Device for calculating risk indicator for mild cognitive impairment, method, and program
The device calculates MCI risk using intestinal microbiota data and subject attributes, addressing the need for effective MCI risk assessment through intestinal microbiota analysis and providing a basis for cognitive function improvement strategies.
Patent Information
- Application Number
- PCT/JP2024/019474
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-31
- Filing Date
- 2024-05-28
- Publication Date
- 2024-12-05
AI Technical Summary
Current methods lack an effective way to calculate the risk of mild cognitive impairment (MCI) using intestinal microbiota data from stool specimens and questionnaires, despite correlations between L-ergothioneine blood concentration and memory ability.
A device and method that extracts data on intestinal microorganisms associated with MCI, classifies them, and uses a model to calculate the probability of MCI risk based on intestinal microbiota data and subject attributes, providing a risk index calculation for MCI.
Enables non-invasive evaluation of MCI risk using stool samples and questionnaires, allowing for personalized risk assessment and potential interventions for improving cognitive function.
Smart Images

Figure JP2024019474_05122024_PF_FP_ABST
Abstract
Description
Mild cognitive impairment risk index calculation device, method, and program
[0001] An embodiment of the present invention relates to a technology for calculating a disease evaluation index using gut microbiota data, and in particular provides a technology for calculating the risk of mild cognitive impairment for a subject who has submitted a stool sample and a questionnaire.
[0002] Research is being conducted on food compositions for maintaining brain health, particularly cognitive function. Patent Document 1 discloses that L-ergothioneine is a food composition useful for improving cognitive function in humans.
[0003] Furthermore, a technology for calculating a disease evaluation index using the human intestinal microbiota has been developed. Patent Document 2 discloses a technology for estimating the risk of a user who has submitted a stool sample contracting a disease such as atopic dermatitis using a structural equation model.
[0004] JP 2019-180364 A Patent No. 7270143 A
[0005] However, paragraph 0056 of Patent Document 1 merely indicates that there is a correlation between blood levels of L-ergothioneine and memory ability. Therefore, when examining a subject's memory ability, it is necessary to collect blood from the subject and measure the blood level of L-ergothioneine.
[0006] The present invention was completed through intensive research focusing on these issues, and its purpose is to provide a technology for calculating the risk of mild cognitive impairment for subjects who submit a stool sample and questionnaire.
[0007] In order to solve the above problem, a first invention is a risk index calculation device for mild cognitive impairment, comprising: an extraction unit that extracts second data on intestinal microorganisms or factors suggested to be associated with mild cognitive impairment from first data on the intestinal microbiota of a subject; and a calculation unit that inputs the second data into a model that calculates the probability that the subject has mild cognitive impairment, using third data on one or more intestinal microorganisms that show differences between the intestinal microbiota of a healthy person and the intestinal microbiota of a person suffering from mild cognitive impairment as variables, and calculates the subject's risk of mild cognitive impairment, wherein the risk of mild cognitive impairment is the probability of having or will suffer from mild cognitive impairment.
[0008] A second invention is a mild cognitive impairment risk index calculation device described in the first invention, in which the second data is classified into a group of intestinal microorganisms that are common in mild cognitive impairment and a group of intestinal microorganisms that are rare in mild cognitive impairment.
[0009] The third invention is a mild cognitive impairment risk index calculation device described in the second invention, in which the intestinal microbial group that is prevalent in mild cognitive impairment and the intestinal microbial group that is rare in mild cognitive impairment are each classified into three subgroups: male, female, and unrelated to both sexes.
[0010] A fourth invention is a mild cognitive impairment risk index calculation device according to the first invention, wherein the second data relates to mucin degradation, IgA protease production, serotonin secretion from enterochromaffin cells, bile acid oxidation, promotion of recovery of intestinal barrier function, production of HDAC inhibitors, or protection of IgA from trypsin degradation.
[0011] A fifth invention is the mild cognitive impairment risk index calculation device described in the first invention, in which the extraction unit also extracts attribute information of the subject, and the calculation unit scores the second data using the attribute information of the subject and / or information on intestinal microorganisms corresponding to factors related to the mild cognitive impairment, and inputs the second data after scoring into the model.
[0012] A sixth invention is a mild cognitive impairment risk index calculation device described in the fifth invention, wherein the subject's attribute information includes background information about the subject at the time of birth and background information about their childhood life.
[0013] A seventh invention is the mild cognitive impairment risk index calculation device according to the first invention, wherein the model is a model created for each gender attribute, or a mixed gender model.
[0014] An eighth invention is a method for calculating a risk index of mild cognitive impairment, in which a computer extracts, from first data regarding the gut microbiota of a subject, second data regarding factors or gut microorganisms that are suggested to be associated with mild cognitive impairment, and inputs the second data into a model that calculates the probability that the subject has mild cognitive impairment, using third data regarding one or more gut microorganisms that show differences between the gut microbiota of a healthy person and the gut microbiota of a person suffering from mild cognitive impairment as a variable, to calculate the subject's risk of mild cognitive impairment, where the risk of mild cognitive impairment is the probability of having or will have mild cognitive impairment.
[0015] A ninth invention is a mild cognitive impairment risk index calculation program that causes a computer to execute the following steps: an extraction step of extracting second data on factors or intestinal microorganisms that are suggested to be associated with mild cognitive impairment from first data on the subject's intestinal microbiota; and a calculation step of inputting the second data into a model that calculates the probability that the subject has mild cognitive impairment, using third data on one or more intestinal microorganisms that show differences between the intestinal microbiota of a healthy person and the intestinal microbiota of a person suffering from mild cognitive impairment as variables, and calculating the subject's risk of mild cognitive impairment, wherein the risk of mild cognitive impairment is the probability of having or will suffer from mild cognitive impairment.
[0016] According to the present invention, it is possible to provide a technique for calculating the risk of a disease that a subject wishes to be evaluated for, for the subject who has submitted a stool sample and a questionnaire.
[0017] 1 is an overall schematic diagram of a disease evaluation index calculation system according to an embodiment of the present invention; FIG. 1 is a flowchart of a disease evaluation index calculation process according to the embodiment; FIG. 1 is a table showing questionnaire items for investigating basic information of subjects and test subjects according to the embodiment; FIG. 1 is a table showing the conditions of a control group according to the embodiment; FIG. 1 is a table showing the conditions of an MCI group according to the embodiment; FIG. 1 is a diagram for explaining the selection of gut microbial variables when the population is female according to the embodiment; FIG. 2 is a table showing the effect size value, application to the association model, and factors related to MCI for each gut bacterium in FIG. 6; FIG. 2 is a diagram for explaining the selection of gut microbial variables when the population is male according to the embodiment; FIG. 3 is a table showing the effect size value, application to the association model, and factors related to MCI for each gut bacterium in FIG. 8; FIG. 3 is a hypothesis conceptual diagram of the influence of gut bacteria related to MCI according to the embodiment; FIG. 4 is an enlarged view of a portion of FIG. 10; FIG. 5 is a diagram showing the relationship between age and occupancy rate of gut bacteria according to the embodiment (the population is female); FIG. 6 is a diagram for explaining added points according to MCI-related factors according to the embodiment; FIG. 7 is a diagram showing the results of estimating parameters of the association model according to the embodiment (the population is female). FIG. 1 is a diagram for explaining a score estimation model after adding points and a risk calculation model according to this embodiment (the population is female). FIG. 2 is a diagram showing the results of estimating parameters of an association model according to a modified example (the population is male). FIG. 3 is a diagram for explaining a score estimation model after adding points and a risk calculation model according to a modified example (the population is male). FIG. 4 is a diagram showing the results of an ROC analysis of the risk value of MCI according to this embodiment (the population is female). FIG. 5 is a diagram showing the results of an ROC analysis of the risk value of MCI according to a modified example (the population is male).
[0018] The present invention will be described with reference to the accompanying drawings, in which like reference numerals are used to designate common parts in the drawings, and redundant description will be omitted.
[0019] (Disease evaluation index calculation system) Fig. 1 is an overall schematic diagram of a disease evaluation index calculation system according to an embodiment of the present invention. The disease evaluation index calculation system includes a phase for creating a gut microbiota database (DB) from stool samples submitted by a large number of subjects, a phase for creating a model for one or more diseases (specific diseases), and a phase for calculating an evaluation index for the risk of a specific disease. Here, the large number of subjects are participants who provide samples (stool samples and questionnaire response data) used to create the model.
[0020] Below, mild cognitive impairment (hereinafter abbreviated as MCI), which is related to a decline in cognitive function, will be described as an example of a specific disease. The risk of MCI indicates the probability of having or developing MCI. Furthermore, the gut microbiota is a collection of intestinal microorganisms that include intestinal bacteria (bacteria, archaea, etc.) living in the intestine, as well as mold, yeast, etc.
[0021] (Gut Microbiota DB Creation Phase) In the gut microbiota DB creation phase, a large number of subjects (more than 10,000 people) defecate in a toilet and collect stool samples using a stool collection kit. Next, a DNA extraction company that receives the stool samples inputs the stool samples into the gut microbial DNA extraction device 100, which outputs a DNA solution related to the gut microbiota.
[0022] The analysis company receives the DNA solution and inputs it into the gut microbiota analyzer 200 to analyze the gut microbiota. The analysis company stores the analysis result data in the gut microbiota DB 300. The analysis begins by creating an ASV (Amplicon Sequence Variant) sequence unique to each subject's gut microbiota. Next, using a publicly available external DNA sequence database (e.g., the Ribosomal Database Project database), a taxon name is assigned to the gut microbiota data of each subject. The taxon name in this embodiment includes taxon names of taxonomic classes such as family, genus, and species, as well as taxon names used in publicly available external DNA sequence databases.
[0023] The gut microbiota DB 300 associates ID (identification number) information of each subject with the gut microbiota data of that subject. Although the DNA extraction company and the analysis company have been described separately, the same company may perform both the extraction and analysis work.
[0024] Next, the subjects who have submitted stool samples answer a questionnaire using their ID information and submit it to a questionnaire collection company, which then stores the completed questionnaire response data in the questionnaire DB 400.
[0025] The questionnaire survey surveys detailed information including the subject's demographic information, body mass index (BMI), bowel frequency, lifestyle habits (drinking, smoking, exercise frequency), disease status, sleep status, CES-D score, whether or not they have received treatment for Helicobacter pylori, history of hospitalization or surgery, and use of prescription or over-the-counter medications. In this embodiment, the subject's demographic information refers to basic information such as gender, age, and childhood living area, and details of which will be described later.
[0026] For female subjects only, menstrual status and whether pregnant or breastfeeding will be investigated. The disease status will be investigated by asking about the name of the disease and whether or not the patient is currently undergoing treatment for 13 categories of disease (MCI, atopic dermatitis, bone and / or joint disease, bronchial asthma, diabetes, dyslipidemia, gastrointestinal disease, heart disease, hypertension, kidney disease, liver disease, lower back and / or joint pain, and other diseases). In other words, people suffering from a disease include those who currently have the disease and / or those who have had the disease in the past. Regarding the use of prescription and over-the-counter drugs, a survey was conducted to determine whether or not the person was taking drugs from 16 categories (drugs for gastric / duodenal ulcers / reflux esophagitis, antihypertensive drugs, antihyperlipidemic drugs, antidiabetic drugs, sleeping pills, painkillers / antipyretics, allergy drugs, antianginal drugs, laxatives / constipation drugs, osteoporosis drugs, rheumatoid arthritis drugs, corticosteroids, antibiotics, cold medicines, antithrombotic drugs, and others) at the time of stool collection.
[0027] (Model Creation Phase) In the model creation phase, the model creator first selects a disease to be evaluated. The association model creation device 500 includes an input unit that accepts input of MCI as the selected disease, an extraction unit that extracts ID information of subjects related to MCI for each gender attribute from the questionnaire DB 400 and also extracts gut microbiota data for the corresponding ID information from the gut microbiota DB 300, and a creation unit that creates an association model using the creation method described below. The association model creation device 500 stores the created association model in the association model DB 600.
[0028] Next, the score estimation model creation device 700 extracts the association model from the association model DB 600 and creates a score estimation model by a creation method described later. The created score estimation model is stored in the score estimation model DB 800.
[0029] (Disease evaluation index calculation phase) In the disease evaluation index calculation phase, the evaluation company submits a stool sample collected with a stool collection kit and requests that a person (hereinafter referred to as the subject) who wishes to have their risk of MCI assessed submit a stool sample as described in the gut microbiome DB creation phase. The subject's stool sample, like the stool samples of the many subjects described above, is processed by the intestinal microbial DNA extraction device 100 and the intestinal microbiome analysis device 200. The subject's intestinal microbiome data is stored in the intestinal microbiome DB 300.
[0030] Furthermore, the MCI and the subject's ID information (including the subject's attribute information) are input to the disease evaluation index calculation device 900. Here, the subject's attribute information is also input to the disease evaluation index calculation device 900 from the questionnaire response data submitted by the subject. The questionnaire response data submitted by the subject may be input to the questionnaire DB 400, similar to the questionnaire response data of the multiple subjects described above. In this case, the disease evaluation index calculation device 900 simply extracts the subject's questionnaire response data from the questionnaire DB 400. In this way, the subject is a person (user) who provides a sample (stool sample and questionnaire response data) to have their risk of MCI assessed.
[0031] The disease evaluation index calculation device 900 includes an input unit that accepts input information, an extraction unit that extracts information related to the input information, a point adding unit that adds points to a score estimation model, a calculation unit that calculates an evaluation index using the added score estimation model, and an output unit that outputs an evaluation report. 910 in Fig. 1 indicates the above-mentioned point adding unit.
[0032] The input unit receives input of MCI as a specific disease and ID information of the subject. The extraction unit extracts information related to the ID information of the subject (attribute information and intestinal microbiota data of the subject) from the external intestinal microbiota DB300, and extracts a score estimation model related to MCI from the score estimation model DB800.
[0033] The score adding unit 910 adds points to the score estimation model using the subject's attribute information.The calculation unit then inputs the subject's intestinal microbiota data into the added score estimation model and calculates an evaluation index for MCI risk.The output unit outputs an evaluation report based on the calculated evaluation index.Note that although the explanation has been given with separate businesses performing each phase, the same business may perform all phases.
[0034] 2 is a flowchart of the disease evaluation index calculation process according to this embodiment. This process is divided into two processing groups: a first processing group (S100 to S120) that creates a score estimation model for each disease in advance, and a second processing group (S200 to S240) that calculates an evaluation index for the risk of MCI, which is the disease that the subject wishes to be evaluated for.
[0035] The first processing group includes a step of extracting gut microbiota data for each disease from multiple subjects (S100), a step of creating an association model for each disease (S110), and a step of creating a score estimation model for each disease (S120). The second processing group includes a step of inputting MCI as the disease for which the subject wishes to be evaluated (S200), a step of extracting gut microbiota data from the subject (S210), a step of adding points to the score estimation model using the subject's attribute information and / or information on gut bacteria corresponding to factors related to MCI (S220), a step of estimating scores for latent variables using the score estimation model after adding points (S230), and a step of calculating an assessment index for MCI risk (S240).
[0036] (Explanation of S100) In this embodiment, "MCI" is used as an example of a disease. S100 extracts gut microbiome data for a group of healthy individuals and a group of individuals suffering from MCI (people belonging to these groups correspond to the subjects described above) from the gut microbiome DB 300. FIG. 3 is a table showing questionnaire items for investigating basic information about subjects and examinees. FIGS. 4 and 5 show conditions for extracting each group. These conditions include conditions that apply only to women. Note that the questionnaire items in FIG. 3 are used to investigate not only subjects but also examinees.
[0037] FIG. 3 is a table showing questionnaire items for investigating basic information of subjects and test subjects according to this embodiment. This embodiment relates to a technology for calculating a disease evaluation index for MCI using the subject's gut microbiome data. The gut microbiome of a subject and test subject may be influenced by the subject's age, biological sex (hereinafter simply referred to as "sex"), and lifestyle factors such as the region in which the subject lived during childhood. Questionnaire item numbers 8 and 9 correspond to those related to lifestyle factors. Additionally, item 10, the method of delivery, is asked as a factor in inheriting the mother's gut microbiome. Therefore, information potentially related to the subject's gut microbiome data is referred to as the subject's basic information.
[0038] In this embodiment, attribute information of the subjects and test subjects is investigated using the questionnaire items shown in Figure 3. To investigate the influence of the subjects' gut microbiota and their lifestyle background, for example, the prefecture in which they mainly lived up to the age of three is asked.
[0039] In this embodiment, since it is assumed that all subjects and examinees are Japanese, the subjects' and examinees' countries of birth are not asked. However, the country of birth may be asked as a questionnaire item. The basic information of the subjects and examinees extracted from the questionnaire response data in this way will be referred to below as the subjects' and examinees' "attribute information." The attribute information includes the subjects' and examinees' backgrounds at the time of birth and their childhood backgrounds.
[0040] FIG. 4 is a table showing the conditions of the control group according to this embodiment. The control group (Normal Control, hereinafter abbreviated as NC) is a group of healthy people who meet all of the conditions listed in FIG. 4. FIG. 5 is a table showing the conditions of the MCI group according to this embodiment. The group of people suffering from MCI (hereinafter abbreviated as MCI group) is a group of people who meet all of the conditions listed in FIG. 5. In this embodiment, each group is further divided by gender. The MCI group consists of 26 men in their 40s to 80s and 33 women in their 40s to 80s.
[0041] (Explanation of S110) S110 is a step for creating an association model for MCI. In this embodiment, association models for MCI are created separately for men and women. In this embodiment, the male population used gut microbiota data from 11 MCI subjects in their 70s and 17 disease-free control subjects, while the female population used gut microbiota data from 18 MCI subjects in their 70s and 23 disease-free control subjects. Here, the MCI group was limited to those in their 70s because this is the most common age group among subjects in their 40s to 80s. Note that even if the human gut microbiota contains the same gut microorganisms, their behavior within the body may differ between men and women.
[0042] First, the gut microbiota data extracted in S100 is subjected to centered log-ratio (CLR) conversion to convert the count values of the gut microbiota data into CLR values. This conversion process makes statistical processing easier.
[0043] Next, in this embodiment, a step (S115) of estimating gut microbiome variables is performed. The gut microbiome variables refer to gut microbiomes that are suggested to be associated with MCI. S115 is performed to search for one or more gut microbiomes that represent differences in the gut microbiota between the control group (NC) and the MCI group. In this embodiment, gut bacteria are selected for each male and female population using the effect size between the NC and MCI groups as an index. Compared to the control group, gut bacteria with an effect size greater than 0.2 are classified and defined as gut bacteria abundant in the MCI group, and gut bacteria with an effect size less than -0.2 are classified and defined as gut bacteria rare in the MCI group. Because the population is divided into males and females, the gut bacteria abundant in the MCI group and the gut bacteria rare in the MCI group also differ between males and females (see Figures 6 and 9, described below). Figures 6 and subsequent figures use gut bacteria as an example of gut microbiomes.
[0044] FIG. 6 is a diagram illustrating the selection of gut microbiota variables for a female population according to this embodiment. The horizontal axis represents the effect size, and the vertical axis represents the name of the gut bacteria (genus level). For convenience in creating bar graphs, gut bacteria with a positive effect size are represented as "MCI," and gut bacteria with a negative effect size are represented as "Healthy." Thus, gut bacteria with a positive effect size are abundant in the gut microbiota of the MCI group, while gut bacteria with a negative effect size are rare in the gut microbiota of the MCI group. Therefore, the processing of S110 allows for the estimation of multiple gut bacteria that are presumed to contribute to the differences in the gut microbiota between the control group and the MCI group. Note that although there are actually 51 types of gut bacteria, only 26 types are displayed in FIG. 6 for ease of viewing the bar graph.
[0045] FIG. 7 is a table showing the effect size values, application to the association model, and factors associated with MCI for the 51 types of intestinal bacteria mentioned above. Check marks indicate which intestinal bacteria are applied to the association model and which MCI-related factors they correspond to. MCI-related factors refer to factors or intestinal bacteria (intestinal microorganisms) that are suggested to be associated with MCI. In addition, intestinal bacteria marked with check marks in the "association model" column will be described later in FIG. 14.
[0046] 7 lists 10 MCI-related factors. Specifically, these are: 1) mucin degradation, 2) IgA protease production, 3) promotion of serotonin secretion from enterochromaffin cells (EC cells), 4) bile acid oxidation, 5) promotion of recovery of intestinal barrier function, 6) hydrogen production, 7) production of HDAC inhibitors, 8) protection of IgA from trypsin degradation, 9) intestinal bacteria prevalent in MCI (in men), and 10) intestinal bacteria prevalent in MCI (in women). In this embodiment, the MCI-related factors are classified into 10. However, the MCI-related factors are not limited to these, and may be one or more. Alternatively, other MCI-related factors may be included.
[0047] Among the gut bacteria listed in the "Healthy" effect size column, there are several that are not marked with a check mark for any of the MCI-related factors. These correspond to other MCI-related factors. For example, Prevotella corresponds to "gut bacteria that are low in MCI (for both men and women)" (see Figure 11 below). Furthermore, Flavonifractor, a gut bacteria listed in the "MCI" effect size column but not marked with a check mark for any of the MCI-related factors, corresponds to "gut bacteria that are high in MCI (for both men and women)" (see Figure 11 below). In Figure 7, because the population was female, the item "gut bacteria that are high in MCI (for men)" is not marked with a check mark. Furthermore, please note that although the names of gut bacteria are normally written in italics, they are written in regular typeface due to the character code available in the specification.
[0048] Figures 8 and 9 are figures corresponding to Figures 6 and 7 but for a male population. That is, Figure 8 is a figure for explaining the selection of gut microbial variables for a male population according to this embodiment. Figure 9 is a table showing, for each gut bacterium in Figure 8 , the effect size value, application to the association model described below, and factors related to MCI. In Figure 9 , because the population is male, the item "Gut bacteria common in MCI (for women)" is not marked. Furthermore, the item "Protection of IgA from trypsin degradation" is not marked because the corresponding gut bacterium does not appear in Figure 9 .
[0049] It is then confirmed whether or not all of the multiple intestinal bacteria estimated in S110 correspond to the MCI-related factors described in Fig. 7. In this embodiment, by performing such confirmation, it becomes possible to perform the point-addition process described later.
[0050] First, the MCI-related factors in FIG. 7 will be explained using FIG. 10. FIG. 10 is a hypothetical conceptual diagram of the influence of the intestinal microflora associated with MCI according to this embodiment. In the figure, an upward-pointing bold arrow indicates an increase. A downward-pointing bold arrow indicates a decrease. For example, an upward-pointing bold arrow is attached to "degradation of IgA." In this case, it means "promoting the degradation of IgA."
[0051] In this embodiment, a comparative study was conducted in advance to formulate the hypothesis shown in FIG. 10 . Specifically, the association between MCI and the gut microbiota was investigated by comparing an MCI group (11 men, 18 women) in their 70s with a control group (17 men, 23 women) taking gender into consideration. A separate literature search was conducted to examine known characteristics, such as hydrogen production, of the multiple gut bacteria associated with MCI identified through this comparative study. Some of the multiple gut bacteria have not been described in the literature as being associated with MCI. However, based on some known characteristics and this comparative study, it was hypothesized that the composition of the gut microbiota in the MCI group may cause dysregulation of the gut microbiota, increased intestinal barrier permeability, increased blood-brain barrier permeability, and increased chronic neuroinflammation, and that these abnormalities, if sustained over a long period of time, ultimately lead to cognitive decline.
[0052] In addition, the comparative study was conducted by gender, and if the absolute value of the effect size was 0.2 or higher, it was determined to be associated with the MCI group. Some of the gut bacteria in Figure 10 have "(male) or (female)" written next to their name. This means "(MCI-related factors common in men)" or "(MCI-related factors common in women)."
[0053] In this embodiment, in addition to an analysis of a mixed-gender group, a reanalysis was performed on gender-specific groups to attempt to understand the relationship between MCI and the intestinal microbiota. As a result, it became possible to explain, although hypothetically, part of the mechanism of MCI involving intestinal bacteria, as shown in FIG. 10 . By formulating such a hypothesis, this embodiment makes it possible to set MCI-related factors and construct a model for calculating the probability that a subject has or will have MCI. Then, in this embodiment, as described below, an association model and a score estimation model related to MCI are created, and further, a risk calculation model is created. This risk calculation model corresponds to a model for calculating the probability that a subject has or will have MCI.
[0054] Figure 10 is a conceptual diagram of a hypothesis regarding part of the mechanism of MCI involving intestinal bacteria. Below, hypotheses regarding the influence of the intestinal microbiota on MCI are explained in sections. (1) A decrease in the hydrogen-producing bacteria Roseburia, Megasphaera, Victivallis, Ruminococcus (males), and Agathobacter (females), and an increase in Eggerthella, which oxidizes bile acids in a hydrogen concentration-dependent manner, lead to dysregulation of the intestinal microbiota. (2) A decrease in the hydrogen-producing bacteria Roseburia, Megasphaera, Victivallis, Ruminococcus (males), and Agathobacter (females) indirectly causes an increase in intestinal pH due to reduced acetic acid production, leading to dysregulation of the intestinal microbiota. (3) An increase in Erysipelatoclostridium, which produces IgA protease, and a decrease in Paraprevotella (females), which protects IgA through the degradation of intestinal trypsin, lead to dysregulation of the intestinal flora and increased inflammation in intestinal epithelial cells. (4) An increase in Clostridium XVIII and Ruminococcus 2, which are involved in the degradation of the mucin layer, and a decrease in Roseburia (flagellin), which is involved in the protection of intestinal barrier function, lead to increased intestinal barrier permeability. (5) An increase in Erysipelatoclostridium, which promotes intestinal serotonin secretion, leads to increased permeability of the blood-brain barrier. (6) Although the intestinal bacteria that increase due to dysregulation of the intestinal flora differ between men and women, they contribute to inflammation. (7) A decrease in bacteria that produce HDAC (histone deacetylase) inhibitors (such as valeric acid-producing bacteria: Oscillibacter and Megasphaera, and propionic acid- and butyric acid-producing bacteria: Roseburia) causes epigenetic dysregulation due to excessive HDAC, leading to increased inflammation. (8) A decrease in hydrogen-producing bacteria: Roseburia, Megasphaera, Victivallis, Ruminococcus, and Agathobacter causes a decrease in the removal of reactive oxygen species from hydrogen, leading to increased chronic inflammation.(9) The composition of the gut microbiota in MCI patients leads to dysregulation of the gut microbiota, increased intestinal barrier permeability, increased blood-brain barrier permeability, and increased chronic neuroinflammation, and these abnormalities persist for a long period of time, ultimately leading to a decline in cognitive function.
[0055] In Figure 10, the "gut bacteria group prevalent in MCI" and "gut bacteria group rare in MCI" identified in the comparative study described above are further classified into three subgroups: male, female, and unrelated. Due to space limitations, this classification is shown enlarged in Figure 11. Furthermore, the underlined "item names" and "gut bacteria names" in Figures 10 and 11 are not relevant in the literature survey described above. For example, hydrogen is known to be involved in MCI. However, although the gut bacteria Megasphaera, Victivallis, and Agathobacter (female) are known to have hydrogen-producing ability, this study was the first to reveal that these gut bacteria are associated with MCI. Therefore, these gut bacteria are underlined.
[0056] In this embodiment, MCI-related factors can be set based on a hypothesis of the mechanism of MCI related to intestinal bacteria. More specifically, items in which the names of intestinal bacteria shown in Figures 10 and 11 are listed can be set as MCI-related factors.
[0057] (Point Addition Process 1: Point Addition Using Attribute Information) Figure 12 shows the relationship between age and occupancy rate for all female subjects for Bifidobacterium, one of the intestinal bacteria with a negative effect size in women in Figure 6. In Figure 12, Healthy is the average value for each age in the control group, and MCI is the average value for each age in the MCI group. At younger ages, there are no MCI patients. Therefore, for ages where data on MCI patients cannot be measured, an estimated value is used as the occupancy rate for that age (shown as MCI (estimated value) in Figure 12). The occupancy rate refers to the proportion of a certain intestinal bacterium in the intestinal microflora.
[0058] The onset of MCI is associated with the accumulation of brain damage. Furthermore, the intestinal microflora changes with age. As will be described later, in this embodiment, a scoring process may be performed using attribute information such as the subject's age.
[0059] As can be seen from Figure 12, the judgment of whether the occupancy rate of Bifidobacterium is appropriate differs depending on the age of the female subject. For example, assume that the subject is 60 years old. The average values for 60-year-olds in Figure 12 are Healthy 6.9% and MCI 3.2%. Therefore, if the occupancy rate of Bifidobacterium in the subject's intestinal flora is 6.9% or higher, no additional points are added. On the other hand, if the occupancy rate of the subject is 3.2% or lower, additional points are added to Bifidobacterium.
[0060] (Scoring Process 2: Adding Points to Intestinal Bacteria Corresponding to Factors Associated with MCI) In this embodiment, if a subject possesses all of the MCI-related factors among the multiple intestinal bacteria estimated in S110, a score may be added for the intestinal bacteria corresponding to the MCI-related factors among the multiple intestinal bacteria estimated in S110. FIG. 13 is a diagram for explaining the addition of points according to the bacteria possessed according to this embodiment. This diagram shows a case where the subject in the MCI group is a female. For this reason, "intestinal bacteria common in MCI (for men)" and the like are excluded from the MCI-related factors. This diagram shows that for subjects in the MCI group who have been assigned scores (a number different from 0, representing the occupancy rate or the number of sequences) for all MCI-related factors (i.e., when the subject's intestinal microbiota possesses all of the MCI-related factors), a score is added to the corresponding bacteria possessed. Note that in this embodiment, if a subject possesses all of the MCI-related factors, a score is added because it is believed that the subject's involvement in MCI is accelerated. However, a score may also be added if the subject possesses some of the MCI-related factors.
[0061] Returning to the process of S110, the step of creating an association model using the intestinal bacteria estimated in S115 will be described. In this embodiment, the association model is created using the CLR-transformed intestinal microbiota data of the NC and MCI groups and the MCI incidence status data.
[0062] FIG. 14 is a diagram showing the results of estimating the parameters of the association model according to this embodiment. The population is female. Here, a structural equation model (SEM) is used as one of the association models. A structural equation model represents the relationship between observed variables and latent variables, and the relationship between latent variables. Therefore, an association model is a statistical analysis model constructed from observed variables that are found to have a certain level of correlation with a specific disease and latent variables that are related to the observed variables.
[0063] Association model 1 is the area surrounded by a solid line. Rectangles represent observed variables, "e" represents the residual variance of the observed variables, ellipses represent latent variables, and the numbers on the arrows from latent variables to observed variables or other latent variables represent standardized parameters of the association model. For example, the observed variable MCI is an observed variable related to the disease MCI and is defined as a binary categorical variable representing MCI incidence status. If a subject has MCI, it is "1," and if not, it is "0."
[0064] When creating the association model, we assumed that there were two latent variables that explain the MCI status variable: Lv1 has a positive effect on MCI, and Lv2 has a negative effect.
[0065] First, all of the intestinal bacteria shown in Figure 6 are assigned as observed variables (also called intestinal microorganism variables) of latent variable Lv1 or Lv2. This is the initial association model. Starting from this initial association model, the model is modified. In modifying the model, observed variables are deleted until no negative components appear in each variance-covariance matrix calculated during the structural equation modeling process.
[0066] After that, observed variables are deleted so that the p-value (significance probability) of each parameter is less than 0.05, and from the constructed models, the one with GFI (Goodness of Fit Index) and AGFI (Adjusted GFI) values close to 1, RMSEA (Root Mean Square Error of Approximation) value close to 0, and the absolute value of the path coefficient from the latent variable to the disease incidence status variable being the largest is adopted as the final association model.
[0067] The created association model is fed with data from the combined population of NC and MCI groups (i.e., gut microbiota data and MCI incidence data), and each parameter in the association model is calculated. This allows us to examine how, when focusing on the NC and MCI groups, the gut microbiota, set as an observed variable, explains variables representing MCI incidence status via the assumed latent variables (Lv1, Lv2).
[0068] The standardized path coefficient from latent variable Lv1, which is hypothesized to have a positive effect on MCI, to the MCI prevalence variable is 0.52 (p<0.01), and the standardized path coefficient from latent variable Lv2, which is hypothesized to have a negative effect on MCI, to the MCI prevalence variable is -0.52 (p<0.01). The standardized factor loading from latent variable Lv1 to each observed variable (gut microbiome variable) of Lv1, for example, Ruthenibacterium, is 0.65. The standardized factor loading from latent variable Lv2 to each observed variable (gut microbiome variable) of Lv2, for example, Megamonas, is 0.85. All standardized factor loadings from the latent variables to each observed variable (gut microbiome variable) are significant at p<0.05. The value of -0.41 on the arrow between the latent variables represents the magnitude of the correlation between Lv1 and Lv2.
[0069] The residual variance of the MCI prevalence status variable was 0.24, and in this association model, the two hypothesized latent variables (Lv1, Lv2) explained approximately 76% of the variance of the MCI prevalence status variable.
[0070] In this embodiment, S115 is an example of a method for selecting observed variables for the association model, and the gut microbial variables are estimated using the effect size as an index. S115 is not an essential configuration.
[0071] The relationship between the observed variables related to the latent variables (Lv1, Lv2) in FIG. 14 and the intestinal bacteria marked with a check mark in the "Association Model" column in FIG. 7 will be described. The observed variables related to latent variable Lv1 in FIG. 14 are entered as values in the "MCI" effect size column in FIG. 7 and marked with a check mark in the "Association Model" column. Similarly, the observed variables related to latent variable Lv2 are entered as values in the "Healthy" effect size column and marked with a check mark in the "Association Model" column. In this way, it can be seen that the intestinal bacteria marked with a check mark in the "Association Model" column in FIG. 7 are applied to the association model of this embodiment.
[0072] There are various methods for selecting observed variables for the association model. For example, some statistical test (such as the Wilcoxon rank-sum test) can be performed between the control group and the MCI group, and the intestinal bacteria that show a significant difference in abundance between the two groups can be used as observed variables for the association model. Alternatively, intestinal bacteria that have already been reported as having the potential to be biomarkers for the disease of interest can be used as observed variables for the association model.
[0073] (Explanation of S120) S120 is a step of creating a model that estimates latent variable scores when the MCI prevalence status is unknown. When creating the association model, data from a combined population of NC and MCI groups (i.e., gut microbiota data and MCI prevalence status data) is used. However, the subjects are those who wish to be assessed for MCI risk, and their disease prevalence status is unknown. Therefore, a model that estimates latent variable scores when the MCI prevalence status is unknown is needed.
[0074] In order to estimate the scores of Level 1 and Level 2 when the MCI status of a female subject is unknown, a portion of the measurement equation model (dotted line in FIG. 14 ) is extracted from the association model 1 to create a score estimation model 2. For this reason, the same values as the values of each parameter in the measurement equation model are set as the values of each parameter in the score estimation model 2. Here, the measurement equation model is a model that shows the influence of the latent variables on the observed variables of intestinal bacteria (intestinal microorganism variables). The score estimation model 2 of this embodiment is a linear multiple regression model that uses the observed variables of intestinal bacteria (intestinal microorganism variables) as explanatory variables and the scores of the latent variables (Lv1 and Lv2) as response variables.
[0075] In this embodiment, the subject's attribute information and / or information on intestinal bacteria corresponding to factors related to MCI is used to add points to the score estimation model 2, and then the subject's Lv1 and Lv2 scores are estimated. Hereinafter, the estimated values of the latent variables Lv1 and Lv2 are represented by Lv1est and Lv2est, respectively.
[0076] (Explanation of S200) S200 is a step in which the subject inputs the disease for which the subject wishes to be evaluated. In this embodiment, a score estimation model for MCI is created separately for each gender. Therefore, the subject also inputs their gender. The subject is female.
[0077] (Explanation of S210) S210 is a step of extracting the subject's intestinal microbiota data and attribute information. The subject submits a stool sample, and the subject's intestinal microbiota data is stored in the intestinal microbiota DB 300. S210 extracts the subject's intestinal microbiota data from the intestinal microbiota DB 300 using the subject's ID information. Furthermore, the subject's attribute information (such as age and the area where the subject lived during childhood) is extracted from the questionnaire response data submitted by the subject.
[0078] (Explanation of S220) S220 is a step of adding points to the score estimation model using attribute information of the subject and / or information on intestinal bacteria corresponding to factors related to MCI. The attribute information of the subject is input to the disease evaluation index calculation device 900.
[0079] The point-adding unit 910 calculates a composite variable x (x1 = attribute, x2 = living area, ...) from the subject's attribute information. Next, the composite variable x is input to a predetermined point-adding function y = f(x), and the value y output by the point-adding function is added to each intestinal microbial variable that is related to Level 1 and has a check mark in the "MCI-related factor" column in Figure 7 (for example, Erysipelatoclostridium in Figure 14).
[0080] 15 is a diagram illustrating a score estimation model and a risk calculation model after adding points according to this embodiment. Each gut microbiome variable (e.g., Erysipelatoclostridium') with an apostrophe in FIG. 15 is a gut microbiome variable (observed variable) after adding points. For example, Erysipelatoclostridium' = Erysipelatoclostridium + y.
[0081] Among the variables related to Level 1, Flavonifractor is not marked in the "MCI-related Factor" column in Figure 7, so no points are added and no apostrophe is added. In this embodiment, points are added to some of the variables related to Level 1, but depending on the settings of the MCI-related factors, points may be added to all variables related to Level 1. Furthermore, the MCI-related factors may be set so that points are added to variables related to all latent variables (Level 1 and Level 2).
[0082] The area surrounded by a dotted line in Figure 15 is a score estimation model 2' after adding points. Also, the area surrounded by a dashed line in Figure 15 is a risk calculation model 3', which will be described later. Note that in the overall part 1' in Figure 15, the variable values of each intestinal microbial variable in the association model in Figure 14 have been changed. For example, the variable value of the intestinal microbial variable Erysipelatoclostridium has been changed to "Erysipelatoclostridium + y".
[0083] (Explanation of S230) S230 is a step of estimating scores for latent variables using a score estimation model 2' after adding points for MCI. Specifically, second data related to factors or intestinal microorganisms suggested to be associated with MCI is extracted from first data related to the subject's intestinal microbiome. The second data related to factors or intestinal microorganisms suggested to be associated with MCI is input into each observed variable (e.g., Erysipelatoclostridium') of the score estimation model 2', and the subject's scores for the latent variables Lv1 and Lv2 of the score estimation model 2' are estimated. The estimated values of the latent variables Lv1 and Lv2 are set as Lv1est and Lv2est. In other words, S230 inputs the second data into a model (score estimation model 2') that uses the added-point intestinal microorganisms as a variable and calculates the probability that the subject has a disease, and estimates the subject's scores for the latent variables Lv1 and Lv2 of the score estimation model 2'. Here, the intestinal microorganisms to which points have been added may be referred to as the second data after the addition of points. Furthermore, the first data corresponds to data relating to the intestinal microbiota of the subject. The second data corresponds to, for example, the intestinal microorganisms of the subject with a check mark in the "association model" column in FIG. 7 . The third data corresponds to the intestinal microorganisms of the subject with a check mark in the "association model" column. That is, the third data corresponds to one or more intestinal microorganisms that show differences between the intestinal microbiota of a healthy person and the intestinal microbiota of a person suffering from mild cognitive impairment.
[0084] (Explanation of S240) S240 is a step of calculating an evaluation index for the risk of MCI. Specifically, the risk of MCI for the subject is calculated using the path coefficients from the latent variables Lv1 and Lv2 to the observed variable MCI in the risk calculation model 3′ in FIG. 15 and the estimated values (Lv1est, Lv2est) of the latent variables Lv1 and Lv2. That is, S240 calculates the risk of MCI for the subject in the risk calculation model 3′ based on the probability calculated by the score estimation model 2′ (the subject's scores for the latent variables Lv1 and Lv2 in the score estimation model 2′).
[0085] In this embodiment, the method for constructing the MCI probability estimation model uses a logistic regression model in which the latent variables Lv1 and Lv2 are used as explanatory variables and a binary categorical variable representing the subject's MCI incidence status is used as the objective variable. This logistic regression model corresponds to risk calculation model 3' (dashed line in FIG. 15).
[0086] The disease evaluation index calculation device 900 outputs an evaluation report to the subject based on the calculated risk of MCI.
[0087] The above is the flowchart of the disease evaluation index calculation process in Figure 2. When the population is female, in association model 1 related to MCI, the correlation between the factors of latent variables Lv1 and Lv2 in Figure 14 is -0.41. On the other hand, when the population is male, it is assumed that there is a correlation between Lv1 and Lv2 and there is one latent variable. Below, the case where there is one latent variable will be described as a modified example.
[0088] (Modification: Case where there is one latent variable) FIG. 16 is a diagram showing the results of estimating the parameters of association model 1 according to a modification. The population is male. FIG. 17 is a diagram for explaining score estimation model 2' after adding points and risk calculation model 3' according to a modification. The modification shows the case where there is one latent variable. All of the variables related to Level 1 in FIG. 17 have a check mark in one of the items in the "MCI-related factors" column in FIG. 9, so all have been added points and an apostrophe has been added.
[0089] Even when there is one latent variable, S110 (creating an association model), S120 (creating a score estimation model), S220 (adding points using the subject's attributes), S230 (estimating the score of the latent variable), and S240 (calculating the risk) in Figure 2 can be processed as described above.
[0090] (Risk Calculation Example) Calculation of the risk of MCI will be described separately for this embodiment (the population is female) and a modified example (the population is male). Figure 18 is a diagram showing the results of an ROC analysis of the risk value of MCI according to this embodiment (the population is female). The risk calculation according to this embodiment (the population is female) and the accuracy of the calculated risk will be described.
[0091] For a group of female subjects, the combined NC and MCI group was stratified and randomly divided into an 80% training group and a 20% validation group. An association model was trained using the latent variable scores (Level 1 and Level 2) of the training group and information on MCI incidence status, and an MCI incidence probability estimation model was constructed to estimate the MCI incidence probability from the latent variable scores. The latent variable score estimates (Level 1est and Level 2est) of the validation group were applied to the constructed MCI incidence probability estimation model, and the accuracy of the MCI incidence probability estimation model was examined using receiver operating characteristic (ROC) analysis. The MCI incidence probability is the risk of developing MCI, and the results of the ROC analysis can be used as a disease evaluation index.
[0092] The method for constructing the MCI probability estimation model uses a logistic regression model in which the latent variables Lv1 and Lv2 are used as explanatory variables and a binary categorical variable representing the subject's MCI incidence status is used as the objective variable. This logistic regression model corresponds to risk calculation model 3' in Figure 15.
[0093] FIG. 18 shows the ROC curve for this logistic regression model. The vertical axis represents "Sensitivity." As the probability of determining that a person who actually has MCI has the disease increases, the value on the vertical axis approaches 1. The horizontal axis represents "1-Specificity." As the probability of determining that a person who does not have MCI does not have the disease increases, the specificity value approaches 1. Since the horizontal axis subtracts the specificity from 1, as the value on the horizontal axis approaches 0, the probability of determining that a person who does not have MCI does not have the disease increases. The solid line represents the results of a model using Level 1 and Level 2 as explanatory variables (this embodiment described in FIGS. 14 and 15 (the population is female)). AUC represents the Area Under the Curve value, which is 0.87 in this embodiment (the population is female).
[0094] In Figure 18, the diagonal line (dotted line) represents a probability of 0.5, and in the case of a model (solid line) with Lv1 and Lv2 as explanatory variables, the area enclosed by the solid and dotted lines is large, allowing one to visually (or intuitively) determine that the risk estimation method has high accuracy. Furthermore, AUC is an index of a risk estimation method, and an AUC value close to 1 can be objectively determined to be a good risk estimation method. This embodiment can be determined to have high accuracy as an indirect risk estimation method, excluding direct risk estimation methods such as blood tests.
[0095] 19 is a diagram showing the results of an ROC analysis of the risk value of MCI for the modified example (the population is male). The AUC is 0.75. Therefore, similar to the present embodiment (the population is female), the modified example can be determined to have high accuracy as an indirect risk estimation method, excluding direct risk estimation methods such as blood tests.
[0096] (Effects) According to this embodiment, an association model for each disease (e.g., association model 1 in FIG. 14 ) is created using gut microbiota data and disease incidence data of a large number of subjects, and a score estimation model (e.g., score estimation model 2 in FIG. 14 ) is created in advance from the measurement equation portion of the association model. For subjects who wish to be assessed for risk of MCI as a specific disease, individual processing can be performed according to the stool sample and questionnaire response data that the subject submits in the same way as a large number of subjects.
[0097] Using a stool sample submitted by the subject, gut microbiota data of the subject is extracted. Furthermore, using questionnaire response data submitted by the subject, attribute information of the subject (including the subject's background at birth and childhood background) is extracted. Using the subject's attribute information, points are added to a score estimation model for MCI (e.g., score estimation model 2 in FIG. 14 ). By inputting the subject's gut microbiota data into the observed variables of a score estimation model after point addition (e.g., score estimation model 2' after point addition in FIG. 15 ), the subject's score for the latent variable can be estimated. The estimated subject's score and a risk calculation model (e.g., risk calculation model 3' in FIG. 15 ) can be used to calculate the subject's risk of MCI.
[0098] The subject can then easily and non-invasively receive an assessment report on the subject's own risk of developing or having developed MCI, and the business providing the assessment report can design, consider, and propose prebiotics to the subject.
[0099] For example, by including in the subject's assessment report the subject's risk of MCI as well as prescriptions for functional foods or other products based on the subject's gut microbiome data, assessment companies can provide subjects with solutions for improving cognitive function.
[0100] (Solution for Improving Cognitive Function) In addition to the risk of MCI, the assessment report for the subject may also introduce functional foods that are expected to improve the cognitive function of the subject. In this embodiment, the cognitive function improving foods are introduced according to the gut microbiota data of the subject.
[0101] In this embodiment, from among these cognitive function improving foods, functional foods (cognitive function improving foods) that increase the intestinal microbiota that contributes to improving MCI are identified based on the subject's intestinal microbiota data, and a report on how to take the identified functional food is provided (or prescribed) to the subject. The subject regularly takes the identified functional food according to the intake method. Then, by again using the cognitive function index calculation system of this embodiment as a follow-up observation several months later, the cognitive function improving solution can be used. The cognitive function improving solution can provide the subject with an evaluation report of before and after taking the functional food.
[0102] For example, the observed variable MCI in association model 1 in Figure 14 is an observed variable related to the disease MCI and is defined as a binary categorical variable representing the MCI incidence status, but is not limited to this. The observed variable MCI may be set to a continuous value from 0 to 100, and an intervention test may be conducted to confirm whether MCI increases or decreases when a subject ingests a functional food that affects gut microbial variables that may be factors (risk factors or preventive factors) for MCI.
[0103] Although the embodiments (including modifications) of the present invention have been described above, two or more of these embodiments may be combined and implemented. Alternatively, one of these embodiments may be partially implemented. Furthermore, two or more of these embodiments may be partially combined and implemented.
[0104] Furthermore, the present invention is not limited to the above-described embodiments of the invention, and various modifications within the scope of the claims and within the scope that can be easily conceived by a person skilled in the art are also included in the present invention.
[0105] For example, although it is assumed that there are one or two latent variables, there may be three or more. Furthermore, in this embodiment, ID information of subjects related to a disease is extracted from the questionnaire DB 400 for each attribute of gender to create an association model, but this is not limited to this. A mixed gender model may also be created without extracting each attribute of gender.
[0106] In addition, in S230, when estimating the score of the latent variable, the score estimation model after the point addition is used, but this is not limited to this. The score estimation model before the point addition may be used without the point addition process.
[0107] Additionally, gut microbes that have been suggested to be associated with MCI may be:Acidaminococcus, Adlercreutzia, Agathobacter, Akkermansia, Anaeromassilibacillus, Anaerostipes, Anaerotignum, Anaerotruncus, Bacteroides, Bifidobacterium, Blautia, Catenibacterium, Clostridium IV, Clostridium sensu stricto, Clostridium XIVa, Clostridium XIVb, Clostridium XVIII, Coprobacillus, Coprobacter, Coprococcus, Dialister, Dysosmobacter, Eggerthella, Eisenbergella, Enterocloster, Enterococcus, Erysipelatoclostridium, Faecalibacterium, Flavonifractor, Fournierella, Frisingicoccus, Fusobacterium, Gordonibacter, Holdemanella, Holdemania, Hungatella, Ihubacter, Intestinibacter, Intestinimonas, Lactococcus, Lawsonibacter, Leuconostoc, Ligilactobacillus, Longicatena, Massilimicrobiota, Megamonas, Megasphaera, Merdimonas, Negativibacillus, Neglecta, Oscillibacter, Parabacteroides, Paraprevotella, Parasutterella, Prevotella, Rombooutsia, Roseburia, Rothia, Ruminococcus, Ruminococcus 2, Ruthenibacterium, Sellimonas, Slackia, Streptococcus, Sutterella, Turicibacter, Unclassified (DNA sequence without taxonomic name), Veillonella, Victivallis.
[0108] While this embodiment uses an association model, the present invention is not limited to this and may be performed by machine learning. When using an association model, the disease evaluation index calculation device according to this embodiment includes an input unit for inputting one or more diseases that a subject wishes to be examined for, a first database for storing data on the intestinal microbiota resulting from analysis of stool samples from a plurality of subjects, a second database for storing questionnaire response data on the diseases from the subjects, an extraction unit for extracting first intestinal microbiota data of healthy individuals and second intestinal microbiota data of individuals suffering from the diseases using predetermined extraction conditions, and an extraction unit for inputting the first intestinal microbiota data, the second intestinal microbiota data, and the questionnaire response data, and extracting a first observation variable related to a plurality of intestinal microorganisms that indicates a difference between the first intestinal microbiota data and the second intestinal microbiota data, a second observation variable that indicates the disease incidence status, and one or more observation variables related to the first observation variable and the second observation variable. and a first creation unit that creates an association model composed of a latent variable of the first observation variable and a parameter between the first and second observation variables and / or a parameter between the plurality of latent variables from the one or more latent variables, a second creation unit that creates a score estimation model in which the first observation variable is an explanatory variable and the score of the latent variable is a target variable, a score adding unit that inputs attribute information of a subject who wishes to be evaluated for the disease and adds points to some of the explanatory variables of the score estimation model using the attribute information of the subject, an estimation unit that inputs gut microbiota data of the subject to the explanatory variables of the added-point score estimation model and estimates the score of the latent variable, and a calculation unit that calculates the subject's risk of the disease using the parameter from the one or more latent variables to the second observation variable and / or the parameter between the plurality of latent variables and the estimated score. Of these configurations, the first creation unit, the second creation unit, the score adding unit, the estimation unit, and the calculation unit may be performed by machine learning.
[0109] REFERENCE SIGNS LIST 100 Intestinal microbial DNA extraction device 200 Intestinal microbiota analysis device 300 Intestinal microbiota DB 400 Questionnaire DB 500 Association model creation device 600 Association model DB 700 Score estimation model creation device 800 Score estimation model DB 900 Disease evaluation index calculation device 910 Point addition unit
Claims
1. A risk index calculation device for mild cognitive impairment comprising: an extraction unit that extracts second data on intestinal microorganisms or factors suggested to be associated with mild cognitive impairment from first data on the subject's intestinal microbiota; and a calculation unit that inputs the second data into a model that calculates the probability that the subject has mild cognitive impairment using third data on one or more intestinal microorganisms that show a difference between the intestinal microbiota of a healthy person and the intestinal microbiota of a person suffering from mild cognitive impairment as a variable, and calculates the subject's risk of mild cognitive impairment, wherein the risk of mild cognitive impairment is the probability of having or will have mild cognitive impairment.
2. A mild cognitive impairment risk index calculation device as described in claim 1, wherein the second data is classified into a gut microbial group that is prevalent in the mild cognitive impairment and a gut microbial group that is prevalent in the mild cognitive impairment.
3. A mild cognitive impairment risk index calculation device as described in claim 2, wherein the intestinal microbial group that is prevalent in mild cognitive impairment and the intestinal microbial group that is low in mild cognitive impairment are each classified into three subgroups: male, female, and common to both sexes.
4. A mild cognitive impairment risk index calculation device as described in claim 1, wherein the second data relates to mucin degradation, IgA protease production, serotonin secretion from enterochromaffin cells, oxidation of bile acids, promotion of recovery of intestinal barrier function, production of HDAC inhibitors, or protection of IgA from trypsin degradation.
5. A mild cognitive impairment risk index calculation device as described in claim 1, wherein the extraction unit also extracts attribute information of the subject, and the calculation unit scores the second data using the attribute information of the subject and / or information on intestinal microorganisms corresponding to factors related to the mild cognitive impairment, and inputs the second data after scoring into the model.
6. A mild cognitive impairment risk index calculation device as described in claim 5, wherein the subject's attribute information includes background information of the subject at the time of birth and background information of the subject's life during childhood.
7. A mild cognitive impairment risk index calculation device as described in claim 1, wherein the model is a model created for each gender attribute, or a mixed gender model.
8. A method for calculating a risk index of mild cognitive impairment, comprising: a computer extracting, from first data regarding the subject's gut microbiota, second data regarding gut microbes or factors suggested to be associated with mild cognitive impairment; inputting the second data into a model that calculates the probability that the subject has mild cognitive impairment using third data regarding one or more gut microbes that show a difference between the gut microbiota of a healthy person and the gut microbiota of a person suffering from mild cognitive impairment as a variable; and calculating the subject's risk of mild cognitive impairment, wherein the risk of mild cognitive impairment is the probability of having or will have mild cognitive impairment.
9. A program for calculating a risk index for mild cognitive impairment, comprising: an extraction step of extracting, from first data on the subject's intestinal microbiota, second data on intestinal microbes or factors suggested to be associated with mild cognitive impairment; and a calculation step of inputting the second data into a model that calculates the probability that the subject has mild cognitive impairment, using third data on one or more intestinal microbes that show a difference between the intestinal microbiota of a healthy person and the intestinal microbiota of a person suffering from mild cognitive impairment as a variable, and calculating the subject's risk for mild cognitive impairment, wherein the risk for mild cognitive impairment is the probability of having or will have mild cognitive impairment.
Citation Information
Patent Citations
Food composition for cognitive function improvement, cognitive function improver, and method for producing the same
JP2019180364A
Gut microbiota-related methods for treating dementia and age-dependent cognitive decline
JP2023508968A
Disease evaluation index calculation system, method, and program
JP7270143B1
Composition for preventing or treating deterioration in brain function or maintaining or improving brain function
WO2022118944A1
