KR102510936B1 - Apparatus and method for extract range of antibody data based on gut microorganism data - Google Patents

Abstract

An apparatus for deriving a range of immune indicators in the blood based on intestinal microbial data collects a plurality of sample data for feces of an object, and includes a first number of feature variables extracted from each of the collected plurality of sample data. A collection unit that acquires feature data; a first extractor that extracts second feature data including a second number of feature variables smaller than the first number from the first feature data based on a predefined algorithm; and a second feature data A second extraction unit for extracting third feature data including a third number of feature variables less than the second number from the second feature data based on genetic analysis for, excluding at least some of the third feature variables A third extractor for extracting a fourth number of feature variables less than the third number from the third feature data based on the classification performance of the machine learning-based blood immune index range prediction model for the third feature data and a feces of the object A blood immune index range derivation unit for extracting a fourth number of feature variables from and inputting the extracted fourth number of feature variables into a machine learning-based blood immune index range prediction model to derive a range of immune index ranges in blood. there is.

Description

Apparatus and method for deriving range of immune indicators in blood based on intestinal microbial data

The present invention relates to an apparatus and method for deriving a range of immune indicators in the blood based on intestinal microbial data.

It is known that the microbiota plays an important role in maintaining host (human) immunity and homeostasis of metabolites. The intestinal microflora and the host exchange chemical signals, and the intestinal microflora exerts a hypertrophic effect on the host system, such as the expression of immune cells, the production of neurotransmitters, and the production of short chain fatty acids (SCFA).

Meanwhile, IgE is well known as a type of immunoglobulin involved in the development of allergic diseases. Although IgE accounts for 0.05% of total Ig, it has the ability to induce the strongest inflammatory response because it promotes the secretion of histamine related to tissue inflammation in cooperation with mast cells in allergic reactions.

Since the concentration of serum total IgE tends to be high in patients with various allergies (e.g., patients with atopic dermatitis, allergic rhinitis, bronchial asthma, etc.), total IgE in serum is measured as one of the methods used to diagnose allergic diseases. . In addition, it is known that the concentration of total serum IgE increases with the onset and exacerbation of allergic diseases.

Korean Patent Publication No. 2019-0004586 (published on January 14, 2019)

An object of the present invention is to provide an apparatus and method for deriving a range of immune indicators in the blood based on intestinal microbial data. However, the technical problem to be achieved by the present embodiment is not limited to the technical problems described above, and other technical problems may exist.

As a technical means for achieving the above-described technical problem, the apparatus for deriving a range of immune indicators in blood based on intestinal microbial data according to the first aspect of the present invention collects a plurality of sample data for feces of an object, and a collection unit that obtains first feature data including a first number of feature variables extracted from each of a plurality of collected sample data; a first extraction unit extracting second feature data including a second number of feature variables less than the first number from the first feature data based on a predefined algorithm; a second extraction unit extracting third feature data including a third number of feature variables less than the second number from the second feature data based on genetic analysis of the second feature data; After excluding at least some of the third number of feature variables, based on the classification performance of the machine learning-based blood immune index range prediction model for the third feature data, from the third feature data Less than the third number A third extractor for extracting four feature variables and extracting the fourth number of feature variables from the object's feces, and applying the extracted fourth feature variables to the machine learning-based blood immune index range prediction model It may include a blood immune index range derivation unit for deriving a range of immune index in blood by inputting the input.

A method for deriving a range of immune indicators in the blood based on intestinal microbial data according to a second aspect of the present invention collects a plurality of sample data for feces of an object, and first extracted from each of the collected plurality of sample data obtaining first feature data including the number of feature variables; extracting second feature data including a second number of feature variables smaller than the first number from the first feature data based on a predefined algorithm; extracting third feature data including a third number of feature variables less than the second number from the second feature data based on genetic analysis of the second feature data; After excluding at least some of the third number of feature variables, based on the classification performance of the machine learning-based blood immune index range prediction model for the third feature data, from the third feature data Less than the third number Extracting four feature variables, extracting the fourth number of feature variables from feces of an object, and inputting the extracted fourth feature variables into the machine learning-based blood immune index range prediction model to obtain blood It may include deriving a range of immune indicators.

The above-described means for solving the problems is only illustrative and should not be construed as limiting the present invention. In addition to the exemplary embodiments described above, there may be additional embodiments described in the drawings and detailed description.

According to any one of the above-described means for solving the problems of the present invention, the present invention can provide a device and method for deriving a range of immune indicators in blood based on intestinal microbial data.

Through this, the present invention can determine whether the subject's blood immune index belongs to the high-risk group based on the range of blood immune index derived through the subject's intestinal microbial data.

In addition, the present invention can help reduce the incidence of allergic diseases and effectively treat allergic diseases by diagnosing and predicting the risk group of the subject's blood immune indicators at an early stage so that the subject can properly manage them.

1 is a block diagram of an apparatus for deriving a blood immune index range according to an embodiment of the present invention.
2a to 2d are views for explaining a method of deriving a range of immune indicators in blood according to an embodiment of the present invention.
3 is a flowchart illustrating a method of deriving a range of immune indicators in blood based on intestinal microbial data according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail so that those skilled in the art can easily practice the present invention with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" to another part, this includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween. . In addition, when a certain component is said to "include", this means that it may further include other components without excluding other components unless otherwise stated.

In this specification, a "unit" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Further, one unit may be realized using two or more hardware, and two or more units may be realized by one hardware.

In this specification, some of the operations or functions described as being performed by a terminal or device may be performed instead by a server connected to the terminal or device. Likewise, some of the operations or functions described as being performed by the server may also be performed in a terminal or device connected to the corresponding server.

Hereinafter, specific details for the implementation of the present invention will be described with reference to the accompanying configuration diagram or process flow chart.

1 is a block diagram of an apparatus for deriving a blood immune index range according to an embodiment of the present invention.

Referring to FIG. 1, the blood immune index range deriving apparatus 10 includes a collection unit 100, a first extraction unit 110, a second extraction unit 120, a third extraction unit 130, and a blood immune index range An extraction unit 140 may be included. However, the apparatus 10 for deriving the range of immune indicators in blood shown in FIG. 1 is only one implementation example of the present invention, and various modifications are possible based on the components shown in FIG. 1 .

Hereinafter, FIG. 1 will be described with reference to FIGS. 2A to 2D.

The collection unit 100 may collect a plurality of sample data about feces of an object.

The collection unit 100 may obtain first feature data including a first number of feature variables from each of a plurality of collected sample data. For example, the collecting unit 100 includes a first number (eg, 726) of feature variables from each of the first sample data, the second sample data, and the third sample data obtained by subdividing feces of the object collected from the object. It is possible to obtain the first feature data to be. Here, the feature variables refer to various enteric bacteria variables included in each sample data.

The first extractor 110 extracts a second feature including a second number less than the first number (eg, 726), for example, 150 to 200 feature variables, from the first feature data based on a predefined algorithm. data can be extracted. At this time, the second number may be, for example, 172.

Here, the predefined algorithm may include arcsine square root transformation.

The first extractor 110 performs arcsine square root transformation on a plurality of first feature data for a plurality of sample data, and the probability that a plurality of identical feature variables included in the plurality of first feature data is equal to or less than a specific score is a predetermined probability If it exceeds, the second feature data may be extracted by removing the same feature variable from the first feature data for each of a plurality of sample data.

For example, when arcsine square root transformation is performed on the A feature variable for each of a plurality of sample data, the score for the A feature variable of the first sample data is a first score (eg, 0 point), and the second sample data When the score for the A feature variable of is the first score and the score for the A feature variable of the third sample data is the first score, the probability that the A feature variable has the first score in the plurality of sample data is a predetermined probability (eg, 85%), the first extractor 110 may remove the A feature variable from the first feature variable included in the first feature data for each sample data.

The classification performance measurement unit (not shown) inputs the second feature data including the second number of feature variables extracted from the first feature data of each of a plurality of sample data into a machine learning-based blood immune index range prediction model and classifies performance score can be measured.

For example, referring to FIG. 2A , the classification performance measuring unit (not shown) converts the second feature data including the second number of feature variables into the ensemble-based blood immune index range prediction model 20 and the DNN-based blood A classification performance score for whether the range of an immune index (eg, IgE) in blood is classified into one of a plurality of predetermined ranges (lower range, upper range) by inputting into each of the immune index range prediction models 22 can be measured. Here, the predetermined lower range may include, for example, a blood immune index concentration of less than 150 UI/ml, and a preset upper range may include, for example, a blood immune index concentration of 300 UL/ml or more. . When the blood immune index is classified into one of a predetermined lower range or upper range, it can be estimated that the probability of having an allergic disease is high.

Looking at the classification performance score for the range of blood immune indicators classified through the ensemble-based blood immune indicator range prediction model (20), the classification performance score of the algorithm using a random forest is 0.179, and Adaboost The classification performance score of the algorithm using ) is 0.589 points, the classification performance score of the algorithm using gradient boost is 0.375 points, and the classification performance score of the algorithm using XGBoost is 0.411 points.

The classification performance score for the range of immune indicators in the blood classified through the DNN-based predictive model for the range of immune indicators in the blood (22) is 0.411 points.

When examining such classification performance scores, it can be seen that the ensemble-based blood immune index range prediction model 20 using Athaboost shows the highest performance in classifying the range of blood immune indicators.

Returning to FIG. 1 again, the second extractor 120 may express the second feature data as a visualization graph (see FIG. 2B ) and perform genetic analysis on the second feature data expressed as the visualization graph. Here, the visualization graph may include, for example, a Volcano plot graph. A volcano plot graph is a graph that effectively visualizes genes showing a difference in expression level between two groups.

The second extractor 120 may perform DEG (Differentially Expressed Gene) analysis through a visualization graph in which the second feature data is visualized. Here, the DEG analysis is a method of analyzing whether the average expression level of the same gene is significantly different under different conditions. Specifically, DEG analysis is a method of measuring gene expression values, processing them statistically, selecting gene candidates with significant expression between a control group and a comparison group, and then analyzing gene expression patterns according to specific conditions or treatment groups.

The second extractor 120 extracts third feature data including a third number less than the second number, for example, 60 to 100 feature variables, from the second feature data based on genetic analysis of the second feature data. can be extracted. Here, the third number may be, for example, 81.

The second extractor 120 extracts a P-value and a Fold-change from the DEG analysis for each of a plurality of feature variables included in the second feature data, and extracts a third value based on the extracted P-value and Fold-change. Feature data can be extracted. Here, Fold-change is information indicating how many times the average expression level in the experimental group for a gene is the average expression level in the control group, and P-value indicates whether the difference in average expression level between the two groups is a statistically significant value. It is information.

Specifically, the second extractor 120 converts third feature data including a feature variable having a P-value less than a preset threshold among P-values for each feature variable of the derived second feature data into second feature data. can be extracted from Here, the number of feature variables having a P-value less than the predetermined threshold has a third number less than the second number.

For example, the second extractor 120 converts third feature data including a feature variable having a P-value of less than 0.5 among P-values for each of a plurality of feature variables included in the second feature data into the second feature data. can be extracted from the data.

After the third extractor 130 excludes at least some of the third number of feature variables included in the extracted third feature data, the classification performance of the machine learning-based blood immune index range prediction model for the third feature data A fourth number less than the third number from the third feature data based on . For example, 40 to 60 feature variables may be extracted. Here, the fourth number may be, for example, 51.

The third extractor 130 inputs characteristic variables excluding at least some of the third number of characteristic variables included in the third characteristic data into the blood immune index range prediction model, and when the number of characteristic variables is to a certain extent, blood It is possible to determine whether the immune index range prediction model is good at classifying immune indicators in blood (ie, classifying into a predetermined lower range or a predetermined upper range).

Specifically, the third extractor 130 classifies a plurality of feature variables included in the third feature data while excluding at least some feature variables having a low fold-change, and classifies the measured machine learning-based blood immune index range prediction model Based on the performance, a fourth feature variable may be extracted from the third feature data.

The third extractor 130 takes the absolute value of the fold-change for each of the third number of feature variables included in the third feature data, and then sequentially extracts the feature variables having the small absolute value of the fold-change one by one into the third feature data. It is possible to measure the classification performance of a machine learning-based blood immune index range prediction model while excluding feature data.

For example, the third extractor 130 extracts 80 feature variables excluding the feature variable having the lowest fold-change absolute value among 81 feature variables (the third number of feature variables) as a machine learning-based immune index in the blood. Classification performance is measured by inputting them into the range prediction model, and then inputting 79 feature variables excluding the feature variable with the lowest absolute value of fold-change among the 80 feature variables into the machine learning-based blood immune index range prediction model for classification Classification performance can be measured by measuring performance, and then inputting 78 feature variables excluding the feature variable with the lowest absolute value of fold-change among the 79 feature variables into a machine learning-based blood immune index range prediction model. In this way, the third extractor 130 may measure classification performance while reducing the number of feature variables until the number of feature variables becomes sufficiently small.

Here, the machine learning-based blood immune index range prediction model is an ensemble-based blood immune index range prediction model, which is based on random forest, Adaboost, gradient boost, and XGBoost and DNN. A blood immune index range prediction model may be included.

Referring to FIGS. 2C and 2D together, according to the present invention, the top 51 feature variables 201 having high fold-change absolute values among 81 feature variables (the third number of feature variables) are exemplarily based on machine learning It was confirmed that the blood immune index range prediction model showed the best classification performance when input to the blood immune index range prediction model.

Looking at the classification performance score for the range of blood immune indicators classified through the ensemble-based blood immune indicator range prediction model (24) in which 51 feature variables (201) are input, classification of the algorithm using a random forest The performance score is 0.741 points, the classification performance score of the algorithm using Adaboost is 0.714 points, the classification performance score of the algorithm using gradient boost is 0.589 points, and the classification performance score of the algorithm using XGBoost is The score was 0.571.

The classification performance score for the range of immune indicators in the blood classified through the DNN-based blood immune indicator range prediction model 26 in which 51 feature variables 201 are input is 0.482 points.

In summary, among the ensemble-based blood immune indicator range prediction model (24) and the DNN-based blood immune indicator range prediction model (26), the ensemble-based blood immune indicator range prediction model (24) using a random forest is It was confirmed that it showed the highest classification performance with 0.741 (203) points.

Returning to FIG. 1 again, the blood immune index range extractor 140 extracts the fourth number of feature variables from the object's feces, and converts the extracted fourth number of feature variables into a machine learning-based blood immune index range prediction model You can derive the range of immune indicators in the blood by entering

Through this, the present invention is based on the range of the blood immune index derived through the blood immune index range prediction model, the subject's blood immune index is in the high-risk group (ie, the risk group whose blood immune index belongs to the preset lower range or upper range) You can determine whether you belong to it or not.

In addition, the present invention can help reduce the incidence of allergic diseases and effectively treat allergies by diagnosing and predicting a high-risk group of immune indicators in the blood at an early stage so that the subject can properly manage immunity.

On the other hand, those skilled in the art, each of the collection unit 100, the first extraction unit 110, the second extraction unit 120, the third extraction unit 130, and the blood immune marker range extraction unit 140 are implemented separately, or , it will be fully understood that one or more of these may be integrated and implemented.

3 is a flowchart illustrating a method of deriving a range of immune indicators in blood based on intestinal microbial data according to an embodiment of the present invention.

Referring to FIG. 3, in step S301, the apparatus 10 for deriving a range of immune indicators in the blood collects a plurality of sample data for feces of an object, and includes a first number of feature variables extracted from each of the collected plurality of sample data It is possible to obtain the first feature data to be.

In step S303, the blood immune index range deriving device 10 may extract second feature data including a second number of feature variables smaller than the first number from the first feature data based on a predefined algorithm. Here, the predefined algorithm may include arcsine square root transformation.

In step S305, the apparatus 10 for deriving a range of immune indicators in blood extracts third feature data including a third number of feature variables less than the second number from the second feature data based on the genetic analysis of the second feature data. can

In step S307, the apparatus 10 for deriving a range of immune indicators in the blood excludes at least some of the third number of feature variables, and then, based on the classification performance of the machine learning-based predictive model for the range of immune indicators in the blood for the third feature data, the third A fourth number of feature variables smaller than the third number may be extracted from the feature data.

In step S309, the apparatus 10 for deriving the range of immune indicators in the blood extracts the fourth number of characteristic variables from the feces of the object, and inputs the extracted fourth number of characteristic variables into the machine learning-based predictive model for the range of immune indicators in the blood. A range of immune indicators can be derived.

In the above description, steps S301 to S309 may be further divided into additional steps or combined into fewer steps, depending on the implementation of the present invention. Also, some steps may be omitted if necessary, and the order of steps may be changed.

An embodiment of the present invention may be implemented in the form of a recording medium including instructions executable by a computer, such as program modules executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The above description of the present invention is for illustrative purposes, and those skilled in the art can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the claims to be described later rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and equivalent concepts thereof should be construed as being included in the scope of the present invention. .

10: Blood immune index range derivation device
100: collection unit
110: first extraction unit
120: second extraction unit
130: third extraction unit
140: Blood immune index range derivation unit

Claims (12)

In the device for deriving the range of immune indicators in the blood based on intestinal microbial data,
a collection unit that collects a plurality of sample data for feces of an object and obtains first feature data including a first number of feature variables extracted from each of the collected plurality of sample data;
a first extraction unit extracting second feature data including a second number of feature variables less than the first number from the first feature data based on a predefined algorithm;
a second extraction unit extracting third feature data including a third number of feature variables less than the second number from the second feature data based on genetic analysis of the second feature data;
When deriving the range of immune indicators in blood by inputting the third feature data including an arbitrary number of feature variables excluding at least some of the third number of feature variables into a machine learning-based blood immune index range prediction model A third extractor for extracting a fourth number of feature variables less than the third number based on the classification performance of the machine learning-based blood immune index range prediction model; and
The fourth number of feature variables are extracted from the feces of another object, and the third feature data including the fourth number of feature variables extracted from the feces of the other object is converted to the machine learning-based blood immune index range prediction model Blood immune index range derivation unit that derives the range of immune index in blood by inputting
A device comprising a.
According to claim 1,
The predefined algorithm includes an arcsine square root transformation,
The first extractor performs arcsine square root transformation on a plurality of first feature data extracted from each of the plurality of sample data, and a probability that a plurality of identical feature variables included in the plurality of first feature data is equal to or less than a specific score is determined. and extracting the second feature data by removing the same feature variable from the plurality of first feature data when the set probability is exceeded.
According to claim 1,
Wherein the second extractor expresses the second feature data as a visualization graph, and performs the genetic analysis on the second feature data expressed as the visualization graph.
According to claim 3,
The second extraction unit performs DEG (Differentially Expressed Gene; DEG) analysis through the visualization graph,
The second extractor extracts a P-value and a Fold-change from the DEG analysis for each of a plurality of feature variables included in the second feature data, and extracts the P-value and Fold-change based on the extracted P-value and Fold-change. 3 apparatus, which extracts feature data.
According to claim 4,
The second extractor extracts the third feature data including a feature variable having a P-value less than a predetermined threshold among the derived P-values.
According to claim 4,
The third extractor measures classification performance of the machine learning-based blood immune index range prediction model while excluding at least some feature variables having a low fold-change from a plurality of feature variables included in the third feature data which is, the device.
In the method for deriving the range of immune indicators in the blood based on intestinal microbial data,
collecting a plurality of sample data for feces of the object and obtaining first feature data including a first number of feature variables extracted from each of the collected plurality of sample data;
extracting second feature data including a second number of feature variables smaller than the first number from the first feature data based on a predefined algorithm;
extracting third feature data including a third number of feature variables less than the second number from the second feature data based on genetic analysis of the second feature data;
After excluding at least some of the third number of feature variables, based on the classification performance of the machine learning-based blood immune index range prediction model for the third feature data, from the third feature data Less than the third number Extracting 4 feature variables; and
Extracting the fourth number of feature variables from the object's feces, and inputting the extracted fourth number of feature variables to the machine learning-based blood immune index range prediction model to derive a range of immune indicators in blood
A method for deriving a range of immune indicators in the blood, comprising a.
According to claim 7,
The predefined algorithm includes an arcsine square root transformation,
Extracting the second feature data
The arcsine square root transform of a plurality of first feature data extracted from each of the plurality of sample data, and a probability that a plurality of identical feature variables included in the plurality of first feature data is equal to or less than a specific score exceeds a predetermined probability. case, extracting the second feature data by removing the same feature variable from the plurality of first feature data.
According to claim 7,
The step of extracting the third feature data is
Expressing the second characteristic data as a visualization graph, and performing the genetic analysis on the second characteristic data represented by the visualization graph.
According to claim 9,
The step of extracting the third feature data is
Performing DEG (Differentially Expressed Gene; DEG) analysis through the visualization graph; and
For each of the plurality of feature variables included in the second feature data, a P-value and a Fold-change are extracted from the DEG analysis, and the third feature data is extracted based on the extracted P-value and Fold-change. Which comprises the step of, how to derive the range of immune indicators in the blood.
According to claim 10,
The step of extracting the third feature data is
Extracting the third feature data including a feature variable having a P-value less than a predetermined threshold among the derived P-values, a method for deriving a range of immune indicators in blood.
According to claim 10,
The step of extracting the fourth number of feature variables is
Measuring classification performance of the machine learning-based blood immune index range prediction model while excluding at least some of the feature variables having a low fold-change from a plurality of feature variables included in the third feature data. In, how to derive the range of data.

Images