WO2023142722A1 - 一种基于智能优化算法的繁殖指导方法及装置 - Google Patents

一种基于智能优化算法的繁殖指导方法及装置 Download PDF

Info

Publication number
WO2023142722A1
WO2023142722A1 PCT/CN2022/138187 CN2022138187W WO2023142722A1 WO 2023142722 A1 WO2023142722 A1 WO 2023142722A1 CN 2022138187 W CN2022138187 W CN 2022138187W WO 2023142722 A1 WO2023142722 A1 WO 2023142722A1
Authority
WO
WIPO (PCT)
Prior art keywords
population
mating
individuals
generation
genotype
Prior art date
Application number
PCT/CN2022/138187
Other languages
English (en)
French (fr)
Inventor
杨之乐
张玉倩
赵世豪
郭媛君
王尧
Original Assignee
深圳先进技术研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳先进技术研究院 filed Critical 深圳先进技术研究院
Publication of WO2023142722A1 publication Critical patent/WO2023142722A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
    • G16B20/40—Population genetics; Linkage disequilibrium
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding

Definitions

  • the present invention relates to the field of artificial intelligence, in particular to a breeding guidance method and device based on an intelligent optimization algorithm.
  • transgenic mice In the field of biomedicine, transgenic mice have unique advantages and practical value, and play an important role; especially the multi-genotype mouse disease models obtained through mating are indispensable in drug development and mechanism research.
  • the process of breeding and obtaining multi-genotype mouse models often requires a lot of time and economic costs; although it is possible to formulate breeding strategies with reference to Mendelian genetic laws and finally obtain desired genotype mouse models, how to best There is no sound guiding strategy for the fastest and most efficient breeding.
  • mice When conducting biomedical experiments, it is often necessary to obtain some pre-expected gene types, and these genes need to be obtained through continuous mating of experimental organisms (take mice as an example). Since the mating process of mice is strictly in accordance with Mendel's law of inheritance, that is, the genotype obtained after each generation of mice is obtained according to a certain probability, it often takes a lot of time and economic cost to obtain the desired genotype . In addition, because there is no reasonable guidance strategy, many mating links in the mating process of mice are useless, which undoubtedly further increases the related costs.
  • the embodiment of the present invention provides a breeding guidance method and device based on an intelligent optimization algorithm, which can provide the fastest and most efficient mating strategy guidance for population reproduction.
  • a kind of breeding guidance method based on intelligent optimization algorithm comprising the following steps:
  • the initial parameters include genotype, expected genotype, lifespan information and lethal gene information;
  • the mating strategy of selecting the group of male and female individuals with the fastest reproduction and the most efficient is carried out until the selection times reach the preset selection times.
  • the total cost calculation formula is:
  • C represents the total cost required to achieve the desired gene, this cost includes time cost and economic cost
  • Ng represents the current generation
  • n represents the generation required to achieve the final desired gene
  • T(Ng) represents the completion of each generation of mating
  • ⁇ *C mouse *N mouse (Ng) represents the economic cost required to complete the mating of each generation
  • C mouse represents the economic cost required to complete the mating of one generation, including food and labor
  • N mouse (Ng) It indicates how many can participate in mating in this generation.
  • ⁇ and ⁇ are weight coefficients, which can be preset and adjusted according to actual requirements. That is, if more attention is paid to time cost rather than economic cost, ⁇ can be greater than ⁇ .
  • Constraint processing is carried out on the population, which includes life span restriction, lethal gene restriction and population quantity restriction.
  • life-deactivated individuals For the lifespan of individuals during the mating process of the population. Individuals exceeding the preset number of years are regarded as life-deactivated individuals in the population; the expression of life-deactivated individuals is:
  • Mouse life represents the current survival status of the mouse
  • 1 represents the individuals in the population that can mate normally
  • 0 represents that the population has exceeded the preset survival algebra, and it is set to be inactivated
  • Mouse Ng represents the current individual that has survived Algebra
  • life max indicates the maximum number of generations that an individual can survive.
  • lethal gene constraints are:
  • lethal gene inactivated individuals The genotypes that lead to the loss of fertility and death of individuals in the population are considered lethal, and individuals with lethal genes are regarded as lethal gene inactivated individuals; the expression of lethal gene inactivated individuals is:
  • Gene death represents a lethal gene, and if the lethal gene appears, individuals carrying this genotype are all regarded as inactivated individuals.
  • population size constraint is specifically:
  • the number of individuals in the preset population will exceed the preset number of individuals, and it will be regarded as a number of inactivated individuals.
  • the number of preset individuals in the population will exceed the preset number, and it will be regarded as a number of inactivated individuals. Specifically:
  • the social learning particle swarm optimization algorithm is used to select the mating strategy of the fastest and most efficient male and female individuals.
  • the expression of the social learning particle swarm optimization algorithm is:
  • ⁇ x i,j (t+1) r 1 (t) ⁇ x i,j (t)+r 2 (t)I i,j (t)+ ⁇ r 3 (t)C i,j (t)
  • x i,j (t) represents the jth dimension of individual i in generation t
  • ⁇ xi ,j (t+1) is the amount learned from individuals better than i
  • I i,j (t ) represents the difference between the jth dimension of the i individual and the corresponding dimension of the individual that is better than the i individual
  • C j,t represents the jth dimension of the i individual and the average value of the jth dimension of all individuals in the current population
  • r 2 (t) and r 3 (t) are all random numbers between 0 and 1
  • ? is a number related to the problem dimension and population size
  • ⁇ is set to 1.
  • the mating strategy of selecting the group of male and female individuals with the highest fitness is carried out until the number of selections reaches the preset number of selections and further includes:
  • the mating strategy of the group of male and female individuals with the highest fitness is output to the display module for display.
  • a breeding guidance device based on an intelligent optimization algorithm comprising:
  • the data acquisition module is used to acquire the initial parameters of the first generation of the population, the initial parameters include genotype, expected genotype, life span information and lethal gene information;
  • the first acquisition module is used to divide the population into two groups of male and female, and obtain the genotype of the second generation of the population generated by the random mating of the two groups of male and female,
  • a query module used to query whether there is a desired gene in the genotype of the second generation
  • the cost calculation module is used to calculate the time and economic cost of obtaining the expected genotype during the entire mating process if there is an expected gene, and initialize the initial parameters; if no expected gene is found, continue to make the first generation of the population randomly mate until appearance of desired genes;
  • the second acquisition module is used to repeatedly acquire the genotype of the second generation of the population generated by random mating of two groups of male and female until the number of the population reaches the preset scale;
  • the strategy selection module is used to select the mating strategy of the group of male and female individuals with the highest fitness when the number of the population reaches the preset size until the number of selections reaches the preset number of selections.
  • the breeding guidance method and device based on the intelligent optimization algorithm in the embodiment of the present invention includes: disclosing a guidance method for the optimal breeding strategy of the expected gene of the population based on the intelligent optimization algorithm, so as to obtain the time and economic cost in the process of the expected gene
  • the intelligent optimization algorithm of social learning particle swarms is used to optimize the mating process by sampling the mating of the population, so as to guide the experimenters which genotypes should be used in the mating process of each generation
  • the populations are put together, so as to achieve the purpose of guiding the mating of the population.
  • Fig. 1 is the flow chart of the abdominal body image liver segmentation method based on deep learning of the present invention
  • FIG. 2 is a block diagram of the deep learning-based abdominal body image liver segmentation device of the present invention.
  • a kind of breeding instruction method based on intelligent optimization algorithm comprising the following steps:
  • S100 Obtain the initial parameters of the first generation of the population, the initial parameters include genotype, expected genotype, lifespan information and lethal gene information.
  • the initialization data must be obtained, including the genotype of the first generation of mice and the genotype of the expected mouse, as well as information such as the lifespan and lethal genes of the mice.
  • the mating process of mice strictly follows the Mendelian law of inheritance, and the genotypes of the next generation are generated strictly according to the relevant probability.
  • this method combines reality and specifies that the number of mice obtained by each mating follows a Gaussian distribution.
  • S200 divide the population into two groups of male and female, and obtain the genotype of the second generation of the population generated by random mating of the two groups of male and female.
  • the population of the algorithm is defined as all mice that can participate in mating.
  • the mating process first divides the population into two populations, male and female, and then adopts the principle of random mating to randomly select individuals from the two populations for mating.
  • the male and female probability of producing offspring after mating also strictly follows Mendel's law of inheritance, that is, the probability of male or female is 1/2.
  • S300 Query whether there is a desired gene in the genotype of the second generation.
  • step S300 includes:
  • the total cost calculation formula is:
  • C represents the total cost required to achieve the desired gene, this cost includes time cost and economic cost
  • Ng represents the current generation
  • n represents the generation required to achieve the final desired gene
  • T(Ng) represents the completion of each generation of mating
  • the time required, combined with the actual cost is usually a constant
  • ⁇ *C mouse *N mouse (Ng) represents the economic cost required for each generation to complete mating
  • C mouse represents the cost of each mouse to complete one generation of mating
  • N mouse (Ng) indicates how many mice can participate in mating in this generation
  • ⁇ and ⁇ are weight coefficients, which can be preset and adjusted by the experimenter according to actual requirements, that is, if more emphasis is placed on Time cost rather than economic cost can make ⁇ greater than ⁇ .
  • step S400 it also includes:
  • Carry out constraint processing on the population, and the constraint processing includes life span constraint, lethal gene constraint and population quantity constraint; among them,
  • Mouse life represents the current survival status of the mouse
  • 1 means that the mouse is alive and can mate normally
  • 0 means that the mouse has exceeded the maximum number of survival generations and has died.
  • Mouse Ng indicates the number of generations the mouse has survived
  • life max indicates the maximum number of generations the mouse can survive.
  • the lethal gene constraint is:
  • mice with this genotype must be inactivated; therefore, the genotypes that lead to individual loss of fertility and death in the population are considered lethal, and individuals with lethal genes are regarded as lethal gene inactivated individuals; the expression of lethal gene inactivated individuals is:
  • Gene death represents a lethal gene, and if the lethal gene appears, individuals carrying this genotype are all regarded as inactivated individuals.
  • mice The specific constraints on the number of mice are:
  • the number of individuals in the preset population will exceed the preset number of individuals as inactivated individuals, and the individuals with more parameter mating times will be prioritized in turn as inactivated individuals.
  • mice in addition to lifespan constraints and lethal gene constraints, it should also include constraints on the maximum number of mice. Due to the limitations of experimental conditions, the number of mice participating in mating cannot be unlimited. When the total number of mice is greater than the extreme value, it should be used Mice that exceed the extreme number are inactivated, and mice that participate in more mating times should be inactivated preferentially.
  • S500 Obtaining the genotypes of the second generation of the population generated by the random mating of the male and female groups repeatedly until the population reaches a preset size.
  • the optimization process can be started; when setting the iterative process of the algorithm, as long as the expected genotype does not appear , it is necessary to continue to arrange mice to mate; the expected genotype appears, and an individual is officially obtained.
  • each data and parameter should be initialized, and then the next individual is obtained until the specified or preset population size is reached; After specifying the population size, SLSPSO (Social Learning Particle Swarm Optimization) can start the iterative optimization process, so as to finally obtain the optimal mating strategy.
  • SLSPSO Social Learning Particle Swarm Optimization
  • the invention adopts the social learning particle swarm algorithm (SLPSO) to select the mating strategy of the fastest and most efficient male and female individuals.
  • SLPSO is a heuristic intelligent optimization method, which allows individuals in the population to continuously learn from individuals that are better than themselves (higher fitness value) to perform iterative optimization, and finally obtain the optimal result.
  • the expression of SLPSO is:
  • ⁇ x i,j (t+1) r 1 (t) ⁇ x i,j (t)+r 2 (t)I i,j (t)+ ⁇ r 3 (t)C i,j (t)
  • x i,j (t) represents the jth dimension of individual i in generation t
  • ⁇ xi ,j (t+1) is the amount learned from individuals better than i
  • I i,j (t ) represents the difference between the jth dimension of the i individual and the corresponding dimension of the individual that is better than the i individual
  • C j,t represents the jth dimension of the i individual and the average value of the jth dimension of all individuals in the current population
  • r 2 (t) and r 3 (t) are all random numbers between 0 and 1
  • ? is a number related to the problem dimension and population size.
  • ⁇ is set to a value of 1; mapped in the present invention, each individual mainly includes the process from the initial stage to the final desired genotype Mating protocol (i.e. recording parent and offspring genotypes for each mating).
  • step S600 it also includes:
  • the entire learning framework will be optimized according to the above-mentioned steps, until the set maximum number of training times is reached, then the mating strategy of the group of mice with the highest fitness can be output, and the results can be visualized to guide the experimenters Arrange the mice to mate according to the secondary strategy; the highest fitness includes the fastest and most efficient mating strategy for mouse reproduction.
  • mice because the laboratory is most commonly used to conduct experiments on mice; therefore, the embodiments of the present application are described using mice as an example, but it does not mean that the method of the present invention can only be used for mice.
  • a breeding guidance device based on an intelligent optimization algorithm which is characterized in that it includes:
  • the data acquisition module 100 is used to acquire the initial parameters of the first generation of the population, the initial parameters include genotype, expected genotype, lifespan information and lethal gene information;
  • the first acquisition module 200 is used to divide the population into two groups of male and female, and obtain the genotype of the second generation of the population generated by random mating of the two groups of male and female,
  • the query module 300 is used to query whether there is a desired gene in the genotype of the second generation
  • the cost calculation module 400 is used to calculate the time and economic cost of obtaining the expected genotype in the entire mating process if there is an expected gene, and initialize the initial parameters; if no expected gene is found, continue to make the first generation of the population randomly mate, until the desired gene appears;
  • the second acquisition module 500 is used to repeatedly acquire the genotype of the second generation of the population generated by random mating of two groups of male and female until the number of the population reaches a preset scale;
  • the strategy selection module 600 is used to select the mating strategy of the group of male and female individuals with the highest fitness when the number of the population reaches the preset size until the selection times reach the preset selection times.
  • This application discloses an optimal strategy guidance method for population expected gene reproduction based on an intelligent optimization algorithm, with the goal of obtaining the time and economic cost in the process of expected genes, and considering the constraints of population life, lethal genes, and quantity, for population mating
  • the intelligent optimization algorithm of sampling social learning particle swarms is used to complete the optimization of the mating process, so as to guide the experimenters which genotype populations should be put together in each generation of mating process, so as to achieve the purpose of guiding population mating.

Landscapes

  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Software Systems (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biotechnology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Genetics & Genomics (AREA)
  • Chemical & Material Sciences (AREA)
  • Proteomics, Peptides & Aminoacids (AREA)
  • Bioethics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physiology (AREA)
  • Databases & Information Systems (AREA)
  • Epidemiology (AREA)
  • Analytical Chemistry (AREA)
  • Public Health (AREA)
  • Ecology (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种基于智能优化算法的繁殖指导方法及装置，涉及人工智能领域，该方法包括：一种基于智能优化算法的种群期望基因繁殖最优策略指导方法，以得到期望基因过程中的时间和经济成本为目标，考虑种群的寿命、致死基因及数量等约束，对种群交配采样社会学习粒子群的智能优化算法来完成对交配过程的优化，以指导实验人员在每一代交配过程中应该把哪些基因型的种群放在一起，从而达到对种群交配进行指导目的。

Description

一种基于智能优化算法的繁殖指导方法及装置 技术领域
本发明涉及人工智能领域，具体而言，涉及一种基于智能优化算法的繁殖指导方法及装置。
背景技术
在生物医学领域，转基因小鼠具有独特的优越性和实用价值，发挥着重大作用；尤其是通过交配获得的多基因型小鼠疾病模型，在药物研发和机制研究中不可或缺。然而，在实际操作中，繁育获取多基因型小鼠模型的过程往往需要很大的时间、经济成本；尽管可参照孟德尔遗传定律制定繁育策略并最终获得期望基因型小鼠模型，但是如何最快、最高效地繁育尚无合理指导策略。
在进行生物医学相关实验时，往往需要得到一些预先期望的基因类型，而这些基因则需要通过对实验生物(以小鼠为例)来进行不断地交配而最终得到。由于在小鼠交配过程严格按照孟德尔遗传定律，即小鼠没代交配后所获得的基因型是按照一定概率得到的，因此在得到期望基因型的过程中往往需要很大的时间、经济成本。此外，由于没有合理的指导策略，在进行小鼠交配的环节中，有很多的交配环节都是无用功，这无疑进一步增加了相关成本。
通常，研究人员在通过小鼠交配来得到期望基因型时，往往都是凭经验的人为安排小鼠每一代的交配，而这通常会获得大量本身不期望的基因型，随机性太大，相应的会造成很多额外的时间和经济成本；因此，需要设计一种有效的小鼠交配指导策略，来为得到期望基因过程中繁殖小鼠，提供合理的交配指导。
发明内容
本发明实施例提供了一种基于智能优化算法的繁殖指导方法及装置，能为种群的繁殖提供最快、最高效的交配策略指导。
根据本发明的一实施例，提供了一种基于智能优化算法的繁殖指导方法，包括以下步骤：
获取种群初代的初始参数，初始参数包括基因型、期望基因型、寿命信息及致死基因信息；
将种群分为雌雄两组，获取雌雄两组随机进行交配产生种群的第二代的基因型；
查询第二代的基因型中是否有期望基因；
若有期望基因，则计算整个交配过程中得到期望基因型的时间和经济成本，并初始化初始参数；若未发现有期望基因，则继续使种群的初代随机进行交配，直至出现期望基因；
重复获取雌雄两组随机进行交配产生种群的第二代的基因型，直至种群的数量达到预设规模；
当种群的数量达到预设规模时，进行选取繁殖最快、最高效的那组雌雄个体的交配策略，直至选取次数达到预设选取次数。
进一步地，若有期望基因，则计算整个交配过程中得到期望基因型的时间和经济成本具体为：
通过总成本计算公式计算整个交配过程中得到期望基因型的时间和经济成本；
总成本计算公式为：
Figure PCTCN2022138187-appb-000001
其中，C表示达到期望基因时所需要的总成本，此成本包括时间成本和经济成本，Ng表示当前的代数，n表示达到最终期望基因时所需要的代数，T(Ng)表示完成每代交配所需的时间，β*C mouse*N mouse(Ng)表示每代完成交配所需的经济成本，C mouse表示完成一代交配时所需要花费的经济成本，包括食物、人工，N mouse(Ng)则表示该代一共有多少只可以参与交配，α和β分别为权重系数，可由根据实际要求进行预设调整，即若更注重时间成本而非经济成本，则可使α大于 β。
进一步地，在若有期望基因，则计算整个交配过程中得到期望基因型的时间和经济成本之前还包括：
对种群的进行约束处理，约束处理包括寿命约束、致死基因约束及种群数量约束。
进一步地，寿命约束为：
在种群进行交配的过程中考虑个体的寿命问题，超过预设年限的个体在种群中视为寿命失活个体；寿命失活个体的表达式为：
Figure PCTCN2022138187-appb-000002
式中，Mouse life代表小鼠当前的存活状态，1表示种群中的可以正常交配的个体，0表示种群中已经超过预设生存代数，其设置为已经失活，Mouse Ng表示当前个体已经存活的代数，life max表示个体可以存活的最大代数。
进一步地，致死基因约束为：
种群中导致个体丧失生育能力及死亡的基因型被视为致死基于，将有致死基因的个体视为致死基因失活个体；致死基因失活个体的表达式为：
Figure PCTCN2022138187-appb-000003
其中，Gene death表示致死基因，若该致死基因出现，则携带该基因型的个体都视为失活个体。
进一步地，种群数量约束具体为：
预设种群个体数量，将超过预设的个体数量，视为数量失活个体。
进一步地，种群预设个体的数量，将超过预设数量，视为数量失活个体具体为：
依次优先将参数交配次数多的个体视为失活个体。
进一步地，采用社会学习粒子群算法来选取繁殖最快、最高效的雌雄个体的交配策略，社会学习粒子群算法表达式为：
Figure PCTCN2022138187-appb-000004
Δx i,j(t+1)＝r 1(t)Δx i,j(t)+r 2(t)I i,j(t)+φr 3(t)C i,j(t)
其中，x i,j(t)代表第t代时，i个个体的第j个维度，Δx i，j(t+1)是从优于i个个体身上学习的量，I i,j(t)代表i个个体的第j个维度与比i个个体优秀的个体对应维度的差值，C j,t则代表i个个体的第j个维度与当前种群所有个体第j个维度的平均值的差值，r 1(t)、r 2(t)与r 3(t)均为0到1之间的随机数，？为一个与问题维度和种群规模相关的数，φ设定值为1。
进一步地，在当种群的数量达到预设规模时，进行选取适应度最高的那组雌雄个体的交配策略，直至选取次数达到预设选取次数之后还包括：
当选取次数达到预设选取次数时，则输出适应度最高的那组雌雄个体的交配策略至显示模块进行显示。
一种基于智能优化算法的繁殖指导装置，包括：
数据获取模块，用于获取种群初代的初始参数，初始参数包括基因型、期望基因型、寿命信息及致死基因信息；
第一获取模块，用于将种群分为雌雄两组，获取雌雄两组随机进行交配产生种群的第二代的基因型，
查询模块，用于查询第二代的基因型中是否有期望基因，
成本计算模块，用于若有期望基因，则计算整个交配过程中得到期望基因型的时间和经济成本，并初始化初始参数；若未发现有期望基因，则继续使种群的初代随机进行交配，直至出现期望基因；
第二获取模块，用于重复获取雌雄两组随机进行交配产生种群的第二代的基因型，直至种群的数量达到预设规模；
策略选取模块，用于当种群的数量达到预设规模时，进行选取适应度最高的那组雌雄个体的交配策略，直至选取次数达到预设选取次数。
本发明实施例中的基于智能优化算法的繁殖指导方法及装置，方法包括：公开了一种基于智能优化算法的种群期望基因繁殖最优策略指导方法，以得到 期望基因过程中的时间和经济成本为目标，考虑种群的寿命、致死基因及数量等约束，对种群交配采样社会学习粒子群的智能优化算法来完成对交配过程的优化，以指导实验人员在每一代交配过程中应该把哪些基因型的种群放在一起，从而达到对种群交配进行指导目的。
附图说明
此处所说明的附图用来提供对本发明的进一步理解，构成本申请的一部分，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。在附图中：
图1为本发明基于深度学习的腹部体影像肝脏分割方法的流程图；
图2为本发明基于深度学习的腹部体影像肝脏分割装置的模块图。
具体实施方式
为了使本技术领域的人员更好地理解本发明方案，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分的实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都应当属于本发明保护的范围。
需要说明的是，本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本发明的实施例能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备 固有的其它步骤或单元。
参见图1，根据本发明一实施例，提供了一种基于智能优化算法的繁殖指导方法，包括以下步骤：
S100：获取种群初代的初始参数，初始参数包括基因型、期望基因型、寿命信息及致死基因信息。
首先要得到初始化的数据，要得到初代小鼠的基因型以及期望小鼠的基因型，以及小鼠的寿命和致死基因等信息。其中小鼠交配的过程严格遵循孟德尔遗传定律，所产生的下一代基因型严格按照相关概率产生，此外本方法结合实际，指定每一次交配所得到的小鼠的数量遵循一个高斯分布。
S200：将种群分为雌雄两组，获取雌雄两组随机进行交配产生种群的第二代的基因型。
具体地，算法的种群定义为所有可以参与交配的小鼠。交配过程首先将种群分为公母两个种群，然后采取随机交配的原则，从两个种群中随机抽取个体进行交配。交配后产生子代的公母概率也严格遵循孟德尔遗传定律，即公或母的概率均为1/2。
S300：查询第二代的基因型中是否有期望基因。
具体地，步骤S300包括：
S301：通过总成本计算公式计算整个交配过程中得到期望基因型的时间和经济成本；
总成本计算公式为：
Figure PCTCN2022138187-appb-000005
其中，C表示达到期望基因时所需要的总成本，此成本包括时间成本和经济成本，Ng表示当前的代数，n表示达到最终期望基因时所需要的代数，T(Ng)表示完成每代交配所需的时间，结合实际该成本通常是一个常数；β*C mouse*N mouse(Ng)表示每代完成交配所需的经济成本，C mouse表示每只小鼠在完成一代交配时所需要花费的经济成本，包括食物、人工等，N mouse(Ng)则表示 该代一共有多少只可以参与交配，α和β分别为权重系数，可由实验人员根据实际要求进行预设调整，即若更注重时间成本而非经济成本，则可使α大于β。
S400：若有期望基因，则计算整个交配过程中得到期望基因型的时间和经济成本，并初始化初始参数；若未发现有期望基因，则继续使种群的初代随机进行交配，直至出现期望基因。
具体地，在步骤S400之前还包括：
对种群的进行约束处理，约束处理包括寿命约束、致死基因约束及种群数量约束；其中，
寿命约束为：
由于小鼠的寿命是有限的，因此在进行小鼠交配的过程中必须考虑小鼠的寿命问题，超过最大寿命的小鼠必须在种群中失活；在种群进行交配的过程中考虑个体的寿命问题，超过预设年限的个体在种群中视为寿命失活个体；寿命失活个体的表达式为：
Figure PCTCN2022138187-appb-000006
式中，Mouse life代表小鼠当前的存活状态，1表示小鼠活着且可以正常交配，0表示小鼠已经超过最大生存代数，已经死亡。Mouse Ng表示当前小鼠已经存活的代数，life max表示小鼠可以存活的最大代数。
致死基因约束为：
由于在交配过程中，有一些基因型的小鼠会导致小鼠丧失生育能力，也有一部分基因会导致小鼠直接死亡，这些基因型都被认定为是致死基因，出现这种基因型的小鼠都必须失活；因此，种群中导致个体丧失生育能力及死亡的基因型被视为致死基于，将有致死基因的个体视为致死基因失活个体；致死基因失活个体的表达式为：
Figure PCTCN2022138187-appb-000007
其中，Gene death表示致死基因，若该致死基因出现，则携带该基因型的个体都视为失活个体。
小鼠数量约束具体为：
预设种群个体数量，将超过预设的个体数量，视为数量失活个体，且依次优先将参数交配次数多的个体视为失活个体。
具体地，除了寿命约束以及致死基因约束以外，还应包括小鼠最大数量的约束，由于实验条件的限制，参与交配的小鼠不可能是无限的，当小鼠总数大于极值时，应使超出极值数目的小鼠失活，且应该让参与交配次数越多的小鼠越优先失活。
S500：重复获取雌雄两组随机进行交配产生种群的第二代的基因型，直至种群的数量达到预设规模。
具体地，算法框架的起始阶段，在得到期望基因型以及初代小鼠的基因型和其他相关信息后，即可开始寻优过程；在设定算法的迭代过程时，只要期望基因型没有出现，就必须继续安排小鼠交配；期望基因型出现，一个个体才算正式得出，此时应将各数据和参数进行初始化，再得到下个个体，直到达到指定或预设的种群规模；达到指定种群规模后即可使SLSPSO(社会学习粒子群算法)开始迭代寻优过程，从而最终得到最优的交配策略。
S600：当种群的数量达到预设规模时，进行选取繁殖最快、最高效的那组雌雄个体的交配策略，直至选取次数达到预设选取次数。
本发明采用社会学习粒子群算法(SLPSO)来选取繁殖最快、最高效的雌雄个体的交配策略。SLPSO是一种启发式智能优化方法，其通过让种群中的个体不断向比自己优秀的个体(适应度值更高)学习来进行迭代寻优，最终得到最优结果。SLPSO的表达式为：
Figure PCTCN2022138187-appb-000008
Δx i,j(t+1)＝r 1(t)Δx i,j(t)+r 2(t)I i,j(t)+φr 3(t)C i,j(t)
其中，x i,j(t)代表第t代时，i个个体的第j个维度，Δx i，j(t+1)是从优于i个个体身上学习的量，I i,j(t)代表i个个体的第j个维度与比i个个体优秀的个体对应维度的差值，C j,t则代表i个个体的第j个维度与当前种群所有个体第j个维度的平均值的差值，r 1(t)、r 2(t)与r 3(t)均为0到1之间的随机数，？为一个与问题维 度和种群规模相关的数，本算法为了加快计算时间，φ设定值为1；映射于本发明中，每个个体主要是包含从初始阶段到最终得到期望基因型过程中的交配方案(即记录每一次交配的父代和子代基因型)。
实施例中，在步骤S600之后还包括：
S601：当选取次数达到预设选取次数时，则输出适应度最高的那组雌雄个体的交配策略至显示模块进行显示。
具体地，整个学习框架将按照上述步骤的模式进行寻优，直到达到设定的最大训练次数时，即可输出使适应度最高的那组小鼠交配策略，并将该结果可视化，指导实验人员按照次策略安排小鼠进行交配；适应度最高包括小鼠繁殖繁殖最快、最高效的交配策略。
需要说明的是，由于实验室最为常用的为小鼠进行实验；因此，本申请的实施例以小鼠为例进行说明，但不意味着限制本发明方法仅能用于小鼠。
参见图2，根据本发明一实施例，提供了一种基于智能优化算法的繁殖指导装置，其特征在于，包括：
数据获取模块100，用于获取种群初代的初始参数，初始参数包括基因型、期望基因型、寿命信息及致死基因信息；
第一获取模块200，用于将种群分为雌雄两组，获取雌雄两组随机进行交配产生种群的第二代的基因型，
查询模块300，用于查询第二代的基因型中是否有期望基因，
成本计算模块400，用于若有期望基因，则计算整个交配过程中得到期望基因型的时间和经济成本，并初始化初始参数；若未发现有期望基因，则继续使种群的初代随机进行交配，直至出现期望基因；
第二获取模块500，用于重复获取雌雄两组随机进行交配产生种群的第二代的基因型，直至种群的数量达到预设规模；
策略选取模块600，用于当种群的数量达到预设规模时，进行选取适应度最高的那组雌雄个体的交配策略，直至选取次数达到预设选取次数。
本申请公开了一种基于智能优化算法的种群期望基因繁殖最优策略指导方法，以得到期望基因过程中的时间和经济成本为目标，考虑种群的寿命、致死基因及数量等约束，对种群交配采样社会学习粒子群的智能优化算法来完成对交配过程的优化，以指导实验人员在每一代交配过程中应该把哪些基因型的种群放在一起，从而达到对种群交配进行指导目的。
以上所述仅是本发明的优选实施方式，应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。

Claims (10)

  1. 一种基于智能优化算法的繁殖指导方法，其特征在于，包括以下步骤：
    获取种群初代的初始参数，所述初始参数包括基因型、期望基因型、寿命信息及致死基因信息；
    将所述种群分为雌雄两组，获取雌雄两组随机进行交配产生所述种群的第二代的基因型；
    查询所述第二代的基因型中是否有所述期望基因；
    若有所述期望基因，则计算整个交配过程中得到所述期望基因型的时间和经济成本，并初始化所述初始参数；若未发现有所述期望基因，则继续使所述种群的初代随机进行交配，直至出现所述期望基因；
    重复获取雌雄两组随机进行交配产生所述种群的第二代的基因型，直至所述种群的数量达到预设规模；
    当所述种群的数量达到所述预设规模时，进行选取繁殖最快、最高效的那组雌雄个体的交配策略，直至选取次数达到预设选取次数。
  2. 根据权利要求1所述的基于智能优化算法的繁殖指导方法，其特征在于，所述若有所述期望基因，则计算整个交配过程中得到所述期望基因型的时间和经济成本具体为：
    通过总成本计算公式计算整个交配过程中得到所述期望基因型的时间和经济成本；
    所述总成本计算公式为：
    Figure PCTCN2022138187-appb-100001
    其中，C表示达到期望基因时所需要的总成本，此成本包括时间成本和经济成本，Ng表示当前的代数，n表示达到最终期望基因时所需要的代数，T(Ng)表示完成每代交配所需的时间，β*C mouse*N mouse(Ng)表示每代完成交配所需的经济成本，C mouse表示完成一代交配时所需要花费的经济成本，包括食物、人工，N mouse(Ng)则表示该代一共有多少只可以参与交配，α和β分别为权重系数，可由根据实际要求进行预设调整，即若更注重时间成本而非经济成本，则可使α大于 β。
  3. 根据权利要求1所述的基于智能优化算法的繁殖指导方法，其特征在于，在若有所述期望基因，则计算整个交配过程中得到所述期望基因型的时间和经济成本之前还包括：
    对所述种群的进行约束处理，所述约束处理包括寿命约束、致死基因约束及种群数量约束。
  4. 根据权利要求3所述的基于智能优化算法的繁殖指导方法，其特征在于，所述寿命约束为：
    在所述种群进行交配的过程中考虑个体的寿命问题，超过预设年限的个体在所述种群中视为寿命失活个体；所述寿命失活个体的表达式为：
    Figure PCTCN2022138187-appb-100002
    式中，Mouse life代表小鼠当前的存活状态，1表示所述种群中的可以正常交配的个体，0表示所述种群中已经超过预设生存代数，其设置为已经失活，Mouse Ng表示当前个体已经存活的代数，life max表示个体可以存活的最大代数。
  5. 根据权利要求3所述的基于智能优化算法的繁殖指导方法，其特征在于，所述致死基因约束为：
    所述种群中导致个体丧失生育能力及死亡的基因型被视为致死基于，将有所述致死基因的个体视为致死基因失活个体；所述致死基因失活个体的表达式为：
    Figure PCTCN2022138187-appb-100003
    其中，Gene death表示致死基因，若该致死基因出现，则携带该基因型的个体都视为失活个体。
  6. 根据权利要求3所述的基于智能优化算法的繁殖指导方法，其特征在于，所述种群数量约束具体为：
    预设所述种群个体数量，将超过预设的个体数量，视为数量失活个体。
  7. 根据权利要求6所述的基于智能优化算法的繁殖指导方法，其特征在于， 所述种群预设个体的数量，将超过预设数量，视为数量失活个体具体为：
    依次优先将参数交配次数多的个体视为失活个体。
  8. 根据权利要求1所述的基于智能优化算法的繁殖指导方法，其特征在于，采用社会学习粒子群算法来选取所述繁殖最快、最高效的雌雄个体的交配策略，所述社会学习粒子群算法表达式为：
    Figure PCTCN2022138187-appb-100004
    Δx i,j(t+1)＝r 1(t)Δx i,j(t)+r 2(t)I i,j(t)+φr 3(t)C i,j(t)
    其中，x i,j(t)代表第t代时，i个个体的第j个维度，Δx i，j(t+1)是从优于i个个体身上学习的量，I i,j(t)代表i个个体的第j个维度与比i个个体优秀的个体对应维度的差值，C j,t则代表i个个体的第j个维度与当前种群所有个体第j个维度的平均值的差值，r 1(t)、r 2(t)与r 3(t)均为0到1之间的随机数，？为一个与问题维度和种群规模相关的数，φ设定值为1。
  9. 根据权利要求1所述的基于智能优化算法的繁殖指导方法，其特征在于，在所述当所述种群的数量达到所述预设规模时，进行选取适应度最高的那组雌雄个体的交配策略，直至选取次数达到预设选取次数之后还包括：
    当选取次数达到预设选取次数时，则输出适应度最高的那组雌雄个体的交配策略至显示模块进行显示。
  10. 一种基于智能优化算法的繁殖指导装置，其特征在于，包括：
    数据获取模块，用于获取种群初代的初始参数，所述初始参数包括基因型、期望基因型、寿命信息及致死基因信息；
    第一获取模块，用于将所述种群分为雌雄两组，获取雌雄两组随机进行交配产生所述种群的第二代的基因型，
    查询模块，用于查询所述第二代的基因型中是否有所述期望基因，
    成本计算模块，用于若有所述期望基因，则计算整个交配过程中得到所述期望基因型的时间和经济成本，并初始化所述初始参数；若未发现有所述期望基因，则继续使所述种群的初代随机进行交配，直至出现所述期望基因；
    第二获取模块，用于重复获取雌雄两组随机进行交配产生所述种群的第二代的基因型，直至所述种群的数量达到预设规模；
    策略选取模块，用于当所述种群的数量达到所述预设规模时，进行选取适应度最高的那组雌雄个体的交配策略，直至选取次数达到预设选取次数。
PCT/CN2022/138187 2022-01-26 2022-12-09 一种基于智能优化算法的繁殖指导方法及装置 WO2023142722A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210095564.2A CN114550822A (zh) 2022-01-26 2022-01-26 一种基于智能优化算法的繁殖指导方法及装置
CN202210095564.2 2022-01-26

Publications (1)

Publication Number Publication Date
WO2023142722A1 true WO2023142722A1 (zh) 2023-08-03

Family

ID=81674546

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/138187 WO2023142722A1 (zh) 2022-01-26 2022-12-09 一种基于智能优化算法的繁殖指导方法及装置

Country Status (2)

Country Link
CN (1) CN114550822A (zh)
WO (1) WO2023142722A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114550822A (zh) * 2022-01-26 2022-05-27 深圳先进技术研究院 一种基于智能优化算法的繁殖指导方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101843214A (zh) * 2009-03-25 2010-09-29 李晓方 农作物多基因型杂交一代种群的构建与种子生产方法
JP2016077273A (ja) * 2014-10-22 2016-05-16 静岡県 所望の形質を有する豚の作出方法
CN112215259A (zh) * 2020-09-17 2021-01-12 温州大学 基因选择方法和装置
CN113051148A (zh) * 2019-12-26 2021-06-29 南京邮电大学 一种基于双基因链遗传算法的组合测试用例生成方法
CN114550822A (zh) * 2022-01-26 2022-05-27 深圳先进技术研究院 一种基于智能优化算法的繁殖指导方法及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101843214A (zh) * 2009-03-25 2010-09-29 李晓方 农作物多基因型杂交一代种群的构建与种子生产方法
JP2016077273A (ja) * 2014-10-22 2016-05-16 静岡県 所望の形質を有する豚の作出方法
CN113051148A (zh) * 2019-12-26 2021-06-29 南京邮电大学 一种基于双基因链遗传算法的组合测试用例生成方法
CN112215259A (zh) * 2020-09-17 2021-01-12 温州大学 基因选择方法和装置
CN114550822A (zh) * 2022-01-26 2022-05-27 深圳先进技术研究院 一种基于智能优化算法的繁殖指导方法及装置

Also Published As

Publication number Publication date
CN114550822A (zh) 2022-05-27

Similar Documents

Publication Publication Date Title
Zenger et al. Genomic selection in aquaculture: application, limitations and opportunities with special reference to marine shrimp and pearl oysters
Brakefield Evo-devo and constraints on selection
Jennions et al. Why do females mate multiply? A review of the genetic benefits
Jablonka Inheritance systems and the evolution of new levels of individuality
Lehtonen et al. Evolutionary and ecological implications of sexual parasitism
WO2023142722A1 (zh) 一种基于智能优化算法的繁殖指导方法及装置
Anderson Large-scale parentage inference with SNPs: an efficient algorithm for statistical confidence of parent pair allocations
Dunham Utilization of transgenic fish in developing countries: potential benefits and risks
McAndrew et al. Application of genetics and genomics to aquaculture development: current and future directions
Ahrens et al. Standing genetic variation and compensatory evolution in transgenic organisms: a growth-enhanced salmon simulation
Veen et al. anyFish: an open-source software to generate animated fish models for behavioural studies
Kumar et al. Red flour beetle (Tribolium castaneum): From population genetics to functional genomics
Stearns Why sex evolved and the differences it makes
Urban et al. The repeated evolution of stripe patterns is correlated with body morphology in the adaptive radiations of East African cichlid fishes
Schneider et al. Gene flow between arrhenotokous and thelytokous populations of Venturia canescens (Hymenoptera)
Weller et al. Relaxed feeding constraints facilitate the evolution of mouthbrooding in Neotropical cichlids
LIVELY Trematode infection and the distribution and dynamics of parthenogenetic snail populations
Meffert et al. A test of speciation via sexual selection on female preferences
Yasui et al. The origination events of gametic sexual reproduction and anisogamy
Santos‐Santos et al. Ontogenetic divergence generates novel phenotypes in hybrid cichlids
Jaffe Emergence and maintenance of sex among diploid organisms aided by assortative mating
Aikio et al. Mating preference in the invasion of growth enhanced fish
Wiley Natural Selection
Lin et al. Computer simulation and mathematical modeling of the interactions between ecological and sexual selection to reveal the mechanism of sympatric speciation
KoudandÉ et al. A model for population growth of laboratory animals subjected to marker‐assisted introgression: how many animals do we need?

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22923511

Country of ref document: EP

Kind code of ref document: A1