JP6928948B2 - Halitosis judgment device and program - Google Patents

Description

The present invention relates to a halitosis determination device and a program.

Conventionally, there is known a technique of detecting the concentration of a specific gas contained in the exhaled breath of a subject and determining whether or not halitosis is generated in the oral cavity of the subject (for example, Patent Document 1).
Further, there is known a technique of culturing bacteria in the oral cavity based on the saliva of a subject and analyzing the cultured bacteria to determine whether or not halitosis is generated in the oral cavity of the subject (for example,). Patent Document 2).

Japanese Unexamined Patent Publication No. 2003-329630 Japanese Unexamined Patent Publication No. 2002-236124

However, in the technique described in Patent Document 1, when determining whether or not halitosis is occurring, the subject moves to a place where a device capable of detecting exhaled breath is installed, or a device capable of detecting exhaled breath. Was sometimes required to be moved to the location where the subject was examined. Further, in the technique described in Patent Document 2, when determining whether or not halitosis is occurring, it is required to culture the bacteria in the oral cavity of the subject, and it is difficult to shorten the time required for the determination. There was a case.
The present invention has been made in view of the above problems, and an object of the present invention is to provide a halitosis determination device and a program capable of determining the presence or absence of halitosis of a subject based on the saliva of the subject.

One aspect of the present invention is a detection unit that detects a type of short base sequence having five bases, which is a base sequence of DNA existing in the bacterial flora of the saliva of a subject, and the short base sequence detected by the detection unit. For each type of base sequence, a ratio acquisition unit for acquiring the ratio of the base sequence of the type present in the bacterial flora, the type of bacteria existing in the bacterial flora in the oral cavity of the subject, and the bacterial flora of the bacteria. Reference information indicating the correspondence with the ratio in each combination of the type and the ratio, the type of the short base sequence detected by the detection unit, and the ratio of the short base sequence acquired by the ratio acquisition unit for each type. Based on the above, the type of bacteria existing in the oral flora of the subject and the proportion of the bacteria in the bacterial flora are determined, and the information indicating the type, which is the result of the determination, is used as the bacterial species information. , A determination unit that outputs information indicating the ratio as ratio information, teacher bacterial species information indicating the type of bacteria existing in the bacterial flora in the oral cavity, and the ratio of the bacteria indicated by the teacher bacterial flora in the bacterial flora. A storage unit that stores a learned model learned by machine learning and a determination unit so as to output halitosis information indicating whether or not there is halitosis in the oral cavity based on the teacher ratio information shown for each type. said species information but for outputting an acquisition unit that acquires a pre-Symbol percentage information, and the species information acquired by the acquiring unit, and the ratio information acquired by the acquiring unit to the learned model It is a halitosis determination device including an output unit that outputs halitosis information indicating whether or not the subject has a halitosis that has been input and calculated.

Further, in the halitosis determination device of one aspect of the present invention, the machine learning is deep learning.

Further, in the halitosis determination device of one aspect of the present invention, the teacher bacterial species information and the teacher ratio information are information based on the results of analyzing saliva of a plurality of subjects.

Further, one aspect of the present invention is a teacher bacterial species information indicating the type of bacteria existing in the bacterial flora in the oral cavity and a teacher ratio indicating the ratio of the bacteria indicated by the teacher bacterial species information in the bacterial flora for each type. A computer equipped with a storage unit that stores a learned model learned by machine learning so as to output halitosis information indicating whether or not there is halitosis in the oral cavity based on the information, in the bacterial flora of the subject's saliva. A detection step of causing a detection unit that detects the type of a short base sequence having five bases in the base sequence of an existing DNA to detect the type of the short base sequence, and the short base sequence detected in the detection step. For each type of base sequence, the ratio acquisition step of acquiring the ratio of the base sequence of the type present in the bacterial flora, the type of bacteria existing in the bacterial flora in the oral cavity of the subject, and the bacterial flora of the bacteria. Reference information indicating the correspondence with the ratio in each combination of the type and the ratio, the type of the short base sequence detected in the detection step, and the type of the short base sequence acquired in the ratio acquisition step. Based on the proportion of the bacteria, the type of bacteria present in the oral flora of the subject and the proportion of the bacteria in the bacterial flora are determined, and the information indicating the type, which is the result of the determination, is used as the bacterial species. A determination step that outputs information indicating the ratio as information , a bacterial species information indicating the type of bacteria existing in the bacterial flora in the oral cavity of the subject, and a bacterial flora of the bacteria indicated by the bacterial species information in the bacterial flora. The acquisition step for acquiring the ratio information indicating the ratio for each type, the bacterial species information acquired in the acquisition step, and the ratio information acquired in the acquisition step were input to the trained model and calculated. This is a program for executing an output step of outputting halitosis information indicating whether or not the subject has halitosis.

According to the present invention, the presence or absence of bad breath of the subject can be determined based on the saliva of the subject.

It is a figure which shows the outline of the halitosis determination system of this embodiment. It is a functional block diagram which shows an example of the structure of the bacterial species ratio detection apparatus which detects the bacterial species information and the ratio information based on the existence ratio of the short base sequence of this embodiment. It is the first figure which shows an example of correspondence information of this embodiment. It is a functional block diagram which shows an example of the structure of the bacterial species ratio detection apparatus which detects the bacterial species information and the ratio information based on the abundance ratio of the 16S rRNA gene of this embodiment. It is a 2nd figure which shows an example of correspondence information of this embodiment. It is a functional block diagram which shows an example of the structure of the halitosis determination device of this embodiment. It is a flow chart which shows an example of the operation of the halitosis determination system of this embodiment.

[Embodiment]
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram showing an outline of the halitosis determination system 1 of the present embodiment.
As shown in FIG. 1, the halitosis determination system 1 includes a halitosis determination device 10 and a bacterial species ratio detection device 20.
The halitosis determination device 10 is a device for determining the presence or absence of halitosis of a determination target person (hereinafter, subject ET). The bacterial species ratio detection device 20 acquires a collection (bacterial flora) of bacteria in the oral cavity of the subject ET, and determines the type of bacteria existing in the oral cavity of the subject ET and the ratio of the bacteria present in the oral cavity. It is a device that detects each type. The bacterial flora can be obtained from, for example, the tongue coating of the subject ET or the saliva of the subject ET (hereinafter, saliva SV). In one example of this embodiment, the bacterial species ratio detecting device 20 detects the type of bacteria and the ratio of the bacteria present in the oral cavity based on the saliva SV of the subject ET. In the following description, information that is detected by the bacterial species ratio detection device 20 and indicates the type of bacteria existing in the oral cavity of the subject ET will be referred to as bacterial species information KB. Further, the information detected by the bacterial species ratio detection device 20 and indicating the ratio of the bacteria indicated by the bacterial species information KB in the oral cavity of the subject ET for each type of bacteria is described as the ratio information RB. The halitosis determination device 10 determines the presence or absence of halitosis in the subject ET based on the bacterial species information KB detected by the bacterial species ratio detection device 20 and the ratio information RB.

Here, the method for detecting the bacterial species information KB and the ratio information RB by the bacterial species ratio detecting device 20 is a method based on the abundance ratio of a short base sequence (hereinafter, short base sequence) and a method based on the abundance ratio of the 16S rRNA gene. There is. First, the bacterial species ratio detection device 20 (hereinafter referred to as the bacterial species ratio detection device 20a) that detects the bacterial species information KB and the ratio information RB based on the abundance ratio of the short base sequence will be described, and then the abundance ratio of the 16SrRNA gene will be described. The bacterial species ratio detecting device 20 (hereinafter referred to as the bacterial species ratio detecting device 20b) that detects the bacterial species information KB and the ratio information RB based on the above will be described. In the following description, among the bacterial species ratio detecting devices 20, the configurations included in the bacterial species ratio detecting device 20a are marked with “a” at the end of the reference numerals. Further, "b" is added to the end of the code in the configuration provided in the bacterial species ratio detecting device 20b. Further, when it is not distinguished which of the bacterial species ratio detecting devices 20 has the configuration, "a" or "b" is omitted.

<About the bacterial species ratio detection device 20a>
Hereinafter, the details of the bacterial species ratio detecting device 20a for detecting the bacterial species information KB and the ratio information RB based on the abundance ratio of the short base sequence will be described with reference to the drawings.
FIG. 2 is a functional configuration diagram showing an example of the configuration of the bacterial species ratio detecting device 20a that detects the bacterial species information KB and the ratio information RB based on the abundance ratio of the short base sequence of the present embodiment.
The bacterial species ratio detecting device 20a includes a sequencer 21a, a control unit 22a, and a storage unit 23a. The storage unit 23a is realized by, for example, a ROM (Read Only Memory), a flash memory, an SD card, a RAM (Random Access Memory), a register, or the like. For example, short base information 231a and correspondence information 232a are stored in advance in the storage unit 23a.

The short base information 231a is information indicating the base sequence of Deoxy Ribonucleic Acid (hereinafter referred to as DNA) to be analyzed by the sequencer 21a, and is information indicating the short base sequence represented by the combination of five nucleotides. In one example of this embodiment, the analysis target of the sequencer 21a is a bacterium existing in saliva SV (in the oral cavity of the subject ET).
The sequencer 21a analyzes the base sequence to be analyzed indicated by the short base information 231a among the DNA base sequences of the bacteria existing in the saliva SV (in the oral cavity of the subject ET). Such sequence comparison is called alignment. The sequencer 21a supplies information indicating the type of the obtained short base sequence (hereinafter, short base sequence information SB) to the control unit 22a after alignment. In addition, the sequencer 21a provides the control unit 22a with information indicating the amount (ratio) of the obtained short base sequence in saliva SV for each type of the short base sequence (hereinafter referred to as short base ratio information RSB). Supply.

The control unit 22a realizes each functional unit by executing a program stored in the storage unit 23a by a processor such as a CPU (Central Processing Unit). The control unit 22a realizes, for example, a determination unit 221a and an output unit 222a as its functional units. The control unit 22a may be realized by hardware such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), or by cooperation between software and hardware. You may.

The determination unit 221a acquires the short base sequence information SB and the short base ratio information RSB from the sequencer 21a. Based on the short base sequence information SB, the short base ratio information RSB, and the corresponding information 232a, the determination unit 221a determines the bacteria (bacterial species information KB) existing in the oral cavity of the subject ET and the bacteria in the oral cavity. It is determined that the ratio exists (ratio information RB). Correspondence information 232a is information indicating the correspondence between the short base sequence information SB and the short base ratio information RSB, and the bacterial species information KB and the ratio information RB.

Hereinafter, the details of the corresponding information 232a will be described with reference to the drawings. FIG. 3 is a first diagram showing an example of the corresponding information 232a of the present embodiment.
As shown in FIG. 3, the correspondence information 232a is associated with the short base sequence information SB, the short base ratio information RSB, the bacterial species information KB, and the ratio information RB.
Here, when the short base sequence information SB of a certain subject ET and the short base ratio information RSB are known, the type of bacteria existing in the oral cavity of the subject ET (bacterial species information KB) and the bacteria present in the oral cavity. The ratio (ratio information RB) to be used is uniquely determined. The correspondence between the short base sequence information SB and the short base ratio information RSB and the bacterial species information KB and the ratio information RB is as follows: the short base sequence indicated by the short base information 231a and the bacteria in the saliva SV in which the base sequence exists. It is performed based on the analysis result of.

In the example shown in FIG. 3, n types of short base sequence information SBs and a number (n) of short base ratio information RSBs corresponding to the short base sequence information SBs are associated with each other as short base information 231a. .. n is a natural number. Specifically, FIG. 3 shows that "GGACC (short base sequence information SB1)" is present in saliva SV at a ratio of 0.25% (short base sequence information RSB1), and "CCCCC (short base sequence information SB2)" is present. ) ”Is present at a ratio of 0.5% (short base ratio information RSB2), and…,“ CCTAG (short base sequence information SBn) ”is present at a ratio of 0.25% (short base ratio information RSBn). Is shown.
Further, in the example shown in FIG. 3, the above-mentioned short base sequence information SB1 to short base sequence information SBn and short base ratio information RSB1 to short base ratio information RSBn include bacterial species information KB1 to bacterial species information KB4 and ratio information. RB1 to ratio information RB4 are associated with short base information 231a. Specifically, in the short base information 231a shown in FIG. 3, the short base sequence of the short base sequence information SB1 to the short base sequence information SBn is present in the saliva SV at the ratio of the short base ratio information RSB1 to the short base ratio information RSBn. In this case, actinomycetemcomitans (bacterial species information KB1) is present in the oral cavity of the subject ET at a ratio of 19% (ratio information RB1), and nucleatum (bacterial species information KB2) is 29% (ratio information RB2). Gingivalis (bacterial species information KB3) is present at a ratio of 21% (ratio information RB3), and mutans (bacterial species information KB4) is present at a ratio of 31% (ratio information RB4). show. The correspondence information 232a is associated with the bacterial species information KB and the ratio information RB for each combination of the short base sequence information SB and the short base ratio information RSB. Correspondence information 232a is an example of reference information.

Returning to FIG. 2, the determination unit 221a searches for the corresponding information 232a using the acquired short base sequence information SB and short base ratio information RSB as search keys. The determination unit 221a determines (extracts) the bacterial species information KB and the ratio information RB associated with the short base sequence information SB and the short base ratio information RSB that match the search key from the corresponding information 232a. The determination unit 221a supplies the determined bacterial species information KB and the ratio information RB to the output unit 222a.
The output unit 222a supplies the halitosis determination device 10 with the bacterial species information KB and the ratio information RB acquired from the determination unit 221a.

In the above description, the case where the short base information 231a is information indicating a combination of five nucleotides has been described, but the present invention is not limited to this. The short base information 231a may be information indicating a base sequence of any length as long as it is information indicating a combination of at least 5 or more nucleotides.
Further, the sequencer 21a may be configured to supply the base sequence information indicating the DNA base sequence of the bacteria existing in the saliva SV to the control unit 22a. In this case, the control unit 22a may include a functional unit (for example, a detection unit) that detects the type of the short base sequence based on the base sequence information acquired from the sequencer 21a and the short base information 231a. Further, the control unit 22a may include a functional unit (for example, a ratio acquisition unit) for acquiring the amount (ratio) of the short base sequence in saliva SV for each type of detected short base sequence.

<About the bacterial species ratio detection device 20b>
Hereinafter, the details of the bacterial species ratio detecting device 20b for detecting the bacterial species information KB and the ratio information RB based on the abundance ratio of the 16S rRNA gene will be described with reference to the drawings.
FIG. 4 is a functional configuration diagram showing an example of the configuration of the bacterial species ratio detection device 20b that detects the bacterial species information KB and the ratio information RB based on the abundance ratio of the 16S rRNA gene of the present embodiment.
The bacterial species ratio detecting device 20b includes a sequencer 21b, a control unit 22b, and a storage unit 23b. For example, known base information 231b and correspondence information 232b are stored in advance in the storage unit 23b.

The known base information 231b is information indicating the DNA base sequence to be analyzed by the sequencer 21b, and is information indicating the base sequence of the known 16S rRNA gene. In the following description, the base sequence of the known 16S rRNA gene will be referred to as a known base sequence.
The sequencer 21b analyzes the base sequence to be analyzed indicated by the known base information 231b among the DNA base sequences of the bacteria existing in saliva SV (in the oral cavity of the subject ET). The sequencer 21b aligns and supplies information indicating the type of the obtained known base sequence (hereinafter, known base sequence information EB) to the control unit 22b. In addition, the sequencer 21b aligns and provides the control unit 22b with information indicating the amount (ratio) of the obtained known base sequence in saliva SV for each type of the known base sequence (hereinafter, known base ratio information REB). Supply.

The control unit 22b realizes each functional unit by executing a program stored in the storage unit 23b by a processor such as a CPU. The control unit 22b realizes, for example, a determination unit 221b and an output unit 222b as its functional units.

The determination unit 221b acquires the known base sequence information EB and the known base ratio information REB from the sequencer 21b. Based on the known nucleotide sequence information EB, the known nucleotide ratio information REB, and the corresponding information 232b, the determination unit 221b has the bacteria (bacterial species information KB) existing in the oral cavity of the subject ET and the bacteria in the oral cavity. It is determined that the ratio exists (ratio information RB). Correspondence information 232b is information indicating the correspondence between the known base sequence information EB and the known base ratio information REB, and the bacterial species information KB and the ratio information RB.

Hereinafter, the details of the corresponding information 232b will be described with reference to the drawings.
FIG. 5 is a second diagram showing an example of the corresponding information 232b of the present embodiment.
As shown in FIG. 5, the correspondence information 232b is associated with the known base sequence information EB, the known base ratio information REB, the bacterial species information KB, and the ratio information RB.
Here, when the known nucleotide sequence information EB of a certain subject ET and the known nucleotide ratio information REB are known, the type of bacteria existing in the oral cavity of the subject ET (bacterial species information KB) and the bacteria present in the oral cavity. The ratio (ratio information RB) to be used is uniquely determined. The correspondence between the known base sequence information EB and the known base ratio information REB and the bacterial species information KB and the ratio information RB is the known base sequence indicated by the known base information 231b and the bacteria in the saliva SV in which the known base sequence exists. It is performed based on the analysis result of.

In the example shown in FIG. 5, m known base sequence information EBs and (m) known base ratio information REBs corresponding to the known base sequence information EBs are associated with each other as known base information 231b. m is a natural number. Specifically, FIG. 5 shows that "GGACC (known nucleotide sequence information EB1)" is present in saliva SV at a ratio of 0.25% (known nucleotide sequence information REB1), and "CCCCC (known nucleotide sequence information EB2)" is present. ) ”Is present at a ratio of 0.5% (known base ratio information REB2), and…,“ CCTAG (known base sequence information EBm) ”is present at a ratio of 0.25% (known base ratio information REBm). Is shown.
Further, in the example shown in FIG. 5, the above-mentioned known base sequence information EB1 to known base sequence information EBm and known base sequence information EB1 to known base sequence information EBm include bacterial species information KB1 to bacterial species information KB4 and known bases. The ratio information REB1 to the known base ratio information REB4 are associated with the known base information 231b. Specifically, in the known base information 231b shown in FIG. 5, the known base sequence of the known base sequence information EB1 to the known base sequence information EBm exists in the saliva SV at the ratio of the known base ratio information REB1 to the known base ratio information REBm. In this case, actinomycetemcomitans (bacterial species information KB1) is present in the oral cavity of the subject ET at a ratio of 19% (ratio information RB1), and nucleatum (bacterial species information KB2) is 29% (ratio information RB2). Gingivalis (bacterial species information KB3) is present at a ratio of 21% (ratio information RB3), and mutans (bacterial species information KB4) is present at a ratio of 31% (ratio information RB4). show. The correspondence information 232b is associated with the bacterial species information KB and the ratio information RB for each combination of the known base sequence information EB and the known base ratio information REB. Correspondence information 232b is an example of reference information.

Returning to FIG. 4, the determination unit 221b searches for the corresponding information 232b using the acquired known base sequence information EB and the known base ratio information REB as search keys. The determination unit 221b determines (extracts) the bacterial species information KB and the ratio information RB associated with the known base sequence information EB and the known base ratio information REB that match the search key from the corresponding information 232b. The determination unit 221b supplies the determined bacterial species information KB and the ratio information RB to the output unit 222b.
The output unit 222b supplies the halitosis determination device 10 with the bacterial species information KB and the ratio information RB acquired from the determination unit 221b.

The sequencer 21b may be configured to supply the base sequence information indicating the DNA base sequence of the bacteria existing in the saliva SV to the control unit 22b. In this case, the control unit 22b may include a functional unit (for example, a detection unit) that detects the type of the known base sequence based on the base sequence information acquired from the sequencer 21b and the known base information 231b. Further, the control unit 22b may include a functional unit (for example, a ratio acquisition unit) for acquiring the amount (ratio) of the known base sequence in the saliva SV for each type of the detected known base sequence.
Further, the bacterial species ratio detecting device 20 may have a configuration having both the functions of the bacterial species ratio detecting device 20a and the bacterial species ratio detecting device 20b.

Hereinafter, the details of the halitosis determination device 10 will be described with reference to the drawings.
FIG. 6 is a functional configuration diagram showing an example of the configuration of the halitosis determination device 10 of the present embodiment.
As shown in FIG. 6, the halitosis determination device 10 includes a control unit 11 and a storage unit 12.
The learned model 121 is stored in advance in the storage unit 12. The trained model 121 includes teacher bacterial species information (hereinafter, teacher bacterial species information TKB) indicating the type of bacteria existing in the bacterial flora in the oral cavity, and teacher bacterial species information TKB among the bacteria existing in the bacterial flora. This is a model in which the parameters of the activation function of the hidden layer are machine-learned so as to output the result information RT based on the teacher ratio information TRB indicating the ratio of the indicated bacteria for each type of bacteria. The result information RT is information indicating whether or not there is bad breath in the oral cavity. The result information RT shows, for example, "1" when there is bad breath in the oral cavity and "0" when there is no bad breath in the oral cavity. The case where there is halitosis in the oral cavity is, for example, a case where the concentration of a gas (for example, methyl mercaptan) that causes halitosis in the exhaled breath is equal to or higher than a predetermined threshold value. Therefore, in the result information RT, when it is estimated that the concentration of the gas (hereinafter, specific gas) that causes halitosis contained in the exhaled breath is equal to or higher than a predetermined threshold value based on the bacterial species information KB and the ratio information RB. When it indicates "1" and is estimated to be lower than a predetermined threshold value, it indicates "0". Machine learning is, for example, SVM (Support Vector Machine) or deep learning. In an example of this embodiment, a case where the trained model 121 is trained by deep learning will be described.

The trained model 121 is trained based on the saliva SVs of a plurality of subject ETs for which the result information RT has been acquired in advance by detecting the concentration of the specific gas contained in the exhaled breath, for example. Specifically, in the trained model 121, based on the saliva SV, the bacterial species information KB and the proportion information RB acquired by the bacterial species ratio detecting device 20 having the above-described configuration are used as the teacher bacterial species information TKB and the teacher ratio information, respectively. It is learned as TRB.
The trained model 121 calculates the result information RT based on, for example, the teacher bacterial species information TKB and the teacher ratio information TRB, and the parameters of the activation function of the hidden layer. Further, in the trained model 121, when the result information RT calculated based on the teacher bacterial species information TKB and the teacher ratio information TRB is incorrect, the parameter of the activation function of the hidden layer is adjusted by the error back propagation method.

The result information RT may be information indicating the value of the concentration of the specific gas contained in the exhaled breath. In this case, the trained model 121 is trained based on the saliva SVs of a plurality of subject ETs for which the result information RT indicating the concentration of the specific gas contained in the exhaled breath has been acquired in advance.

The control unit 11 realizes each functional unit by executing a program stored in the storage unit 12 by a processor such as a CPU. The control unit 11 realizes, for example, the acquisition unit 111, the calculation unit 112, and the output unit 113 as its functional units.

The acquisition unit 111 acquires the bacterial species information KB and the ratio information RB from the bacterial species ratio detecting device 20. The acquisition unit 111 supplies the acquired bacterial species information KB and ratio information RB to the calculation unit 112.
The calculation unit 112 calculates the result information RT based on the acquired bacterial species information KB and ratio information RB and the learned model 121. The calculation unit 112 supplies the calculated result information RT to the output unit 113.
The output unit 113 acquires the result information RT from the calculation unit 112. The output unit 113 outputs the acquired result information RT. The output unit 113 outputs, for example, an image showing the result information RT to a display device such as a display panel, and displays the image. Further, the output unit 113 transmits the result information RT to, for example, a server device that stores the test result of the subject ET.

Hereinafter, the operation of the halitosis determination system 1 will be described with reference to the drawings.
FIG. 7 is a flow chart showing an example of the operation of the halitosis determination system 1 of the present embodiment.
The sequencer 21 included in the bacterial species ratio detection device 20 determines the type of DNA base sequence existing in the saliva SV and the ratio of the type of base sequence present in the saliva SV for each type based on the saliva SV of the subject ET. Analyze (step S110). Specifically, the sequencer 21a analyzes the short base sequence information SB and the short base ratio information RSB based on the saliva SV of the subject ET and the short base information 231a. Further, the sequencer 21b analyzes the known base sequence information EB and the known base ratio information REB based on the saliva SV of the subject ET and the known base information 231b. The determination unit 221 determines the bacterial species information KB and the ratio information RB based on the information analyzed by the sequencer 21 and the correspondence information 232 (step S120). Specifically, the determination unit 221a determines the bacterial species information KB and the ratio information RB associated with the acquired short base sequence information SB and short base ratio information RSB from the corresponding information 232a. Further, the determination unit 221 determines the bacterial species information KB and the ratio information RB to which the acquired known base sequence information EB and the known base ratio information REB are associated with the corresponding information 232b. The output unit 222 outputs the bacterial species information KB and the ratio information RB determined by the determination unit 221 to the halitosis determination device 10 (step S130).

The acquisition unit 111 included in the halitosis determination device 10 acquires the bacterial species information KB and the ratio information RB from the bacterial species ratio detecting device 20 (step S140). The calculation unit 112 calculates the result information RT based on the bacterial species information KB and the ratio information RB acquired by the acquisition unit 111 and the learned model 121 (step S150). The output unit 113 outputs the result information RT calculated by the calculation unit 112 (step S160).

As described above, the halitosis determination system 1 of the present embodiment includes a halitosis determination device 10 and a bacterial species ratio detection device 20. The halitosis determination device 10 includes a control unit 11, and realizes the acquisition unit 111, the calculation unit 112, and the output unit 113 as its functional units. The acquisition unit 111 is information detected by the bacterial species ratio detection device 20, and includes bacterial species information KB indicating the type of bacteria existing in the oral bacterial flora (saliva SV in this example) of the subject ET, and bacteria. The ratio information RB indicating the ratio of the bacteria in the saliva SV indicated by the species information KB in the saliva SV for each type is acquired. The calculation unit 112 calculates based on the bacterial species information KB acquired by the acquisition unit 111, the ratio information RB acquired by the acquisition unit 111, and the learned model 121. The output unit 113 outputs halitosis information (result information RT in this example) indicating whether or not the subject ET has halitosis, which is information indicating the calculation result of the calculation unit 112. Here, the trained model 121 includes a teacher bacterial species information TKB indicating the type of bacteria existing in the saliva SV, and a teacher ratio information TRB indicating the proportion of the bacteria in the saliva SV indicated by the teacher bacterial species information TKB for each type. It is a trained model trained by machine learning so as to output the result information RT based on.

As a result, the halitosis determination device 10 of the present embodiment determines whether or not the subject ET has halitosis based on the saliva SV. Therefore, the halitosis determination device 10 of the present embodiment can determine whether or not the subject ET has halitosis without using a device that detects the gas contained in the exhaled breath. Further, the halitosis determination device 10 of the present embodiment cultures the bacteria existing in the oral cavity and has halitosis in order to determine whether or not there is halitosis based on the bacteria existing in the saliva SV and the ratio of the bacteria. Compared with the method of determining whether or not the subject ET has bad breath, it can be determined in a short time.

Further, the trained model 121 of the present embodiment is a model learned by deep learning. Here, even when bacteria that generate gas that causes halitosis are present in the oral cavity, it is not always the case that there is halitosis (the concentration of gas that causes halitosis in the exhaled breath increases). Specifically, the conditions under which halitosis occurs are related to the type of bacteria present in the oral cavity and the ratio of each type of bacteria present in the oral cavity. However, it is difficult to analyze all combinations of the types of bacteria and the ratios for each type as reference information (short base information 231a and known base information 231b).
The halitosis determination device 10 of the present embodiment can calculate the result information RT by the learned model 121 and determine whether or not the subject ET has halitosis by a simple process.

Further, the teacher bacterial species information TKB and the teacher ratio information TRB of the present embodiment are information based on the results of analyzing saliva SVs of a plurality of subject ETs. Specifically, the teacher bacterial species information TKB is a plurality of bacterial species information KB detected by the bacterial species ratio detecting device 20 based on a plurality of saliva SVs. Further, the teacher ratio information TRB is a plurality of ratio information RBs detected by the bacterial species ratio detection device 20 based on a plurality of saliva SVs. When the bacterial flora used when acquiring the teacher bacterial species information TKB and the teacher ratio information TRB is saliva SV, the presence or absence of halitosis can be determined in a short period of time as compared with the case where the bacterial flora is exhaled. ..

Further, the bacterial species ratio detecting device 20a of the present embodiment includes a sequencer 21a, a control unit 22a, and a storage unit 23a. The control unit 22a realizes a determination unit 221a and an output unit 222a as its functional units. The sequencer 21a detects the type of short base sequence present in saliva SV. In addition, the sequencer 21a acquires the ratio of the short base sequence of the detected type in the saliva SV for each type of the detected short base sequence. The determination unit 221a includes the bacterial species information KB and the ratio information RB based on the short base information 231a, the short base sequence information SB, and the short base ratio information RSB for each short base sequence indicated by the short base sequence information SB. To judge.

Here, the DNA base sequence contained in the saliva SV is affected by fluctuations in the number of copies, chimera formation, errors due to the polymerase chain reaction method, horizontal gene transfer between bacteria, etc., and the bacterial species information KB and the ratio information RB are determined. It may be difficult to specify it as the base sequence used when performing. The bacterial species ratio detecting device 20a of the present embodiment determines the bacterial species information KB and the ratio information RB based on the short base sequence contained in the saliva SV. Since the short base sequence has a small number of nucleotides to be determined (5 bases in this example), it is affected by fluctuations in the number of copies, chimera formation, errors due to the polymerase chain reaction, horizontal gene transfer between bacteria, and the like. Can be reduced. Therefore, the bacterial species ratio detection device 20 of the present embodiment accurately determines the bacterial species information KB and the ratio information RB, and the determined bacterial species information KB and the ratio information RB are used to inform the halitosis determination device 10 of the bad breath of the subject ET. The presence or absence can be determined.

Further, the bacterial species ratio detecting device 20b of the present embodiment includes a sequencer 21b, a control unit 22b, and a storage unit 23b. The control unit 22b realizes the determination unit 221b and the output unit 222b as its functional units. The sequencer 21b detects the type of known base sequence present in saliva SV. In addition, the sequencer 21b acquires the ratio of the known base sequence of the detected type in the saliva SV for each type of the detected known base sequence. The determination unit 221b includes the bacterial species information KB and the ratio information RB based on the known base information 231b, the known base sequence information EB, and the known base ratio information REB for each known base sequence indicated by the known base sequence information EB. To judge.

Here, when the types of base sequences used for determining the bacterial species information KB and the ratio information RB are enormous, it is difficult to reduce the time required for the process of determining the bacterial species information KB and the ratio information RB. There is a possibility. The bacterial species ratio detection device 20b of the present embodiment has a known base sequence (for example, 16SrRNA gene) of a bacterium that exists in the oral cavity and may generate a specific gas among the DNA base sequences contained in saliva SV. Based on this, the bacterial species information KB and the ratio information RB are determined. Therefore, the bacterial species ratio detection device 20 of the present embodiment determines the bacterial species information KB and the ratio information RB by a simple process, and the determined bacterial species information KB and the ratio information RB are used to connect the subject ET to the halitosis determination device 10. The presence or absence of bad breath can be determined.

The halitosis determination device 10 may have a configuration for re-learning the learned model 121. In this case, the control unit 11 of the halitosis determination device 10 includes an input unit, a determination unit, and an update unit as its functional units. The input unit acquires test information indicating whether or not the subject ET has bad breath.
The input unit is, for example, a device for inputting character information such as a keyboard, a pointing device such as a mouse or a touch panel, and the like. The determination unit determines whether or not the result information RT and the inspection information match.
When the determination unit determines that the result information RT and the inspection information do not match, the update unit relearns (updates) the trained model 121. The re-learning of the trained model 121 is to update the teacher data (teacher bacterial species information TKB and teacher ratio information TRB) to the latest one and relearn the model. Specifically, the update unit sets the bacterial species information KB and the ratio information RB of the test subject ET in which the result information RT and the test information do not match as the teacher bacterial species information TKB and the teacher ratio information TRB. In addition, the update unit adjusts the parameters of the activation function of the hidden layer of the trained model 121 by the error back propagation method so that the result information RT based on the teacher bacterial species information TKB and the teacher ratio information TRB matches the test information. do. Here, the update of the update unit may be performed every time the determination unit determines that the result information RT and the inspection information do not match, and the bacteria determined that the result information RT and the inspection information do not match. It may be performed when the species information KB and the ratio information RB are acquired in an updatable number of parameters of the activation function of the hidden layer. Further, in updating the update unit, the bacterial species information KB and the ratio information RB of the test subject ET in which the result information RT and the test information match may be used as the teacher bacterial species information TKB and the teacher ratio information TRB. ..
Further, the halitosis determination device 10 may be configured to include a functional unit (importance addition unit) that weights the bacterial species information KB and the known base ratio information REB. In this case, the importance addition unit gives high weight to the bacterial species information KB and the known base ratio information REB, which are important when determining the presence or absence of bad odor, among the bacterial species information KB and the known base ratio information REB. Low weighting is applied to non-important bacterial species information KB and known base ratio information REB.

<Modification example>
In the above description, the case where the halitosis determination device 10 outputs the result information RT based on the bacterial species information KB and the ratio information RB has been described, but the present invention is not limited to this. The halitosis determination device 10 may be configured to output the result information RT based on the known base sequence information EB and the known base ratio information REB.
In this case, the trained model 121 is trained based on the saliva SVs of a plurality of subject ETs for which the result information RT has been acquired in advance. Specifically, the trained model 121 provides teacher data (hereinafter, teacher) on the known base sequence information EB and the known base ratio information REB acquired by the bacterial species ratio detection device 20b having the above-described configuration based on the saliva SV. It is learned as known base sequence information TEB and teacher known base ratio information TREB). The trained model 121 calculates the result information RT based on the teacher known base sequence information TEB and the teacher known base ratio information TREB, and the parameters of the activation function of the hidden layer. Further, in the trained model 121, when the result information RT calculated based on the teacher known base sequence information TEB and the teacher known base ratio information TREB is incorrect, the parameters of the activation function of the hidden layer are adjusted by the error back propagation method. NS.
In this case, the bacterial species ratio detecting device 20b supplies the known base sequence information EB and the known base ratio information REB to the halitosis determination device 10 instead of the bacterial species information KB and the ratio information RB. The halitosis determination device 10 outputs the result information RT based on the acquired known base sequence information EB and known base ratio information REB, and the learned model 121.
As a result, in the halitosis determination system 1 of the modified example, the process of detecting the bacterial species information KB and the proportion information RB by the bacterial species ratio detecting device 20b can be omitted, so that the processing related to the output of the result information RT can be performed at a higher speed. Can be done.

Each part of the halitosis determination device 10 and the bacterial species ratio detection device 20 in each of the above embodiments may be realized by dedicated hardware, or may be realized by a memory and a microprocessor. There may be.

Each part of the halitosis determination device 10 and the bacterial species ratio detection device 20 is composed of a memory and a CPU (central processing unit), in order to realize the functions of each part of the halitosis determination device 10 and the bacterial species ratio detection device 20. The function may be realized by loading and executing the program of.

Further, a program for realizing the functions of each part included in the halitosis determination device 10 and the bacterial species ratio detection device 20 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into the computer system. , You may perform the process by executing. The term "computer system" as used herein includes hardware such as an OS and peripheral devices.

In addition, the "computer system" includes the homepage providing environment (or display environment) if the WWW system is used.
Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, or a storage device such as a hard disk built in a computer system. Further, a "computer-readable recording medium" is a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, and dynamically holds the program for a short period of time. In that case, it also includes the one that holds the program for a certain period of time, such as the volatile memory inside the computer system that is the server or client. Further, the above-mentioned program may be a program for realizing a part of the above-mentioned functions, and may be a program for realizing the above-mentioned functions in combination with a program already recorded in the computer system.

Although the embodiment of the present invention has been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment and may be appropriately modified without departing from the spirit of the present invention. can. The configurations described in each of the above-described embodiments may be combined.

1 ... Smell odor determination system, 10 ... Smell odor determination device, 20, 20a, 20b ... Bacterial species ratio detection device, 21, 21a, 21b ... Sequencer, 11, 22a, 22b ... Control unit, 12, 23a, 23b ... Storage unit, 111 ... Acquisition unit, 112 ... Calculation unit, 113, 222, 222a, 222b ... Output unit, 221, 221a, 221b ... Judgment unit, 121 ... Learned model, 231a ... Short base information, 231b ... Known base information, 232, 232a, 232b ... Correspondence information, KB, KB1, KB2, KB3, KB4 ... Bacterial species information, RB, RB1, RB2, RB3, RB4 ... Ratio information, EB, EB1, EB2, EB3 ... Known base sequence information, REB, REB1 , REB2, REB3 ... Known base ratio information, SB, SB1, SB2, SB3 ... Short base sequence information, RSB, RSB1, RSB2, RSB3 ... Short base ratio information, RT ... Result information, TKB ... Teacher bacterial species information, TRB ... Teacher ratio information, ET ... subject, SV ... saliva

Claims (4)

A detector that detects the type of short base sequence that is the base sequence of DNA existing in the saliva flora of the subject and has five bases.
For each type of the short base sequence detected by the detection unit, a ratio acquisition unit for acquiring the ratio of the base sequence of the type present in the bacterial flora, and a ratio acquisition unit.
Reference information indicating the correspondence between the type of bacteria existing in the bacterial flora in the oral cavity of the subject and the ratio of the bacteria in the bacterial flora for each combination of the type and the ratio, and the short length detected by the detection unit. Based on the type of the base sequence and the ratio of each type of the short base sequence acquired by the ratio acquisition unit, the type of bacteria existing in the bacterial flora in the oral cavity of the subject and the bacterial flora in the bacterial flora. A determination unit that determines the ratio, outputs the information indicating the type as the result of the determination as bacterial species information, and outputs the information indicating the ratio as the ratio information.
Halitosis in the oral cavity based on teacher bacterial species information indicating the type of bacteria existing in the bacterial flora in the oral cavity and teacher ratio information indicating the proportion of bacteria in the bacterial flora indicated by the teacher bacterial species information in the bacterial flora for each type. A storage unit that stores the trained model learned by machine learning so as to output halitosis information indicating whether or not there is
Said species information the determination unit outputs an acquiring unit that acquires a pre-Symbol percentage information,
The bacterial species information acquired by the acquisition unit and the ratio information acquired by the acquisition unit are input to the trained model, and the calculated halitosis information indicating whether or not the subject has halitosis is output. Output section and
Halitosis determination device. What is machine learning?
Deep learning,
The halitosis determination device according to claim 1. The teacher bacterial species information and the teacher ratio information are
Information based on the results of saliva analysis of multiple subjects,
The halitosis determination device according to claim 1 or 2. Halitosis in the oral cavity based on teacher bacterial species information indicating the type of bacteria existing in the bacterial flora in the oral cavity and teacher ratio information indicating the proportion of bacteria in the bacterial flora indicated by the teacher bacterial species information in the bacterial flora for each type. To a computer equipped with a storage unit that stores a trained model learned by machine learning so as to output halitosis information indicating whether or not there is
A detection step in which a detection unit that detects a type of short base sequence having five bases in the base sequence of DNA existing in the saliva bacterial flora of a subject detects the type of the short base sequence.
For each type of the short base sequence detected in the detection step, a ratio acquisition step of acquiring the ratio of the base sequence of the type present in the bacterial flora, and a ratio acquisition step.
Reference information indicating the correspondence between the type of bacteria existing in the bacterial flora in the oral cavity of the subject and the ratio of the bacteria in the bacterial flora for each combination of the type and the ratio, and the said detection detected in the detection step. Based on the type of the short base sequence and the ratio of each type of the short base sequence acquired in the ratio acquisition step, the type of bacteria existing in the bacterial flora in the oral cavity of the subject and the bacterium of the bacterium. A determination step in which the proportion in the flora is determined, the information indicating the type, which is the result of the determination, is used as bacterial species information, and the information indicating the proportion is output as the proportion information.
An acquisition step of acquiring bacterial species information indicating the type of bacteria existing in the bacterial flora in the oral cavity of a subject and proportion information indicating the proportion of bacteria in the bacterial flora indicated by the bacterial species information in the bacterial flora for each type.
Output that inputs the bacterial species information acquired in the acquisition step and the ratio information acquired in the acquisition step into the trained model and outputs halitosis information indicating whether or not the calculated subject has halitosis. Steps and
A program that executes.

Images