JP2016214109A - Estimation method of intestine age, and estimation device of intestine age - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an estimation method of an intestine age and an evaluation device of the intestine age which can estimate the intestine age.SOLUTION: An estimation method of an intestine age comprises an estimation step for estimating an intestine age of an object person by use of an artificial neural network. The neural network is constructed by using an intestinal bacterial flora database in which following items are associated with each other: an identifier which identifies each of a plurality of subjects having different ages; a value representing respective ratios of all intestinal bacteria extracted from each of the subjects to a plurality of intestinal bacteria in all kinds of bacterium classification levels extracted from the subjects; and a value representing an age of each of the subjects.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an intestinal age estimation method and an intestinal age estimation apparatus.

  In recent years, attention has been focused on intestinal age as interest in health has increased. Intestinal age is the age (or age) determined based on the state of the intestine. A large number of bacteria exist in the intestinal tract of humans and animals to form an intestinal flora. The types of bacteria that make up the intestinal flora vary depending on physiology, disease, drugs, food, lifestyle, stress, aging, interactions of enterobacteria, and the like.

  In order to investigate such an intestinal condition, a phylogenetic microarray (phylogenetic microarray) was used for samples of faeces of a young group of 30 years old, an elderly group of 70 years old, and an elderly group of 100 years or older. Main factor analysis (PCA) is performed using Human Intestinal Tract Chip (HITChip) to examine 16S rRNA gene amplicon, and the intestinal flora is analyzed. And the composition distribution of the elderly group of 100 years old or more and the intestinal microflora of the offspring are compared. Moreover, the analysis regarding the relationship of a composition is performed about a test subject's inflammatory state and intestinal microflora (for example, refer nonpatent literature 1).

  In addition, the base sequence of bacteria contained in the feces of young and elderly persons is analyzed, and the composition distribution is analyzed for the intestinal flora of young and elderly persons (for example, non- Patent Document 2).

“Through Ageing, and Beyond: Gut Microbiota and Inflammatory Status in Seniors and Centenarians”, Elena Biagi, Lotta Nylund, etc., PLoS ONE, May 2010 Volume 5 Isseu 5 e10667,2010 “Composition, variability, and temporal stability of the intestinal microbiota of the elderly”, Marcus J. Claesson, Siobhan Cusack, et al., PNAS, www.pnas.org/cgi/doi/10.1073/pnas.1000097107,2010

  However, with the techniques described in Non-Patent Document 1 and Non-Patent Document 2, the intestine age of the subject could not be estimated.

  The present invention has been made in view of the above points, and an object of the present invention is to provide an intestinal age estimation method and an intestinal age estimation apparatus that can estimate the intestinal age.

(1) In order to achieve the above object, an intestinal age estimation method according to an aspect of the present invention includes an estimation step of estimating an intestinal age of a subject using an artificial neural network. The artificial neural network includes an identifier for identifying each of a plurality of subjects having different ages, and a plurality of intestinal bacteria at a bacterial classification level including all types extracted from the plurality of subjects. For each of the plurality of subjects, a value representing a ratio to all intestinal bacteria at the bacterial classification level extracted for each of the plurality of subjects is associated with a value representing the age of each of the plurality of subjects. An intestinal age estimation method constructed using an intestinal microbiota database.

(2) In the intestinal age estimation method according to one aspect of the present invention, in the estimation step, for each of a plurality of intestinal bacteria having a plurality of bacterial classification levels including all types extracted from the plurality of subjects. The intestinal age of the subject may be estimated by inputting a value representing the ratio of all the intestinal bacteria extracted from the subject to the artificial neural network.

(3) In the intestinal age estimation method according to one aspect of the present invention, a classification step of separating the ages of the plurality of subjects included in the intestinal microbiota database into a plurality of age groups, and the classification For each of the plurality of age groups classified by a process, an average value calculating step for calculating an average value of the values representing the ratio for each of the intestinal bacteria, and the average value calculated by the average value calculating step A difference between the ratio of each of the subjects belonging to the age group is calculated for each of the enteric bacteria, and a total calculation step of calculating a sum of the calculated differences for each of the subjects; From a plurality of subjects belonging to the age group, a predetermined number of subjects are selected from each of the age groups in ascending order of the sum calculated in the sum calculation step. Using the selection step, the intestinal bacterial information and age of the subject selected by the selection step, a reconstruction step for reconstructing the intestinal flora database, and the intestines reconstructed by the reconstruction step The artificial neural network may be constructed using an internal bacterial flora database.

(4) In the intestinal age estimation method according to one aspect of the present invention, the number of inputs of the artificial neural network includes a plurality of intestines at a bacterial classification level including all types extracted from the plurality of subjects. It is the total number of types of internal bacteria, and the number of outputs of the artificial neural network may be the total number of the plurality of age groups classified by the classification step.

(5) Moreover, in the intestinal age estimation method according to an aspect of the present invention, in the classification step, the plurality of subjects are separated every 10 years, and in the estimation step, the intestinal age of the subject is determined. It may be estimated every 10 years.
(6) Moreover, in the estimation method of the intestinal age which concerns on 1 aspect of this invention, the said intestinal microflora database is the said of the enteric bacteria extracted from the test subjects 0 to 104 years old including before and during weaning. You may make it build based on the value which shows a ratio.

(7) In the intestinal age estimation method according to one aspect of the present invention, the artificial neural network may be a feedforward neural network having a hidden layer.

(8) Moreover, in the intestinal age estimation method according to one aspect of the present invention, the intestinal microflora database includes the intestinal bacteria extracted from the subject with respect to all the intestinal bacteria extracted from the subject. The acquisition step of acquiring each ratio for each subject, and the total number m of the subjects (m is an integer of 2 or more), using the ratio of each intestinal bacterium for each subject acquired in the acquisition step, An extraction step of extracting a total number n of types of a plurality of enteric bacteria (n is an integer of 2 or more) at a bacterial classification level including all types extracted from the plurality of subjects; The total number m of the subjects, the total number n of the types of the plurality of enteric bacteria, and a value representing the ratio of all the intestinal bacteria in the bacterial classification level extracted for each of the plurality of subjects. And A configuration data construction step of constructing configuration data of rows × n rows, information indicating the age of each of the plurality of subjects is acquired, and information indicating the acquired age for each subject is configured in the configuration data configuration step The database may be constructed by a database construction step of constructing an intestinal microbiota database in association with the configured data.

(9) In order to achieve the above object, an intestinal age estimation apparatus according to an aspect of the present invention includes an identifier for identifying each of a plurality of subjects having different ages, and all types extracted from the plurality of subjects. For each of a plurality of enteric bacteria at a bacterial classification level comprising: a value representing a ratio to all enteric bacteria at a bacterial classification level extracted for each of the plurality of subjects; and an age of each of the plurality of subjects. Using a database storing an intestinal microflora database associated with a value to represent, and an artificial neural network constructed using the intestinal microflora database stored in the database, An estimation unit that estimates the intestinal age of the subject.

According to (1) and (9), the intestinal age of the subject can be estimated by the artificial neural network constructed using the intestinal flora database. In addition, each classification level of bacteria is a classification level such as a gate, a class, an eye, a family, a genus, and a species. The test subject is a person who provided a sample to construct a database of intestinal microflora. The target person is a person whose intestinal age is estimated. The subject may be the subject.
According to (2), the intestinal age of the subject can be estimated based on the state of the intestinal bacteria of the subject.

According to (3), in the artificial neural network constructed to estimate the intestinal age, the intestinal flora of the subject that becomes noise when estimating the intestinal age can be excluded. As a result, according to the present embodiment, the intestinal age of the subject can be accurately estimated.
According to (4), the probability of belonging to each age group can be calculated by inputting the intestinal flora information detected from the subject to the artificial neural network.

According to (5), it is possible to accurately estimate the intestinal age by estimating which of the age groups for every 10 years the subject's intestinal bacterial composition belongs.
According to (6), the intestinal age of the subject can be accurately determined using an artificial neural network based on the intestinal microbiota database by a plurality of subjects in each age group from before weaning to 104 years old. Can do.

According to (7), it is possible to accurately estimate the intestinal age other than the actual age.
According to (8), an intestinal microbiota database can be constructed using intestinal bacteria extracted from subjects other than the subject or the subject. As a result, in this embodiment, the age to which the subject belongs by inputting the intestinal flora information of the subject to the artificial neural network constructed using the intestinal flora database constructed in this way. Groups can be estimated with high accuracy.

It is a block diagram of an intestinal age estimation device according to an embodiment. It is a flowchart of the process which the intestinal age estimation apparatus which concerns on embodiment performs. It is a figure showing an example of the construction procedure of the intestinal microflora database which concerns on embodiment. It is a figure showing an example of the construction procedure of the neural network which concerns on embodiment. It is a figure showing an example of composition of a neural network concerning an embodiment. It is a flowchart of the estimation process of the intestinal age which concerns on embodiment. It is a flowchart of the intestinal bacteria extraction process which the intestinal bacteria extraction apparatus which concerns on embodiment performs. It is a figure showing an example of a structure of the neural network constructed | assembled by the 1st test example.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<Configuration of Intestinal Age Estimation Device 1>
First, the configuration of the intestinal age estimation apparatus 1 will be described.
FIG. 1 is a block diagram of an intestinal age estimation apparatus 1 according to this embodiment. As shown in FIG. 1, the intestinal age estimation device 1 includes an intestinal bacteria extraction device 11, a database generation unit 12, a database 13, a neural network construction unit 14, and an estimation unit 15.

  The intestinal bacterium extraction apparatus 11 extracts intestinal bacteria (also referred to as enteric bacteria) from samples of subjects and subjects. Here, the test subject is a person who provided a sample to construct a database of gut microbiota. The target person is a person whose intestinal age is estimated. In addition, a subject may be a subject after database construction. The extracted extraction result includes an identifier for identifying the subject or the subject, the extracted bacteria name, and information indicating the composition ratio of each bacteria. The intestinal bacteria extraction apparatus 11 outputs the extracted result of the extracted subject or subject to the database generation unit 12 and the estimation unit 15. The method for extracting enteric bacteria will be described later. Moreover, the intestinal bacteria extraction apparatus 11 may be comprised by several apparatuses, several applications, etc. so that it may mention later.

  The database generation unit 12 constructs an intestinal flora database by associating the age (attribute) of the input subject or subject with the extraction result output from the intestinal bacteria extraction apparatus 11, and constructs the intestinal flora The database is stored in the database 13. A database construction method will be described later. In addition, after constructing the intestinal flora database, the database generation unit 12 extracted intestinal bacteria other than those included in the intestinal flora database while extracting the intestinal flora of the subject. In this case, the types of bacteria constituting the intestinal flora may be increased. The database generating unit 12 may store the intestinal flora information of the subject in the database 13 when estimating the intestinal age.

  As shown in Table 1, the database 13 stores an intestinal microflora database in which information indicating age (hereinafter referred to as age information) and a constituent ratio of the intestinal microflora are associated with the identifier of the subject. ing. In addition, the composition ratio of the intestinal microflora is extracted from each subject in all types of bacteria n (n is an integer of 2 or more) extracted from m (m is an integer of 2 or more) subjects. It is the share of the number of intestinal bacteria. For this reason, the ratio with respect to the enteric bacteria extracted only from the other test subject is 0%.

  In addition, the classification level of bacteria in Table 1 is at least one level among gates, classes, eyes, families, genera, species, and the like. For example, as shown in Table 1, when the identifier is “1”, the age (attribute) is “weaning”, the composition ratio of the first bacteria is “0.057 (%)”, and the composition ratio of the second bacteria is “1.23 (%)”,..., “0 (%)” is associated with the composition ratio of the nth bacteria. The age of the subject stored in the database 13 is information on multiple intestinal flora (hereinafter referred to as intestinal flora information) for each age group, for example, during weaning (0 years old) to 104 years old. Has been. The database 13 may store an intestinal flora database based on intestinal flora information for each bacterial classification level.

  The database 13 stores information on age groups classified by the neural network construction unit 14 as shown in Table 2. Here, the age group is one of a set of predetermined ages such as 0 to 10 years old, 10 to 19 years old, ..., 90 to 99 years old, 100 years old or more. For example, the age group to which a 42-year-old subject belongs is 40-49 years old.

The neural network construction unit 14 reads intestinal flora information for m persons stored in the database 13 and age information associated with the intestinal flora information. As shown in Table 2, the neural network construction unit 14 classifies the read age information into a plurality of age groups, and stores information on the classified age groups in the database 13.
Moreover, the neural network construction unit 14 selects, for example, 10 subjects from among the groups belonging to each age group based on the average value of the composition ratios of the bacterial classification levels in the intestinal bacteria in each group. . The selection method will be described later. The neural network construction unit 14 optimizes the model of intestinal microbiota information of the subjects of each selected group, constructs a neural network using the values calculated by the optimization, and constructs the constructed neural network Is stored in the database 13. A method for constructing the neural network will be described later.

  The neural network construction unit 14 constructs the neural network from the intestinal microbiota information stored in the database 13 when the total number of subjects stored in the database 13 exceeds a predetermined value after constructing the neural network. The neural network may be reconstructed by reselecting the target person for each age group. For example, when the intestinal flora information of 386 subjects is stored in the database 13 and the intestinal flora information of the subject is sequentially added by the database generation unit 12 for 133 persons, the neural network construction unit 14 The neural network may be reconstructed.

  The estimation unit 15 inputs the intestinal microbiota information of the subject output from the intestinal bacteria extraction device 11 to the neural network stored in the database 13, and the intestinal age of the subject is in any age group. Estimate whether the probability of belonging to a group is high. A method for estimating the intestinal age will be described later. The estimation unit 15 outputs the estimated determination result to a printer, a display unit (not shown), or the like.

  Note that the intestinal age estimation apparatus 1 or the external apparatus may generate health advice, meal advice, or the like to the subject based on the determination result. In this case, the intestinal age estimation apparatus 1 or the external apparatus may be connected to a database in which advice relating to health, advice relating to meals, and the like are associated with each age estimated as the actual age.

<Outline of processing procedure performed by intestinal age estimation apparatus 1>
Next, an outline of a processing procedure performed by the intestinal age estimation apparatus 1 will be described.
FIG. 2 is a flowchart of processing performed by the intestinal age estimation apparatus 1 according to this embodiment.
(Step S1) The intestinal bacterium extraction apparatus 11 performs an intestinal bacterium extraction process from each of samples of subjects of a plurality of age groups.
(Step S <b> 2) The database generation unit 12 performs an intestinal microbiota database construction process by associating the extraction result extracted by the intestinal bacteria extraction device 11 with the input subject attributes.

(Step S3) The neural network construction unit 14 reads out age information for m people from the database 13, and generates two or more age groups from the read age information for m people.
(Step S4) The neural network construction unit 14 reads intestinal flora information for m persons from the database 13, and uses the read intestinal flora information for m persons to estimate the intestinal age. Build a network. Subsequently, the neural network construction unit 14 stores information on the constructed neural network in the database 13.

(Step S5) The intestinal bacteria extraction apparatus 11 performs an intestinal bacteria extraction process for the subject.
(Step S <b> 6) The estimation unit 15 calculates the intestinal age of the subject by inputting the intestinal flora information of the subject output from the intestinal bacteria extraction device 11 into the neural network read out from the database 13 and calculating it. The estimation process is performed.
The intestinal age estimation apparatus 1 repeats the processes of steps S5 to S6 for each subject.

<Construction of intestinal flora database>
Next, an example of a method for constructing the intestinal flora database performed by the database generation unit 12 in step S2 of FIG. 2 will be described.
FIG. 3 is a diagram illustrating an example of a procedure for constructing the intestinal flora database according to the present embodiment.
(Step S11) The database generation unit 12 acquires the composition ratios of the gut microbiota of all subjects. Table h101 in FIG. 3 is a constituent ratio of the intestinal microflora of the first subject, Table h102 in FIG. 3 is a constituent ratio of the intestinal microflora of the second subject, and Table h103 in FIG. , The composition ratio of the gut microbiota of the mth subject. Here, the composition ratio of the intestinal microflora is the ratio of each intestinal bacterium to all the intestinal bacteria in one subject. Note that the intestinal bacteria to be extracted and the number of types of intestinal bacteria may vary among subjects. For example, the first to 100th intestinal bacteria may be detected from the first subject, and the first, third,..., 100, 101, 102 intestinal bacteria may be extracted from the second subject.

(Step S12) The database generation unit 12 extracts the total number m of subjects and the total number n of types of intestinal bacteria using the information on the composition ratios of the gut microbiota of all subjects acquired in step S11. . Note that the total number n of bacteria to be extracted is at least one of the bacteria classification levels (gate, class, eye, family, genus, species, etc.). For example, when there are 100 types of enteric bacteria extracted from the first subject and 80 types of enteric bacteria different from the 100 types are extracted from other subjects, the total number of enteric bacteria types n is 180 types. The database generation unit 12 stores the extracted total number m of subjects and the total number n of enteric bacteria in the database 13.

(Step S13) The database generation unit 12 uses the information of the total number of subjects extracted in Step S12, the total number of types of intestinal bacteria, and the composition ratios of the gut microbiota of all subjects acquired in Step S11, As shown in Table h111 of FIG. 3, m × n configuration data is constructed. In the m × n configuration data, the database generation unit 12 inserts 0% into the composition ratio of enteric bacteria extracted from other subjects as shown in the table h111 of FIG.

(Step S14) The database generation unit 12 acquires age (attribute) information of each subject.
(Step S15) The database generation unit 12 associates the age information acquired in Step S14 with the m × n configuration data constructed in Step S13, and m rows × (n + 1) as shown in the table h121 of FIG. Build an intestinal microbiota database of rows.

<Neural network construction>
Next, an example of a neural network construction method performed by the neural network construction unit 14 in step S4 of FIG. 2 will be described. In the following description, in step S3 of FIG. 2, before weaning, during weaning, after weaning, 9 years old, 10 years old (10-19 years old),... A case where the age group of is generated will be described. FIG. 4 is a diagram illustrating an example of a construction procedure of the neural network according to the present embodiment.

(Step S21) The neural network construction unit 14 reads the total number m of subjects and the total number of types of enteric bacteria stored in the database 13. The total number to be read is a total number corresponding to each classification level of the bacteria selected in step s23.
(Step S22) The neural network construction unit 14 selects one age group.

(Step S23) The neural network construction unit 14 sequentially selects one of the enteric bacteria. In addition, the level of enteric bacteria selected is each classification level of bacteria. Here, each classification level of bacteria is a classification level such as a gate, a class, an eye, a family, a genus, and a species as described above. Subsequently, the neural network construction unit 14 calculates the average value of the constituent ratios for m persons for each intestinal bacterium.
(Step S24) The neural network construction unit 14 calculates the difference between the average value calculated in Step S23 and the composition ratio of each subject for each enteric bacterium.

(Step S25) The neural network construction unit 14 calculates the sum of the differences calculated in step S24 for each subject. Here, the larger the sum value, the more the sample deviates from the average value, and the smaller the sum value, the closer to the average value.
(Step S26) In each age group, the neural network construction unit 14 selects a predetermined number of subjects, for example, 10 subjects in ascending order of the total value. When the total number (number of people) included in the group is less than 10, all of them may be adopted as samples, or only samples whose total sum is a predetermined value or less may be selected.

(Step S27) The neural network construction unit 14 determines whether or not all age groups have been selected. If the neural network construction unit 14 determines that all age groups have been selected (step S27; YES), it proceeds to the process of step S29 and determines that all age groups have not been selected (step S27; YES). Step S27; NO), the process proceeds to step S28.
(Step S28) The neural network construction unit 14 selects one of the unselected age group, and returns the process to step S23.

(Step S29) The neural network construction unit 14 reconstructs the intestinal flora database as shown in Table 3 using the composition ratio of the intestinal flora of the subject selected for each age group in Step S26.

  In Table 3, for example, when 10 subjects are selected from each age group, the number of subjects included in the reconstructed intestinal microbiota database is 130 (= 10 × 13 groups). Therefore, the reconstructed intestinal microbiota database is {(total number of age groups) × (predetermined number of people)} (row) × {(total number of types of enteric bacteria) +1} (column). In addition, the group of Table 3 includes information such as age ranges, for example, before weaning, 20-29 years old, and the like.

(Step S30) The neural network construction unit 14 uses the gut microbiota database reconstructed in step S29 to perform optimization by using, for example, the R (version 3.0.3) software caret package, and calculates size and decay. . The caret package is a package in which model evaluation such as machine learning and general prediction model creation is included. Also, “size” is the number of hidden layers, and “decay” is a parameter of a penalty term for suppressing overfitting.
Subsequently, the neural network construction unit 14 uses the classified group as the objective variable (output), the bacteria classification level data of the intestinal flora as the explanatory variable (input), and uses the calculated size = 3 and decay, for example, A feedforward neural network is constructed using the R software nnet package, and information relating to the constructed feedforward neural network is stored in the database 13. The nnet package is a package for constructing a neural network.
This completes the construction of the neural network.

  In the example described above, an example of enteric bacteria selected by the neural network construction unit 14 to construct the neural network in step S23 will be further described. For example, when selecting at the genus level, the neural network construction unit 14 sequentially selects one from all the genus-level intestinal bacteria extracted from n subjects. When selecting at the family level, the neural network construction unit 14 sequentially selects one of all the family-level intestinal bacteria extracted from n subjects. When selecting at the eye level, the neural network construction unit 14 sequentially selects one of all eye-level enteric bacteria extracted from n subjects. When selecting at the mesh level, the neural network construction unit 14 sequentially selects one from all the mesh-level intestinal bacteria extracted from n subjects. When selecting at the portal level, the neural network constructing unit 14 sequentially selects one of all the intestinal bacteria at the portal level extracted from n subjects.

  When the intestinal flora database is at the genus level and the genus level is selected in step S23, a neural network is constructed based on the constituent ratio of the intestinal flora at the genus level. When the gut microbiota database is at the family level and the family level is selected in step S23, a neural network is constructed based on the constituent ratio of the gut microbiota at the family level. When the intestinal flora database is at the eye level and the eye level is selected in step S23, a neural network is constructed based on the constituent ratio of the intestinal flora at the eye level. When the intestinal flora database is at the net level and the net level is selected in step S23, a neural network is constructed based on the composition ratio of the intestinal flora at the net level. When the intestinal flora database is at the gate level and the gate level is selected in step S23, a neural network is constructed based on the constituent ratio of the intestinal flora at the gate level.

  In addition, although the example mentioned above demonstrated the example which reconstructs an intestinal microflora database using the intestinal microflora information of the test subject selected for every age group, it is not restricted to this. For example, the neural network construction unit 14 updates the intestinal flora database by associating the information indicating the selection with the information indicating the age group with the intestinal flora information of the subject selected for each age group. By doing so, the database may be reconstructed.

<Configuration of neural network>
Here, an example of the configuration of the neural network will be described.
FIG. 5 is a diagram illustrating an example of the configuration of the neural network according to the present embodiment. The neural network shown in FIG. 5 is constructed by dividing the 386 subjects into the above-mentioned 13 age groups and using intestinal microbiota information of subjects belonging to each age group as shown in a test example described later. It is. In addition, this is an example in which the total number of types of enteric bacteria extracted from 386 subjects was 187 types.

  As a result of calculation using the caret package, it was calculated that size = 3 and decay = 0.00316 are optimum as a neural network for the intestinal flora database. As a result of constructing a neural network using this value using the R software nnet package, a feedforward neural network having the configuration shown in FIG. 5 was obtained.

In FIG. 5, b1 and b2 are biases, and i1 to in are inputs. Since the total number n of enterobacteria types is 187, there are 187 inputs. H1 to h3 are hidden layers. The feedforward neural network shown in FIG. 5 has three hidden layers. Further, o1 to op are outputs. Note that p is the total number of age groups. For this reason, the feedforward neural network shown in FIG. 5 has 13 outputs for each age group.
Note that the values of the biases b1 and b2 are values calculated by, for example, the nnet package.

  In this way, the reason for using bias and hidden layers in the neural network is that the intestinal flora may be out of balance due to unbalanced diet, stress, etc., and the actual age and intestinal age may not necessarily match It is. For example, even if the actual age is 42 years old, there may be a subject whose intestinal age is estimated to be in his 20s, and there may be a subject whose intestinal age is estimated as being in his 80s.

  The weight values between nodes in the feedforward neural network shown in FIG. 5 were calculated using the nnet package, and the results shown in Table 4 were obtained.

In Table 4, “connection” indicates connection between nodes. For example, the weight value from b1 to h2 is −3.31, and the weight value from b1 to h2 is 1.47.
In FIG. 5, for example, the probability of belonging to the age group before weaning is output from o1, the probability of belonging to the age group being weaned is output from o2, and the probability of belonging to the age group of 9 to 9 years after weaning is output from o3. Is output. From o4, the probability of belonging to the age group of the 10s is output, from o5, the probability of belonging to the age group of the 20s, and from o6, the probability of belonging to the age group of the 30s. From o7, the probability of belonging to the age group of the 40s is output, from o8, the probability of belonging to the age group of the 50s, and from o9, the probability of belonging to the age group of the 60s. From o10, the probability of belonging to the age group of the 70s is output, from o11, the probability of belonging to the age group of the 80s, is output, from o12, the probability of belonging to the age group of the 90s, and from o13 Is a probability that belongs to an age group of 100 years or older.

<Estimation of intestinal age>
Next, an example of the intestinal age estimation processing method performed by the estimation unit 15 in step S6 of FIG. 2 will be described.
FIG. 6 is a flowchart of intestinal age estimation processing according to the present embodiment.
(Step S51) The estimation unit 15 acquires the constituent ratio of the intestinal flora of the subject.
(Step S <b> 52) The estimating unit 15 expands the composition ratio of the intestinal bacterial flora of the subject to the total number n of types of intestinal bacteria stored in the database 13. As a result, 0% is inserted into the intestinal bacteria that have not been extracted from the subject. For example, when the total number of types of enteric bacteria extracted from a plurality of subjects is 186, and the number of types of enteric bacteria extracted from the subject is 100, the estimation unit 15 has 86 (= 186-100). Add intestinal bacteria information. And the estimation part 15 inserts 0% in the component ratio of 86 types of bacteria added.

(Step S <b> 53) The estimation unit 15 reads information related to the neural network from the database 13.
(Step S54) The estimation unit 15 inputs the value of the composition ratio of the first bacteria to i1 of the neural network shown in FIG. 5, and inputs the value of the composition ratio of the nth bacteria to. Thereby, the estimation part 15 calculates the probability (o1-op) of each age group to which a subject's intestinal age belongs using the neural network shown in FIG.

(Step S55) The estimation unit 15 selects the largest value among the probabilities calculated in Step S54. For example, when the probability of belonging to the 40s is the largest value, the estimation unit 15 selects an age group of the 40s. Subsequently, based on the selected result, the estimation unit 15 outputs, for example, information indicating “your gut age is in the 40s” to the external device.

  In the above-described example, the example in which the largest value is selected in step S55 has been described. However, the present invention is not limited to this. The estimation unit 15 may output all the results calculated in step S54 to the external device as estimation results. Alternatively, data obtained by visualizing all calculated values for each age group by color classification, density classification, and the like may be generated, and the generated data may be estimated and output to an external device.

<Method for extracting enteric bacteria>
Next, an example of an intestinal bacteria extraction method performed by the intestinal bacteria extraction apparatus 11 will be described.
FIG. 7 is a flowchart of intestinal bacteria extraction processing performed by the intestinal bacteria extraction apparatus 11 according to the present embodiment.

(Step S101) The intestinal bacteria extraction apparatus 11 measures about 20 mg of stool for each subject or each subject and suspends it in 450 ul of an extract (100 mM Tris / HCl, 4 mM EDTA, pH 9.0).
(Step S102) The intestinal bacteria extraction apparatus 11 mixes 50 ul of 10% SDS solution, 300 mg of 0.1 mm diameter glass beads, and 500 ul of TE saturated phenol (for example, Wako Pure Chemical Industries) into the extract suspended in Step S101. For example, a crushing process is performed at a power level of 5 seconds for 30 seconds using FastPrep FP 100A (manufactured by Funakoshi).

(Step S103) The intestinal bacterium extraction apparatus 11 obtains 400 ul of supernatant after centrifugation at 14,000 g for 5 minutes with respect to the solution subjected to the crushing process in Step S102. Subsequently, the intestinal bacterium extraction apparatus 11 added and mixed 250 ul of phenol / chloroform solution (for example, Wako Pure Chemicals) to the obtained supernatant, and centrifuged at 14,000 g for 5 minutes to obtain 250 ul of supernatant. .

(Step S104) The intestinal bacterium extraction apparatus 11 adds 250 ul of 2-propanol to the supernatant obtained in Step S103, and dissolves the isopropanol-precipitated material with 200 ul of Tris-EDTA buffer (pH 8.0). Create a solution.

(Step S105) The intestinal bacterium extraction apparatus 11 is Tru357F [(5′-CGCTCTTCGATCTCTGTACGGRAGGCAGCAG-3 ′ (SEQ ID NO: 1)), which is a first primer set for amplifying the third to fourth variable regions of the bacterial 16S rRNA gene. And Tru806R [5'-CGCTCTTCCCATCTGACGGAACTACHVGGGTWTCTAAT-3 '(SEQ ID NO: 2)] and Fwd [5'-X1NNNNNNNN-X2-, which is a 2nd primer set required for analysis using, for example, the next-generation sequencer Miseq (manufactured by Illumina). 3 ′] and Rev [5′-X3-NNNNNNNN-X4-3 ′], for example, primer combining by Oligoprimer making service of Life Technologies To.
X1, X2, X3, X4, and N represent the following sequences.
X1: AATGATACCGGCGACCACCGAGATCTACAC (SEQ ID NO: 3)
X2: ACACTCTTTCCCTACACGACGCTCTTCCCATCCTG (SEQ ID NO: 4)
X3: CAAGCAGAAGACGGCATACGAGAT (SEQ ID NO: 5)
X4: GTGACTGGAGTTCAGACGTGTGCTCTTCGATCTGAC (SEQ ID NO: 6)
N is a base sequence for identifying n subjects consisting of arbitrary bases

(Step S106) The intestinal bacterium extraction apparatus 11 prepares a reaction solution having a total volume of 25 μl including the template DNA solution and the first primer set using, for example, TaKaRa Ex Taq HS kit (manufactured by Takara Bio Inc.). Subsequently, the intestinal bacterium extraction apparatus 11 is used, for example, by Veriti 200 (manufactured by Life Technologies), for example, 94 ° C. for 3 minutes, then 94 ° C. for 30 seconds, 50 ° C. for 30 seconds, 72 ° C. for 30 seconds, 72 ° C. A 10 minute PCR (Polymerase Chain Reaction) reaction is performed.

(Step S107) The intestinal bacteria extraction apparatus 11 performs electrophoresis on the PCR product obtained in Step S106 on a 1% agarose gel, and confirms the band pattern.
(Step S108) The intestinal bacterium extraction apparatus 11 performs a PCR reaction using the PCR product 1ul obtained in step S106 as a template and using the 2nd primer set in the same manner as described above. However, the number of PCR cycles is not 20 but 8 times.

(Step S109) The intestinal bacteria extraction apparatus 11 performs electrophoresis on the PCR product obtained in Step S108 on a 1% agarose gel, and confirms the band pattern.
(Step S110) After the treatment of Step S109, the intestinal bacterium extraction apparatus 11 performs purification using, for example, QIAquick 96 PCR Purification Kit (manufactured by Qiagen) and, for example, Quant-iP PicoGreen dsDNA Assay kit (manufactured by Life Technologies). To measure the concentration. Subsequently, the intestinal bacterium extraction apparatus 11 supplies a mixture of DNA solutions having the same concentration to, for example, Miseq v3 Reagent kit (manufactured by Illumina), and performs sequence analysis with Miseq.

(Step S111) The intestinal bacterium extraction apparatus 11 removes the Phyx (control sequence at the time of sequencing) and the human genome sequence from the fastq file obtained in step S110.
(Step S112) After the process of step S111, the intestinal bacteria extraction apparatus 11 removes Q17 or lower base from the base sequence 3 ′ side.

(Step S113) The intestinal bacteria extraction apparatus 11 removes the lead that has become 150 bp or less after the process of Step S112. Subsequently, the intestinal bacterium extraction apparatus 11 removes leads having a Q25 or less and unpaired leads as a whole.
(Step S114) After the process of step S113, the intestinal bacterium extraction apparatus 11 converts the obtained paired end sequence into, for example, fastq-join (version 1.1.2-301) (http://code.google.com). / P / ea-utils / wiki / FastqJoin).

(Step S115) After the process of step S114, the intestinal bacteria extraction apparatus 11 removes the chimera sequence using, for example, UCIME.
(Step S116) Finally, for example, 4,979 ± 1,800 sequences are obtained per sample, and the intestinal bacterium extraction apparatus 11 converts them into, for example, QIIME software (version 1.8.0) (http: /// Each sequence having 97% homology in qime.org/) is defined as OTU (Operational Taxonomy Unit). Subsequently, the intestinal bacterium extraction apparatus 11 uses a representative sequence of each OTU, for example, Greengenes database 12_10 (http://greengenes.secondgenome.com/downloads/database/12_10), for example, BLAST (Basic Locent Local). By searching using Search Tool, configuration data regarding the bacterial classification level (gate, class, eye, family, genus) of the intestinal flora is generated. The configuration data is a table of m rows × o columns when the type of enteric bacteria detected from m persons is o. BLAST is, for example, an algorithm for finding a similar sequence group with a score equal to or higher than a predetermined threshold, and a program that implements the algorithm. Subsequently, the database generation unit 12 stores the age of the subject in the database 13 in association with the configuration data at the bacterial classification level of the intestinal flora generated by the intestinal bacteria extraction apparatus 11.

  This is the end of the intestinal bacteria extraction processing procedure. In addition, the procedure of the intestinal bacteria extraction process described above, the solution to be used, the volume of the solution to be used, the application to be used, and the like are examples, and are not limited thereto.

[First Test Example]
Next, a first test example will be described. The subject in the first test example is a man in his 40s.
(Procedure 1) The intestinal bacteria extraction apparatus 11 uses the stool of 367 subjects for the database, and obtains configuration data at the bacterial classification level of the intestinal flora by the processing of steps S101 to S116 shown in FIG. Generate. Subsequently, the database generation unit 12 associates the age of each of the 386 subjects with the configuration data at the bacterial classification level of the intestinal microflora generated by the intestinal bacterial extraction device 11, and creates an intestinal microflora database. The constructed gut microbiota database is stored in the database 13. That is, in the database 13, identifiers and ages are stored in association with configuration data of 386 (number of subjects) rows × 187 (number of types of bacteria) columns. Intestinal bacteria extracted from 386 subjects were those shown in the first column of Tables 5 to 14. As shown in Tables 5 to 14, the bacterial classification level of the extracted intestinal bacteria is a genus level. In Tables 5 to 14, “[...]” Represents a name that was once used because the classification is currently unclear. “Gen.” Represents that the genus level classification is unknown or unidentified at present, “fan.” Represents that the class level classification is currently unknown or unidentified, and “ord.” . ”Indicates that the eye level classification is currently unknown or unidentified.

(Procedure 2) The neural network construction unit 14 reconstructs the intestinal microflora database by selecting subjects in the processes of steps S11 to S15 in FIG. 3 for the intestinal microbiota database constructed in Procedure 1. To do. The neural network construction unit 14 selected 10 subjects from each of the 13 age groups. However, before weaning, during weaning, there were less than 10 subjects belonging to the age group of 100 years or older. 117 people were selected.

(Procedure 3) The neural network constructing unit 14 optimizes the model for the intestinal flora database reconstructed by the procedure 2 using, for example, the R software package, and size = 3 and decay = 0. .00316 is calculated. Subsequently, the neural network constructing unit 14 constructs a feedforward neural network by using the calculated size = 3 and decay = 0.00316, for example, using the R software nnet package. An example of the constructed feedforward neural network is shown in FIG. FIG. 8 is a diagram illustrating an example of a configuration of a neural network constructed according to the first test example. In the example shown in FIG. 8, b1 and b2 are biases as in FIG. 5, i1 to i186 are inputs corresponding to 186 types of bacteria, and o1 to o13 are outputs corresponding to age groups.

(Procedure 4) The intestinal bacteria extraction apparatus 11 extracts the intestinal bacterial flora by the processing of steps S101 to S116 shown in FIG. 7 using the stool of a male subject in their 40s. The ratios shown in the second column of Tables 5 to 13 are the constituent ratios of the intestinal bacteria extracted from the subject.

(Procedure 5) The neural network construction unit 14 inputs the intestinal flora information of the subject obtained in step 4 to the neural network constructed in step 3, and selects the age group to which the intestinal age of the subject belongs. presume. The second column of Table 15 shows the estimation results of the intestinal age of subjects in their 40s according to the first test example.

  As shown in the second column of Table 15, the result of inputting the composition ratio of the intestinal flora of the subject in the first test example to the neural network is as follows: o1 = 0, o2 = 0, o3 = 0.0066, o4 = 0.0747, o5 = 0.0697, o6 = 0.0869, o7 = 0.6004, o8 = 0.1613, o9 = 0.0004, o10 = 0.0001, o11 = 0, o12 = 0, o13 = 0. As described above, the intestinal age of the subject was the largest at 40.04% in the 40s (40-49 years). Thus, it was suggested that the intestinal flora composition of the subject in the first test example is equivalent to the actual age.

[Second Test Example]
Next, a second test example will be described. The subject in the second test example is a woman in her 50s.
Procedures 1 to 3 are the same as in the first experimental example.
(Procedure 4 ′) The intestinal bacteria extracting apparatus 11 extracts the intestinal bacterial flora by the processing of steps S101 to S116 shown in FIG. 7 using feces of a female subject in their 50s.

(Procedure 5 ′) The neural network construction unit 14 inputs the intestinal flora information of the subject obtained in step 4 ′ into the neural network constructed in step 3, and the age to which the intestinal age of the subject belongs. Estimate groups. The third column of Table 15 shows the estimation results of the intestinal age of subjects in their 50s according to the second test example.

  As shown in the third column of Table 15, the result of inputting the composition ratio of the intestinal flora of the subject in the second test example to the neural network is as follows: o1 = 0, o2 = 0.0008, o3 = 0. 2307, o4 = 0.7029, o5 = 0, o6 = 0.0406, o7 = 0.0002, o8 = 0.133, o9 = 0, o10 = 0.114, o11 = 0, o12 = 0, o13 = 0. Thus, the intestinal age of the subjects was the largest in their 10s (10-19 years) at 70.29%, followed by 23.07% after weaning to 9 years. Thus, it was suggested that the intestinal flora composition of the subject in the second test example was considerably younger than the actual age.

[Third test example]
Next, a third test example will be described. The subject in the second test example is a woman in her 50s.
Procedures 1 to 3 are the same as in the first experimental example.
(Procedure 4 ″) The intestinal bacterial extraction apparatus 11 extracts the intestinal bacterial flora by the processing of steps S101 to S116 shown in FIG. 7 using feces of a female subject before weaning.

(Procedure 5 ″) The neural network construction unit 14 inputs the intestinal flora information of the subject obtained in Step 4 ′ to the neural network constructed in Step 3, and the intestinal age of the subject belongs. Estimate age groups. The third column of Table 15 shows the estimation result of the intestinal age of the subject before weaning according to the third test example.

  As shown in the fourth column of Table 15, the result of inputting the composition ratio of the intestinal flora of the subject in the second test example to the neural network is as follows: o1 = 0, o2 = 0, o3 = 0.0029, o4 = 0.0958, o5 = 0.1083, o6 = 0.0706, o7 = 0.3770, o8 = 0.3355, o9 = 0.0096, o10 = 0.0002, o11 = 0, o12 = 0, o13 = 0. Thus, the intestinal age of subjects was the largest in the 40s (40-49 years) at 37.70%, and then was 33.55% in the 50s (50-59 years). Thus, it was suggested that the intestinal flora composition of the subject in the third test example is considerably older than the actual age.

  As described above, the intestinal age estimation method of the present embodiment uses an artificial neural network (for example, a feedforward neural network) to estimate the intestinal age of the subject (for example, the step of FIG. 2). S2 and intestinal age estimation method including step S54 in FIG. 6, wherein the artificial neural network extracts an identifier for identifying each of a plurality of (m) subjects having different ages and a plurality of subjects. For each of a plurality of subjects for each of a plurality of intestinal bacteria at the bacterial classification level including all types (for example, Methanobrewibacter, Actinomyces, ...) A value (composition ratio or composition ratio) representing the ratio to all intestinal bacteria at the extracted bacteria classification level is associated with a value representing the age of each of a plurality of subjects. It is constructed using intestinal flora database.

  With this configuration, in the present embodiment, the intestinal age of the subject can be estimated by an artificial neural network constructed using the intestinal microbiota database.

  In the intestinal age estimation method of the present embodiment, in the estimation step, a plurality of intestinal bacteria (for example, 187 at the genus level) including all types extracted from a plurality (m) of subjects. For each of the types, the values (composition ratios or composition ratios) representing the ratios to all intestinal bacteria extracted from the subject are input to an artificial neural network (for example, a feedforward neural network). The intestinal age of the person is estimated.

  With this configuration, in this embodiment, the intestinal age of the subject can be estimated based on the state of the intestinal bacteria of the subject.

  In the intestinal age estimation method of the present embodiment, the ages of a plurality of (m) subjects included in the intestinal flora database are set to a plurality of age groups (for example, before weaning, during weaning, after weaning −9 A value representing the ratio (composition ratio or composition ratio) for each of a plurality of age groups classified by the classification process (step S3 in FIG. 2) and the classification process (separated by age,..., 100 years old). The average value calculating step (step S23 in FIG. 4) for calculating the average value of each enteric bacterium, the difference between the average value calculated by the average value calculating step and the ratio of each subject belonging to the age group A total calculation step (step S25 in FIG. 4) for calculating each enteric bacterium (step S24 in FIG. 4) and calculating the sum of the calculated differences for each subject, and a plurality of members belonging to the age group A selection step (step S26 in FIG. 4) for selecting from each of the age groups a predetermined number of subjects in ascending order of the total sum calculated by the sum calculation step, and the intestines of the subject selected by the selection step. The reconstruction process (step S29 in FIG. 4) for reconstructing the intestinal flora database using the information and age of the intestinal bacteria, and the intestinal flora database reconstructed by the reconstruction process, A neural network (for example, a feedforward neural network) is constructed (step S30 in FIG. 4).

  Thus, in this embodiment, the intestinal bacterial flora database is obtained by extracting the intestinal bacterial flora of the average subject in the age group from a plurality of subjects belonging to each age group. The neural network is constructed using the reconstructed intestinal microbiota database. Thereby, in this embodiment, in the neural network constructed for estimating the intestinal age, it is possible to exclude the intestinal flora of the subject that becomes noise when estimating the intestinal age. As a result, according to the present embodiment, the intestinal age of the subject can be accurately estimated.

  In the intestinal age estimation method of the present embodiment, the number of inputs (i1 to in) of the artificial neural network (for example, feedforward neural network) is all types extracted from a plurality of (m) subjects. The total number of types of a plurality of enteric bacteria at the bacterial classification level including (for example, 187 in the case of a genus level), the number of outputs (o1 to op) of the artificial neural network is a plurality of age groups classified by the classification process The total number of

  With this configuration, in this embodiment, the intestinal flora information detected from the subject can be input to the neural network, whereby the probability of belonging to each age group can be calculated.

In the intestinal age estimation method of the present embodiment, in the classification step, a plurality of subjects are separated every 10 years, and in the estimation step, the subject's intestinal age is estimated every 10 years.
With this configuration, in the present embodiment, it is possible to accurately estimate the intestinal age by estimating which age group of every 10 years the subject's intestinal bacteria configuration belongs to.

In the intestinal age estimation method of the present embodiment, the intestinal flora database is based on a value indicating the ratio of intestinal bacteria extracted from subjects 0 to 104 years old, including before weaning and during weaning. Built.
Thereby, in this embodiment, using the neural network based on the intestinal microbiota database by each subject of each age group from before weaning to 104 years old stored in the database 13, the intestinal age of the subject Can be determined with high accuracy.

In the intestinal age estimation method of the present embodiment, the artificial neural network is a feedforward neural network having a hidden layer.
Thereby, in this embodiment, it can estimate accurately including intestinal ages other than real age.

  Further, in the intestinal age estimation method of the present embodiment, the intestinal microflora database acquires, for each subject, each ratio of the intestinal bacteria extracted from the subject to all the intestinal bacteria extracted from the subject. Using the acquisition process (step S11 in FIG. 3) and the ratio of each enteric bacterium acquired by the acquisition process, the total number m of subjects (m is an integer of 2 or more) and a plurality of subjects An extraction step (step S12 in FIG. 3) for extracting the total number n (n is an integer of 2 or more) of a plurality of types of enteric bacteria at the bacterial classification level including all the extracted types, and the extraction step. The total number of subjects m, the total number n of types of a plurality of intestinal bacteria, and the value representing the ratio of all the intestinal bacteria in the bacterial classification level extracted for each of the plurality of subjects, m rows × n lines The configuration data construction step (step S13 in FIG. 3) for constructing the composition data, the information indicating the age of each of the plurality of subjects are acquired, and the information indicating the age for each subject is acquired in the configuration data construction step. And a database construction step (steps S14 and S15 in FIG. 3) for constructing an intestinal microbiota database in association with the configured data.

  Thereby, in this embodiment, an intestinal microflora database can be constructed | assembled using the intestinal bacteria extracted from subjects other than a test subject or a subject. As a result, in this embodiment, the age to which the subject belongs by inputting the intestinal flora information of the subject to the artificial neural network constructed using the intestinal flora database constructed in this way. Groups can be estimated with high accuracy.

  In addition, in the example mentioned above, although the database production | generation part 12 demonstrated the example which constructs an intestinal microflora database by all the types n of enteric bacteria extracted from m test subjects, it is not restricted to this. In addition to all types n of intestinal bacteria extracted from m subjects, the database generation unit 12 stores intestinal bacteria based on the results of searching past experimental data and papers in advance. May be included.

In addition, in the example mentioned above, although the example which used the example of age as an attribute was demonstrated, it is not restricted to this. The attribute may be sex, weight, height, place of residence, hometown, nationality, and the like. Further, the attributes associated with the intestinal flora information may be plural (for example, age and nationality). For example, when the attribute is body weight, the body weight is classified into several groups (for example, less than 10 kg, 10-19 kg, 20-29 kg,..., 90-99 kg, 100 kg or more), and subjects are extracted for each weight group. An intestinal microbiota database may be constructed.
In addition, the information stored in the database 13 may be a plurality of data such as nationalities and regions.

  In the above-described example, an example in which the intestinal age estimation device 1 estimates the intestinal age based on the composition ratio of the human intestinal flora is described, but the present invention is not limited thereto. The intestinal age estimation apparatus 1 may estimate the intestinal age of an animal based on the composition ratio of the intestinal flora of the animal.

  Further, a part or all of the intestinal age estimation apparatus 1 of the present embodiment may be performed by an application installed by a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and a computer.

DESCRIPTION OF SYMBOLS 1 ... Intestinal age estimation apparatus, 11 ... Intestinal bacteria extraction apparatus, 12 ... Database production | generation part, 13 ... Database, 14 ... Neural network construction part, 15 ... Estimation part

Claims (9)

An estimation method of intestinal age, including an estimation step of estimating an intestinal age of a subject using an artificial neural network,
The artificial neural network is
An identifier for identifying each of a plurality of subjects having different ages;
For each of a plurality of intestinal bacteria at a bacterial classification level including all types extracted from the plurality of subjects, a ratio to all the intestinal bacteria at the bacterial classification level extracted for each of the plurality of subjects is represented. Value and
A value representing the age of each of the plurality of subjects;
An intestinal age estimation method constructed using an intestinal microbiota database associated with. In the estimation step,
For each of a plurality of intestinal bacteria of a plurality of bacterial classification levels including all types extracted from the plurality of subjects, a value representing a ratio to all intestinal bacteria extracted from the subject is stored in the artificial neural network. The intestinal age estimation method according to claim 1, wherein the intestinal age of the subject is estimated by inputting. A classification step of separating the age of the plurality of subjects included in the gut microbiota database into a plurality of age groups;
For each of the plurality of age groups classified by the classification step, an average value calculating step for calculating an average value of the values representing the ratio for each of the enteric bacteria,
The difference between the average value calculated by the average value calculation step and the ratio of each of the subjects belonging to the age group is calculated for each of the enteric bacteria, and the sum of the calculated differences is calculated for each of the subjects. A sum calculation step for calculating
A selection step of selecting from each of the age groups by a predetermined number of people in descending order of the value of the sum calculated from the plurality of subjects belonging to the age group,
Reconstructing the intestinal flora database using the information and age of the intestinal bacteria of the subject selected by the selection step;
The intestinal age estimation method according to claim 1 or 2, wherein the artificial neural network is constructed using the intestinal microbiota database reconstructed by the restructuring step. The number of inputs of the artificial neural network is the total number of types of a plurality of enteric bacteria at a bacterial classification level including all types extracted from the plurality of subjects,
The intestinal age estimation method according to claim 3, wherein the number of outputs of the artificial neural network is the total number of the plurality of age groups classified by the classification step. In the classification step, the plurality of subjects are separated every 10 years,
The intestinal age estimation method according to claim 3 or 4, wherein in the estimation step, the intestinal age of the subject is estimated every 10 years.   The said intestinal microflora database is constructed | assembled based on the value which shows the said ratio of the enteric bacteria extracted from the test subject of 0 to 104 years old including before weaning and during weaning. The intestinal age estimation method according to claim 1.   The intestinal age estimation method according to any one of claims 1 to 6, wherein the artificial neural network is a feedforward neural network having a hidden layer. The intestinal flora database is
An acquisition step of acquiring, for each subject, the ratio of each of the intestinal bacteria extracted from the subject to all the intestinal bacteria extracted from the subject;
Including the total number m of the subjects (m is an integer of 2 or more) and all types extracted from the plurality of subjects using the ratio of each enteric bacterium obtained by the obtaining step. An extraction step of extracting a total number n (n is an integer of 2 or more) of a plurality of types of enteric bacteria at the bacterial classification level;
The total number m of the subjects extracted by the extraction step, the total number n of the types of the plurality of intestinal bacteria, and the ratio to all the intestinal bacteria at the bacterial classification level extracted for each of the plurality of subjects. A configuration data constructing step of constructing m rows × n rows of configuration data using values to represent;
The information indicating the age of each of the plurality of subjects is acquired, and the acquired information indicating the age of each subject is associated with the configuration data constructed in the configuration data construction step to construct an intestinal flora database The intestinal age estimation method according to any one of claims 1 to 7, which is a database constructed by a database construction step. An identifier for identifying each of a plurality of subjects having different ages, and each of a plurality of intestinal bacteria at a bacterial classification level including all types extracted from the plurality of subjects, extracted for each of the plurality of subjects. A database that stores an intestinal microflora database in which a value representing a ratio to all intestinal bacteria at a bacterial classification level and a value representing the age of each of the plurality of subjects are associated;
Using an artificial neural network constructed using the intestinal microbiota database stored in the database, an estimation unit that estimates the intestinal age of the subject,
An intestinal age estimation apparatus comprising:

Images