Machine-learning-based genotype and environment interaction method and application thereof

Abstract

The present invention relates to the technical field of bioinformatics. Specifically disclosed are a machine-learning-based genotype and environment interaction method and the application thereof. The method comprises the following steps: step one, collecting environmental data of each growth and development period within a growth and development cycle of crops; step two, calculating an environmental index within a target growth and development period; step three, calculating a mean value of environmental indices and a comparative mean value of the environmental indices, and determining an environmental index of a growth and development period that has the greatest influence on the mean value of the environmental indices, i.e., an environmental index with the highest correlation; step four, performing calculation to obtain a phenotypic plasticity value of a target gene; step five, calculating a latent functional gene-environment influence parameter; and step six, determining whether a latent functional gene is an important latent functional gene under environmental influence. The present invention can mine critical factors that influence a growth process and phenotypic variation of crops, so as to formulate a cross-environment prediction strategy, optimize a variety selection path and help a breeder make production decisions, thereby advancing the plant breeding progress.

Description

A genotype-environment interaction method based on machine learning and its application Technical Field

[0001] The present invention relates to the technical field of bioinformatics, and in particular to a genotype and environment interaction method based on machine learning and its application. Background Art

[0002] In the fields of biology and genetic breeding, especially crop breeding, phenotype refers to the external characteristics of an organism, such as shape, structure, size, and color, which are determined by the genotype and environment. The phenotype group refers to all the characteristics of an organism. It is not limited to agronomic traits, but should also pay more attention to the physiological state of the plant.

[0003] A Chinese invention patent, published under the patent number CN110459265B, discloses a method for improving the accuracy of genome-wide prediction (GP). The method includes: (1) phenotypic and genotypic identification of a target crop population, and then, based on a genome-wide association study (GWAS) of the entire population, identifying the four single-nucleotide polymorphisms (SNPs) with the greatest effect; (2) using the four SNPs with the greatest effect as fixed effects and adding a genotype-by-environment interaction component to the GP model to maximize prediction accuracy.

[0004] Phenotypic variation is jointly promoted by genetics, environment, and their interactions. To cultivate crops with high yield and strong adaptability to new and changing climates, it is imperative to analyze the role of environmental factors. Although significant progress has been made in improving accuracy, existing genome and genotype × environment (G×E) prediction models lack interpretability. If the relative contributions of genes and environment cannot be accurately quantified and the specific underlying factors cannot be identified, many long-standing biological questions cannot be answered. Therefore, it is very necessary to establish a comprehensive framework with environmental dimensions for the analysis and prediction of complex traits. Summary of the Invention

[0005] The purpose of the present invention is to provide a genotype and environment interaction method based on machine learning and its application to solve the technical problems in the above background.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A genotype-environment interaction method based on machine learning and its application, comprising the following steps:

[0008] Step 1: Collect environmental data at each growth stage in the crop growth cycle;

[0009] Step 2: Calculate the environmental index during the target growth period based on environmental data;

[0010] Step 3: Based on the environmental indexes of all reproductive periods within the reproductive cycle, calculate the mean of the environmental index and the comparative mean of the environmental index, and determine the environmental index of the reproductive period that has the greatest impact on the mean of the environmental index, that is, the environmental index with the highest correlation;

[0011] Step 4: Calculate the phenotypic plasticity value of the target gene based on the highest correlation environmental index and the phenotype of the target gene;

[0012] Step 5: Calculate the environmental impact parameters of potential functional genes based on the phenotypic plasticity values;

[0013] Step 6: Based on the environmental impact parameters of the potential functional genes, determine whether the potential functional genes are important potential functional genes affected by the environment;

[0014] If the environmental impact parameter of the potential functional gene is less than the environmental impact parameter threshold of the potential functional gene, the potential functional gene is judged not to be a potential functional gene significantly affected by the environment;

[0015] If the environmental impact parameter of the potential functional gene is ≥ the environmental impact parameter threshold of the potential functional gene, the potential functional gene is determined to be a potential functional gene that is significantly affected by the environment.

[0016] As a further solution of the present invention: the environmental data includes: effective accumulated temperature, photosynthetically active radiation, effective moisture and soil pH.

[0017] As a further solution of the present invention: the specific calculation method of the environmental index is:

[0018] The effective accumulated temperature is marked as W n , photosynthetically active radiation is marked as G n , effective moisture is marked as S n , soil pH is marked as T n , and perform data processing; where n is different reproductive periods, and takes 1, 2, 3, ..., R, and R is a positive integer;

[0019] By formula: Calculate the environmental index Z n , where a1, a2, a3, and a4 are preset scaling factors, and a1, a2, and a3 are not equal to 0.

[0020] As a further solution of the present invention: the specific calculation method of the environmental index mean is:

[0021] A1: The default environment index with the highest correlation is Z i , where i = 1, 2, 3, ..., R, R is a positive integer;

[0022] A2: Calculate the mean environmental index based on the environmental index of all reproductive periods within the reproductive cycle;

[0023] By formula Calculate the mean value of the environmental index Where n is the different reproductive stages.

[0024] As a further solution of the present invention: the specific calculation method of the environmental index comparison mean is:

[0025] According to the mean of environmental index By formula Calculate the comparative mean of environmental index Where n is the different reproductive stages.

[0026] As a further solution of the present invention: the method for determining the environmental index with the highest correlation is:

[0027] The mean of the environmental index Comparison with the mean of environmental index The difference calculation is performed to obtain the index difference, and the index difference is compared and analyzed to determine the environmental index with the highest correlation. The environmental index with the highest correlation is the group with the largest index difference.

[0028] As a further solution of the present invention: based on the environmental index with the highest correlation and the phenotype of the target gene, the phenotypic plasticity value of the target gene is obtained in combination with the least squares method. The specific method is:

[0029] B1: By changing the highest correlation environment index Z e , thus obtaining different phenotypes of the target gene, and marking the phenotype of the target gene as X e , where e is a different environmental index, which takes 1, 2, 3, ..., R, and R is a positive integer;

[0030] B2: Based on multiple sets of data points (Z1, X1), (Z2, X2), ..., (Z e , X e ), find a straight line that minimizes the sum of the vertical distances from all data points to this line, and this straight line is the phenotypic plasticity value of the target gene.

[0031] As a further solution of the present invention: the specific calculation method of the environmental impact parameter of the potential functional gene is:

[0032] C1: Obtain the potential functional gene of the target gene and mark the potential functional gene as D j , where j is a different potential functional gene, and takes 1, 2, 3, ..., R, where R is a positive integer;

[0033] Among them, potential functional genes include: gene sequence, haplotype, SNP (single nucleotide polymorphism);

[0034] C2: Change the highest correlation environment index Z within the calibration range e , and record the number of potential functional gene changes and the sum of the magnitude of changes in potential functional genes

[0035] The calibration range is: the range of changes in the environmental index with the highest correlation that causes only a single change in the gene phenotype;

[0036] The ratio of potential functional gene changes is the ratio of the number of functional gene changes to the number of changes in the environmental index with the highest correlation;

[0037] C3: Ratio of the number of potential functional gene changes to The sum of the change amplitudes F of potential functional genes when they change is used for data processing, using the formula: Calculate the environmental impact parameters of potential functional genes Among them, b1 and b2 are weight proportional factors, and both are greater than 0.

[0038] As a further solution of the present invention: presetting the potential functional gene environmental impact parameter threshold value as Y1, the potential functional gene environmental impact parameter Compare and analyze with the threshold value Yl of the environmental impact parameter of the potential functional gene to determine whether the potential functional gene is a potential functional gene that is significantly affected by the environment;

[0039] like This indicates that the environment has little effect on the potential functional gene, and the potential functional gene is not a potential functional gene that is significantly affected by the environment;

[0040] like This indicates that the environment has a great influence on the potential functional gene, and determines that the potential functional gene is a potential functional gene that is significantly affected by the environment.

[0041] Beneficial effects of the present invention:

[0042] (1) The present invention fully mines environmental information by utilizing artificial intelligence algorithms to analyze the phenotypic plasticity of important agricultural traits during critical growth periods, analyze the interaction between genes and the environment, and predict the phenotypes of important agricultural traits;

[0043] (2) The present invention utilizes the interaction between genotype and environment to select varieties that adapt to climate change, matches genotype and environmental type, explores the key factors affecting crop growth process and phenotypic variation, formulates cross-environment prediction strategies, optimizes variety selection paths, and helps breeders make production decisions, thereby promoting the process of plant breeding. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The present invention will be further described below with reference to the accompanying drawings.

[0045] FIG1 is a schematic diagram of the steps of the method of the present invention;

[0046] FIG2 is a diagram of the least squares method of the present invention;

[0047] FIG3 is a schematic diagram of the steps for determining potential functional genes that are significantly affected by the environment in the present invention. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0049] Example 1

[0050] Referring to Figures 1 and 2 , the present invention provides a genotype-environment interaction method based on machine learning, comprising the following steps:

[0051] Step 1: Collect environmental data at each growth stage in the crop growth cycle;

[0052] Among them, environmental data include: effective accumulated temperature, photosynthetically active radiation, effective moisture and soil pH;

[0053] The effective accumulated temperature is marked as W n , photosynthetically active radiation is marked as G n , effective moisture is marked as S n , soil pH is marked as T n , where n is different reproductive periods and takes 1, 2, 3, ..., R, where R is a positive integer;

[0054] It should be noted that the crop growth cycle is the time from sowing to seed maturity, expressed in days. For some crops such as hemp, potatoes, sugarcane, and green manure, the term refers to the time from sowing to harvesting the main product.

[0055] Growth period refers to the different growth stages of crops, which are divided into several periods according to the order and morphological characteristics of their organs during the entire growth process; for example, winter wheat is divided into seedling stage, three-leaf stage, tillering stage, overwintering stage, greening stage, jointing stage, booting stage, heading stage, flowering stage, and maturity stage;

[0056] Effective accumulated temperature is the sum of the effective temperatures of crops during their growth period, that is, the sum of the differences between the average daily temperature and biological zero degrees Celsius during the growth period. Effective accumulated temperature reflects the heat demand of biological growth and development.

[0057] Photosynthetically active radiation is the solar radiation that can be used by green plants for photosynthesis. Its wavelength range is between 380 and 710 nanometers. Photosynthetically active radiation is the main energy source for biomass formation and the main factor affecting crop photosynthesis.

[0058] Available water refers to the amount of water in the soil that can be absorbed and utilized by crops. When there is too much or too little water in the soil, it will affect the growth and yield of crops.

[0059] Soil pH is the degree of acidity or alkalinity of the soil. Soil pH is one of the important factors affecting soil fertility. It not only affects the effectiveness of soil nutrients, but also affects the activity of microorganisms in the soil, thereby affecting the growth and yield of crops.

[0060] Step 2: Calculate the environmental index during the target growth period based on the environmental data. The specific calculation method is:

[0061] The effective accumulated temperature, photosynthetically active radiation, effective moisture and soil pH are processed and the formula is used: Calculate the environmental index Z n , where a1, a2, a3, and a4 are preset scaling factors, and a1, a2, and a3 are not equal to 0;

[0062] Step 3: Based on the environmental indexes of all reproductive periods within the reproductive cycle, calculate the mean of the environmental index and the comparative mean of the environmental index, and determine the environmental index of the reproductive period that has the greatest impact on the mean of the environmental index, that is, the environmental index with the highest correlation. The specific method is as follows:

[0063] A1: The default environment index with the highest correlation is Z i , where i = 1, 2, 3, ..., R, R is a positive integer;

[0064] A2: Calculate the mean environmental index based on the environmental index of all reproductive periods within the reproductive cycle;

[0065] By formula Calculate the mean value of the environmental index Where n is different reproductive periods;

[0066] A3: After calculating the environmental index excluding the one with the highest correlation, compare the mean of the environmental index;

[0067] According to the mean of environmental index By formula Calculate the comparative mean of environmental index Where n is different reproductive periods;

[0068] A4: Average the environmental index Comparison with the mean of environmental index Perform difference calculation to obtain index difference, and compare and analyze the index difference to determine the environmental index with the highest correlation. The environmental index with the highest correlation is the group with the largest index difference.

[0069] Step 4: Based on the highest correlation environmental index and the phenotype of the target gene, the phenotypic plasticity value of the target gene is obtained by combining the least squares method. The specific method is as follows:

[0070] B1: By changing the highest correlation environment index Z e , thus obtaining different phenotypes of the target gene, and marking the phenotype of the target gene as X e , where e is a different environmental index, which takes 1, 2, 3, ..., R, and R is a positive integer;

[0071] B2: Based on multiple sets of data points (Z1, X1), (Z2, X2), ..., (Z e , X e ), find a straight line that minimizes the sum of the vertical distances from all data points to this line, and this straight line is the phenotypic plasticity value of the target gene.

[0072] Example 2

[0073] Based on Example 1, as shown in FIG3 , the present invention is a genotype and environment interaction method based on machine learning, further comprising: calculating the environmental impact parameter of the potential functional gene according to the phenotypic plasticity value, and determining whether the potential functional gene is a potential functional gene that is significantly affected by the environment;

[0074] The purpose of determining whether a potential functional gene is an environmentally significant potential functional gene is to explore key factors that affect crop growth and phenotypic variation, thereby developing cross-environmental prediction strategies, optimizing variety selection paths, and helping breeders make production decisions, thereby promoting plant breeding.

[0075] C1: Obtain the potential functional gene of the target gene and mark the potential functional gene as D j, where j is a different potential functional gene, and takes 1, 2, 3, ..., R, where R is a positive integer;

[0076] Among them, potential functional genes include: gene sequence, haplotype, SNP (single nucleotide polymorphism);

[0077] C2: Change the highest correlation environment index Z within the calibration range e , and record the number of potential functional gene changes and the sum of the magnitude of changes in potential functional genes

[0078] The calibration range is: the range of changes in the environmental index with the highest correlation that causes only a single change in the gene phenotype;

[0079] The ratio of potential functional gene changes is the ratio of the number of functional gene changes to the number of changes in the environmental index with the highest correlation;

[0080] C3: Ratio of the number of potential functional gene changes to The sum of the change amplitudes F of potential functional genes when they change is used for data processing, using the formula: Calculate the environmental impact parameters of potential functional genes Among them, b1 and b2 are weight proportional factors, and both are greater than 0;

[0081] C4: The threshold of the environmental impact parameter of the potential functional gene is preset as Yl, and the environmental impact parameter of the potential functional gene is set as Compare and analyze with the threshold value Yl of the environmental impact parameter of the potential functional gene to determine whether the potential functional gene is a potential functional gene that is significantly affected by the environment;

[0082] like This indicates that the environment has little effect on the potential functional gene, and the potential functional gene is not a potential functional gene that is significantly affected by the environment;

[0083] like This indicates that the environment has a great influence on the potential functional gene, and determines that the potential functional gene is a potential functional gene that is significantly affected by the environment.

[0084] Example 3

[0085] A machine learning-based genotype-environment interaction method for application in environmental processing.

[0086] The working principle of the present invention includes the following steps: step 1: collecting environmental data of each growth period within a crop growth cycle; step 2: calculating an environmental index within a target growth period according to the environmental data; step 3: calculating an environmental index mean and an environmental index comparison mean according to the environmental indexes of all growth periods within the growth cycle, and determining the growth period environmental index that has the greatest impact on the environmental index mean, that is, the environmental index with the highest correlation; step 4: calculating a phenotypic plasticity value of the target gene according to the environmental index with the highest correlation and the phenotype of the target gene; step 5: calculating an environmental impact parameter of a potential functional gene according to the phenotypic plasticity value; and step 6: determining whether the potential functional gene is a potential functional gene that is significantly affected by the environment according to the environmental impact parameter of the potential functional gene; if the environmental impact parameter of the potential functional gene is less than a threshold value of the environmental impact parameter of the potential functional gene, determining that the potential functional gene is not a potential functional gene that is significantly affected by the environment; and if the environmental impact parameter of the potential functional gene is greater than or equal to the threshold value of the environmental impact parameter of the potential functional gene, determining that the potential functional gene is a potential functional gene that is significantly affected by the environment.

[0087] The above is a detailed description of an embodiment of the present invention. However, the content described is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A genotype-environment interaction method based on machine learning, characterized in that: The following steps are involved: Step 1: Collect environmental data at each growth stage in the crop growth cycle; Step 2: Calculate the environmental index during the target growth period based on environmental data; Step 3: Based on the environmental indexes of all reproductive periods within the reproductive cycle, calculate the mean of the environmental index and the comparative mean of the environmental index, and determine the environmental index of the reproductive period that has the greatest impact on the mean of the environmental index, that is, the environmental index with the highest correlation; Step 4: Calculate the phenotypic plasticity value of the target gene based on the highest correlation environmental index and the phenotype of the target gene; Step 5: Calculate the environmental impact parameters of potential functional genes based on the phenotypic plasticity values; Step 6: Based on the environmental impact parameters of the potential functional genes, determine whether the potential functional genes are important potential functional genes affected by the environment; If the environmental impact parameter of the potential functional gene is less than the environmental impact parameter threshold of the potential functional gene, the potential functional gene is judged not to be a potential functional gene significantly affected by the environment; If the environmental impact parameter of the potential functional gene is ≥ the environmental impact parameter threshold of the potential functional gene, the potential functional gene is determined to be a potential functional gene that is significantly affected by the environment.

2. A genotype and environment interaction method based on machine learning according to claim 1, characterized in that: The environmental data include: effective accumulated temperature, photosynthetically active radiation, effective moisture and soil pH.

3. The genotype-environment interaction method based on machine learning according to claim 1, characterized in that: The specific calculation method of the environmental index is: The effective accumulated temperature is marked as W n , photosynthetically active radiation is marked as G n , effective moisture is marked as S n , soil pH is marked as T n , and perform data processing; where n is different reproductive periods, and takes 1, 2, 3, ..., R, and R is a positive integer; By formula: Calculate the environmental index Z n , where a1, a2, a3, and a4 are preset scaling factors, and a1, a2, and a3 are not equal to 0.

4. A genotype and environment interaction method based on machine learning according to claim 3, characterized in that: The specific calculation method of the environmental index mean is: A1: The default environment index with the highest correlation is Z i , where i = 1, 2, 3, ... R, R is a positive integer; A2: Calculate the mean environmental index based on the environmental index of all reproductive periods within the reproductive cycle; By formula Calculate the mean value of the environmental index Where n is the different reproductive stages.

5. The genotype and environment interaction method based on machine learning according to claim 4, characterized in that: The specific calculation method of the environmental index comparison mean is: According to the mean of environmental index By formula Calculate the comparative mean of environmental index Where n is the different reproductive stages.

6. The genotype and environment interaction method based on machine learning according to claim 5, characterized in that: The method for determining the environmental index with the highest correlation is: The mean of the environmental index Comparison with the mean of environmental index The difference calculation is performed to obtain the index difference, and the index difference is compared and analyzed to determine the environmental index with the highest correlation. The environmental index with the highest correlation is the group with the largest index difference.

7. The genotype and environment interaction method based on machine learning according to claim 6, characterized in that: Based on the highest correlation environmental index and the phenotype of the target gene, the phenotypic plasticity value of the target gene is obtained by combining the least squares method. The specific method is as follows: B1: By changing the highest correlation environment index Z e , thus obtaining different phenotypes of the target gene, and marking the phenotype of the target gene as X e , where e is a different environmental index, and the value of e is 1, 2, 3, ..., R, and R is a positive integer; B2: Based on multiple sets of data points (Z1, X1), (Z2, X2), ..., (Z e , X e ), find a straight line that minimizes the sum of the vertical distances from all data points to this line, and this straight line is the phenotypic plasticity value of the target gene.

8. The genotype-environment interaction method based on machine learning according to claim 1, characterized in that: The specific calculation method of the environmental impact parameter of the potential functional gene is: C1: Obtain the potential functional gene of the target gene and mark the potential functional gene as D j , where j is a different potential functional gene, and takes 1, 2, 3, ..., R, where R is a positive integer; Among them, potential functional genes include: gene sequence, haplotype, SNP; C2: Change the highest correlation environment index Z within the calibration range e , and record the number of potential functional gene changes and the sum of the magnitude of changes in potential functional genes The calibration range is: the range of changes in the environmental index with the highest correlation that causes only a single change in the gene phenotype; The ratio of potential functional gene changes is the ratio of the number of functional gene changes to the number of changes in the environmental index with the highest correlation; C3: Ratio of the number of potential functional gene changes to The sum of the change amplitudes F of potential functional genes when they change is used for data processing, using the formula: Calculate the environmental impact parameters of potential functional genes Among them, b1 and b2 are weight proportional factors, and both are greater than 0.

9. The genotype-environment interaction method based on machine learning according to claim 1, characterized in that: The threshold value of the environmental impact parameter of the potential functional gene is preset as Yl, and the environmental impact parameter of the potential functional gene is set as Compare and analyze with the threshold value Yl of the environmental impact parameter of the potential functional gene to determine whether the potential functional gene is a potential functional gene that is significantly affected by the environment; like This indicates that the environment has little effect on the potential functional gene, and the potential functional gene is not a potential functional gene that is significantly affected by the environment; like This indicates that the environment has a great influence on the potential functional gene, and determines that the potential functional gene is a potential functional gene that is significantly affected by the environment.

10. An application of the genotype and environment interaction method based on machine learning as described in any one of claims 1 to 9 in environmental processing.

Images