factors.
These factors will include reflecting the relative health fitness status of individuals
that is essentially resistance to, and/or tolerance of, multiple and potentially unknown
factors and such traits usually have low heritability (Falconer and Mackay, 1996;
Sae-Lim et al., 2013). Estimated heritability for survival (S) in our study as with
many earlier studies, suggests that there is very limited potential to improve survival
100 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
rate using a directional selection approach. Alternatively, developing disease-
resistant strains could offer an indirect way of providing the opportunity to improve
this important trait and in this regard, there has been significant recent progress with
respect to improving survival in shrimp farming (Cuéllar-Anjel et al., 2012; Moss et
al., 2013).
3.4.3 Genetic Correlations between Growth and Survival
The genetic correlation between growth estimates at different ages in our study
was high (0.95), indicating selecting for body weight at the younger age will improve
growth rate at the later age simultaneously. This result accords well with other
estimates for penaeid shrimp species in the literature. Reported genetic correlations
for growth traits at different ages have ranged from 0.71 to 0.95 in P. vannamei
(Campos-Montes et al., 2013; Pérez‐Rostro and Ibarra, 2003b; Zhang et al., 2017)
and 0.63 ± 0.19 in P. monodon for body weight at 16 weeks and 24 weeks (Coman et
al., 2010), respectively. Recognition of this association can be used to optimize
selection for P. vannamei in commercial breed improvement programs. Improving
growth rate is the most important commercial trait targeted in breeding of farmed
aquatic species, and the majority of earlier programs that reported their genetic gains
from selection work refer to growth rate as the only trait targeted (Gjedrem and Rye,
2018). Growth can be measured by estimating body weight at specific ages, from
morphological traits (e.g. total length, body length, etc.) or by sampling daily growth
rate. Body weight and specific morphological traits are most useful due to the
practicality of data measurement, and the ease of collecting accurate data.
Furthermore, recognition of high phenotypic and genetic correlations between body
weight and other morphological growth traits at specific ages (i.e. rp>0.95; rg>0.95),
means that we can often select based on measurements of only a single trait for
101 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
growth, thereby saving significant time and cost (Campos-Montes et al., 2017;
Krishna et al., 2011; Pérez‐Rostro and Ibarra, 2003b; Sun et al., 2015a).
In our study, we consider BW1 represents the most appropriate time to apply
selection in real commercial programs due to high heritability and mean body size
closest to real harvest size in shrimp farming as most farmers harvest their shrimps at
~20 g. Another advantage in penaeid shrimps of choosing this younger age is to
avoid later impacts of heterogeneous growth related to sex (females usually mature at
much larger size than do males). In P. vannamei however, both sexes have similar
growth rates before body size reaches 25 g. So by employing selection at a mean
weight of ~20g this essentially eliminates any gender effect on estimates of genetic
parameters or Estimated Breeding Values (EBVs).
The genetic correlation between survival rate and growth rate was low and not
significantly different from zero in our study. While this result accords with some
previous reports on penaeid shrimps, others have reported both positive (Gitterle et
al., 2005c; Zhang et al., 2017) and negative (Kenway et al., 2006; Zhang et al., 2017)
associations between the two traits.
3.4.4 Effects of Strain on Growth and Survival
This is the first report of differences in growth and survival rate in domesticated P.
vannamei strains in China. Clearly, the SA_1 strain showed a superior growth rate
while LA showed the lowest growth rate. To our knowledge, differences in growth
rates among strains observed here are consistent with anecdotal evidence from the
shrimp farming industry in China. In the current study, variation in mean body
weight among strain combinations was much higher than that seen with other aquatic
species subject to diallel crossing, including common carp (Nielsen et al., 2010) and
102 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
European sea bass (Vandeputte et al., 2014). This probably results from the
significantly different genetic backgrounds of individual strains (Ren et al., 2018),
the different breeding strategies applied to strains before they were introduced into
China, and to a some extent, the different management practices employed on strains
in various hatcheries in China following their introductions. While no significant
positive heterosis was evident for body weight among the different crosses, the result
is consistent with other aquatic species (Bentsen et al., 1998; Gjerde and Refstie,
1984).
NA_1 in the current study showed a significantly lower survival rate compared
with the other strains examined. This reflects the significant reduction of available
genetic resources for NA_1 since introduction to China in 2014, possibly related to
the issue of low pond survival. Prior to this, NA_1 made a major contribution to P.
vannamei broodstock production in China. High inbreeding rates in recent times
however, has probably contributed to the low survival rate of this strain seen
currently. In our earlier study, we showed that NA_1 culture resources in China
generally showed comparatively low genetic diversity levels based on microsatellite
markers (Ren et al., 2018). High inbreeding levels commonly impact negatively on
fitness traits related to survival and growth rates. Survival rate is however,
significantly impacted by performance in different culture environments due to G X
E effects. Therefore, the strain differences in survival rate detected here may not
necessarily be a reflection of any specific commercial farming environment.
3.4.5 Implications for Further Study
The main objective of the current study was to initiate a base breeding line for
Pacific white shrimp in China and to exploit additive genetic variance in locally
adapted culture lines. Since the sample size of most purebred strains or crosses
103 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
compared here was less than 15 replicates, we could not investigate potential
heritability values for 10 individual crosses of mating type. To date, there has only
been a single program in penaeid shrimps that has directly exploited heterosis
between strains to improve culture performance (Goyard et al., 2008a). Exploiting
additive genetic variation however, could also benefit the shrimp farming industry in
China over the long term. Therefore, we recommend that initial selection be
undertaken within strains (Ponzoni et al., 2013) to homogenize growth performance
in early generations to conserve more genetic variation and to achieve genetic gains
over longer time frames.
While the current study employed recirculating water tanks as culture
environments to estimate impact of family selection on growth rate, future
application of an improved line under commercial farming environmental conditions
will require genotype-by-environment (G×E) interactions to be assessed. After
reviewing studies of G×E interactions on growth traits in 38 aquatic species, average
genetic correlation was reported to be 0.72 (Sae‐Lim et al., 2016) and was even
higher in earlier experiments conducted on penaeid shrimps for growth performance
in different culture environments (Gitterle et al., 2005c; Ibarra and Famula, 2008b;
Tan et al., 2017a). This suggests that the base line developed in our study has a high
chance of also being successful after selective breeding for fast growth traits in a
variety of different shrimp farming environments.
Our test tanks provided a reliable and excellent clean water system for
domestication of P. vannamei. At the BW2 data collection stage, mean body weight
for the top 30% of female and male broodstock were 41.22 ± 5.26g and 37.05 ±
4.59g respectively, a mean size that in general, accords with maturation size of SPF
104 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
stocks used to produce next generation families for genetic improvement work.
Furthermore, our experimental tank system can also be used for maturation and
meets bio-secure SPF environmental requirements as a nucleus-breeding centre
(NBC) for cultured SPF P. vannamei broodstock according to current quarantine
regulations in China.
105 Chapter 3: Genetic parameters for growth and survival traits in a base population of Pacific white shrimp(Penaeus vannamei) developed from domesticated strains in China
3.5 CONCLUSIONS
In conclusion, our study shows that there is substantial genetic variation for
growth traits in the L. vannamei synthetic breeding line established in our study in
China. Synthesising a base population using domesticated Chinese culture strains
from multiple diverse source populations captured broad genetic variation for rate of
growth that can be exploited in a future breed improvement program. In the future, to
improve the accuracy of estimating genetic parameters, random additive effects
should exclude maternal effects and random environmental effects common to sibs.
These goals can be achieved by a nested mating design and via communal rearing of
each family shortly after hatching and subsequently rebuilding pedigree information
from molecular marker data.
107 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
Chapter 4: Comparison of Reproductive Performance of
Domesticated P. vannamei Females Reared in Recirculating Tanks and
Earthen Ponds: An Evaluation of Reproductive Quality of Spawns in
Relation to Female Body Size and Spawning Order
109 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
ABSTRACT
Based on estimated levels of heritability for growth traits in our nucleus
population, we had determined that appropriate levels of genetic variation were
available in our line to exploit fast growth traits. The next step in the genetic
improvement program therefore, was to optimize the quality of reproductive
performance of candidate broodstock, and to progress plans for the long term
conservation of genetic variation in the nucleus population, facilitating the
development of an effective seed dissemination strategy. Here we investigated the
relative reproductive performance of female broodstock reared under two common
rearing systems: i) recirculating tanks (RT) and ii) earthen ponds (EP), and evaluated
the relative quality of individual reproductive performance (RT vs EP), the quality of
reproductive females in relation to individual body size of spawners, and female
reproductive quality relative to spawning order. For this analysis, broodstock under
two culture conditions (RT and EP) using nauplii produced from spawning of a
single batch to eliminate any potential effects from the genetic resources used or age.
were sourced a single night’s cohort of nauplii, with the aim of eliminating any
potential impacts from genetic resource or age. Individuals were reared as
broodstock in either RT or EP rearing systems for the experiment. In total, we scored
reproductive parameters for 156 spawning females (107 RT-reared females and 49
EP-reared females) in two replicate maturation tanks over a 30 day test period in the
second month after unilateral eyestalk ablation. No significant difference (P>0.05)
was observed between RT-reared females and EP-reared females for number of eggs
per spawn (RT= 23.34 ± 0.72 × 104, EP= 22.45 ± 0.67 × 104), number of nauplii per
spawn (RT= 19.85 ± 0.85 × 104, EP= 19.53 ± 0.83 × 104), hatch rate of eggs per
110 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
spawn (RT= 0.83 ± 0.02, EP = 0.85 ± 0.02) and relative fecundity - number of eggs
per g of female (RT= 5.51 ± 0.15× 103, EP= 5.78 ± 0.19 × 103). We recorded 136
and 101 spawning events for RT and ER females, respectively. EP-reared females
(1.93 ± 0.23) showed a significantly higher (P<0.01) spawn frequency compared
with RT-reared females (1.34 ± 0.12). Females under the two treatments showed a
similar pattern for larger body size spawners producing higher numbers of eggs and
nauplii per spawn. Of interest, we observed that while large sized RT-reared females
recorded a higher mean spawn frequency, medium-sized females showed double the
spawn frequency compared with small or large sized females in the EP treatment. No
evidence observed for quality of individual female reproductive performance for
multiple spawning individuals compared with first or second spawning only
individuals for all reproductive parameters evaluated (P>0.05).
111 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
4.1 INTRODUCTION
In most breed improvement programs for farmed aquatic species, the next stage
after development of a breeding nucleus is to identify the best breeding candidates
and to optimise their reproductive performance prior to implementing selection
(Gjedrem and Thodesen, 2005). In penaeid shrimp breeding, specific tasks involved
in this step include: (a) rearing offspring of individuals from the nucleus to sexual
maturation; (b) providing the best broodstock to multipliers; and (c) supplying
nauplii or post larvae (usually PL5 or PL10) to the nursery sector - or if a nursery
stage is not included then juvenile shrimp are supplied directly to growout farmers.
In part, this sequence requires that broodstock used to produce juveniles show good
individual reproductive performance as this is essential, not only to preserve genetic
resources in the nucleus and to accumulate optimal breeding traits in live animals
across generations, but also facilitates dissemination of quality seed to the growout
industry. This is because the majority of profit generated can then be fed back into
investment in the breeding program. From the perspective of a seed multiplier,
important parameters determining relative individual female reproductive quality
include; the number of eggs per spawn (NE), the number of nauplii per spawn (NN),
the hatch rate of eggs (HR) and the proportion of females in the broodstock
population that spawn per night (this also equates to spawn frequency (SF) of
females), total nauplii numbers produced and the associated profit that is possible.
While just about all broodstock used currently in P. vannamei farming around the
world are sourced from captive domesticated spawners (Andriantahina et al., 2012b;
Benzie, 2009; Ceballos-Vázquez et al., 2010; Ibarra et al., 2007b), this is not true for
most other farmed penaeid species (i.e. P. monodon, P. japonica, P. paulensis, P.
112 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
indicus) that still rely fully or partly on broodstock sourced from wild populations
(Arnold et al., 2013; Boucard et al., 2004; Jiang et al., 2009; Marsden et al., 2013;
Peixoto et al., 2011; Peixoto et al., 2008; Preston et al., 2004). Use of domesticated
stocks maintained under appropriate bio-secure control conditions (that may include
specific pathogen-free (SPF) management) can effectively address most problems
associated with employing wild broodstock (Cock et al., 2017; Cock et al., 2009;
Gjedrem and Rye, 2018). Moreover, availability of fully domesticated spawners can
be more economic compared with the costs associated with collecting wild stock,
their year-round availability, and relative quality of their performance. Recently, the
availability of fully domesticated culture lines of Pacific white shrimp in the Asia-
Pacific region has led to significant growth in production of farmed prawn there
(Kumar and Engle, 2016; Lightner et al., 2009b).
Relative reproductive performance of domesticated stock however, needs to be
evaluated appropriately before broodstock are released to the seed production sector.
There has been significant controversy about the relative reproductive performance
of domesticated lines over the past 40 years, resulting from a range of factors
including impact of age, size and/or genetic background (Aquacop, 1979; Arcos et al.,
2005a; Arnold et al., 2013; Browdy, 1998; Coman et al., 2006; Marsden et al., 2013;
Medina et al., 1996; Menasveta et al., 1993; Peixoto et al., 2008; Preston et al., 1999;
Primavera and Posadas, 1981; Wen et al., 2015).
Female reproductive performance in penaeids can be impacted by a number of
factors including: individual physical status, environmental culture water factors,
nutrition and additive genetic composition (Benzie, 1997; Ibarra et al., 2007b). The
same stock grown in different culture conditions can also show different reproductive
performance due to both effects of varying environments and differences in
113 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
nutritional factors. Well managed recirculating tank (RT) systems provide a stable
high quality water environment in indoor bio-secure conditions that should result in
lower mortality and minimum water pollution. For these reasons, they have been
considered to be an ideal rearing system for closing the life cycle of penaeid shrimp
in genetic improvement programs, and also for producing mature SPF broodstock for
the industry (Chen et al., 1991; Crocos and Coman, 1997; Duy et al., 2012; Otoshi et
al., 2003). Tank-reared broodstock however, often do not show comparable
reproductive performance compared with stocks reared in earthen ponds or even that
of wild populations (Andriantahina et al., 2012b; Arnold et al., 2013; Coman et al.,
2006; Otoshi et al., 2003). This issue needs to be further investigated to help meet
demands of the seed production sector.
Earthen ponds (EP) are widely used for rearing domesticated P. vannamei stocks
in the shrimp farming industry (Briggs et al., 2004). Small entrepreneurial family
holders in China first learned about maturation of P. vannamei broodstock in earthen
ponds after unilateral eyestalk ablation was introduced by a Taiwanese technician in
the late 20th century (pers. comm., Aibing Gao, President of the Pacific white Shrimp
Seed Association of Xiamen), and this development pioneered shrimp farming in
China. For decades, this method of nauplii production has contributed more than
50% to total nauplii supply in the seed sector in China. This practice is now used
widely in China where, an annual production of more than one million tons of Pacific
white shrimp has been produced for decades (FAO, 2016c). Several anecdotal stories
about nauplii production in earthen ponds however, have indicated that problems still
exist. Farmers using this system prefer small and medium sized female broodstock
rather than choosing large individuals because they consider that smaller-sized
114 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
mature females show better reproductive performance. Results of scientific studies
on the relationship between body size and individual reproductive performance in
penaeid shrimps in contrast, have suggested the reverse relationship (Andriantahina
et al., 2012b; Arcos et al., 2003a; Arnold et al., 2013; Ceballos-Vázquez et al., 2010;
Ibarra et al., 2007b; Peixoto et al., 2003). Moreover, technicians running hatcheries
often claim that stocks raised in earthen ponds are easier to bring to maturity and
show a higher mating rate compared with imported SPF stocks (generally reared in
recirculating tanks (Otoshi et al., 2003)). To date, there have only been two
comparative studies on relative reproductive performance of P. vannamei broodstock
reared in RT vs EP systems (Andriantahina et al., 2012; Otoshi et al., 2003).
Estimated reproductive parameters were also collected following artificial
insemination that produced significantly low NE, NN, and HR rates compared with
natural mating designs and data available from current commercial nauplii
production.
External body size is the principal criterion for selecting female broodstock in
penaeid shrimps because it is non-invasive and easy to measure relative to the labour
and costs involved (Arcos et al., 2003a; Ibarra et al., 2007b). In general, large female
penaeids are considered better quality spawners because there is evidence for a
positive correlation between individual size and fecundity (NE) (Arcos et al., 2003a;
Emmerson, 1980; Ottogalli et al., 1988; Palacios et al., 1998) and spawn frequency
(SF) (Arnold et al., 2013; Hansford and Marsden, 1995; Menasveta et al., 1994;
Palacios et al., 2000; Wen et al., 2015). For choice of female P. vananmei broodstock,
the recommendation is to use 30 to 45 g individuals (Aquacop, 1983; Bray and
Lawrence, 1991; Otoshi et al., 2003; Wyban and Sweeney, 1991). In small family
hatcheries in China however, farmers usually select females of 25-35g for nauplii
115 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
production. Choice of individual body size of broodstock can be related however, to
the rearing systems used due to effects of culture density, physical environmental
factors in the water used, and nutritional conditions. As a result, no clear advice is
currently available for new producers, so selection of broodstock based on individual
size requires further investigation to determine impacts of rearing systems. In
addition, results of earlier studies that investigated the relationship between body size
and reproductive parameters have varied widely (in some cases they show
contrasting results) in particular, in terms of hatch rate of eggs (HR). Of interest
however, is that most experimental tests of the above parameters have produced
estimates much lower than is achieved currently under commercial production
conditions. This highlights a need to standardize broodstock maturation
environments and nutrition. A starting point for this is to develop optimal
management in experimental test tanks and then later, to trial the procedures at larger
scales.
The primary objective to optimizing quality of larval production in penaeid
shrimp culture is to understand the mechanism(s) behind why a large proportion of
mature females reproduce infrequently or may never spawn, while at the same time
only a very small proportion of mature females spawn multiple times and hence
contribute the majority of nauplii in a hatchery (see reviews by (Arcos et al., 2003a;
Ibarra et al., 2007b)). For more than 40 years, studies have tried to manipulate a
variety of factors to improve the rate of multiple spawning in penaeid species.
Factors that have been considered include: phenotypic traits (Arcos et al., 2003a;
Hoang et al., 2002; Menasveta et al., 1994; Palacios and Racotta, 2003; Palacios et
al., 1999a); physiology and biochemistry (Arcos et al., 2003b; Palacios and Racotta,
116 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
2003; Peixoto et al., 2004); nutrition (Coman et al., 2007; Goodall et al., 2016; Hoa
et al., 2009); additive genetic components (Arcos et al., 2005b; Ibarra et al., 2009;
Macbeth et al., 2007); hormonal levels and functional gene expression (Huerlimann
et al., 2018; Treerattrakool et al., 2014; Tsutsui et al., 2005). To date however, no
studies have fully addressed this question or provided practical ways to improve or
optimize spawning frequency. For P. vannamei, the two culture systems (RT and EP)
widely used for domesticated stocks, may potentially impact spawning frequency. To
date however, no studies have investigated the impact of culturing shrimp in RT vs
EP environments on spawning frequency in P. vannamei stocks.
While multiple spawning of individual mature females is considered to be a
desirable trait to improve overall hatchery reproductive capacity, it is widely agreed
that there should be no compromise made on offspring quality when considering this
trait. Producers have also suggested that the egg quality of multiple spawners should
not necessarily deteriorate from the first spawn (Arcos et al., 2004; Ibarra et al.,
2007b; Palacios and Racotta, 2003), but they acknowledge that time factors have
always potentially impacted the quality of their comparisons. Reproductive
exhaustion of broodstock individuals during nauplii production in penaeid shrimps is
however, recognised to be a relatively common phenomenon (Palacios et al., 1998;
Palacios et al., 1999b; Wyban, 1997), in particular when maturation conditions have
not been managed appropriately. Again, this highlights that it is important to
consider how optimal the maturation conditions employed have been during any
experimental tests of female reproductive quality.
In the current study we reared P. vannamei broodstock under two culture
conditions (RT and EP) using nauplii produced from spawning of a single batch to
eliminate any potential effects from the genetic resources used or age, and compared
117 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
(a) differences in individual mature female reproductive performance under RT and
EP culture treatments, (b) the relationship between females size and the quality of
their reproductive parameters, and (c) female reproductive quality relative to
spawning order, under optimal maturation test environment conditions. Results
generated here can be applied to developing improved genetic management of
cultured P. vannamei lines and to optimise seed production for the shrimp farming
industry in China.
4.2 METHODS AND MATERIALS
4.2.1 Experimental Animal
Shrimp nauplii used in the study came from a single mass spawning (~ 80 full
families) on a single night in a commercial hatchery owned by the Beijing Shuishiji
Biotech Ltd. at Wanning, Hainan Province, China. Larval culture and the nursery
phase occurred from 1st July to 20th July 2017, using identical procedures as
described in Chapter 3. Post larvae (PL10 stage) were sampled randomly and
transferred to either EP or RT for growout.
4.2.2 Broodstock Rearing Procedure in Earthen Ponds
Individuals were stocked into 0.8 ha earthen ponds (Figure 4.1a) in a commercial
shrimp farm owned by Beijing Shuishiji Biotech Ltd. at Wanning, Hainan. Initially,
PLs were stocked at a density of 25 individuals per m2 and fed with a commercially
protein. Feeding ratio over the first five months of the growout stage was
approximately 10% biomass initially, a level that was decreased at a steady rate to
2% of biomass by the end of the cycle. At the end of the five month culture period,
118 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
shrimp were collected (Figure 4.1b) at random and transferred to another earthen
pond and supplied with enhancement nutrition for three months to reach pre-
maturation stage. Management and feeding strategies during this time was almost
identical to that of the first growout stage, except that shrimp were also supplied with
fresh squid meal twice per week.
Figure 4.1 a) Earthen ponds for experimental broodstock trials; b) shrimp after five
month culture period (size of 20.0 ~ 25g); c) shrimp at eight months; d) packaged
broodstock in 10 L nylon bags (temperature at ~ 18 °C ) and transferred to hatchery
for reproductive traits test.
4.2.3 Broodstock Rearing Procedure in Recirculating Tanks
Shrimp were stocked into RT using the same standard procedure used in the
family growth parameter study as described in Chapter 3.
119 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
4.2.4 Design for Experimental Comparisons
When individuals had reached eight months of age (Figure 4.1c), broodstock
from EP/RT were collected at random and transferred (Figure 4.1d) to the hatchery
for acclimation in maturation tanks. Trials used four x 10 m2 RT with two replicates
per treatment. Mature males and females were reared separately at a stocking rate of
eight individuals per m2. Broodstock maturation management was the same as that
employed earlier (Chapter 3). Mature females from EP and RT were tagged with
individual numbered silicon eye rings (Starfish & Ltd.) for source identification and
then mixed together for the experiment. At 10 months of age, test females were
subjected to unilateral eyestalk ablation (Figure 4.2). Reproductive parameters for
females in both the RT and EP treatments were collected one month after eyestalk
ablation, and data recorded for 30 days. Females with mature ovaries (stage IV) were
collected daily at 10:00 AM and transferred to tanks with mature male broodstock.
At 19:00 PM, successfully mated females with attached spermatophores were placed
into individual 500L fibreglass tanks filled with 300L clean seawater. Spawning
environmental conditions were maintained at 28 ± 0.5 °C and a salinity of 32-36 ppt.
At 24:00, all females in the spawning tanks where eggs were released, were then
returned to their maturation tank. Eggs incubated with gentle aeration supplied. In
total, 107 RT females and 49 EP females broodstock were used for the estimation of
individual reproductive parameters.
120 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
Figure 4.2 Test females subjected to unilateral eyestalk ablation.
4.2.5 Evaluation of Reproductive Parameters
Reproductive performance was assessed for a series of standardized parameters
over 30 days. Survival of females under the two treatments (RT vs EP) was recorded
over the one month experimental period. After successful spawning, body weight
after spawning (BW) was estimated using a scale. Individual female fecundity was
measured using two methods; first, number of eggs (NE) per spawn was calculated
using a 200 ml beaker sub-sampling method with three replicates after eggs were
thoroughly homogenized in the spawning tanks with 300 L volume seawater.
Secondly, relative fecundity (FE) data were estimated as the number of eggs per unit
body weight after spawning. The number of nauplii per spawn (NN) was measured
using the approach as used for NE on the second day at 11:30 AM after nauplii had
hatched. Hatchability was measured as the percentage of hatching nauplii per
spawning event as (NN/NE) × 100%. Spawning frequency (SF) of each female was
examined after one month when the experiment had ended. Finally, the number of
121 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
successful spawning events was recorded for each surviving broodstock female at the
end of the experiment.
4.2.6 Statistical Analysis
Percentage data for HR were arcsine transformed prior to analysis (Zar, 1996).
Both percentage data (HR) and back-transformed data (HRat) were used and included
in further statistical analyses. Single factor one-way ANOVAs were performed to
compare reproductive parameters between treatments (RT vs EP). Variables
evaluated included: BW, NE, NN, HR, HRat, FE and SF. In addition, the interaction
between spawning frequency and rearing environment were investigated using a chi-
square test of independence. To evaluate the effect of body size in relation to
reproductive parameters, BW was first divided into three individual female size
classes namely: small (BW<38 g), medium (38-48 g), and large (>48 g) sized
individuals (the threshold for the three size categories was based on criteria of body
size for broodstock selection (see Introduction), with the objective to introduce a
variance for body weight of statistically significant differences among three sized
group within the two treatments.). Following this, BW group data were introduced as
a covariate in an ANCOVA analysis. To assess the quality of reproductive
performance in relation to ‘spawning order’, spawning events were also divided into
three groups, namely: first spawning event over the experimental period, second
spawning event over the experimental period, and multiple spawning events (three or
more) over the experimental period. The three groups for ‘spawning order’ were set
as a covariate in the ANCOVA analysis. Tukey’s post hoc means comparison was
used to assess significance differences between means after ANOVA analyses. Level
122 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
of statistical significance was set at P<0.05. All statistical analysis were performed
using SPSS 23 (IBM).
4.3 RESULTS
4.3.1 Reproductive Performance in Relation to Treatment (RT vs EP)
Means for reproductive parameters from broodstock females reared in the two
culture environments (RT and EP) over a one month test period are presented in
Table 4.1. No statistically significant difference was evident for female survival rate
between the two culture environments, 94% - RT vs 92% - EP, respectively. 136
spawning events for females were recorded in the RT treatment vs 101 spawning
events in the EP environment. In general, reproductive quality parameters namely;
mean number of eggs per spawn (NE), mean number of nauplii per spawn (NN),
mean hatchability rate of eggs per spawning event (HR), mean hatch rate per
spawning event number after arcsine transformation (HRat) and mean fecundity of
each spawning event (FE) were all not significantly different between treatments.
Mean values for NE, NN and HR for females from both RT or EP treatments were
within the range for optimal commercial nauplii production in China (pers. obs.).
Females in the EP treatment however, showed significantly higher spawning
frequency than females in the RT treatment (P<0.01).
123 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
Table 4.1 Comparison of reproductive performance (plus standard errors) of P.
vannamei broodstock reared in two different treatments: earthen ponds (EP) vs
recirculating tanks (RT). Bold type indicates a significant difference (p<0.05).
Of the 101 females in the RT treatment over the 30 day experimental trial,
approximately 30% of individuals did not spawn, while another 30% spawned only
once. 23% spawned twice and 17% spawned three or more times (Figure 4.1 a). In
the EP treatment, 20%, 30% and 14% of females did not spawn, spawned once, or
spawned twice, respectively. The proportion however, of multiple spawners (38%) in
the EP treatment was significantly higher (χ2(3, 0.05), 8.392, P = 0.039) than that in the
RT treatment (Figure 4.1 b).
Reproductive parameters
Broodstock sources
Recirculating tanks (107) Earthen ponds(49)
Number of spawn records 136 101
Survival rate per month 0.94 ± 0.02 0.92 ± 0.04
Body weight during spawning (g) 42.80 ± 0.81 39.67 ± 0.65
Number of eggs per spawning (x104) 23.34 ± 0.72 22.45 ± 0.67
Number of nauplii per spawning (x104) 19.85 ± 0.85 19.53 ± 0.83
124 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
A significant interaction was also evident for females undergoing multiple
spawning events between treatments (RT vs EP) and female body weight with (BW)
(RT (42.80 ± 0.81 g) vs EP (39.67 ± 0.65) (Table 4.2).
125 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
Figure 4.1 a) Pie charts showing the number of spawns for 101 female P. vannamei
broodstock in the recirculating tank treatment (RT) over a one month trial (SF,
number of spawning events); b) Number of spawns for 45 females in the earthen
pond treatment (EP) over a one month trial (SF, number of spawning events).
126 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
4.3.2 Effect of Body Size on Individual Reproductive Performance
Hatchery managers in general, currently suggest that body weight of female
broodstock of P. vannamei should be between 30 to 45g (Aquacop, 1983; Robertson
et al., 1993; Wyban and Sweeney, 1991). Mature female body weight at spawning
(BW) was divided into three bodyweight classes in the current study with 38 g and
48 g used as cut-off weights for dividing broodstock females into three size groups
represented by small (<38 g), medium (38-48 g) and large size broodstock female
bodyweight classes (>48 g). A statistically significant interaction was evident for
BW with multiple spawns among the three size classes in both the RT and EP
treatments (Table 4.2). SF for the same body size classes (small, medium, large)
between treatments (RT and EP) however, was not different (Table 4.2).
Large size class females in both treatments (RT and EP) produced significantly
more eggs per spawn than did females in either the small or medium size classes
(P<0.01). No statistically significant differences were evident for NN in either the
small or medium female size classes between treatments. In parallel, no significant
differences were evident for either HR or HRat comparisons, a result indicating that
egg hatchability was not impacted by individual size class of broodstock female
(Table 4.2). Even given a relatively higher mean BW for females in the RT
treatment and a tendency for more frequent spawning (SF), no statistically significant
difference was observed. One interesting result however, was that females in the
medium size class for BW in the EP treatment produced more than twice the number
of total spawning events (SF) compared with small and large size classes in the same
treatment (EP) (P<0.01). While no significant difference (P>0.05) was observed for
the effect of body size on FE for females in the RT treatment, small class females did
127 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
show a significantly higher FE (P<0.01) than medium class females in the EP
treatment (Table 4.2).
Table 4.2 Comparison of mean reproductive performance (plus standard errors) of
different size classes of female broodstock reared in earthen ponds (EP) vs
recirculating tanks (RT). Superscript letters indicate significant differences within
and between treatments (rearing conditions) for each reproductive parameter.
Reproductive
parameters
Recirculating tanks Earth ponds
Small n=46 Medium n=53 Large n=37 Small n=51 Medium n=39 Large n=11
BW < 38g BW, 38~48 g BW > 48 g BW < 38g BW, 38~48 g BW > 48 g
BW g 32.43 ± 0.52 a 43.45 ± 0.40 b 54.76 ± 0.83 c 34.47 ± 0.33 a 42.98 ± 0.51 b 52.04 ± 1.25 c
NE (x104) 18.19 ± 0.92 a 24.54 ± 1.06 bc 28.04 ± 1.39 c 22.52 ± 0.91 ab 20.78 ± 1.05 ab 27.99 ± 2.01 c
NN (x104) 15.00 ± 1.09 a 20.92 ± 1.32 ab 24.32 ± 1.74 b 19.93 ± 1.06 ab 17.55 ± 1.35 a 24.74 ± 2.74 b
SF monthly 1.00 ± 0.20 a 1.42 ± 0.21 a 1.60 ± 0.22 ab 1.29 ± 0.33 a 2.75 ± 0.35 b 1.25 ± 0.37 a
FE (x103)/g 5.61 ± 0.27 ab 5.68 ± 0.25 ab 5.16 ± 0.27 b 6.56 ± 0.27 b 4.88 ± 0.27 a 5.38 ± 0.37 ab
4.3.3 Reproductive Parameters in Relation to Spawning Order
A summary of reproductive parameters in relation to spawning order are
presented in Table 4.3. In general, the pattern shows that individual females that
spawned more eggs in both treatments possessed comparatively greater reproductive
quality compared with females that spawned only once or twice. The trend suggests
that multiple spawners show better reproductive quality in terms of mean number of
nauplii (NN: 22.61×104, RT; 22.72×104, EP) produced, and higher fecundity
128 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
estimates (FE: 6.26×103, RT; 6.23×103, EP), even though the pattern was not
significantly different.
Table 4.3 Comparison of mean reproductive parameters (plus standard errors) for
different spawn frequency (spawning once only (1), twice only (2), or three or more
times (3+)) for female broodstock reared in earthen ponds (EP) vs recirculating tanks
(RT). Superscript letters indicate significant differences within and between
treatments (rearing conditions) for each reproductive parameter.
For the current study, mean survival rate and means of reproductive parameters of
broodstock were very similar to that reported by commercial hatcheries in China.
Results suggest that our culture conditions were comparable and hence we had
provided near optimal maturation conditions for our experiment.
129 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
4.4.1 Comparative Reproductive Performance of Broodstock in the RT and EPTreatments
Overall, we observed very similar results for NE, NN, and HR reproductive
parameters in the two culture test environment treatments (RT vs EP). For spawning
frequency (SF) however, EP stocks spawned at a significantly higher rate than in the
RT treatment. Results for NE, NN and HR traits reported here were also much higher
than results reported in two earlier studies that compared reproductive parameters in
domesticated P. vannamei broodstock in tanks vs EP (Andriantahina et al., 2012b;
Otoshi et al., 2003). These differences may largely reflect use of different breeding
approaches (natural matings – the current study vs artificial insemination – published
studies). Mean NE, NN, and HR estimates reported here are closer to optima
proposed for P. vannamei broodstock performance standards for NE (20×104) and
HR (85%) (Zeigler et al., 2015). Our results are also consistent with another early
study that indicated that broodstock of P. vannamei reared in RT do not show
compromised reproductive parameters compared with wild spawned or pond reared
females (Otoshi et al., 2003). This difference is also reflected in reports of a shorter
inter-spawn period for pond vs tank housed females (Andriantahina et al., 2012).
Differences in SF for females in the RT vs EP treatments are in line with
observations made by some hatchery technicians in China who report that in general,
stocks reared in EP are easier to mature and show higher mating rates per night than
their SPF counterparts (reared in tanks).
Studies of other penaeid species have reported similar findings with culture stocks
trialled in EP, generally showing better reproductive performance compared with
those maintained in tank systems. This effect may be a result of differences in
environmental factors and incorporation of live food for nutrition in EP that may
130 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
enhance the productive performance of female shrimp. For P. esculentus, while
broodstock reared in ponds were sufficient for hatchery production in terms of
reproductive performance at a commercial scale, tank-reared females showed
significantly lower spawning rates and lower mean numbers of eggs per spawning
event (NE) (Keys and Crocos, 2006). This suggests that RT environments were
unlikely to be favoured for commercial scale hatchery nauplii production. For P.
monodon, a significant improvement in key reproductive parameters (NE, NN and
HR) was observed in females reared over five months in EP then transferred to RT
systems compared with females reared in RT systems over their complete rearing
period (Coman et al., 2013).
Improving the proportion of multiple spawners in a broodstock population has
been recognized as a key factor for optimizing nauplii production in penaeid species
(Coman and Crocos, 2003; Ibarra et al., 2007b; Racotta et al., 2003). In the current
study over a one month trial, a third of the RT-reared females did not spawn, and a
third only spawned a single time (Figure 4.1a). In contrast, EP-reared females
showed a significantly higher SF than that observed in the RT treatment, with only
20.00% failing to spawn, and almost 40% spawning three times or more (Figure
4.1b). This result is similar to those in a recent report from a commercial P.
vannamei nauplii hatchery in Mexico over a 36 day test period where 48% females
did not spawn, 18% spawned once, 15% spawned twice, while 19% spawned three
times or more (Arcos et al., 2003a). In another study of spawning record for 29 days
on 161 eyestalk ablated females, 44% did not spawn and 14% spawned four times or
more (Arcos et al., 2004). These findings together highlight that multiple spawners
are likely to only represent a relatively small proportion of the total female spawning
131 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
population but they do make a very significant contribution to total nauplii
production.
4.4.2 Impacts of Female Body Size on Reproduction Performance
Individual body size is the principle criterion widely used to select broodstock in
penaeid shrimp hatcheries. Results of examining the relationships between
reproductive parameters and individual body size here show clearly that female body
size has a significant impact on reproductive performance for the following traits;
NE, NN, SF, and FE, while there is little or no impact for HR or HRat.
There was a tendency for the large class females in both the EP and RT culture
environments in our study to produce higher NE or NN estimates than smaller
females. This result is also consistent with earlier studies in other penaeid shrimps
where fecundity (NE) has been correlated positively with individual spawner size
(Andriantahina et al., 2012b; Coman et al., 2013; Emmerson, 1980; Hansford and
Marsden, 1995; Marsden et al., 2013; Ottogalli et al., 1988; Palacios et al., 1998;
Peixoto et al., 2008; Wen et al., 2015). It is relatively difficult however, to directly
compare our results for HR with other studies because reported HR ranges vary
widely, particularly in early studies. It is worth noting however, that the HR
estimates reported here were all within the recognised current optimal range for
commercial nauplii production of P. vannamei stocks in China. HR is known to be
closely linked to the relative physiological condition of individual female broodstock
and management of the maturation environment in test tanks.
It is interesting to note that while larger females in general tended to show a
higher SF rate, females in the medium size class group in the EP treatment had a SF
mean of more than double that of small and large class females in the same treatment,
132 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
respectively. This phenomenon potentially could be explained by different strategies
for allocating energy among size classes. We hypothesise that, after maturation,
females in the EP medium size class likely directed more energy towards
reproduction rather than to allocating resources for further growth (i.e. growth rate
slowed and individuals spawned multiple times while those individuals in the larger
size class continued growing and spawned less frequently). It is likely that these
observed differences for SF in interactions between body size group and treatment
are also reflected in the different selection criteria used for P. vannamei broodstock
currently in the shrimp farming industry in China where large sized female SPF
individuals (raised in tanks) are considered better stock while small entrepreneur
hatcheries using their own culture lines prefer small and medium size class females
as broodstock. Our FE results add weight to this observation for a preference for
small and medium size class females in smaller entrepreneurial hatcheries, because
high FE of small-medium size results in greater egg production. This is because no
relationship was evident between body size and FE for females raised in an RT
environment whereas FE of small size females in the EP treatment was higher than
the other two size groups in this treatment. SF has also been reported to be positively
correlated with large female size and this size class for females also showed a higher
spawning frequency (Andriantahina et al., 2012b; Arcos et al., 2003a; Menasveta et
al., 1994; Palacios et al., 2000).
The minimum size of adult SPF females currently supplied to farmers in China
ranges from 35g to 45g. Threshold body size (38g) between small and medium size
classes in our study, in general accords well with the recommended size for P.
vannamei female stocks as breeders. In general, 30-45g individuals can be used for
nauplii production in a hatchery (Aquacop, 1983), even though some animal breeders
133 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
have advised use of even larger females of up to of 45g because they may perform
better (Robertson et al., 1993; Wyban and Sweeney, 1991).
4.4.3 Quality of Reproductive Performance in Relation to Spawning Order
It was clear from our results that no compromise was evident for NE, NN, HR, or
FE reproductive parameters in multiple spawners, or even that multiple spawners
were better in terms of mean NN or FE. Our results also support some earlier studies
that show offspring quality was not negatively impacted by spawning order for a
variety of key reproductive parameters including fecundity, fertilization rate,
hatchery, or biochemical variables that in general, reflect reproductive quality (Arcos
et al., 2004; Arcos et al., 2003a; Palacios and Racotta, 2003; Peixoto et al., 2004).
In contrast, a series of earlier studies reported that a deterioration in the
reproductive capacity of broodstock females can result from reproductive exhaustion
and that this is correlated with spawning order in penaeid species (Emmerson, 1980;
Hansford and Marsden, 1995; Marsden et al., 1997; Mendoza et al., 1997; Palacios et
al., 1999a; Palacios et al., 1999b). Differences for results between studies however,
may result from time factors. It is quite common for female penaeids to show a
decline in reproductive capacity under captive maturation conditions after unilateral
eyestalk ablation (Bray et al., 1990; Menasveta et al., 1993; Palacios et al., 1998;
Palacios et al., 1999b; Wyban, 1997). In general, experiments that test spawning
quality in relation to spawning order are undertaken over a relatively long time frame
(30-40 days). Reproductive data on multiple spawns as a result, are often collected
later over the experimental time period experiment than is data for ‘first order’ or
‘second order’ spawns. As a consequence, the time factor for measuring ‘exhaustion’
effects are very different and could significantly impact results between earlier
134 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
studies and more recent ones that have used natural spawning. In particular, if
maturation tank conditions were sub-optimal or diet had been insufficient to supply
adequate nutritional requirements.
In our study, data were collected during the second month after a female had
experienced unilateral eyestalk ablation, so production of nauplii occurred over a
stable period. Furthermore, mortality rates of broodstock and estimates of
reproductive parameters in the current study indicate that near optimal maturation
conditions were supplied to the broodstock tested. As a consequence, this would
likely minimise any impacts of test time on potential for reproductive exhaustion.
Again, this highlights the difficulties with dealing with domesticated penaeid
broodstock studies and how to establish the best, uniform standard experimental
conditions that will allow meaningful comparisons to be made between different
studies.
135 Chapter 4: Comparison of Reproductive Performance of Domesticated P. vannamei FemalesReared in Recirculating Tanks and Earthen Ponds: An Evaluation of Reproductive Quality of Spawnsin Relation to Female Body Size and Spawning Order
4.5 CONCLUSIONS
In conclusion, results here indicate that no significant differences were evident for
the majority of reproductive performance traits tested between female Pacific white
shrimp broodstock reared in RT vs EP environments. Females in the EP treatment
however, produced more nauplii per individual than females raised in an RT
environment and this resulted from a significantly higher SF rate while no evidence
was observed for reproductive exhaustion related to the number of consecutive
spawns. Nauplii production in hatcheries therefore, potentially can be optimized by
employing different strategies in relation to female broodstock body size selection.
When RT-reared stocks are used, selecting larger body size females should result in
higher nauplii production levels, while for small-scale farmers who use EP-reared
stocks, use of female broodstock in the medium size class range should maximize
nauplii production. In the next chapter, the additive genetic components for
reproductive traits examined here in our culture line will be assessed to decide if
these traits can be improved via genetic selection to maximise reproductive output in
cultured P. vannamei stocks in China.
137 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
Chapter 5: Quantitative Genetic Assessment of Female
Reproductive Traits in a Domesticated Pacific White Shrimp (Penaeus
vannamei) Line in China
138 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
ABSTRACT
Earlier (Chapter 4), we reported that optimising P. vannamei broodstock
reproductive performance in combination with selection for large body size could
improve productivity of females in two different culture environments. In parallel,
seed production can be improved if genetic selection is applied to key reproductive
traits when substantial additive genetic variation is present that could be exploited in
a breeding program. Despite the commercial importance of reproductive traits to the
seed production sector, to date few quantitative genetic studies have been conducted
on these traits in farmed penaeid shrimp culture lines. Here, we investigated genetic
parameters for some important reproductive traits that directly impact nauplii
production in Pacific white shrimp hatcheries in China. Our objectives were to
improve broodstock reproductive quality, and to anticipate any potential impacts on
reproductive performance when selecting for increased body weight by assessing
genetic correlations between post-spawning body weight and specific female
reproductive traits. Data were collected on 595 females from 78 fullsib families over
30 days, with a total of 1,113 spawning events recorded. Traits studied included:
body weight after spawning (WAS), number of eggs per spawn (NE), number of
nauplii per spawn (NN), egg hatching rate per spawn (HR), number of eggs produced
relative to female weight (g) (FE), and spawn frequency over 30 days (SF).
Heritability estimated high for WAS (h2 = 0.64 ± 0.10) and moderate for NE (0.26 ±
0.07), NN (0.18 ± 0.06), and SF (0.15 ± 0.06), respectively. On contrast, h2 for HR
(0.04 ± 0.03) and FE (0.05 ± 0.04) were low. The genetic correlations between
growth trait (WAS) with NE, NN and SF were 0.93 ± 0.10, 0.84 ± 0.10, and 0.57 ±
0.18, respectively. While the genetic correlation between WAS and HR was low
(0.02 ± 0.33), a negative genetic correlation was found between WAS and FE (-0.50
139 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
± 0.27). Overall, we concluded that it is possible to improve key female reproductive
traits (i.e. NE, NN, and SF) in cultured white shrimp lines via genetic selection, but it
was unlikely for HR or FE. The genetic relationship between growth rate and key
female reproductive traits indicates that selection for fast growth can in parallel
enhance production in the seed sector, with little or no compromise on egg quality.
140 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
5.1 INTRODUCTION
Reproductive characteristics constitute a set of commercially important traits that
are yet to receive much attention when genetic improvement is applied to farmed
aquatic species (Gjedrem, 2012; Gjedrem and Rye, 2018; Nguyen, 2016). This is
especially true for penaeid shrimp species that possess many unique reproductive
characteristics; in particular, at maturation in hatcheries, many females may spawn
relatively infrequently or may never spawn, while a small proportion of females
spawn multiple times, hence these females are likely to contribute the majority of
nauplii produced (Arcos et al., 2003a; Benzie, 1997; Ibarra et al., 2007b).
Developing knowledge about genetic parameters (heritability and genetic
correlations) for key reproductive traits will be essential for designing better breeding
strategies and for improving broodstock reproductive capacity via genetic selection.
For example, a recent experimental study of a small marine copepod crustacean
(Parvocalanus crassirostris) reported that total egg production was increased by
24.5% following selection over five generations with heritability (h2) for this trait
estimated at 0.38 (Alajmi et al., 2014).
While application of genetic selection methodologies have increased productivity
significantly in a number of farmed aquatic species (De Verdal et al., 2018; Gjedrem
et al., 2012), most breed improvement programs have focused primarily on growth
traits (Gjedrem and Rye, 2018; Janssen et al., 2017; Ren et al., 2018; Sae‐Lim et al.,
2016). There are many examples of selection for improving growth traits however,
that have also reported undesirable correlated effects on other fitness traits in
domesticated animals, particularly with respect to metabolic, reproductive and health
status traits (Rauw et al., 1998). Significant evidence for correlated negative effects
141 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
have been reported in poultry and livestock animals. As an example, negative
correlations were reported in pigs for a number of reproductive traits when
individuals were selected for high meat production efficiency. This resulted in a
significant reduction in both fertility and litter size (Rauw et al., 1998). In contrast,
studies in a wide variety of animal species (including model species like mice,
rabbits, dogs, sheep, pigs and fish) have reported positive relationships between
growth traits and some reproductive traits (e.g. litter size and fecundity) (Bünger et
al., 2005). While the practice of genetic selection works essentially via a ‘black box’
approach, the opportunity still remains to understand, to anticipate and to prevent any
potentially negative impacts of selection based on developing knowledge about
genetic correlations (rg) between growth traits and important reproductive traits
(Rauw et al., 1998).
To date, few studies have investigated genetic parameters for reproductive traits in
farmed aquatic taxa, and of those studies that have been conducted, most have
focussed on farmed salmonid species (Gall and Huang, 1988; Gall and Neira, 2004;
Hao and Chen, 2008; L'Abée‐Lund and Hindar, 1990; Neira et al., 2006; Su et al.,
1997) and tilapias (Thoa et al., 2017; Trọng et al., 2013a; b). In salmonid species, the
majority of productive traits that have been investigated show relatively high levels
of additive genetic variance with most genetic correlations indicating positive
relationships between growth-related traits and the reproductive traits examined. As
an example, Gall and Neira (2004) reported h2 estimates for weight, number of green
eggs, and number of eyed eggs in coho salmon (Oncorhynchus kisutch) ranging
between 0.32 to 0.42. In domesticated rainbow trout (Oncorhynchus mykiss), h2
estimates for female egg number and egg size were 0.32 and 0.28, respectively (Gall
142 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
and Huang, 1988), while h2 estimates for traits including spawning date, egg size,
number of eggs, and egg volume were all moderate to high (Su et al., 1997). In
contrast in a Nile tilapia (Oreochromis niloticus) breeding program in Vietnam, h2
estimates for both fecundity traits and fertility were low (Trọng et al., 2013a), while
spawning success (h2 = 0.20-0.22) was an exception (Trọng et al., 2013b). Likewise,
Thoa et al. (2017) reported low h2 estimates for number of fry at hatching, total fry
weight, and fry mortality in a base population of red tilapia (Oreochromis spp.)
(Thoa et al., 2017). In addition, a case study that applied selection on reproductive
traits in female channel catfish to produce hybrid catfish embryos reported realized
h2 estimates for fecundity ranging from 0.10 to 0.42 while h2 for percentage hatch
and fry/kg were low (Gima et al., 2014).
While a number of studies have demonstrated the potential capacity to improve
seed production via genetic selection (Arcos et al., 2003a; Benzie, 1997; Ibarra et al.,
2007b), there have been few studies published on quantitative genetic analyses of
reproductive traits in penaeid shrimps. In particular, there has been a noticeable lack
of studies that have investigated capacity of females to spawn multiple times. This
trait has the potential to double or even triple nauplii production by individual
females. Until recently however, there has only been a single report that has
investigated genetic parameters for this trait in penaeids (Ibarra et al., 2005).
Reproductive traits in farmed penaeid species (specifically fecundity-related traits),
are among those that can contribute the most to increasing profitability of the
hatchery sector but to date, only a single study has reported h2 estimates for the above
traits under commercial hatchery conditions (Arcos et al., 2004). While there have
been two studies estimating genetic parameters for reproductive traits in Pacific
white shrimp (P. vannamei) females following artificial insemination (Caballero‐
143 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
Zamora et al., 2015; Tan et al., 2017b), more information will be required for the
seed production sector because these studies indicate that there are differences in
reproductive trait data when applying artificial insemination vs natural mating in
penaeid species. Moreover, high levels of additive genetic variance in reproductive
traits for mean oocyte number, diameter, ovary maturity stage (Arcos et al., 2005b),
and high genetic correlations between levels of vitellogenin in haemolymph and
mean diameter of oocytes have been reported for Pacific white shrimp (Ibarra et al.,
2009). In black tiger shrimp (P. monodon), h2 estimates for days to spawn, number of
eggs, number of nauplii, and hatching rate ranged from 0.18 to 0.47 indicating that
these traits can be improved to increase individual reproductive output via genetic
selection (Macbeth et al., 2007).
The aims of the current study therefore were: (1) to examine genetic variance for
key reproductive traits in female P. vannamei broodstock under commercial
maturation conditions using the base line developed in a complete 4×4 diallel cross,
and (2) to estimate genetic correlations between individual body weight after
spawning with specific reproductive traits. Knowledge generated in this study can be
applied to improve production of nauplii via genetic selection and to minimize any
negative impacts on reproductive performance that may result from selection for fast
growth.
144 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
5.2 MATERIALS AND METHODS
5.2.1 Experimental Families
The experiment was conducted at a Beijing Shuishiji Biotech. Ltd. maturation
facility in Wanning (Hainan Province), China. A base family line of P. vannamei
was produced following a complete 4×4 diallel cross of four domesticated culture
lines obtained in China between June and July 2015. Details of the strains and
maintenance of families generated are described in Ren et al. (2018) and in Chapter 3.
When shrimp had reached maturation stage (age of 10 month) 660 test animals were
randomly selected from each family (in total 78 families, for details of family
production see Chapter 3 Methods Section) and individuals transferred to maturation
tanks (10 m2). Females were then subjected to unilateral eyestalk ablation to induce
ovary maturation, and broodstock males and females reared separately with 110
individuals per tank.
Number-coded silicon eye rings were tagged on the remaining eyestalk for
individual female identification. In total, data from 595 females representing 78 full-
sib families were recorded in the study. A biological water recirculating system was
used to maintain water quality at an exchange rate of 600% to 800% per day. Daily
feed composition consisted of a mixture of fresh meal diet (50% polychaetes, 30%
squid and 20% mussels) delivered at a rate of approximately 5% of total biomass per
tank. Tank water was maintained at 28 ± 2 °C and 31-35 ppt salinity. Data collection
commenced on Jun 3rd and continued over 30 days.
5.2.2 Measurement of Reproductive Traits
Females with mature ovaries (stage IV) were collected daily at 10:00 and
transferred to tanks containing male broodstock. At 19:00, successfully mated
females with attached spermatophores were placed in individual 500 L fibreglass
145 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
tanks filled with 300 L of clean seawater. Environmental conditions for spawning
were maintained at 28 ± 0.5 °C at a salinity of 32-36 ppt. At 24:00, all females in
spawning tanks were returned to maturation tanks and eggs remaining in the
spawning tanks were incubated with gentle aeration.
Body weight after successful spawning (WAS) was measured at 12:00. Five
reproductive traits were measured and recorded: i) number of eggs (NE) per spawn
measured in three replicate 200 ml breaker samples after being thoroughly
homogenized in seawater taken from the spawning tanks (300 L in volume); ii)
number of nauplii (NN) per spawning measured in the same way as for NE after
nauplii had hatched on the second day at 11:30; iii) relative fecundity, measured as
the number of eggs per unit body weight (FE), calculated by dividing the number of
eggs per spawn by WAS; iv) egg hatching rate per spawn (HR); and v) female
spawning frequency (SF), calculated as the total number of successful spawning
events per individual female at the end of the experimental period.
5.2.3 Statistical Analysis
Prior to formal data analysis, raw HR percentage data were arcsine transformed
(Zar, 1996). For comparative purposes, both the original data (HR) and arcsine
transformed data (HRat) were used in subsequent statistical analyses. For all other
traits (WAS, NE, NN, FE and SF), untransformed values were used in the data
analysis. Genetic variance and covariance components and h2 for targeted traits were
estimated using an animal model applying the restricted maximum likelihood
(REML) methodology in WOMBAT (Meyer, 2007). For SF, the linear mixed animal
model can be written as follows:
y = Xβ + Zα + e, (1)
146 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
Where:
y is a vector of observations for a reproductive trait per spawn event (WAS,
NE, NN, HR, HRat, FE, or SF),
β is a vector of fixed effects consisting of maturation tanks, the recorded
spawn event batches, and regression coefficient of age of shrimps,
α is the vector of random additive genetic effects of the animals,
e is the vector of random residual errors and
X and Z are known incidence matrices relating observations to the fixed effects
mentioned above, and animal effects, respectively. Both a and e follow a normal
distribution with mean zero and variance Aσa2 and Iσe2, respectively. Here, σa2 and
σe2 are the additive genetic and residual error variances while A is the numerator
relationship matrix based on pedigree information.
In this experiment, estimates of WAS, NE, NN, HR, HRat and FE resulted from
multiple observations for some female individuals. Repeated measures for these traits
were across 30 days, with these data recorded from different spawning order, i.e. first
spawn, second spawn, up to sixth spawn for some females. Therefore, we also used a
repeatability animal model (Model 2) to account for replication within individuals:
y = Xβ + Z1α + Z2pe + e, (2)
Where:
y is a vector of observations for a reproductive trait per spawn event (WAS,
NE, NN, HR, HRat, or FE),
β is a vector of fixed effects consisting of maturation tanks, the recorded
147 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
spawn event batches, and regression coefficient of age of shrimps,
α is the vector of random additive genetic effects of the animals,
pe is the vector of random maternal permanent environment effects
contributed by individual females to their offspring families,
e is the vector of random residual errors and
X, Z1 and Z2 are known incidence matrices relating observations to the fixed effects
mentioned above, animal effects and permanent environment effects, respectively. a,
pe and e follow a normal distribution with mean zero and variance Aσa2 , Iσpe2 and
Iσe2, respectively. Here, σa2, σpe2 and σe2 are the additive genetic, permanent
environment and error variances while A is the numerator relationship matrix based
on pedigree information.
In Model (1), h2 was calculated as the ratio of additive genetic variance to total
phenotype variance (h2= σα2/ σp2). In Model (2), heritability (h2) and repeatability
(rep) were calculated as follows:
�2 �σ�2
σ�2 � σ�鍘2 � σ鍘2� �鍘� �
σ�2 � σ�鍘2
σ�2 � σ�鍘2 � σ鍘2
In the primary analysis, the results were similar between the linear mixed animal
model and the repeatability animal model, which were also supported by comparing
the log-likelihood ratio of the models. Therefore, a multivariate mixed animal model
was fitted to estimate the genetic and phenotype correlation between examined traits,
expressed in matrix notation as:
148 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
�����th�tt�th�th��t
� �� � �� � � � (3)
Where yWAS, yNE, yNN, yHR, yFE and ySF are the same as defined in Model
1, respectively. Total phenotypic variance (σp2) was estimated as the sum of additive
animal genetic variance (σα2) and random residual components (σe2). Genetic or
phenotypic correlation between two traits was calculated as: � � �12
�12 �2
2where σ12
was the genetic or phenotypic covariance between two traits, and σ12 and σ22 were
either additive genetic variances of trait 1 and 2, or phenotypic variances of the two
traits, respectively.
5.3 RESULTS
5.3.1 Descriptive Statistics
Over the 30 day experiment, 950 successful spawning records were observed out
of a total of 1113 spawning events. The number, mean values, minimum and
maximum values, standard deviations and coefficients of variation for each trait are
presented in Table 5.1. Mean body weight after spawning (WAS) was 39.66 ± 8.44 g
(ranging from 18.19 to 70.63 g). Mean number of eggs (NE, 225.15 × 103) and
number of nauplii (NN, 194.63 × 103) per spawning were in general terms,
comparable with that obtained during commercial Pacific white shrimp nauplii
production in China (pers. obs.). The high mean hatch rate (HR = 84.60%) indicates
149 Chapter 5: Quantitative Genetic Assessment of Female Reproductive Traits in a DomesticatedPacific White Shrimp (Penaeus vannamei) Line in China
that the experimental broodstock management protocols employed had been
appropriate.
Table 5.1 Descriptive statistics of reproductive traits for female Penaeus vannamei.
WAS = Weight after spawning, NE = Number of eggs per spawning, NN = Numberof nauplii per spawning, HR = egg hatching rate %, HRat= Arcsine transformed HR,FE = Number of eggs per unit weight, SF = Number of spawns during the 30 dayexperiment period.
Traits Unit N Mean Minimum Maximum Standarddeviation
Coefficientvariation (%)
WAS g 947 39.66 18.19 70.63 8.44 21.27NE ×103 949 225.1