KR20200041533A - Real-time breeding value and genetic parameter evaluation system - Google Patents

Abstract

The present invention relates to a method and apparatus for evaluating genetic ability (breeder value or genetic parameter) in real time in a relational database in which blood relations of an individual, phenotype information, and the like are stored. More particularly, the present invention, finds a gender error which can occur in family line information, an error of duplicate object information, and a cycle error in which the offspring are marked as parents of the ancestor generation, and collectively corrects or removes them, so that a user can more easily evaluate genetic ability.

Description

Real-time breeding value and genetic parameter evaluation system

The present invention is an invention for a system for estimating breeding prices for efficient breeding of agricultural and livestock products.

In order to improve livestock, the estimation of breeding prices for phenotypes to be improved has been established as a basis for selection and selection, which proved to be more effective than past phenotype-based breeding methods. To this end, in the field of breeding, the breeding price can be estimated using a mixed model based on an animal model, and the standard error is calculated accordingly to be used as an index for breeding the phenotype of the interested group. I have done

Since the above breeders are estimated values from the groups used in the analysis and are relative values within the groups, the same individuals are subject to changes in the group genetic gene pool due to the influx or migration of new individuals to the group for which the mixed model was calculated. The value of the breeding price of is different. In other words, whenever the information of the group is updated, the breeding cost of each individual breeder and the genetic parameter of the group are measured again.

In the case of breeders, it is possible to increase the accuracy only by estimating the breeder's value by integrating as many kinship groups as possible (statistical aspects), and farmers and improvers who are not specialized institutions with practical analytical skills have limitations in handling data (technical) Personnel side) Evaluation of the individuals possessed through the analysis agency was carried out, and as a result, it was inconvenient to develop and improve a breeding strategy suitable for the actual site.

In order to solve the mixed model for breeder estimation, a Numeric Relationship Matrix (NRM) is required. To do this, household information (Pedigree or PED data format) consisting of the entity name, father name and mother name is required. It should be flawless without common household errors, such as duplicate object errors, gender errors, and circular errors. To this end, there are libraries of commercial software and free software, but it is not suitable for real-time analysis by processing error objects together with error checking. In order to solve the above problem, blood relation information and phenotype information of each object are managed in a relational database. In addition, after enabling periodic data update, a method was devised to extract current information of individuals meeting specific conditions at a specific time, and to perform a kinetic matrix correction and genetic ability evaluation in real time . It was made possible for the general user to conveniently perform the genetic ability evaluation.

In a first aspect of the present invention, in a method for providing breeding price and genetic parameter evaluation information for improving a variety of agricultural and livestock products, a database including phenotype information of agricultural and livestock products, individual additional information, system information, species information, and household information Preparing a step ; Selecting a line of interest and a phenotype of interest ; Calculating information necessary for calculating a breeding price and calculating a genetic parameter including household information and additional information about an individual of interest and phenotype of interest; Estimating breeding prices and genetic parameters by reflecting the phenotype information, household information, and individual additional information; And displaying the estimated breeding price and genetic parameter.

The second aspect of claim 1, wherein the information necessary for calculating the breeding price and calculating the genetic parameter includes estimating a breeding value and a genetic parameter, characterized in that it comprises a blood relation matrix. Here, the blood relation matrix is characterized in that the error in the household information is corrected.

The present invention aims to evaluate the genetic ability (breeder or genetic parameter) in real time in a relational database in which an individual's blood relations, phenotype information, and the like are stored. In particular, it aims to make it easier for the user to evaluate genetic abilities by finding and correcting or removing the gender errors that can occur in household information, errors in duplicate object information, and cycle errors where offspring are marked as the parents of the ancestors. do.

 Through the present invention, it is possible to subjectively breed such as selecting excellent individuals through genetic ability evaluation in general livestock breeding farms and institutions that do not possess analysis technology. In the field, the genetic composition of the changed group is immediately reflected due to the shipment of livestock, the creation of new descendants, and the influx of individuals from other groups, and the data of the individual and the group is managed immediately. Together, real-time analysis is possible, allowing for flexible breeding strategies in response to rapid changes. In addition, it can be applied to selection and mating planning to contribute to improving the overall group performance when considering the average improvement of various traits of interest.

1 is a schematic diagram of the overall process for evaluating genetic ability.
2 is an example of a user's screen for selecting a phylogeny and phenotype to enter an animal model.
3 is an example of an arbitrary family tree and the household calibration process.
4 is an example of a screen on which the calculated genetic parameters and breeding prices are displayed.
5 is an example of a basic statistics table and chart for the selected phylogeny and phenotype.

The present invention provides a method for providing breeding price and genetic parameter evaluation information for improving a variety of agricultural and livestock products , preparing a database including phenotype information of agricultural and livestock products, individual additional information, system information, species information, and household information; Selecting a line of interest and a phenotype of interest; calculating information necessary for calculating a breeder's price and calculating a genetic parameter including household information and individual additional information for the line of interest and the phenotype of interest; the phenotype information, household information, and individual additional information Estimating breeding price and genetic parameters by reflecting;

It provides a method for providing a breeding price and genetic parameter evaluation information, and an apparatus for performing the same, including the step of displaying the estimated breeding price and genetic parameter.

In addition, the breeder and genetic parameter information providing apparatus according to an embodiment of the present invention,

An evaluation item storage unit that stores and manages a database including phenotypic information of agricultural and livestock products, individual additional information, system information, species information, and household information in memory;

An evaluation item selection unit for selecting a line of interest and a phenotype of interest ;

An evaluation item calculating unit for calculating and storing and managing information necessary for breeding price calculation and genetic parameter calculation including household information and additional information on individuals of interest and phenotype of interest;

An evaluation item evaluation unit that estimates a breeding price and a genetic parameter by reflecting the phenotype information, household information, and individual additional information and stores it in the memory ;

An evaluation result providing unit displaying the estimated breeding price and genetic parameter;

It includes a control unit, memory, and user interface .

The control unit is in charge of overall control of each component of the breeder and genetic parameter evaluation information providing device. It may also be implemented to include all or part of the functionality of other components.

The user interface is a means for inputting and outputting various information or a database of a breeder and genetic parameter evaluation information providing device, and includes various means for a user to input or output information or data to the device, such as a keyboard, a button, a touch screen, and a communication port. can do.

The evaluation result providing unit performs a process of displaying the evaluation results, and may include a conventional electronic display device.

The controller prepares a database including phenotype information of agricultural and livestock products, object additional information, system information, species information, and household information, and selects or inputs a system of interest and a phenotype of interest, and receives household information about the system of interest and phenotype of interest, and and calculating the information required for breeders calculated and estimated genetic parameters containing the object side information, and reflects the phenotypic information, family information, and the object side information, and estimates the breeders and genetic parameters, displaying the estimated breeding values and genetic parameters Overall control can be performed so that can be performed.

In the memory, information or programs necessary for the operation of the breeder and genetic parameter evaluation information providing apparatus may be stored, and information necessary for calculating the breeding price and calculating the genetic parameters, including household information about the line of interest and phenotype of interest, and additional object information, The estimated breeding price and genetic parameters may be stored by reflecting the phenotype information, household information, and individual additional information.

The term “ genetic parameter ” is a statistical quantity, but a population that is estimated as this is a numerical value obtained by measurement in a sample when interpreting a genetic quantity trait. Refers to the genetic property of the group.

In addition, the term " relationship coefficient " is an index that indicates the similarity of the genetic composition between two individuals. It is used to indicate the degree of blood relationship between individuals in a breed or lineage of a livestock or an experimental animal. .

The term " breeding value " refers to the mean value of a particular quantitative trait in the next generation of a particular hybrid, or it is the value of the parent, that is, the value of the donor to pass the gene to the next generation.

In addition, it indicates the extent to which “ heritability ” parental traits are transmitted to offspring. It is often expressed as the ratio of variance due to genetic differences among individuals among the whole phenotype variance. In other words, the genetic power is calculated as the genetic dispersion / phenotype dispersion , and so-called quantitative traits such as height and weight are dominated by genes with multiple micro-effects (polygenes), and also express the effects of the environment mixed together. . Therefore, the quantitative trait change becomes a continuous distribution, and the distribution is called a phenotypic variance . In genotype dispersion, variance according to genotype differences among phenotypic dispersions is called genome dispersion (or genome dispersion) . The genetic variance is further divided into additive genetic variance, dominance variance, and epistatic variance.

"Environmental dispersion" refers to the part of phenotypic dispersion caused by environmental changes in which individuals in a group are placed.

"Phenotype" is a trait of interest that can affect the economic value of improving the corresponding phenotype of agricultural and livestock, and refers to a continuous economic trait. The phenotype of agricultural and livestock products can be either mountain lion, dorsal fat, or weight, and cows can be milk production, and sheep can be hair weight, mountain lion weight, or body weight. For poultry, for chickens or ducks, it can be either an antelope, weight or egg weight (egg weight), and for a horse or racehorse, it can be a jumping ability, an antelope, or a racing ability.

In the case of crops including corn, rice, wheat, barley, soybean, potato, and sweet potato, harvest index (quantity / total production), quantity breeding (number of fruits), or photosynthesis efficiency (leaf width, angle, etc.) You can.

"Additional information of an individual" means fixed effects that are put into an expression to accurately obtain only the variance (explanatory power) of breeders among the phenotype, that is, to correct the model. have.

Remedy, it can be the mother's age, age group (brothers born together), feed intake, season of birth, year of birth, sunshine, gender.

Analysis module for preparing database and statistical analysis

This method requires a database of livestock phenotype information, household information, and additional information about the animals, a pedigree correction module that checks and eliminates self-generated household information errors, and packages of statistical software R. It consists of a process of providing basic statistics, genetic parameters, and breeders for Breed) and line (Line or Stock), and this process is performed through the memory unit and the control unit.

Calculating the information needed to calculate breeders' prices and genetic parameters ;

In order to correct the phenotype, household information and model expression of the selected livestock system, additional information of the object to be used as a fixed effect is extracted from the database. This process is performed through the evaluation item storage.

division Pawnant Component Database Article Integration Information
Database Static phenotype required to solve animal model equations (fixed phenotype measurements in one measurement)
E.g. trial age, trial difficulties, etc.) and
Dynamic phenotypes (phenotypes whose measurements change over time, e.g. weight, shank length, etc.),
Additional information of the individual (example) gender, date of birth, search, age of slaughter, presence of horns, etc.)
Lineage and species information
Household information (object name, father's name, mother's name) Pedigree correction module Algorithm to check household errors and remove objects Algorithm to find and delete error information by receiving household information extracted from the database in PED data format R statistics software pedigree package Package that corrects household information in order to calculate the number of households and produce a relationship matrix by receiving household information from which errors have been removed
<Source> https://cran.r-project.org/web/packages/pedigree/index.html nadiv package Package that creates blood-related matrix using corrected household information
<Source> https://cran.r-project.org/web/packages/nadiv/index.html , http://github.com/matthewwolak/nadiv rrBLUP package Animal set by the analyst with blood-related matrix, phenotype data, and additional information
<Source> https://potatobreeding.cals.wisc.edu/software/
https://cran.r-project.org/web/packages/rrBLUP/index.html

Steps to correct errors in household information

Genetic capacity analysis using household information requires prior inspection and arrangement to create a Numeric Relationship Matrix (NRM). If the wrong lineage is used as it is, the matrix may not be created, and even if it is made, the error of several wrong individuals results in an incorrect genetic ability evaluation in all the individuals related to the blood relationship.

The lineage error correction module included in the real-time genetic ability analysis method of the present invention largely finds three types of error-related information and further deletes individual information related to errors collectively in accordance with the management and analysis efficiency of the livestock breed management database. Correction.

All household information is formatted in PED format (individual ID, SireID, DamID), extracted from the database, and then modified in bulk through the following procedure.

1. Duplicate object error

Duplicate information means that the same entity appears multiple times in the first column of the extracted household information. There are several kinds of sub-types for simple-looking duplicate errors, and the processing method for each subtype is as follows.

1-1 No missing parent

It is determined whether the avi information and the parent information of the object match the avi information and the mother information of the other object, and if both the avi information and the mother information match, the object information is deleted, and if only one of the avi information or the mother information matches Absent information or mother information that is not used can be replaced with missing values to correct errors in household information.

More specifically, when both IDs of Sire and Dam of the duplicate object are the same, the exact same object is repeatedly input, so only one data is left and deletion is performed.

<How to correct errors when duplicate objects have the same parent ID>

If both IDs of Sire and Dam of the duplicated object are different, it is difficult to distinguish which data is the actual data if the parental ID is different and the object appears more than once. By selecting only one of several pieces of information, it is possible to prevent the kinship relationship in the matrix from being fragmented, but since it cannot be determined by guessing, in this case, all family information of all the duplicated objects is erased to prevent incorrectly calculated kinetic matrix calculation.

<How to correct errors when the parent IDs of duplicate objects are all different>

If the ID of one of the two parents of a duplicate object is different, and if the ID of their Sire (or Dam) is fixed in two or more duplicate objects, the information of the registrant is judged to be reliable, so the Dam (or Sire) is not deleted. By replacing only the ID of the missing value, it is possible to prevent the blood relationship from being broken.

<If only one ID of the parent of the duplicate objects is different>

1-2 If one of the parents has missing values

 When converging each of the parent information that is not the missing value of the object, if each converges, the information of the object is preserved by combining the non-missing parent information with one object data.

<Error correction method when it is possible to integrate non-missing parent information>

When summing parent information that is not a missing value, when each Sire and Dam do not converge into one, the information of Sire and Dam that are not united as one is expressed as a missing value.

<How to correct errors when it is impossible to integrate parental information that is not missing>

2. Circulation error

 Cyclic errors are those in which the offspring reappear as ancestor parents. In this case, it is one of the serious errors that household information can have, and it is an error that can be confirmed only by examining all the relationships between the ancestors and ancestors who have unspecified number of generations from the part that only parents and offspring should consider. In this case, the applicant was able to find a list of individuals judged as errors by applying a graph traversal and search method algorithm among data structures in computational and computer science.

When the information of a specific object matches either the father or the mother information, the corresponding entity information is deleted, and the object with the same parent and father information of the individual is changed to the gender identified in the father and father information of other individuals. and,

When the information of the individual identified as the father or mother of the specific object is the same as the information of the individual identified as a descendant of the specific object, the information on the specific object may be deleted to correct an error in the household information.

More specifically, error handling is required even when the part (or parent) of an object is the same as the object itself. In this case, the object is deleted.

<How to correct an individual's own circular error>

When the parent and the parent have the same ID, it is a circular error and also includes a gender error. At this stage, even if the object with the parent information is not removed, it can be removed from the gender error, so it is collected with one Sire or Dam and the other is filled with missing. At this time, when collecting by one gender, it is possible to find information used as the parent of an existing entity and only if the gender is fixed, and if it is not fixed, the information on the household information can be corrected by replacing the two information with missing values.

 <How to fix the same error among parents among circular errors>

The cyclical error in which the offspring reappears as the ancestor's parent makes the hierarchical structure collapse, making it impossible to distinguish between the ancestors and offspring. In the end, since it is not a part to guarantee where a human error has entered, even if the loss of blood information is large, it is possible to solve the circular error by removing and removing all the individuals participating in the circle error from the household information.

< Example of a circular error in which the offspring reappear as parents of the ancestors>

3. Gender error

Since the data extracted in the PED format has the information of Sire (part) and Dam (parent), all the individuals who have roles as parents have gender information. Sometimes, when a specific object is used as a wealth and then used as a parent of another object, it can be said that the household information has a gender error.

In the case where the parent information of a specific object and the mother information of another object match, the mother information and the mother information of the objects can be replaced with missing values to correct errors in the household information.

More specifically, if the proportion of individuals with two genders is less than 95%, the information used as an abina or mother to prevent confusion in the entire household data because it seems that the gender was not clearly recorded in the above case. You can correct errors in household information by replacing all of them with missing values.

<How to correct errors when it is difficult to see the gender of a specific individual as one gender>

However, if a large proportion of the individuals with two genders appear more than 95%, in the species that can have many descendants, as many records are created as parents, it is efficient to save the individuals as much as possible. To this end, the number and ratio of all genders are calculated for a specific individual, and when the ratio biased to one side is more than 95%, the remaining 5% of sex is judged as an error. However, if the gender record of 5% is changed, the information of the parents with the opposite sex will cause a problem, so the parental information with the error of 5% will be missing.

<How to correct errors when it is possible to correct wrong gender information of less than 5%>

Steps of creating a blood relation matrix

The information necessary for calculating the breeding price and calculating the genetic parameters may include a blood relation matrix . In order to produce a kinship matrix, the PED data format (object name, where the erroneous information is removed) is removed as a second pre-processing step to find the error in the extracted household information, remove the object, correct the pedigree, and create the kinship matrix. After adding the founder information to the household information of the father's name, the mother's name), calculating the number of households, sorting the households in ascending order of households, and then generating new corrected PED data. The 'pedigree' package of R statistical software may be used for the correction, but is not limited thereto. A kinship matrix can be generated from the corrected household data, and the 'nadiv' package of R statistics software can be used as an example.

<Arbitrary pedigree chart and example of household correction>

Steps to provide basic statistics

The basic statistics information of the phenotype of the species and lineage selected for model expression and the additional information extracted from the database can be summarized in tables and charts.

Basic statistics can include individuals in the household, number of fathers and mothers, gender summary, number of households, number of phenotypes, mean, and standard deviation, as in the example below.

Application of animal models

 Generating the relationship matrix using household information and estimating the genetic parameters and breeding value as a mixed model using the relationship. To this end, an animal model considering the genetic effect of an individual may be applied using lineage information in the corresponding system.

As an example, the animal model can be applied as follows.

y = μ + sex + α + ε

       y: Measurement of the trait of interest (dependent variable)

      μ: Overall mean of the blood quality of interest

      sex: sex fixed effect of each individual

      α: Random additive genetic effect of each individual (ex, breeding price)

      ε: Environmental effect

Steps to correct errors in household information

Genetic capacity analysis using household information requires prior inspection and arrangement to create a Numeric Relationship Matrix (NRM). If the wrong lineage is used as it is, the matrix may not be created, and even if it is made, the error of several wrong individuals results in an incorrect genetic ability evaluation in all the individuals related to the blood relationship.

The lineage error correction module included in the real-time genetic ability analysis method of the present invention largely finds three types of error-related information and further deletes individual information related to errors collectively in accordance with the management and analysis efficiency of the livestock breed management database. Correction.

All household information is formatted in PED format (individual ID, SireID, DamID), extracted from the database, and then modified in bulk through the following procedure.

1. Duplicate object error

Duplicate information means that the same entity appears multiple times in the first column of the extracted household information. There are several kinds of sub-types for simple-looking duplicate errors, and the processing method for each subtype is as follows.

1-1 No missing parent

It is determined whether the avi information and the parent information of the object match the avi information and the mother information of the other object, and if both the avi information and the mother information match, the object information is deleted, and if only one of the avi information or the mother information matches Absent information or mother information that is not used can be replaced with missing values to correct errors in household information.

More specifically, when both IDs of Sire and Dam of the duplicate object are the same, the exact same object is repeatedly input, so only one data is left and deletion is performed.

<How to correct errors when duplicate objects have the same parent ID>

If both IDs of Sire and Dam of the duplicated object are different, it is difficult to distinguish which data is the actual data if the parental ID is different and the object appears more than once. By selecting only one of several pieces of information, it is possible to prevent the kinship relationship in the matrix from being fragmented, but since it cannot be determined by guessing, in this case, all family information of all the duplicated objects is erased to prevent incorrectly calculated kinetic matrix calculation.

<How to correct errors when the parent IDs of duplicate objects are all different>

If the ID of one of the two parents of a duplicate object is different, and if the ID of their Sire (or Dam) is fixed in two or more duplicate objects, the information of the registrant is judged to be reliable, so the Dam (or Sire) is not deleted. By replacing only the ID of the missing value, it is possible to prevent the blood relationship from being broken.

<If only one ID of the parent of the duplicate objects is different>

1-2 If one of the parents has missing values

 When converging each of the parent information that is not the missing value of the object, if each converges, the information of the object is preserved by combining the non-missing parent information with one object data.

<Steps to estimate breeding price and genetic parameters>

<Animal Mixed Model (left) and Mixed Model (MME, right)>

Using the blood-related matrix produced earlier, the above-described animal mixing model equation can be solved using the BLUP analysis method to estimate the breeding price, prediction error variance (PEV), and breeding price accuracy.

The REML analysis method can be used to estimate genetic parameters such as genetic variance, environmental variance, and heritability.

Such a process may be performed by the evaluation item evaluation unit.

Since the analysis algorithm is well known, detailed description thereof is omitted here.

Steps for displaying breeder price and genetic parameters

The estimated genetic capacity values can be displayed through the evaluation item providing unit along with the values used in the animal model.

The user can search only the objects of interest from among the entire objects (Search function), and it is possible to sort the objects in the order of the highest breeder and output them for easy use for mating and selection.

Claims (9)

In the method of providing breeding price and genetic parameter evaluation information for improving agricultural and livestock varieties,
Preparing a database including phenotypic information of agricultural and livestock products, individual additional information, lineage information, species information, and household information;
Selecting a line of interest and a phenotype of interest ;
Calculating information necessary for calculating a breeding price and calculating a genetic parameter including household information and additional information about an individual of interest and phenotype of interest;
Estimating breeding prices and genetic parameters by reflecting the phenotype information, household information, and individual additional information;
Displaying the estimated breeding price and genetic parameters;
Breeding value and genetic parameter evaluation information providing method comprising the. The method of claim 1, wherein the information necessary for calculating the breeding price and calculating the genetic parameter includes a blood-related matrix .
The method of claim 1, wherein the phenotype information includes a static phenotype and a dynamic phenotype .
[3] The method of claim 2, wherein the blood relation matrix is corrected for errors in household information. 5. The method of claim 4, wherein the error for the household information includes a duplicate individual error, a circular error, or a gender error. The method of claim 1, wherein the breeding price is calculated through the animal model of Equation 1 below.
<Equation 1>
y = μ + sex + α + ε
y: Measurement of the trait of interest (dependent variable)
μ: Overall mean of the blood quality of interest
sex: sex fixed effect of each individual
α: Random additive genetic effect of each individual (breeding price)
ε: Environmental effect and residual The method of claim 1, wherein the step of displaying the breeder's price and genetic parameters for a predetermined selected phenotype selectively shows only information about the individual of interest. The method of claim 1, wherein the step of displaying the breeding price and the genetic parameter for a predetermined phenotype is characterized by arranging and displaying the individuals in the order of breeding price.
In the apparatus for providing breeding price and genetic parameter information for the improvement of agricultural and livestock varieties,
An evaluation item storage unit that stores and manages a database including phenotypic information of agricultural and livestock products, individual additional information, system information, species information, and household information in memory;
An evaluation item selection unit for selecting a line of interest and a phenotype of interest ;
An evaluation item calculating unit for calculating and storing and managing information necessary for breeding price calculation and genetic parameter calculation including household information and additional information on individuals of interest and phenotype of interest;
An evaluation item evaluation unit that estimates a breeding price and a genetic parameter by reflecting the phenotype information, household information, and individual additional information and stores it in the memory ;
An evaluation result providing unit displaying the estimated breeding price and genetic parameter;
Breeding value and genetic parameter information providing device comprising a.

Images