Grading method and evaluation system for health level of mammals eating transgenic crops

Abstract

The invention relates to a grading method and an evaluation system for health levels of mammals eating transgenic crops, and belongs to the field of bioinformatics and biological big data analysis. According to the method, 26 pieces of physiological index data in 6 dimensions are screened and integrated; creating a data preprocessing module for the limited pairing array; obtaining global optimal mixed probability distribution of the physiological indexes, and determining a health interval; and establishing an algorithm for evaluating the health status of the mammal: establishing a three-level score data output structural body, establishing an input data structural body, obtaining a single physiological score value and a weight thereof, obtaining equal weight of six dimensions, and obtaining four grades of the health status. According to the invention, the limitations of selection of different models of animals, group difference, incomplete indexes, small sample size and the like in a single toxicological experiment are solved, and a health status grading algorithm and evaluation and human-computer interaction visualization system for diversified physiological indexes by integrating limited paired original experimental data obtained based on the single experiment is provided.

Description

Technical Field

[0001] The present invention relates to the fields of bioinformatics and biological big data analysis, and more specifically, to a grading method and an evaluation system for the health level of mammals that eat genetically modified crops, and in particular to a method and a system based on big data analysis and Python language to evaluate and grade the health level of mammals after eating genetically modified crops (through changes in physiological indicators of mammals that ingest genetically modified crops). Background Art

[0002] Genetically modified crops have shown great potential in the global food crisis due to their enhanced stress resistance and higher yields. In order to ensure the high-quality and large-scale application of genetically modified (GM) crops, a comprehensive evaluation of their nutritional value and toxicological safety is essential. Among various GM crops, GM corn (such as DBN9936 and Ruifeng 125 / BT corn) can increase yield by 16.4%, GM rice (such as IPA1-Pro10) can increase yield by 15.9%, and GM soybean (such as ric1a / 2a) can increase yield by more than 20%. The above three GM crops play an important role as staple food and feed. When these GM crops are used as food supply for mammals and humans, people still have concerns about the safety of these GM crops and their by-products (especially the impact on human and animal health).

[0003] At present, the main method for evaluating the health of mammals that eat GM crops is to conduct toxicology experiments using model mammals, such as the 13-week GM crop feeding experiment on rats or mice and the 52-week GM crop feeding experiment on crab-eating macaques. The basic process of the experiment includes detecting physiological indicators of mammals, such as red blood cell concentration (RBC), serum total cholesterol concentration, etc., variance analysis of various physiological indicators, and judging the significance of the differences in various physiological indicators between the experimental group and the control group after variance analysis. The operation is complicated and time-consuming. In addition, the existing independent toxicology experiments have not systematically classified the physiological indicators being tested, and can only draw a qualitative conclusion that eating genetically modified corn, rice and soybeans has no adverse effects on the overall health of mammals. In addition, due to the differences in experimental designs such as research objects and detection methods, as well as the limited sample size of experimental data, the conclusions of a single study have certain limitations [http: / / doi.org / 10.1016 / j.fct.2009.06.014; http: / / doi.org / 10.1016 / j.yrtph.2021.105091; http: / / doi.org / 10.1021 / acs.jafc.2c07142].

[0004] In summary, the existing main methods for evaluating the health of mammals that eat genetically modified crops have problems such as complex operation, high time cost, limitations caused by limited experimental design and sample size, and the inability to conduct systematic and quantitative evaluation. Before the large-scale commercial cultivation of genetically modified crops, there is a need for automated, comprehensive, systematic and quantitative evaluation of the overall health of mammals that eat them. Summary of the invention

[0005] In order to overcome the limitations of the existing use of a single toxicology experiment to evaluate the overall health level of mammals after eating genetically modified crops, and to meet the needs of evaluating the safety of eating genetically modified crops, the present invention solves the limitations of the existing single toxicology experiment evaluation caused by the selection of experimental model animals, group design and limited sample repetitions, as well as the problem of being unable to quantitatively evaluate multiple indicators, and provides a visualization system based on big data analysis and Python language for comprehensive quantitative evaluation of the health level of mammals eating genetically modified crops.

[0006] According to a first aspect of the present invention, a method for grading the health level of mammals that eat genetically modified crops is provided, comprising the following steps:

[0007] (1) The physiological indicators of mammals after eating genetically modified crops are screened according to the plant type p of the genetically modified crops and the animal type a of the mammals to obtain an array A for initialization a,p,n , including n = 1, 2, ... N indicators; for each physiological indicator A a,p,n , establish a Gaussian mixture distribution containing density function weights; for each physiological index A a,p,n Perform expected maximum optimization to obtain the physiological index A a,p,n Mixed probability density distribution of ;

[0008] (2) Define the safety assessment score level, define the mixed probability density 95% two-tailed test probability interval as the scoring health interval, and complete the scoring system initialization;

[0009] (3) Establishing an algorithm module for evaluating the health of mammals that eat genetically modified crops, including: creating a physiological indicator system initialization file based on the physiological indicator system, setting the hierarchical relationship of the physiological indicator system; establishing an initialization score storage output structure for the three-level score, obtaining an input array corresponding to the plant type p of the genetically modified crop and the animal type a of the mammal, matching it with the physiological indicator hierarchy, and cleaning it into an input data structure I a,p ; Score the overall health of mammals after consuming genetically modified crops, and further grade the health level based on the score.

[0010] Preferably, the method for establishing the Gaussian mixture distribution of the density function weights comprises:

[0011] Define the probability density function pdf of the control group K as,

[0012]

[0013] The pdf of the experimental group GM is,

[0014]

[0015]

[0016] in,

[0017] j is A a,p,n The jth group of records in the index, j = 1, 2, ... J a,p,n ;

[0018] μ is the mean of the group corresponding to record j;

[0019] σ is the variance of the group corresponding to record j;

[0020] θ is the normal distribution parameter corresponding to the record including μ and σ;

[0021] X is the data group type, which is the control group K or the experimental group GM;

[0022] ρ a,p,n,j To control the density function for inclusion and exclusion,

[0023]

[0024] The weight of the selection exists for,

[0025]

[0026] in,

[0027] The normalized weight of the group X density weight of record j satisfies 1;

[0028] Record j corresponding to group X sample density weight Calculated according to the experimental sample size, it is defined as:

[0029]

[0030] satisfy in: is the number of samples in group X corresponding to record j.

[0031] Preferably, in step (1), for each physiological indicator A a,p,n The expected maximum optimization is as follows: a,p,n,j exist Global combination, total In the combinations, find the best in the whole domain, that is, the minimum log-likelihood;

[0032]

[0033] st

[0034]

[0035] in,

[0036] The mean and variance of the mixed distribution of group X are,

[0037]

[0038]

[0039] The maximum expected likelihood function is defined as,

[0040]

[0041]

[0042] Preferably, in step (1), A a,p,n The mixed probability density distribution is specifically:

[0043] The optimal solution is numerically obtained using the expected likelihood function The optimal control combination is Optimal mixture distribution parameters Get physiological index A a,p,n A mixed probability distribution of .

[0044] Preferably, the scoring of the overall health of the mammal after eating the genetically modified crop comprises:

[0045] (1) to I a,p For each variable n, calculate its cumulative probability density function cdf:

[0046] PDF Its cdf is,

[0047]

[0048] (2) For variables composed of sub-levels and directly measurable Perform polynomial regression fitting and determine weights;

[0049] (3) For the i-th variable and all its child variables m=1…M, take random numbers of equal length respectively, according to:

[0050]

[0051] in,

[0052] α ml Fit constant coefficients to the corresponding variables;

[0053] const. is the fitting constant;

[0054] ε is a very small number;

[0055] This gives the variable The weight matrix

[0056] (4) For the dimension of virtual variables Equal weighted mixed m=1…M, obtain the distribution of dummy variables; equal weighted mixing meets the basis,

[0057]

[0058] For each set of superordinate dimensions and total scores, a score S is given according to the mixed distribution or fitting weighting; the score is transformed:

[0059]

[0060] (5) The health level is graded according to the score. When O is between 80 and 100, it is defined as good health; when O is between 60 and 79, it is defined as good health; when O is between 6 and 59, it is defined as sub-health; when O is between 0 and 5, it is defined as high-risk unhealthy state.

[0061] According to another aspect of the present invention, there is provided a system for evaluating the health status of mammals that eat transgenic crops, comprising:

[0062] Module for optimizing physiological indicators of mammals after eating genetically modified crops: used to screen the physiological indicators of mammals after eating genetically modified crops according to the plant type p of the genetically modified crops and the animal type a of the mammals, and obtain an array A for initialization a,p,n , including n = 1, 2, ... N indicators; for each physiological indicator A a,p,n , establish a Gaussian mixture distribution containing density function weights; for each physiological index A a,p,n Perform expected maximum optimization to obtain the physiological index A a,p,n Mixed probability density distribution of ;

[0063] Scoring system initialization module: used to define the safety assessment score level, define the mixed probability density 95% two-tailed test probability interval as the scoring health interval, and complete the scoring system initialization;

[0064] The algorithm establishment module for evaluating the health of mammals eating genetically modified crops is used to create a physiological indicator system initialization file based on the physiological indicator system, set the hierarchical relationship of the physiological indicator system; establish an initialization score storage output structure for the three-level score, obtain the input array corresponding to the plant type p of the genetically modified crop and the animal type a of the mammal, match it with the physiological indicator hierarchy, and clean it into the input data structure I a,p ; Score the overall health of mammals after consuming genetically modified crops, and further grade the health level based on the score.

[0065] Preferably, for each physiological indicator A a,p,n The expected maximum optimization is as follows: a,p,n,j exist Global combination, total In the combinations, find the best in the whole domain, that is, the minimum log-likelihood;

[0066]

[0067] st

[0068]

[0069] in,

[0070] The mean and variance of the mixed distribution of group X are,

[0071]

[0072] in,

[0073] j is A a,p,n The jth group of records in the index, j = 1, 2, ... J a,p,n ;

[0074] μ is the mean of the group corresponding to record j;

[0075] σ is the variance of the group corresponding to record j;

[0076] θ is the normal distribution parameter corresponding to the record including μ and σ;

[0077] X is the data group type, which is the control group K or the experimental group GM;

[0078] ρ a,p,n,j To control the density function for inclusion and exclusion,

[0079]

[0080] The weight of the selection exists for,

[0081]

[0082] in,

[0083] The normalized weight of the group X density weight of record j satisfies 1;

[0084] Record j corresponding to group X sample density weight Calculated according to the experimental sample size, it is defined as:

[0085]

[0086] satisfy in: is the number of samples in group X corresponding to record j.

[0087] In general, the above technical solution conceived by the present invention has the following technical advantages compared with the prior art:

[0088] (1) The present invention establishes a physiological indicator data set after mammals eat transgenic crop feed, an integrated health dimension and physiological indicator system, a health level grading and evaluation method for a limited pair array, and a human-computer interactive visualization system. The method includes: screening and integrating 26 physiological indicator data in 6 dimensions; creating a data preprocessing module for a limited pair array, including: obtaining the optimal mixed probability distribution of the entire domain of physiological indicators and determining the health interval; establishing an algorithm for evaluating the health of mammals, including: creating a three-level score data output structure of 26 physiological indicators in 6 dimensions, creating an input data structure, obtaining a single physiological index score and its weight in the corresponding dimension, obtaining the equal weighting of the 6 dimensions, and obtaining a health grading; constructing a human-computer interactive visualization module to obtain a visualized grading evaluation result. The present invention solves the limitations of selecting different model animals, group differences, imperfect indicators, small sample size and limited repetitions in a single toxicology experiment, and provides a grading algorithm and evaluation and human-computer interactive visualization system for diversified health indicators that integrates limited paired original experimental data obtained from a single experiment.

[0089] (2) The present invention provides a comprehensive quantitative grading algorithm based on different small sample original actual health indicators, solving the core problem of comprehensive quantitative evaluation of diversified health indicators at this stage. The present invention provides a human-computer interactive visualization system for fully automatic evaluation and scoring of the overall health of mammals consuming genetically modified crops based on big data analysis and Python language, solving the problems of complex operation, high time cost, and limitations caused by limited experimental design and sample size in existing toxicology experiments, as well as the inability to systematically and quantitatively evaluate, and meeting the needs for quantitative evaluation of the overall health of mammals consuming genetically modified crops.

[0090] (3) The comprehensive quantitative grading algorithm based on the original realized health indicators of different small samples and the human-computer interactive visualization system provided by the present invention can quickly, comprehensively and quantitatively evaluate the health status of mammals after eating genetically modified foods, and has important application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 This is a data framework diagram / physiological indicator framework diagram of a physiological indicator dataset of mammals after eating genetically modified crops provided by an embodiment of the present invention.

[0092] Figure 2 A flowchart of the use of a visualization platform for quantitatively grading and evaluating the overall health of mammals that eat genetically modified crops provided by an embodiment of the present invention.

[0093] Figure 3 A schematic diagram showing a parameter input interface provided in an embodiment of the present invention.

[0094] Figure 4 A schematic diagram showing a file import interface provided in an embodiment of the present invention.

[0095] Figure 5 A schematic diagram showing the main interface of a visualization tool provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0096] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0097] The present invention establishes a data set, a physiological index system, a method for quantitatively estimating the health level, and a human-computer interactive visualization system for evaluating the health level of mammals fed with transgenic crops. The method includes: systematically collecting and screening 26 physiological index data in six dimensions of rats, mice, crab-eating macaques and other mammals that have been disclosed so far after eating transgenic corn, rice, soybeans and other crops; constructing a data import structure for grading and evaluation to perform calculations, and obtaining quantitative grading results and evaluation results. The present invention solves the limitations caused by the selection of experimental model animals, group design and limited sample repetitions in small sample and single toxicology experimental evaluations, as well as the core problem of comprehensive quantitative evaluation of diversified health indicators, and provides a comprehensive quantitative grading algorithm based on different small sample original actual health indicators and a human-computer interactive visualization system. The method provided by the present invention can evaluate the health status of mammals after eating transgenic food, and has important application value.

[0098] The present invention establishes a grading method and evaluation system for the health level of mammals that eat genetically modified crops, including:

[0099] 1. Establish an indicator system and data set for physiological index data of mammals after consuming genetically modified crops, where the indicator system includes: 26 physiological indexes classified into 6 dimensions; the data set includes: extracting raw data from a single toxicology study based on the physiological indicator system, and unifying the raw data units based on international units to obtain a physiological index data set; Figure 1 This is a data framework diagram / physiological indicator framework diagram of the dataset of physiological indicators of mammals after eating genetically modified crops. The physiological indicator framework diagram in the dataset contains 6 dimensions and 26 physiological indicators, among which the body mass index dimension is indicated by weight gain and relative organ mass; the blood cell concentration dimension is indicated by red blood cell (RBC) concentration, white blood cell (WBC) concentration, neutrophil (NEU) concentration, lymphocyte (LYM) concentration, monocyte (MON) concentration, and platelet (PLT) concentration; the cardiovascular system dimension is indicated by serum total cholesterol (CHOL) concentration, and serum lactate dehydrogenase (LDH) activity; the liver function dimension is indicated by serum alanine aminotransferase (ALT) activity, serum aspartate aminotransferase (AST) activity, and serum total bilirubin (TBiL) concentration; the renal function dimension is indicated by serum urea nitrogen (BUN) concentration, serum creatinine (CRE) concentration, and serum glucose (GLU) concentration; the electrolyte function dimension is indicated by serum sodium ion (Na + ) concentration, serum potassium ion (K + ) concentration, serum calcium ion (Ca 2+ ) concentration, serum chloride ion (Cl -) Concentration, serum phosphorus (P) element concentration indicator. Among the dimensions, liver function will affect blood cell concentration, body mass index, electrolyte function and blood cell concentration will jointly affect cardiovascular function.

[0100] 2. Then, a data preprocessing module for limited paired occurrence arrays is established (an information processing method in which all steps are implemented by a computer), which includes data cleaning of the data set, including: removing duplicate records, removing records with missing key fields, filling in records with non-key fields, unifying units, and filling in corrected record magnitudes; filtering data according to plant type (p) and animal type (a) to obtain all arrays A for initialization. a,p,n , including n = 1, 2, ... N indicators; for each physiological indicator A a,p,n , establish a Gaussian mixture distribution containing density function weights; for each physiological index A a,p,n Perform expected maximum optimization: for ρ a,p,n,j exist Global combination, total In the middle combination, the global optimum is obtained, that is, the log-likelihood is minimized;

[0101]

[0102] (This part indicates that the objective function we want to minimize is the sum of two log-likelihood functions. Here, L represents the likelihood function, and are the estimated values ​​of parameters μ and σ respectively. Similarly, is another set of estimates)

[0103] ST is the abbreviation of "subject to", which means "subject to" or "satisfy the condition". "ST" means that in the process of minimizing the sum of log-likelihood functions, the parameter ρ a,p,n,j Must belong to a collection This means that when finding the optimal solution, the values ​​of these parameters must meet certain conditions or restrictions.

[0104]

[0105] (For ρ a,p,n,j exist Global combination, total In the combination, the global optimality is obtained, that is, the log-likelihood is minimized. a,p,n,j is a variable or a set of variables, and represents the set of allowed values ​​for these variables)

[0106] in,

[0107] The mean and variance of the mixed distribution of group X are,

[0108]

[0109] in,

[0110] j is A a,p,n The jth group of records in the index, j = 1, 2, ... J a,p,n ;

[0111] μ is the mean of the group corresponding to record j;

[0112] σ is the variance of the group corresponding to record j;

[0113] θ is the normal distribution parameter corresponding to the record including μ and σ;

[0114] X is the data group type, which can be K or GM;

[0115] ρ a,p,n,j To control the density function for inclusion and exclusion,

[0116]

[0117] The weight of the selection exists for,

[0118]

[0119] in,

[0120] The normalized weight of the group X density weight of record j satisfies

[0121] Combined with experimental data, record the sample density weight of group X corresponding to j Calculated according to the experimental sample size, it is defined as,

[0122]

[0123] satisfy in,

[0124] is the number of samples in group X corresponding to record j.

[0125] Define the security assessment score level and complete the scoring system initialization.

[0126] An algorithm module for evaluating the health of mammals eating genetically modified crops (an information processing method in which all steps are implemented by a computer) is established, including: creating a physiological indicator system initialization file based on the physiological indicator system, setting the hierarchical relationship of the physiological indicator system; establishing an initialization score storage output structure O_(a,p) in the format of a python dictionary, which is passed during the calculation process; cleaning the input data to obtain the input arrays corresponding to the plant type (p) and the animal type (a), matching them with the physiological indicator hierarchy, and cleaning them into the input data structure I_(a,p); a,p For each variable n in , the cumulative probability density function (cdf) is calculated; the overall health of mammals after eating genetically modified crops is scored, and the health levels are further graded according to the scores.

[0127] The present invention establishes a grading method and evaluation system for the health level of mammals that eat genetically modified crops (an information processing method in which all steps are implemented by a computer), wherein the Gaussian mixture distribution of the density function weights comprises:

[0128] Define the control group K probability density function (pdf) as,

[0129]

[0130] The pdf of the experimental GM is,

[0131]

[0132] And further,

[0133]

[0134] The present invention establishes a grading method and evaluation system for the health level of mammals that eat genetically modified crops. The hierarchical relationship of the physiological index system is shown in Table 1:

[0135] Table 1 Hierarchical relationship table of mammal indicator system

[0136]

[0137]

[0138] The present invention establishes a grading method and evaluation system for the health level of mammals eating genetically modified crops (an information processing method in which all steps are implemented by a computer), wherein the initialization score storage output structure O_(a, p) comprises:

[0139] (1) Outermost dictionary key: all variable names. Each outer dictionary key stores a score structure output dictionary.

[0140] (2) Score structure output dictionary structure:

[0141] Parent key: the parent indicator of the outer indicator, the key of the total score "Total" is specially defined as "TOP";

[0142] Child key: child indicator of the outer indicator, "No" is defined when there is no child;

[0143] weights key: weight of the outer layer indicator corresponding to the child level;

[0144] scores key: the non-normalized score of the corresponding sub-level of the outer indicator, with a single score range of 0-100 (%);

[0145] Animal key: The animal category matched by this array. This key is not included when not filtered.

[0146] plant key: The plant category matched by this array. This key is not included when not filtered.

[0147] The present invention establishes a grading method and evaluation system for the health level of mammals that eat genetically modified crops. The input data structure I_(a, p) includes:

[0148] (1) Outermost variable dictionary key: variable name. Each outer dictionary key stores an input data variable.

[0149] (2) Variable data structure:

[0150] col_names key: stores the column name information of data records, which must include ID and STUDY columns, and together define the uniqueness of the data stored under the records key;

[0151] Records key: stores all input data;

[0152] parent key; the variable's parent indicator;

[0153] Child key: child index of the variable;

[0154] Animal key: the animal category entered in this array;

[0155] plant key: The plant category entered in this array.

[0156] The present invention establishes a grading method and evaluation system for the health level of mammals that eat genetically modified crops (an information processing method in which all steps are implemented by a computer), wherein the scoring of the overall health of mammals after eating genetically modified crops includes:

[0157] (1) to I a,p For each variable n in , calculate its cumulative probability density function (cdf):

[0158] PDF Its cdf is,

[0159]

[0160] Its range is 0–100%.

[0161] (2) For variables composed of sub-levels and directly measurable (such as WBC) to perform polynomial regression fitting and determine the weights;

[0162] (3) For the i-th variable and all its child variables m=1…M, take random numbers of equal length respectively, according to,

[0163]

[0164] in,

[0165] α ml Fit constant coefficients to the corresponding variables;

[0166] const. is the fitting constant;

[0167] ε is a very small number;

[0168] This gives the variable The weight matrix

[0169] (4) For the dimension of virtual variables Equal weighted mixed m=1…M, obtain the distribution of dummy variables; equal weighted mixing meets the basis,

[0170]

[0171] For each set of superordinate dimensions and total scores, a score S is given according to the mixed distribution or fitting weighting; the score is transformed:

[0172]

[0173] (5) The health level is graded according to the score. When O is between 80 and 100, it is defined as good health; when O is between 60 and 79, it is defined as good health; when O is between 6 and 59, it is defined as sub-health; when O is between 0 and 5, it is defined as high-risk unhealthy state.

[0174] Example 1: Determination of physiological indicators of mammals after consuming transgenic crops, establishment of data sets and construction of systems, including the following steps:

[0175] Step 1: Determine the literature screening criteria based on the PRISMA specification and select several toxicology experimental cases that meet the criteria from the literature database;

[0176] Step 2: Extract the types of genetically modified crops, dosages, mammal types, genders, and animal health physiological indicators from several toxicology experimental cases;

[0177] Step 3, classify the 25 physiological indicators detected in existing toxicology experiments into a physiological indicator system with six dimensions;

[0178] Step 4, extracting raw data after variance analysis in several toxicology experiments according to the physiological index system constructed in step 3;

[0179] Step 5: Unify the raw data extracted in step 4 based on international units.

[0180] Specifically, the literature collection criteria determined in step 1 include:

[0181] Duplicate literature was removed from different literature databases; the articles must include animal experiments involving GM crops, or whole grains or ground grain powder fed with GM crops; review articles, conference reports and non-research papers were excluded; the study must include a control group (fed with non-GM crops) and a treatment group (fed with GM crops), and provide specific values ​​for physiological and biochemical indicators, body weight, organ weight, etc.; each treatment should have at least 5 replicates;

[0182] Specifically, the step 1 is to screen from the literature database certain toxicology experimental cases, including:

[0183] Relevant studies on the safety of ingestion of three genetically modified crops, corn, rice and soybean, were collected. The included databases include PubMed, Web of Science and Scopus, covering articles published between January 2000 and July 2024. The search terms used were "gene editing", "transgenic", "health", "safety" and "feeding". Based on the PRISMA specifications, the research cases were screened and several toxicology experimental cases were obtained;

[0184] Specifically, the common variables of mammalian physiological indicators and transgenic crop types, dosages and mammalian types described in step 2 include:

[0185] Physiological indices included weight gain, relative organ mass, red blood cell concentration, white blood cell concentration, lymphocyte concentration, neutrophil concentration, monocyte concentration, platelet concentration, total cholesterol concentration, total bilirubin concentration, serum potassium ion, sodium ion, calcium ion, chloride ion and phosphorus ion concentration, and the activities of lactate dehydrogenase, alanine aminotransferase and aspartate aminotransferase;

[0186] Types of genetically modified crops include genetically modified corn, rice, and soybeans;

[0187] The dosage of genetically modified crops includes low dosage: containing 0-20% genetically modified crops; medium dosage: containing 20%-50% genetically modified crops; high dosage: 50%-100%;

[0188] Mammal types included rats, mice, and cynomolgus macaques;

[0189] Specifically, the six dimensions of the physiological indicator system in step 3 include:

[0190] Body mass index, blood cell concentration, cardiovascular system function, liver function, kidney function and electrolyte function;

[0191] Specifically, the data extracted in step 4 includes:

[0192] The mean and variance of physiological parameters of mammals that eat genetically modified crops and the corresponding mean and variance of physiological parameters of mammals that eat non-genetically modified crops;

[0193] Specifically, the units in step 5 are unified as follows:

[0194] The red blood cell unit is unified as "10 12 The units of white blood cells and platelets are unified as "10 9The units of lymphocytes, neutrophils and monocytes are unified as "%". The units of total cholesterol concentration are unified as "mmol / L", the units of serum lactate dehydrogenase activity are unified as "U / L". The units of serum alanine aminotransferase activity and serum aspartate aminotransferase activity are unified as "U / L", the units of total bilirubin concentration are unified as "μmol / L". The units of serum glucose concentration are unified as "mmol / L", serum urea nitrogen concentration are unified as "mmol / L", and serum creatinine concentration are unified as "μmol / L". The units of serum sodium, potassium, calcium, chloride and phosphorus concentrations are unified as "mmol / L".

[0195] Embodiment 2: A grading and evaluation structure is constructed, including: a human-computer interaction interface, a data preprocessing module, an algorithm module and a result visualization module, including the following steps:

[0196] Step 1: Construct a data preprocessing module for health assessment of mammals consuming genetically modified crops;

[0197] Step 2: Construct an algorithm module for evaluating the health of mammals that consume genetically modified crops;

[0198] Step 3: Construct a visualization system module for health assessment of mammals consuming genetically modified crops.

[0199] Specifically, the health assessment data preprocessing module for mammals eating genetically modified crops in step 1 includes:

[0200] Clean the data set, including: removing duplicate records, removing records with missing key fields, filling in records with non-key fields, unifying units, and filling in and correcting record magnitudes;

[0201] Filter the data according to plant type (p) and animal type (a) to obtain all arrays A_(a,p) for initialization, including n=1,2,…N indicators;

[0202] For each physiological index A_(a,p,n), a Gaussian mixture distribution including density function weight is established; for each physiological index A_(a,p,n), the expected maximum optimization is performed;

[0203] Define the safety assessment score level: the mixed probability density 95% two-tailed test probability interval is the scoring health interval, and complete the scoring system initialization.

[0204] Specifically, the health assessment algorithm module for mammals consuming genetically modified crops in step 2 comprises:

[0205] Create physiological indicator system initialization file based on theoretical indicator system and set the hierarchical relationship of physiological indicator system;

[0206] Create an initial score storage output structure O_(a,p) in the format of a python dictionary and pass it during the calculation process;

[0207] The input data is cleaned to obtain the input arrays corresponding to the plant type (p) and animal type (a), matched with the physiological indicator level, and cleaned into the input data structure I_(a,p); for each variable n in I_(a,p), its cumulative probability density function (cdf) is calculated;

[0208] For variables that are directly measurable and consist of sub-levels (such as WBC) to perform polynomial regression fitting and determine the weights; Equal weighted mixed m=1…M, and obtain the distribution of dummy variables; for each set of superordinate dimensions and total scores, a mammal health score is given according to the mixed distribution or fitting weighting, and written into O_(a,p) respectively.

[0209] Specifically, the health assessment visualization system module of mammals consuming genetically modified crops in step 3 includes:

[0210] Parameter input interface, file import interface, result display interface and help interface;

[0211] If you import data through the file import interface, the data is stored in the form of a dictionary. After clicking the "Confirm" button, the data is transferred from the file import interface to the parameter input interface;

[0212] If you import data through the parameter input interface, the data is stored in the form of a dictionary, and the default values ​​are used to complete the parameters that are not entered;

[0213] Click the "Confirm" button on the parameter input interface, and the data will be transferred from the parameter input interface to the result display interface and displayed in the parameter preview box;

[0214] Click the "Confirm" button on the result display interface, and the data will be transferred from the front-end result display interface to the back-end calculation module to execute the algorithm;

[0215] After the backend completes the calculation, it passes the calculation result data back to the front-end result display interface in the form of a dictionary. After the front-end processes the calculation result, it displays it in the form of a pie chart, total health score and individual item chart.

[0216] The purpose of the present invention is to provide a method and system for fully automatic evaluation and quantitative grading of the overall health of mammals after eating genetically modified crops, which solves the problems of complex operation, high time cost and limitations caused by limited experimental design and sample size in existing toxicology experiments, as well as the inability to systematically and quantitatively evaluate.

[0217] Example 3

[0218] according to Figure 2 The present invention provides a visualization platform capable of quantitatively and gradedly evaluating the overall health of mammals that eat genetically modified crops, comprising:

[0219] Step 1: Upload physiological index parameters to the visualization software.

[0220] In this embodiment, the parameter input interface is designed as shown in the following figure ( Figure 3 ); The upper part of this interface is "Parameter Input", where you can manually add input data by clicking the "+" button below the parameter input box; the lower part is "Custom Standard Input", which supports custom dimensions and indicators.

[0221] By clicking the "File Import" button, you can jump to the file import interface.

[0222] Click the "Confirm" button to confirm that the parameter input is complete and return to the result display interface.

[0223] Step 2: Import the physiological index data file into the software. The design of the file import interface is shown in the figure below ( Figure 4 ). In the left frame, "Input Source", "File Settings" and "Data Preview" show the current status of these three items. The top right part supports importing data files by selecting files or manually entering file paths. After the import is completed, a check will be placed after "Input Source" in the left frame to indicate that the step has been completed; in the bottom right part, the input file type, input data delimiter and whether there is a header are set through the drop-down list and input box respectively. After the input and selection are completed, a check will be placed after "File Settings" in the left frame. After both "Input Source" and "File Settings" are completed, the tool will parse the input data and display it in the "Import Preview" box in the middle of the right side.

[0224] Step 3: Obtain the results of the health assessment of mammals after eating genetically modified crops. The result display interface is shown in the figure below ( Figure 5 ). The drop-down list on the left supports selecting the type of animal and plant to which the data belongs. The middle shows a pie chart drawn from the data calculation results, the upper right shows the health score obtained based on the data calculation results, and the lower right shows a single item chart of the specific indicator.

[0225] You can jump to the help interface by clicking the question mark in the upper left corner, or jump to the parameter input interface by clicking the "Parameter Input" button on the right.

[0226] By clicking the "Confirm" button, the algorithm is executed, and the "Total Health Score" and evaluation pie chart are displayed after the calculation is completed. After selecting a single indicator from the drop-down list, the single chart of the specific indicator can be displayed.

[0227] The calculation results can be stored by clicking the "Save As" button in the lower right corner.

[0228] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for grading the health status of mammals that eat genetically modified crops, characterized in that: The following steps are involved: (1) The physiological indicators of mammals after eating genetically modified crops are screened according to the plant type p of the genetically modified crops and the animal type a of the mammals to obtain an array A for initialization a，p，n , including n = 1, 2, ... N indicators; for each physiological indicator A a，p，n , establish a Gaussian mixture distribution containing density function weights; for each physiological index A a，p，n Perform expected maximum optimization to obtain the physiological index A a，p，n Mixed probability density distribution of ; (2) Define the safety assessment score level, define the mixed probability density 95% two-tailed test probability interval as the scoring health interval, and complete the scoring system initialization; (3) Establishing an algorithm module for evaluating the health of mammals that consume genetically modified crops, including: creating a physiological indicator system initialization file based on the physiological indicator system and setting the hierarchical relationship of the physiological indicator system; Establish the initialization score storage output structure of the three-level score, obtain the input array corresponding to the plant type p of the genetically modified crops and the animal type a of the mammals, match them with the physiological indicator level, and clean them into the input data structure I a，p ; Score the overall health of mammals after consuming genetically modified crops, and further grade the health level based on the score.

2. The method for grading the health level of mammals eating genetically modified crops according to claim 1, characterized in that: The method for establishing the Gaussian mixture distribution of the density function weights comprises: Define the probability density function pdf of the control group K as, The pdf of the experimental group GM is, in, j is A a，p，n The jth group of records in the index, j = 1, 2, ... J a，p，n ; μ is the mean of the group corresponding to record j; σ is the variance of the group corresponding to record j; θ is the normal distribution parameter corresponding to the record including μ and σ; X is the data group type, which is the control group K or the experimental group GM; ρ a，p，n，j To control the density function for inclusion and exclusion, The weight of the selection exists for, in, The normalized weight of the group X density weight of record j satisfies Record j corresponding to group X sample density weight Calculated according to the experimental sample size, it is defined as: satisfy in: is the number of samples in group X corresponding to record j.

3. The method for grading the health level of mammals eating genetically modified crops according to claim 2, characterized in that: In step (1), for each physiological index A a，p，n The expected maximum optimization is as follows: a，p，n，j exist Global combination, total In the combinations, find the best in the whole domain, that is, the minimum log-likelihood; in, The mean and variance of the mixed distribution of group X are, The maximum expected likelihood function is defined as, 4. The method for grading the health level of mammals eating genetically modified crops according to claim 3, characterized in that: In step (1), A a，p，n The mixed probability density distribution is specifically: The optimal solution is numerically obtained using the expected likelihood function The optimal control combination is Optimal mixture distribution parameters Get physiological index A a，p，n A mixed probability distribution of .

5. The method for grading the health level of mammals eating genetically modified crops according to claim 1, characterized in that: The overall health of mammals after eating genetically modified crops is scored, including: (1) to I a，p For each variable n, calculate its cumulative probability density function cdf: PDF Its cdf is, (2) For variables composed of sub-levels and directly measurable Perform polynomial regression fitting and determine weights; (3) For the i-th variable and all its child variables m=1…M, take random numbers of equal length respectively, according to: in, α ml Fit constant coefficients to the corresponding variables; const. is the fitting constant; ε is a very small number; This gives the variable The weight matrix (4) For the dimension of virtual variables Equal weighted mixed m=1…M, obtain the distribution of dummy variables; equal weighted mixing meets the basis, For each set of superordinate dimensions and total scores, a score S is given according to the mixed distribution or fitting weighting; the score is transformed: (5) The health level is graded according to the score. When O is between 80 and 100, it is defined as good health; when O is between 60 and 79, it is defined as good health; when O is between 6 and 59, it is defined as sub-health; when O is between 0 and 5, it is defined as high-risk unhealthy state.

6. A system for evaluating the health status of mammals that eat genetically modified crops, characterized in that: include: Module for optimizing physiological indicators of mammals after eating genetically modified crops: used to screen the physiological indicators of mammals after eating genetically modified crops according to the plant type p of the genetically modified crops and the animal type a of the mammals, and obtain an array A for initialization a，p，n , including n = 1, 2, ... N indicators; for each physiological indicator A a，p，n , establish a Gaussian mixture distribution containing density function weights; for each physiological index A a，p，n Perform expected maximum optimization to obtain the physiological index A a，p，n Mixed probability density distribution of ; Scoring system initialization module: used to define the safety assessment score level, define the mixed probability density 95% two-tailed test probability interval as the scoring health interval, and complete the scoring system initialization; The algorithm establishment module for evaluating the health of mammals eating genetically modified crops is used to create a physiological indicator system initialization file based on the physiological indicator system, set the hierarchical relationship of the physiological indicator system; establish an initialization score storage output structure for the three-level score, obtain the input array corresponding to the plant type p of the genetically modified crop and the animal type a of the mammal, match it with the physiological indicator hierarchy, and clean it into the input data structure I a，p ; Score the overall health of mammals after consuming genetically modified crops, and further grade the health level based on the score.

7. The system for evaluating the health status of mammals eating transgenic crops according to claim 6, characterized in that: For each physiological index A a，p，n The expected maximum optimization is as follows: a，p，n，j exist Global combination, total In the combinations, find the best in the whole domain, that is, the minimum log-likelihood; in, The mean and variance of the mixed distribution of group X are, in, j is A a，p，n The jth group of records in the index, j = 1, 2, ... J a，p，n ; μ is the mean of the group corresponding to record j; σ is the variance of the group corresponding to record j; θ is the normal distribution parameter corresponding to the record including μ and σ; X is the data group type, which is the control group K or the experimental group GM; ρ a，p，n，j To control the density function for inclusion and exclusion, The weight of the selection exists for, in, The normalized weight of the group X density weight of record j satisfies Record j corresponding to group X sample density weight Calculated according to the experimental sample size, it is defined as: satisfy in: is the number of samples in group X corresponding to record j.

Images