Optimization of silver scat population breeding strategy and germplasm resource improvement based on genetic algorithm

In the sea, there are various kinds of fishes, which adopt unique reproductive strategies to ensure the survival of their offspring, and improve the germplasm resources suitable for fishes to ensure the excellence of their offspring [1-2]. By studying the reproductive strategies and germplasm improvement of marine fishes, scientists can better understand the working principle of marine ecosystems and provide important basis for the conservation of marine living resources.

Cooperative breeding behavior Certain marine fishes adopt cooperative ways to reproduce and form breeding groups, and this cooperative breeding behavior provides a better guarantee for the survival of juvenile fishes [3-4]. And the reproductive strategies of marine fishes are multiple reproduction, reproductive competition, and solitary reproduction [5]. In order to increase the number of offspring and survival rate, most marine fishes reproduce multiple times to spawn, and this strategy allows negative species to focus on reproduction during the reproductive season and increase the possibility of reproductive success [6]; during the reproductive season, males often compete fiercely with each other for the attention of females and the opportunity to pair up, and individuals with the best qualities tend to have more opportunities to reproduce [7-8]; a few marine fishes reproduce by solitary reproduction, in which females lay eggs without fertilization and are able to hatch them directly into juveniles, and this type of reproduction is extremely important for fishes that lack suitable breeding partners [9]. However, there are many factors affecting fish reproduction, such as anthropogenic, weather, aquaculture environment, probiotics and other inputs, all of which can have an impact on the quantity and quality of reproduction [10-13]. Germplasm resources refer to all the genetic information of a certain biological species, including all varieties, subspecies, populations of the species, and genetic resources of related wild and closely related species [14]. The abundance of germplasm resources is the material basis for securing agricultural production, so germplasm improvement came into being to obtain quality germplasm resources and preserve them in the future. It is through the selection of good varieties and breeding improvement of livestock and poultry germplasm resources to maintain the diversity of livestock and poultry populations and the sustainable use of genes, which can increase the yield profit quality of livestock and poultry and improve their ability to adapt to the environment [15].

Silver drum is a freshwater fish widely distributed in freshwater environments and is highly favored by consumers. Therefore, silver scat culture has become an aquaculture practice with high economic benefits. However, with the development of social economy and the improvement of people’s living standard, there is an increasing demand for food safety and high quality health food. The aquaculture industry also faces problems such as stock breeding and germplasm resource improvement, and genetic algorithm can provide technical support for these problems, which searches for the optimal solution by simulating the natural evolutionary process [16]. By breeding fish stocks and improving germplasm resources in order to realize the sustainable development of fisheries [17].

In this paper, with reference to relevant studies on silver scat population breeding, we selected flow velocity, water depth, water temperature and light duration as the main environmental factors in the optimization of silver scat population breeding strategy, and constructed an optimization model of population breeding environment suitability based on the key factors. The multi-objective optimization NSGA-II algorithm was used for solving the model of environmental suitability of silver drum population breeding, which solved the multi-objective optimization problem in silver drum population breeding strategy and germplasm resources improvement. Subsequently, a normally distributed crossover operator and an adaptive variational operator were introduced into the traditional NSGA-II algorithm to improve the algorithm’s optimization search ability, and an SDR was introduced to replace the Pareto dominance relation in the improved algorithm, which enabled the improved genetic algorithm to solve the optimal strategy for silver scat population breeding. In this study, after solving the optimization model of silver drum breeding using genetic algorithm, the optimal breeding strategy obtained from the solution was used as the experimental group, and the existing breeding strategy was used as the control group to carry out artificial breeding experiments of silver scat. The growth indexes and gonadal development of silver drum parents in the breeding process reflect the implementation effect of the optimal breeding strategy, and then based on the growth indexes of the hatched juveniles to explore the effect of germplasm resources improvement.

2

Multi-objective optimization model for silver drum stock breeding

2.1

Selection of key environmental factors

The silver drum, which inhabits soft river bottoms or shallow areas with abundant aquatic plants, is a batch spawning fish. The spawning and juvenile development period of silverfish occurs from March to May each year, which is the most vulnerable and important stage in the life cycle of the fish. Usually, the population breeding environment with sufficient light and dense water plants can meet the basic conditions for silver drum population breeding. By analyzing the habitat factors affecting the breeding activities of the silver scat, four important indicators, namely, flow rate, water depth, water temperature and light, were selected based on the review of relevant literature, centering on the breeding period (including the spawning period of the parent fish and the development of the fry) using the method of expert empirical judgment. 1)

Flow velocity

Silver drum can spawn in all kinds of water bodies, but prefer slow flowing or still water that is overgrown with aquatic plants. Silver scats produce viscous eggs, and the incubation time of the eggs is about 3-7 days. After spawning and hatching, the fry need to have a stable water flow for 3-4 days. The flow rate should not be too high during the development of the fry, and the suitable range is 0.2-0.4 m/s.

2)

Water depth

Because too shallow and too deep water depth is not suitable for the movement of silver scat, the spawning place of the parent fish should be generally deeper than 1m, the development of the fish fry water depth range of 1-2m, and at the same time require that the water level can not fluctuate too much, so that it is more conducive to the normal growth of the fish fry.

3)

Water temperature

Water temperature is an important environmental factor that changes with time and space, directly affecting the survival, growth, reproduction, and other physiological activities of fish. Water temperature is a crucial factor in the process of fish spawning, which determines the occurrence and duration of fish spawning. The breeding water temperature range of silver drum is 22-26°C, and the best suitable water temperature is 22-24°C. A water temperature of 22°C is necessary for the normal hatching of fish, and when the water temperature is greater than 22°C, the hatching rate, survival rate and emergence rate will be greatly increased.

4)

Light

For fish, light is a necessary condition to stimulate fish to spawn. A certain amount of light can promote parental reproduction, reduce the phenomenon of difficult spawning, shorten the time of spawning and fertilization, and also benefit the quantity, rate, and development of fertilized eggs.

2.2

Optimization model of population breeding strategy

Stock breeding environment suitability HSI is an index that evaluates the suitability of the breeding environment for fish stocks, reflects the habitat quality of the breeding environment, and is the biological basis for stock breeding environment modeling methods. Population breeding environment suitability curve is a curve to quantify the relationship between species reproduction and its habitat influencing factors, often determined by the frequency of occurrence and distribution of species, and the value of 0~1 is used to indicate the degree of suitability of species reproduction to the environmental factors, which is used to quantitatively reflect the preference of species for the breeding environment of the population.

Considering the different needs of silver drum parents and fry for each habitat factor during the breeding period, we obtained the suitable conditions for silver drum reproduction by reviewing the literature, and constructed formulas for the spawning suitability HSI_pf of parents based on the flow rate, water depth, water temperature, and light duration, respectively. In this paper, the comprehensive suitability index HSI for silver drum population reproduction and germplasm resource improvement based on the arithmetic mean method was expressed as: (1) $H S I = \frac{S I_{D} + S I_{S} + S I_{T} + S I_{L}}{4}$

where HSI is the composite stock breeding environment suitability for artificial stock propagation of silver drum, and SI_D, SI_S, SI_T, SI_L is the suitability index for flow rate, water depth, water temperature and light duration, respectively.

3

Genetic algorithm-based approach to solving multi-objective models

3.1

Multi-objective optimization theory

In the optimization of breeding strategies and germplasm resource improvement in silver drum populations, multi-objective optimization involves the simultaneous pursuit of optimal solutions for multiple objectives, while single-objective optimization focuses on the optimization of one objective. Realistic breeding problems often involve multiple mutually constrained objectives, which increases the complexity of the problem. Such problems can be described by the following mathematical model: (2) ${\begin{array}{l} V - m i n f (x) = {[f_{1} (x), f_{2} (x), \dots, f_{p} (x)]}^{T} \\ s . t . x \in X \\ X \subseteq R^{m} \end{array}$

Where V-min minimization vector, that is, each objective function in the optimization objective tends to minimize. x is the decision vector, X is the decision space formed by the decision vector, f_p(x) is the objective vector, f(x) is the objective space formed by the objective vector, and p represents the number of objective functions.

In the multi-objective optimization problem solving process, at the same time to make all the objectives at the same time to achieve the optimal state does not exist, which is the essence of multi-objective optimization, can only be in the trade-offs and compromises between the objectives, as far as possible, to make all the objective functions converge to the optimal, this time the solution is the Pareto optimal solution, also known as the non-dominated solution, for a multi-objective optimization problem, there are usually more than one For a multi-objective optimization problem, there are usually more than one Pareto optimal solution.

The advantages of the NSGA-II algorithm for multi-objective optimization [18] are mainly reflected in the efficient handling of multi-objective problems. NSGA-II is designed to solve multi-objective optimization problems, and it is capable of considering multiple conflicting objectives at the same time and finding the best compromise between them. The algorithm ensures diversity of solutions while quickly converging to the Pareto frontier, i.e., finding a set of solution sets that cannot bias each other by any single objective improvement in the objective space, through the principles of nondominated sorting and congestion comparison. Due to the stochastic nature of genetic algorithms, NSGA-II is not prone to falling into local optimal solutions, which enables it to explore a wider solution space and find more comprehensive solutions. Therefore, in this paper, NSGA-II (Non-dominated Sorting Genetic Algorithm) was selected to solve the multi-objective optimization problem in silver drum population breeding strategy and germplasm resource improvement.

3.2

Improved NSGA-II/SDR Algorithm Establishment

In order to enhance the local search ability of NSGA-II algorithm, the I-NSGA-II algorithm is constructed by carrying out certain optimization of the crossover operator and the mutation operator in the genetic operation to improve the performance of the algorithm in order to solve the multi-objective optimization problems more effectively.

3.2.1

Cross and Variable Operator Improvements

The traditional simulated binary crossover operator (SBX) [19] has a problem because the search space is small and it is easy to fall into local optimal solutions. To solve this problem, by introducing a normally distributed crossover operator, individuals have a greater degree of variability during the crossover process, which helps to increase the diversity of solutions reducing the occurrence of local optimal solutions. Therefore, this improved method can improve the global search ability of the algorithm and thus achieve better optimization results.

The SXB method is formulated as follows: (3) $C_{1 / 2} = \frac{(P_{1, i} + P_{2, i})}{2} \pm ψ \frac{(P_{1, i} - P_{2, i})}{2}$

By invoking the evolutionary formula: (4) $X^{'} = X + Y \cdot N (0, 1)$

If $\frac{(P_{1, i} - P_{2, i})}{2}$ in the original SBX crossover operator is denoted as the search step and the idea of normal distribution is applied to the crossover operator, the improved crossover operator can be expressed as: (5) $C_{1 / 2} = \frac{(P_{1, i} + P_{2, i})}{2} \pm 1.481 \frac{(P_{1, i} - P_{2, i})}{2} \cdot | N (0, 1) |$

where P_1,i, P_2,i the corresponding variables on the two parent chromosomes, and |N(0, 1)| represents a random number uniformly distributed on (0, 1).

Considering the evolutionary characteristics of the NSGA-II algorithm, an adaptive mutation operator is designed, which can accelerate the search speed of the algorithm by appropriately increasing the mutation probability with the following formula: (6) $P_{m} = {\begin{matrix} P_{m \max} \\ P_{m \max} - \frac{(P_{m \max} - P_{m \min}) (f^{*} - f_{a v g})}{f_{\max} - f_{a v g}} \end{matrix}$

Where f_max represents the value of the optimal solution, f_avg represents the average fitness value of the population, and f^* represents the degree of superiority or inferiority of the selected individuals in solving the problem. The fitness value measures how well an individual adapts in the problem domain, a higher fitness value indicates that the individual is better adapted in the current environment, while a lower fitness value indicates that the individual is worse adapted in the current environment.

3.2.2

NSGA-II/SDR Algorithm

In the dominance ordering stage, SDR [20] is introduced in this paper to replace the Pareto dominance relation in the above improved NSGA-II algorithm to improve and optimize the NSGA-II algorithm, which is called NSGA-II/SDR algorithm. The specific operation flowchart of the NSGA-II/SDR algorithm is shown in Fig. 1. The basic operation principle of the NSGA-II algorithm with SDR is to generate the initial population after selection, crossover and mutation steps to mix the offspring and parent populations, compute the target bearing values of each individual solution for normalization and then compute the target vectors of the solution of each individual, compute the angle between it and the neighboring target vectors, and then select the Nth largest angle as the size of the small habitat after sorting the angles. θ, from which a selection operation is carried out based on the SDR dominance relation to generate the next generation of populations, with other steps such as crowding distance remaining, and so on until the termination condition is satisfied. 1)

Initialize the population

Randomly generate an initial parent population N(t) with a population size of N. Sort all individuals of the parent population according to the SDR.

2)

Chromosome coding method

Real number coding is used. For a problem involving n variables, the chromosome is [x_1j, x_2j, x_3j, ..., x_nj], and each x_ij is a real number, where i = (1, 2, 3, …, n), j is the pattern number of the process numbered i.

3)

Selection operator

The tournament selection is used, and the tournament size is set to 2 in each round, and the individuals of tournament size are randomly selected from the population, from which the best-adapted individuals are selected to proceed to the next operation. Repeat the above selection process until the next generation is filled.

4)

Crossover operator

Analog binary crossover is used. Let the two parent individuals be X¹(x₁₁, ..., x_n1), X²(x₁₂, ..., x_n2), by Eq: (7) ${\begin{array}{l} C_{i 1} = 0.5 \times [(1 + β) \cdot x_{i 1} + (1 - β) \cdot x_{i 2}] \\ C_{i 1} = 0.5 \times [(1 - β) \cdot x_{i 1} + (1 + β) \cdot x_{i 2}] \end{array}$

Two offspring individuals are obtained C¹(c₁₁, ..., c_n1), C²(c₁₂, ..., c_n2). Where β is dynamically randomly determined by the distribution factor η as shown in Eq: (8) $β = {\begin{matrix} {(r a n d \times 2)}^{\frac{1}{(1 + η)}} & r a n d \leq 0.5 \\ {(\frac{1}{(2 - r a n d \times 2)})}^{\frac{1}{(1 + η)}} & o t h e r w i s e \end{matrix}$ 5)

Variation operator

Polynomial variation is used. Select random number μ_i ∈ [0, 1) and calculate the value of β_i. The calculation is shown in Eq: (9) $β = {\begin{matrix} {(2 μ_{i})}^{\frac{1}{(1 + η_{μ})}} & μ_{i} \leq 0.5 \\ 1 - {[2 (1 - μ_{i})]}^{\frac{1}{(1 + η_{μ})}} & μ_{i} > 0.5 \end{matrix}$

η_u is set to 20. Calculate the subgeneration as shown in Eq: (10) $x_{i 1}^{t + 1} = x_{i 1}^{t} + β_{i}$

4

Optimization model solution for silver drum stock reproduction

According to the multi-objective optimization model of silver drum population breeding constructed in the previous paper, combined with the improved multi-objective optimization algorithm NSGA-II to solve the model, set the population size of 100, the maximum number of iterations is 200, the crossover probability of 0.8, the probability of variance of 0.2, in MATLAB to solve, after the use of the Euclidean distance function to find out the Euclidean distance value of different silver drum population breeding strategies, as a criterion to evaluate the goodness of different population breeding strategies, according to the actual data of population breeding environment parameters using Matlab on the breeding environment optimization model solution results are shown in Table 1. After solving, the Euclidean distance function was used to find the Euclidean distance values of different breeding strategies of silver drum, which was used as a criterion to evaluate the goodness of different breeding strategies of different populations, and the results of the optimization model of the breeding environment using Matlab according to the actual data of the breeding environment parameters of the populations are shown in Table 1. After executing the optimization program, there are 11 population breeding optimization strategies in the Pareto-optimal solution set, of which 11 breeding strategies are all in line with the Pareto-optimal, and they are all the optimal values under the optimization model and constraints of the silver drum population breeding environment parameters, and one of the 11 solutions can be selected according to the actual demand. Considering the cost and the feasibility of environmental parameter regulation in silver drum population breeding, this paper determines that breeding strategy 5 is the best optimization scheme for silver drum population breeding strategy, which can ensure the accurate regulation of light and flow rate factors. That is, the temperature, flow rate, water depth and light duration in the optimized breeding strategy of silver drum population were set to 22.5°C, 0.21m/s, 1.7m and 15h, respectively.

Table 1.

The results of the pareto optimal solution set analysis

Strategy	Temperature	Current velocity	Water depth	Luminous duration	Euclidean distance
1	24°C	0.38m/s	1.04m	12h	49.96
2	26°C	0.23m/s	1.32m	19h	46.88
3	22°C	0.22m/s	1.41m	14h	41.52
4	22.5°C	0.32m/s	1.08m	15h	39.3
5	22.5°C	0.21m/s	1.7m	15h	43.47
6	22.9°C	0.31m/s	1.1m	12h	41.76
7	25.5°C	0.26m/s	1.84m	19h	47.68
8	22.2°C	0.23m/s	1.57m	15h	44.7
9	24.2°C	0.34m/s	1.36m	12h	56.02
10	22°C	0.33m/s	1.3m	14h	43.82
11	23.1°C	0.31m/s	1.25m	19h	39.27

5

Analysis of stock breeding and germplasm resource improvement

5.1

Experimental materials and methods

5.1.1

Selection of experimental fish species

The silver drum parents used in this study were from a Dalian Aquaculture Co. In January 2023, 320 age-3 silver drum were selected for temporary rearing in a recirculating water aquaculture workshop, and all individuals were injected with PIT microchips for individual identification, and some of them were bred and spawned in March. 90 parents were selected for breeding experiments in May 2023, and the genetic sex identification of 90 fish was performed using SNP markers for aamhr2. The 90 fish were randomly divided into two groups: the optimized breeding strategy group and the original breeding strategy group (i.e., the current strategy adopted by most farmers, with breeding temperature, flow rate, water depth, and light duration of 23.6°C, 0.25 m/s, 1.2 m, and 12 h, respectively). A total of 45 parents in the original breeding strategy group had an initial mean body mass of 1526±211 g and a sex ratio (female:male) of 21:24. 45 parents in the optimized breeding strategy group had an initial mean body mass of 1587±362 g and a sex ratio (female:male) of 22:23. At the beginning of the experiment, there was no significant difference in the body mass of the silver drum parents between the original and optimized breeding strategy groups. The experimental fish were randomly assigned to two independent parental breeding systems. The parental fish breeding culture systems included a recirculating water culture system, a light intelligent control system, and a temperature automatic control system. The recirculating water culture system included a culture tank of about 52 m³ and a recirculating water treatment system of about 11 m³. The circulating water treatment system includes a residual bait feces collector, a biochemical cotton filter tank, a protein separator, a biological filter tank, an oxygen aeration device, a buffer tank, and a UV sterilization device.

5.1.2

Breeding programs

The same feeding management method was adopted for both groups. During the experimental period, both groups of parental fish were fed by satiation feeding, and the parental fish were fed 1-2 times a day at 2.1%-3.8% of the total weight of the parental fish. The bait consisted of compound feed and biological bait. The compound feed was the marine fish compound feed produced by a bioengineering company, and the raw material composition was fish meal, shrimp meal, fish oil, vitamins, minerals and so on. Biological baits included frozen baitfish, frozen squid, live sandworms and crabs. Water quality indicators include temperature, pH, dissolved oxygen (DO), salinity, and more. were detected in real-time using an online monitoring system for water quality during the experiment.

5.1.3

Regulation of environmental parameters for artificial breeding simulations

1)

Flow rate

Inverter frequency regulation is used as a means of debugging high-power pumps for voltage control. The flow meter is used to monitor the flow rate parameters correspond to each voltage frequency. The final record to 0.1m / s as a gradient, from static water to 1m / s in the middle of all the flow parameters corresponding to the voltage frequency, the experiment will be the frequency converter voltage frequency control to the required flow parameters can be.

2)

Water depth

The experimental pool itself is the height of the drain pipe in the middle of the pool for drainage, so this study has prepared a number of 10cm as a gradient in increasing length of the drain pipe, according to the required depth of replacement of the drain pipe can be regulated.

3)

Temperature

The temperature control system includes a temperature-regulating reservoir, automatic temperature-regulating equipment for heating rods and temperature-regulating equipment for air heating and cooling machines, with a temperature-regulating range of 10-30°C and a temperature-regulating accuracy of 0.1°C. The temperature-regulating system can be adjusted according to the temperature of the water.

4)

Lighting

Lighting intelligent control system includes LED lights and intelligent light environment controller. Two LED lights are hung 1.5m away from the breeding water surface, and the opening and closing time and irradiance of the LED lights are controlled by the intelligent light environment controller, which can set the light gradient mode and simulate the sunrise and sunset.

5.2

Analysis of Silver Drum Stocks Reproduction

In this study, the reproduction experiments of silver drum parents were conducted under the optimal breeding conditions and the original breeding conditions obtained by the genetic algorithm solution, respectively, and the experiments were conducted from May to July 2023, which were filmed and recorded by an on-water camera. At the end of the experiment, the silver drum parents in the two experimental groups were measured for body length and body mass, and the whole pool was fished for eggs, and the presence or absence of breeding behavior, the total number of eggs spawned, and the fertilization rate were counted, respectively.

5.2.1

Results of the analysis of growth indicators of parental fish

The results of body length and body mass of silver drum in the optimized breeding strategy group and the original breeding strategy group are shown in Table 2. On the first day of the experiment, the average body lengths of 45 fish in the optimized breeding strategy group and the original breeding strategy group were 306.4±15.8 mm and 306.1±18.7 mm, respectively, and the average body masses were 1523.4±65.9 g and 1524.5±59.89 g, respectively, and there were no significant differences in body lengths and masses between the optimized breeding strategy group and the original breeding strategy group.D15 Optimized breeding strategy The mean body length and mean body mass of females and males among 45 parents in the D15 optimized breeding strategy group increased to 310.8±18.7 mm, 311.9±13.0 mm, 1542.5±58.7 g and 1548.9±48.2 g, respectively, which were not significantly different from those in the original breeding strategy group. On the 45th and 60th days of the experiment, the mean length and mass of the experimental parents in the optimized breeding strategy group were 320.8±19.6mm, 326.5±17.6mm, 1618.7±51.2g and 1644.1±55.9g, respectively, which were significantly different from those of the original breeding strategy group. It was also found that at D1, D15, and D45 the mortality rate was 0% in both the optimized breeding strategy group and the original breeding strategy group, but at D60, five fish died in the original breeding strategy group with a mortality rate of 11.1%.

Table 2.

The results of the body length and the physical deformation of the fish

Group	Day	Body Length		Body Mass		Mortality
Group	Day	Female	Male	Female	Male	Mortality
Optimization group	D1	305.2±13.8	307.5±19.3	1522.3±63.9	1524.5±56.8	0%
	D15	310.8±18.7	311.9±13.0	1542.5±58.7	1548.9±48.2	0%
	D30	318.5±16.6	319.8±19.1	1587.6±92.1	1594.2±45.2	0%
	D45	320.1±18.4	321.4±18.0	1617.6±112.3	1619.8±74.5	0%
	D60	325.6±18.3	327.4±15.3	1642.3±37.5	1645.9±117.3	0%
Original group	D1	305.6±19.1	306.5±16.4	1523.6±106.3	1525.4±95.2	0%
	D15	309.6±16.1	308.5±18.6	1538.6±104.4	1537.8±93.1	0%
	D30	315.6±17.5	316.5±15.5	1542.3±102.2	1543.6±45.1	0%
	D45	317.6±18.2	319.4±16.6	1565.3±114.2	1569.6±90.3	0%
	D60	320.1±18.2	322.6±15.3	1594.6±42.4	1597.8±51.8	11.1%

5.2.2

Analysis of the effects of silver drum reproduction

Oocytes and sperm cells from female and male parents in different groups were collected on days 1, 20, and 45 for tissue section observation. Based on the stage of oocytes and spermatocytes in the ovary and spermatheca sections, the percentage of oocytes and spermatocytes was calculated to investigate the population reproduction of silver drum from the perspective of gonadal development. The percentage of oocytes in the D1, D30, and D60 oocytes of silver drum females in the optimized breeding strategy group and the original breeding strategy group are shown in Fig. 2, and (a) and (b) represent the results of the optimized breeding strategy and the original breeding strategy group, respectively. The results showed that at the first day of the experiment, only time-phase II and III oocytes were present in the ovarian sections of female parental fish in the optimized breeding strategy group and the original breeding strategy group, which accounted for 94.58% and 5.42% and 94.82% and 5.18%, respectively. Optimized breeding strategy group D20 showed time-phase IV oocytes in 7.65% of ovarian sections, and time-phase III and time-phase II oocytes in 45.62% and 46.73%, respectively. In contrast, no time-phase IV oocytes were present in the ovarian sections of the original breeding strategy group D20. On the 45th day of the breeding experiment, the proportion of time-phase IV oocytes in the ovaries of the parents of the optimized breeding strategy group reached 69.45%, while no time-phase IV oocytes were found in the parents of the original breeding strategy group. This indicates that the development of oocytes in the parents of the optimized breeding strategy group was faster than that in the original breeding strategy group, and the optimized breeding strategy could promote the gonadal development of the parents and enhance the breeding effect.

Based on the stage of spermatogenic cells in the tissue sections of spermathecae of male parental fish, the percentage of spermatogenic cells in the optimized breeding strategy group and the original breeding strategy group D1, D20, and D45 were calculated, and the results of the analyses are shown in Fig. 3, with (a) and (b) representing the results of optimized breeding strategy and the original breeding strategy group, respectively.The primary spermatocytes in the section of spermathecae of the D1 spermatheca in the optimized breeding strategy group and in the original breeding strategy group were respectively 24.38% and 27.44%, and 75.62% and 72.56% of spermatogonia, respectively. Tissue sections of D45 spermatogonia of the optimized breeding strategy group contained 56.94%, 12.69%, 4.52%, 1.26% and 24.59% of spermatozoa, spermatocytes, secondary spermatids, primary spermatocytes and spermatogonia, respectively. The percentage of spermatozoa in the D45 spermatheca of the original breeding strategy group (28.96%) was lower than that of the optimized breeding strategy group (56.94%), indicating that the gonadal development of male parental fish under the optimized breeding strategy condition was faster than that under the original breeding strategy condition.

In addition, at the end of the experiment, it was found that the average fertilization rate of silver drum parents in the experimental group of the optimized reproduction strategy was 82.65%, the spawning period generally lasted for 4-9 days, and the average cumulative spawning amount was 954 eggs. In contrast, the average fertilization rate and the average cumulative spawning amount of parents in the original breeding strategy group were lower than those in the optimized breeding strategy group, which were 78.45% and 745 eggs, respectively, which fully proved that the optimized breeding condition factors based on the legacy algorithm solving had an obvious promotion effect on the breeding effect of silver scat population.

5.3

Analysis of improvement of germplasm resources

The improvement of silver scat germplasm resources was reflected from the aspect of juvenile growth and development indexes, and the growth parameters of silver scat juveniles obtained in the breeding experiment after 138 d of culture are shown in Table 3. The final body weight (594.6g > 562.3g), specific growth rate (SGR) (2.59 > 2.36), absolute growth rate (AGR) (4.13 > 3.90), and feed coefficient (FCR) (1.36 > 1.12) of silver drum larvae in the optimized breeding strategy group were significantly higher than those in the original breeding strategy group. The final body weight, specific growth rate, and absolute growth rate of silver drum larvae in the optimized breeding strategy group were significantly higher than those of the original breeding strategy group (P=0.015, 0.012, and 0.006<0.05), and the feed coefficients were not significantly different (P=0.084>0.05). In addition, it can be seen that the survival rate (SR) of silver drum juveniles in the optimized breeding strategy group was the highest (96.58%), which indicates that under the implementation of the optimized artificial population breeding strategy, the survival rate of juveniles bred by silver drum parents increased, and the growth indexes were better than those of the original breeding strategy group, and that the silver drum germplasm resources were improved.

Table 3.

The comparison of the growth of the young fish

Item	Optimization group		Original group		P
Item	Mean	SD	Mean	SD	P
Initial body weigh(g)	24.1	5.9	23.9	6.9	0.632
Final body weigh(g)	594.6	39.6	562.3	45.8	0.015
CVW	11.64		10.52		0.009
SGR(%/d)	2.59	0.04	2.36	0.08	0.012
AGR(g/d)	4.13	0.52	3.90	0.48	0.006
SR(g)	96.58%		92.45%		0.004
FCR	1.36		1.12		0.084

6

Conclusion

In this paper, after constructing an optimization model of environmental suitability for silver scat breeding, the optimal Pareto set of the model was solved using the multi-objective optimization NSGA-II algorithm, and the optimal breeding strategy was obtained by selecting the Euclidean distance value. The optimal breeding strategy was obtained by selecting the optimal Pareto set based on the Euclidean distance value. Then, the silver scat population breeding experiment was set up to verify the implementation effect of the optimal breeding strategy: 1)

The temperature, flow rate, water depth and light duration in the optimal breeding strategy of silver scat population were set to 22.5°C, 0.21m/s, 1.7m and 15h, respectively.

2)

The mortality rate of silver drum parents in the optimized breeding strategy group was 0%, while the parents in the original breeding strategy group showed mortality at the 60th day of the experiment, with a mortality rate of 11.1%. From the gonadal slices, 69.45% of time-phase IV oocytes were found in the ovaries of the female parents in the optimized breeding strategy group on day 45, while no time-phase IV oocytes were found in the parents in the original breeding strategy group. In addition, at the end of the experiment, 56.94% of the spermatozoa of male parents in the optimized breeding strategy group were found, while only 28.96% of the spermatozoa of male parents in the original breeding strategy group were found, which indicated that the optimized breeding strategy could promote the gonadal development of silver drum parents, and thus improve the breeding effect.

3)

Under the implementation of the optimized population breeding strategy, the survival rate of artificially bred silver drum larvae increased, and the growth indexes such as specific growth rate and absolute growth rate were significantly better than those of the original breeding strategy group, which indicated that the germplasm resources had been improved.

The above results show that the genetic algorithm-based breeding strategy of silver drum has achieved good results in the experiments, which lays the foundation for the breeding and germplasm improvement of silver drum.

Lingua:: Inglese

Frequenza di pubblicazione:: 1 volte all'anno
Argomenti della rivista:: Scienze biologiche, Scienze della vita, altro, Matematica, Matematica applicata, Matematica generale, Fisica, Fisica, altro

Feed RSS della rivista

Optimization of silver scat population breeding strategy and germplasm resource improvement based on genetic algorithm

Pan Chen

Liu Cao

Ning Yang

Hai Huang

Pubblicato online: 24 mar 2025

Ricevuto: 31 ott 2024

Accettato: 10 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0785

Parole chiaveNSGA-II algorithm, Crossover operator, SDR algorithm, Silver scat breeding, Germplasm resource improvement

© 2025 Pan Chen et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Parole chiave
NSGA-II algorithm, Crossover operator, SDR algorithm, Silver scat breeding, Germplasm resource improvement