Deep learning-based method for breeding and multiplying penaeus vannamei
By using a deep learning-based method for the propagation of superior varieties of Litopenaeus vannamei, deep learning models are used to analyze the genomic and phenotypic data of shrimp, quickly screening out gene sequences that meet the phenotypic requirements. This solves the problems of phenotypic degeneration and weak disease resistance in Litopenaeus vannamei, improves breeding efficiency and disease resistance, and reduces market prices.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUANGDONG YUEHAI FEED GROUP
- Filing Date
- 2024-08-20
- Publication Date
- 2026-06-05
Smart Images
Figure CN119032879B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of bioinformatics, and in particular to a method for the propagation of superior varieties of Litopenaeus vannamei based on deep learning. Background Technology
[0002] The whiteleg shrimp is the most farmed shrimp species in my country and has a very high market value.
[0003] my country relies heavily on imports for high-quality broodstock shrimp, meaning that broodstock shrimp are purchased from abroad for breeding. However, broodstock breeding suffers from problems such as phenotypic degeneration and weakened disease resistance. When broodstock shrimp phenotypic degeneration or death occurs, they must be imported again. Especially in recent years, foreign broodstock shrimp companies have been selling only shrimp larvae and not broodstock, further eroding the profits of my country's shrimp farming industry.
[0004] Shrimp are very fragile, and the crowded environment in shrimp farms can easily lead to disease outbreaks and large-scale shrimp deaths. This is why shrimp prices are generally higher than other meats in the market.
[0005] For example, in 2009, an outbreak of Vibrio parahaemolyticus infection caused large numbers of shrimp larvae to die due to early mortality syndrome and acute hepatopancreatic necrosis syndrome, resulting in huge economic losses.
[0006] Breeding shrimp varieties with superior traits such as strong disease resistance and rapid growth is considered the fundamental way to solve disease problems.
[0007] In 2019, the Chinese Academy of Sciences completed the genome sequencing of the Litopenaeus vannamei and released the first reference genome data for the shrimp. This laid the foundation for precise molecular breeding of shrimp.
[0008] In molecular breeding of shrimp, it is necessary to assess the degree of association between specific genes and specific traits, and to identify genotypes associated with desirable traits for breeding.
[0009] The traits that need to be selected for in Litopenaeus vannamei (whiteleg shrimp) are generally based on growth rate, disease resistance, meat yield, and feed consumption. Most of these traits are quantitative traits controlled by multiple genes; that is, the expression of shrimp traits is influenced by the combined effects of multiple gene loci. Furthermore, the heritability of some shrimp traits is relatively low, and they are significantly affected by environmental factors.
[0010] According to the reference genome of shrimp, there are a large number of species-specific genes and a large number of tandem repeat genes in shrimp, which makes it difficult to analyze the correlation between genes and traits in individual shrimp.
[0011] Identifying genes or gene loci closely related to traits is crucial for the breeding of superior shrimp. The Pacific white shrimp has 2.4 billion base pairs, of which approximately 24% are simple repeating sequences (i.e., microsatellites). Therefore, genotyping for breeding traits requires extensive DNA sequencing and analysis of Pacific white shrimp.
[0012] To rapidly conduct molecular selection for superior traits of Litopenaeus vannamei, this application provides a deep learning-based method for the propagation of superior Litopenaeus vannamei strains. Summary of the Invention
[0013] To overcome the problems existing in related technologies, this application provides a deep learning-based method for the propagation of improved varieties of Litopenaeus vannamei, including:
[0014] Obtain genomic data of the first-generation population; the genomic data includes numbered gene sequences; the first-generation population is a group of broodstock shrimp bred from the same maternal line but different paternal lines;
[0015] Obtain phenotypic data of the first-generation population;
[0016] A gene ID set is determined based on the degree of correlation between the phenotypic data and the genomic data; the gene ID set is a set of gene sequences associated with the selected traits; and broodstock shrimp for breeding are determined based on the gene ID set.
[0017] In one implementation, obtaining the genomic data of the initial population specifically includes:
[0018] High-throughput whole-genome sequencing was performed on each broodstock shrimp in the first generation population to obtain the complete genome sequence of the first generation population;
[0019] Obtain a reference map of the shrimp genome;
[0020] The whole genome sequence is numbered according to the shrimp genome reference map to obtain the genome data.
[0021] In one implementation, obtaining the phenotypic data of the initial population specifically includes:
[0022] The first-generation population was cultivated in different breeding environments;
[0023] Collect and record the phenotypic data of the first generation population.
[0024] In one implementation, determining the gene ID set based on the degree of association between the phenotypic data and the genomic data specifically includes:
[0025] Training a model for predicting superior shrimp breeds;
[0026] The phenotypic data and genomic data of each labeled shrimp are input into the shrimp breeding prediction model to obtain the set of gene codes associated with the specific trait.
[0027] In one implementation, the training of the shrimp breeding prediction model specifically includes:
[0028] A sample dataset is obtained, which is constructed based on the reference genome data of shrimp and the trait annotation data corresponding to the reference genome data;
[0029] The sample dataset is divided into a training set and a test set;
[0030] Construct the shrimp breeding prediction model, which is a neural network model with a P-NET structure;
[0031] The shrimp breeding prediction model is iteratively trained using the training set to obtain the trained shrimp breeding prediction model.
[0032] In one implementation, the construction of the shrimp reference genome data and the corresponding trait annotation data specifically includes:
[0033] Obtain reference genome data for shrimp;
[0034] The reference genome data is labeled to obtain reference genome data with real trait labels.
[0035] In one embodiment, the network structure of the shrimp breeding prediction model includes an input layer, a gene layer, a biological pathway layer, a fully connected layer, and an output layer.
[0036] Each node in the gene layer represents a numbered gene sequence; the output of the gene layer is connected to the biological pathway layer, where each node represents a biological pathway; the biological pathway layer is used to associate gene expression patterns with known biological pathways; the fully connected layer is used to sense the degree of association between the phenotypic data and the gene sequences, and to send gene sequence sets with an association degree higher than a preset threshold to the output layer.
[0037] In one implementation, determining the broodstock shrimp for improved breeding based on the gene ID set specifically includes:
[0038] Obtain each gene number in the gene number set, and the gene sequence corresponding to the gene number;
[0039] Gene chips are fabricated based on the gene sequences; the next generation of broodstock shrimp or broodstock shrimp with stable phenotypic traits are screened using the gene chips.
[0040] The technical solution provided in this application may include the following beneficial effects:
[0041] This application first cultivates shrimp populations from different families to obtain phenotypic and genomic data of shrimp from different families. Based on the degree of correlation between phenotypic and genomic data, gene sequences that meet the phenotypic requirements are determined, that is, the gene sequences related to specific traits are located. Then, the next generation of shrimp is selected based on the set of these gene sequences, or bred shrimp with stable traits are directly selected.
[0042] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description
[0043] The above and other objects, features and advantages of this application will become more apparent from the more detailed description of exemplary embodiments thereof in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments thereof.
[0044] Figure 1 This is a schematic flowchart illustrating the method for propagating improved varieties of Litopenaeus vannamei as shown in an embodiment of this application. Detailed Implementation
[0045] Preferred embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While preferred embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.
[0046] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.
[0047] It should be understood that although the terms "first," "second," "third," etc., may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.
[0048] Currently, in aquaculture, the breeding of whiteleg shrimp mainly involves family selection, which involves constructing shrimp populations of different families for cultivation and selecting shrimp families that meet specific traits.
[0049] Current family breeding methods all involve phenotypic selection, which means dividing the shrimp into different control groups for breeding, collecting phenotypic data of each shrimp family, and selecting shrimp that meet the requirements as broodstock through generations.
[0050] Phenotypic breeding cannot determine the relationship between genes and traits, or between environment and traits. It is common for shrimp of the same lineage to degenerate in different environments or after several generations of inbreeding.
[0051] Although family breeding can utilize genetic principles and maintain desirable traits in a population through hybridization, it relies on the breeder's experience and is relatively inefficient.
[0052] This application's embodiments are based on genome-wide selection breeding for the selection of superior shrimp breeds. Genome-wide selection breeding requires constructing a genetic model based on the association between genotypes and phenotypes across the entire genome of a shrimp population. This necessitates collecting shrimp DNA sequences and performing association analysis with traits, a data analysis process that consumes a significant amount of time.
[0053] Therefore, this application proposes a deep learning-based method for the propagation of superior varieties of Litopenaeus vannamei, which can complete the data analysis work through a deep learning model and accelerate the breeding process of superior varieties.
[0054] In the embodiments of this application, a breeding population of Litopenaeus vannamei was first established. Fifty paternal Litopenaeus vannamei and one maternal Litopenaeus vannamei were introduced for family breeding. The resulting first-generation population included a population of 50 different paternal lines.
[0055] like Figure 1 As shown in the embodiments of this application, a method for the propagation of improved varieties of Litopenaeus vannamei based on deep learning includes the following steps:
[0056] 101. Obtain the genomic data of the first-generation population;
[0057] The genomic data includes numbered gene sequences; the primary population consists of broodstock shrimp populations bred from the same maternal line but different paternal lines.
[0058] 102. Obtain phenotypic data of the first-generation population;
[0059] 103. Determine a gene ID set based on the degree of correlation between the phenotypic data and the genomic data; the gene ID set is a set of gene sequences associated with the selected trait;
[0060] 104. Determine the breeding shrimp for improved breeding based on the gene number set.
[0061] In step 101, obtaining the genomic data of the initial population specifically includes the following steps:
[0062] 1011. High-throughput whole-genome sequencing was performed on each broodstock shrimp in the first generation population to obtain the whole genome sequence of the first generation population;
[0063] In this embodiment of the application, the whole genome sequence of each shrimp species is numbered and stored separately.
[0064] 1012. Obtain the shrimp gene reference map;
[0065] 1013. The whole genome sequence is numbered according to the shrimp gene reference map to obtain the genome data.
[0066] The shrimp genome reference map is based on the publicly available genome series and annotation information of Litopenaeus vannamei. In this application, the gene sequences of 50 broodstock shrimp were numbered and annotated to obtain genomic data. Gene sequences not annotated in the reference map were recorded as sequences to be annotated and numbered.
[0067] In step 102, the phenotypic data of the first-generation population are obtained, specifically including the following steps:
[0068] 1021. The first-generation population was cultured in different breeding environments;
[0069] 1022. Collect and record the phenotypic data of the first generation population.
[0070] For example, in a culture environment for genotypic screening of anti-Vibrio traits, multiple shrimp individuals from different families are randomly selected, with each family representing a genotype.
[0071] Immersion infection was performed using Vibrio parahaemolyticus. The Vibrio infection experiment lasted for 8 days. The first 96 shrimp to die were taken as the sensitive group samples, and the last 96 surviving shrimp were taken as the resistant individuals. The survival and death times were recorded as phenotypic data.
[0072] In addition to anti-Vibrio traits, phenotypic data can also be collected for growth traits, meat yield traits, or feed conversion ratio traits.
[0073] In this embodiment of the application, the phenotypic types of shrimp from the same family that have been observed are encoded to obtain the phenotypic data.
[0074] In step 103, the gene ID set is determined based on the degree of correlation between the phenotypic data and the genomic data, specifically as follows:
[0075] The gene number set is determined based on the phenotypic data, the genomic data, and the shrimp breeding prediction model.
[0076] Specifically, it includes the following steps:
[0077] 1031. Training a model for predicting superior shrimp breeds;
[0078] 1032. Input the phenotypic data and genomic data of each labeled shrimp into the shrimp breeding prediction model to obtain the gene ID set related to the specific trait.
[0079] In this embodiment of the application, the pre-trained shrimp breeding prediction model compares the genomic data of shrimp populations from various paternal lines to determine the set of gene codes associated with the anti-Vibrio phenotypic trait.
[0080] A gene ID set may include one or more gene IDs, depending on whether the phenotypic trait is a single gene or multiple genes.
[0081] Specifically, the process of training a shrimp breeding prediction model is as follows:
[0082] 310. Obtain the sample dataset;
[0083] The sample dataset is constructed based on the reference genome data of shrimp and the trait annotation data corresponding to the reference genome data;
[0084] The method for obtaining the sample training set in this embodiment of the application is as follows:
[0085] Obtain reference genome data and reference trait data for shrimp;
[0086] The reference genome data is labeled to obtain reference genome data with real trait labels.
[0087] 320. Divide the sample dataset into a training set and a test set;
[0088] 330. Construct the aforementioned shrimp breeding prediction model;
[0089] Specifically, the shrimp breeding prediction model is a P-NET structured neural network model, where the P-NET structure utilizes React... e Gene-pathway relationships in the dataset.
[0090] In this embodiment of the application, the network structure of the shrimp breeding prediction model includes an input layer, a gene layer, a biological pathway layer, a fully connected layer, and an output layer.
[0091] Specifically, each node in the gene layer represents a numbered gene sequence; the output of the gene layer is connected to the biological pathway layer, where each node represents a biological pathway. The biological pathway layer is used to associate gene expression patterns with known biological pathways; the fully connected layer is used to sense the degree of association between the phenotypic data and the gene sequences, and sends the set of gene sequences with an association degree higher than a preset threshold to the output layer.
[0092] In the input layer, the genomic data and phenotypic data of the first-generation population are encoded as: (x1, x2…x i ,l) n .
[0093] Where, x i Let represent the gene sequence of the i-th number, l represent phenotypic data, and n represent the individual shrimp number.
[0094] In the biological pathway layer, each node unit calculates the degree of association between phenotypic data and the i-th numbered gene sequence, extracts gene sequences with high association, and excludes gene sequences that are not related to the current phenotypic.
[0095] The biological pathway layer in this application identifies gene numbers or gene sequences associated with the target trait, while excluding data from other gene sequences.
[0096] In the fully connected layer, the confidence level of each gene number is calculated to determine the degree of association between the phenotypic data and the gene sequence, and the output layer outputs the gene number set.
[0097] The activation function between the fully connected layer and the output layer is:
[0098]
[0099] Where, x in Let be the probability distribution value of the i-th gene sequence of the n-th labeled shrimp individual.
[0100] This application embodiment classifies each gene sequence by calculating the probability distribution between each gene sequence and phenotypic in each shrimp individual, thereby obtaining a gene number set.
[0101] 340. The shrimp breeding prediction model is iteratively trained using the training set to obtain the trained shrimp breeding prediction model.
[0102] After obtaining shrimp individuals with stable traits, this embodiment of the application designs a gene chip based on the gene sequence set of the individual as a probe to select shrimp for the next iteration, specifically including the following steps:
[0103] Obtain each gene number in the gene number set, and the gene sequence corresponding to the gene number;
[0104] Gene chips are fabricated based on the gene sequences; the next generation of broodstock shrimp or broodstock shrimp with stable phenotypic traits are screened using the gene chips.
[0105] This application first cultivates shrimp populations from different families to obtain phenotypic and genomic data of shrimp from different families. Based on the degree of correlation between phenotypic and genomic data, gene sequences that meet the phenotypic requirements are determined, that is, the gene sequences related to specific traits are located. Then, the next generation of shrimp is selected based on the set of these gene sequences, or bred shrimp with stable traits are directly selected.
[0106] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments concerning the apparatus in the above embodiments, and will not be elaborated further here.
[0107] The solution of this application has been described in detail above with reference to the accompanying drawings. In the above embodiments, the descriptions of each embodiment have different emphases; parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. Those skilled in the art should also understand that the actions and modules involved in the specification are not necessarily essential to this application. Furthermore, it is understood that the steps in the method of this application embodiment can be adjusted, combined, and deleted according to actual needs, and the modules in the device of this application embodiment can be combined, divided, and deleted according to actual needs.
[0108] Furthermore, the method according to this application can also be implemented as a computer program or computer program product, which includes computer program code instructions for performing some or all of the steps in the method described above.
[0109] Alternatively, this application may be implemented as a non-transitory machine-readable storage medium (or computer-readable storage medium, or machine-readable storage medium) storing executable code (or computer program, or computer instruction code) that, when executed by a processor of an electronic device (or electronic device, server, etc.), causes the processor to perform some or all of the steps of the methods described above according to this application.
[0110] Those skilled in the art will also understand that the various exemplary logic blocks, modules, circuits, and algorithm steps described in connection with the present application can be implemented as electronic hardware, computer software, or a combination of both.
[0111] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems and methods according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0112] The various embodiments of this application have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.
Claims
1. A method for the propagation of superior strains of Litopenaeus vannamei based on deep learning, characterized in that, include: Obtain genomic data of the first-generation population; the genomic data includes numbered gene sequences; the first-generation population is a group of broodstock shrimp bred from the same maternal line but different paternal lines; Obtain phenotypic data of the first-generation population; The gene ID set is determined based on the degree of correlation between the phenotypic data and the genomic data; the gene ID set is a set of gene sequences associated with the selected traits. The broodstock shrimp for improved breeding were determined based on the gene ID set; The process of determining the gene ID set based on the correlation between the phenotypic data and the genomic data specifically includes: Training a model for predicting superior shrimp breeds; The phenotypic data and genomic data of each labeled shrimp are input into the shrimp breeding prediction model to obtain a set of gene codes related to specific traits; Training a shrimp breeding prediction model specifically includes: A sample dataset is obtained, which is constructed based on the reference genome data of shrimp and the trait annotation data corresponding to the reference genome data; The sample dataset is divided into a training set and a test set; Construct the shrimp breeding prediction model, which is a neural network model with a P-NET structure; The shrimp breeding prediction model is iteratively trained using the training set to obtain the trained shrimp breeding prediction model. The construction of the shrimp reference genome data and the corresponding trait annotation data specifically includes: Obtain reference genome data for shrimp; The reference genome data is labeled to obtain reference genome data with real trait labels; The network structure of the shrimp breeding prediction model includes an input layer, a gene layer, a biological pathway layer, a fully connected layer, and an output layer. Each node in the gene layer represents a numbered gene sequence; the output of the gene layer is connected to the biological pathway layer, where each node represents a biological pathway; the biological pathway layer is used to associate gene expression patterns with known biological pathways; the fully connected layer is used to sense the degree of association between the phenotypic data and the gene sequences, and to send gene sequence sets with an association degree higher than a preset threshold to the output layer.
2. The method for propagating superior varieties of Litopenaeus vannamei based on deep learning according to claim 1, characterized in that, The acquisition of genomic data from the initial population specifically includes: High-throughput whole-genome sequencing was performed on each broodstock shrimp in the first generation population to obtain the complete genome sequence of the first generation population; Obtain a reference map of the shrimp genome; The whole genome sequence is numbered according to the shrimp genome reference map to obtain the genome data.
3. The method for propagating superior varieties of Litopenaeus vannamei based on deep learning according to claim 1, characterized in that, The acquisition of phenotypic data from the first-generation population specifically includes: The first-generation population was cultivated in different breeding environments; Collect and record the phenotypic data of the first generation population.
4. The method for propagating superior varieties of Litopenaeus vannamei based on deep learning according to claim 1, characterized in that, The seed shrimp selected for improved breeding based on the gene ID set specifically include: Obtain each gene number in the gene number set, and the gene sequence corresponding to the gene number; Gene chips are fabricated based on the gene sequences; the next generation of broodstock shrimp or broodstock shrimp with stable phenotypic traits are screened using the gene chips.