JP6804489B2 - Information providing device and information providing method - Google Patents

Description

The present invention relates to an information providing device and an information providing method.

Conventionally, a technique for relating the state of human microRNA (miRNA) to the presence of cancer has been known (see, for example, Patent Document 1). Patent Document 1 discloses a technique for relating the expression ratio of miRNA pairs to the presence of lung cancer.

Special Table 2017-502699

However, in the above-mentioned conventional technique, the state of miRNA can be used to contribute to the detection of cancer, but the state of miRNA cannot be used to contribute to the prevention of cancer.
In recent years, imbalances in miRNA profiles have been known to promote cancer. Therefore, an information providing service that contributes to improving the balance of the user's miRNA profile by presenting the imbalance of the user's miRNA profile to the user can be considered. However, the miRNA profile is complicated because it is a profile of miRNA in which about 2500 types exist, and it is difficult for a general user to understand the miRNA profile.

The present invention has been made in consideration of such circumstances, and an object of the present invention is to provide an information providing device and an information providing method capable of simply presenting the balance state of a user's miRNA profile. It is in.

In order to achieve such an object, the present invention is grasped by the following configuration.
(1) One aspect of the present invention is learned to output a Q value by compressing the number of dimensions of miRNA information to a smaller number based on the miRNA attribute of the information provider and the multidimensional miRNA information. A first neural network and an information output unit that outputs a miRNA score indicating the good or bad balance of the user's miRNA profile corresponding to the Q value using the first neural network from the user's miRNA information. The information output unit is an information providing device that derives the miRNA score from the Q value output from the first neural network by a predefined function .
( 2 ) One aspect of the present invention is an information providing device in which the miRNA attribute includes the race of the information provider in the configuration of the above (1).
( 3 ) One aspect of the present invention is an information providing device in the configuration of ( 2 ) above, wherein the miRNA attribute includes the age and sex of the information provider.
( 4 ) In one aspect of the present invention, in any of the configurations (1) to ( 3 ) above, the first neural network is further subjected to the learning by further adding the biochemical data of the information provider. It is an information providing device.
( 5 ) In one aspect of the present invention, in the configuration of ( 4 ) above, the information output unit corresponds to a Q value from the user's miRNA information and biochemical data by using the first neural network. It is an information providing device that outputs the user's miRNA score.
( 6 ) In one aspect of the present invention, in the configuration of (1) above , ideal eating information is obtained based on the Q value output from the first neural network and the eating / intake information of the information provider for a certain period of time. A second neural network learned to output is further provided, and the information output unit uses the second neural network from the Q value output by using the first neural network from the miRNA information of the user. It is an information providing device that outputs the eating information recommended to the user.

( 7 ) One aspect of the present invention is an information providing method for providing a user's miRNA score by an information providing device having any of the above configurations (1) to ( 6 ).

According to the present invention, it is possible to provide an information providing device and an information providing method capable of simply presenting the balance state of the user's miRNA profile.

It is the schematic which shows the information providing system which concerns on 1st Embodiment of this invention. It is a figure which shows the learning method of the neural network which concerns on Example 1 of the information providing apparatus of 1st Embodiment of this invention. It is a figure which shows the information providing method which concerns on Example 1 of the information providing apparatus of 1st Embodiment of this invention. It is a figure which shows the learning method of the neural network which concerns on Example 2 of the information providing apparatus of 1st Embodiment of this invention. It is a figure which shows an example of the information providing method which concerns on Example 2 of the information providing apparatus of 1st Embodiment of this invention. It is the schematic which shows the information providing system which concerns on 2nd Embodiment of this invention. It is a figure which shows the learning method of the neural network which concerns on embodiment of the information providing apparatus of 2nd Embodiment of this invention. It is a figure which shows the information providing method which concerns on embodiment of the information providing apparatus of 2nd Embodiment of this invention.

Hereinafter, embodiments for carrying out the present invention (hereinafter referred to as “embodiments”) will be described in detail with reference to the accompanying drawings. The same elements are designated by the same reference numerals throughout the description of the embodiments.

[First Embodiment]
FIG. 1 is a schematic view showing an information providing system according to the first embodiment. The information providing system 1 shown in FIG. 1 includes an information providing device 10 and a terminal device 20. The information providing device 10 and the terminal device 20 may send and receive information online by communication, or may input and output information offline. Hereinafter, as an example, the information providing device 10 and the terminal device 20 will be described as transmitting and receiving information online by communication.
The information providing device 10 is configured as a device for providing the miRNA score of the individual by using the miRNA information of the individual.

The miRNA score is information indicating whether the miRNA profile is well-balanced or not. Examples of the miRNA score include a miRNA score, a miRNA age, and a miRNA deviation value. The miRNA score is a score obtained by scoring the good or bad balance of the miRNA profile with a specified perfect score (for example, a perfect score of 100 points). The miRNA age represents the good or bad balance of the miRNA profile by comparing it to age. The miRNA deviation value expresses the good or bad balance of the miRNA profile by the deviation value.

The individual miRNA information is information indicating the state of the individual's miRNA. There are approximately 2500 types of miRNA. The miRNA information may include information on all types of miRNAs, or may include information on some types of miRNAs. The miRNA information is information including information on each miRNA. The miRNA information may be any of the concentration, number of moles, fluorescence intensity or other index-corrected relative value of each miRNA. The expression method of miRNA information may be a logarithmic expression.

The terminal device 20 transmits information such as user miRNA information to the information providing device 10. The terminal device 20 may be a user's terminal device, or may be a terminal device other than the user's terminal device. In addition, instead of the terminal device 20, the server of the information providing organization may transmit information such as the user's miRNA information to the information providing device 10.
The information providing device 10 includes a first neural network 11 and an information output unit 12. The information output unit 12 of the information providing device 10 uses the first neural network 11 to output information such as the user's miRNA score from the information such as the user's miRNA information received from the terminal device 20. The information providing device 10 transmits information such as the user's miRNA score to the terminal device 20.

The information providing device 10 is composed of a memory, a CPU (Central Processing Unit), and the like. The functions of each part of the information providing device 10 are realized by the CPU included in the information providing device 10 executing a computer program. The information providing device 10 may be configured by using a general-purpose computer device, or may be configured as a dedicated hardware device.

Further, an input device, a display device, or the like may be connected to the information providing device 10 as peripheral devices. Here, the input device refers to an input device such as a keyboard and a mouse. The display device refers to a CRT (Cathode Ray Tube), a liquid crystal display device, or the like.
Further, the peripheral device may be directly connected to the information providing device 10, or may be connected via a communication line.
Further, the information providing device 10 may include a database related to the neural network, or may access the database provided externally via a communication line.

The terminal device 20 may be, for example, a mobile terminal device such as a smartphone or a tablet computer (tablet PC), or a stationary terminal device (for example, a stationary personal computer). May be good.

Next, the information providing device 10 according to the first embodiment will be described with reference to some examples.

[Example 1]
The first embodiment of the information providing device 10 according to the first embodiment will be described with reference to FIGS. 2 and 3. FIG. 2 is a diagram showing a learning method of a neural network according to a first embodiment of the information providing device 10 of the first embodiment. FIG. 3 is a diagram showing an information providing method according to the first embodiment of the information providing device 10 of the first embodiment.

<Learning stage>
The learning stage according to the first embodiment will be described with reference to FIG. The learning stage is a stage of learning the first neural network 11 used in the information providing stage described later. The learning stage of the first neural network 11 may be performed by the information providing device 10, or may be performed by a computer different from the information providing device 10.

In FIG. 2, the learning stage (S100) includes three steps S101, S102 and S103. In the learning stage (S100), the first neural network outputs the Q value by compressing the number of dimensions of the miRNA information to a smaller number based on the miRNA attribute of the information provider and the multidimensional miRNA information. 11 learning is performed.

An individual's miRNA attribute is an attribute of the individual that affects the individual's miRNA information. Examples of miRNA attributes include age, gender, race, and the like. The miRNA attributes preferably include at least age and gender. The miRNA attributes may be only age and gender. In addition, nationality may be used as information indicating race.

The Q value is information that represents the multidimensional miRNA information of the information provider input to the first neural network 11 with a smaller number of dimensions.

Multidimensional miRNA information is information indicating the state of a plurality of types of miRNA. It is known that there are approximately 2500 types of human miRNAs. Therefore, the number of dimensions of miRNA information indicating the state of all types of human miRNAs is very large. On the other hand, the types of miRNAs that have a strong correlation with future cancer development risk (cancer prevention) are considered to be limited. Therefore, in the learning stage (S100), the first neural network 11 is learned so that the number of dimensions of the miRNA information is compressed to a smaller number of dimensions and the Q value is output.

(Step S101) As the input information to be input to the first neural network 11, the miRNA attribute of the information provider and the multidimensional miRNA information are prepared. An information provider is a person who provides information for learning a neural network. Next, the miRNA attribute of the information provider and the multidimensional miRNA information are input to the first neural network 11. For example, the age and gender of the information provider and the multidimensional miRNA information are input to the first neural network 11. By using age and gender as miRNA attributes, it is possible to learn the first neural network 11 having different coefficients depending on age and gender.

(Step S102) As a result of inputting the miRNA attribute of the information provider and the multidimensional miRNA information into the first neural network 11, the Q value is output from the first neural network 11.

(Step S103) Learning of the first neural network 11. As a neural network learning method, for example, an autoencoder (self-encoder) may be used. The autoencoder has the "information provider's miRNA attribute and multidimensional miRNA information" reconstructed by applying an inverse encoding (decoding layer) to the Q value of the output of the first neural network 11 (encoding layer). , The first neural network 11 is trained so that the difference between the original input "miRNA attribute of the information provider and the multidimensional miRNA information" is small, and the correct answer data / non-correction data is performed as in the supervised learning. There is no need for a person to give meaning such as correct answer data, and it is possible to automatically calculate a meaningful Q value.

As an example of the learning method according to the present embodiment, self-encoding that returns the Q value of the M dimension (however, M <N) generated from the first miRNA information of the N dimension by the first neural network 11 to the second miRNA information of the N dimension. Is performed, and the first neural network 11 is trained so that the second miRNA information matches the first miRNA information.

<Information provision stage>
The information providing stage according to the first embodiment will be described with reference to FIG. The information providing stage is a stage in which the user's miRNA score is provided from the user's miRNA information by using the first neural network 11 that has been learned by the learning stage described above.

In FIG. 3, the information providing step (S200) includes three steps S201, S202 and S203. In the information providing stage (S200), the information providing device 10 includes a first neural network 11 that has been learned by the learning stage described above. That is, the information providing device 10 is learned to output the Q value by compressing the number of dimensions of the miRNA information to a smaller number of dimensions based on the miRNA attribute of the information provider and the multidimensional miRNA information. 1 The neural network 11 is provided.

(Step S201) The information providing device 10 receives the user's miRNA information from the terminal device 20. The information output unit 12 inputs the user's miRNA information received from the terminal device 20 into the first neural network 11.

(Step S202) As a result of inputting the user's miRNA information into the first neural network 11, a Q value is output from the first neural network 11.

(Step S203) The information output unit 12 converts the Q value output from the first neural network 11 into a miRNA score. As a method of converting the Q value into a miRNA score, for example, a conversion table in which the miRNA score corresponding to the Q value is recorded may be provided in advance, or a function for deriving the miRNA score from the Q value may be defined in advance. May be good.

The information providing device 10 transmits the user's miRNA score to which the Q value output from the first neural network 11 has been converted to the terminal device 20. The terminal device 20 records the user's miRNA score received from the information providing device 10, and displays the user's miRNA score on the display screen in response to the information browsing operation by the user. As a result, the user can view his / her miRNA score on the terminal device 20. The miRNA score can concisely present the good or bad balance of the miRNA profile.

[Example 2]
Example 2 is a modification of Example 1 described above. In Example 2, in addition to the miRNA information, biochemical data is further used to provide a user's miRNA score. The second embodiment of the information providing device 10 according to the present embodiment will be described with reference to FIGS. 4 and 5. FIG. 4 is a diagram showing a learning method of the neural network according to the second embodiment of the information providing device 10 of the first embodiment. FIG. 5 is a diagram showing an example of an information providing method according to the second embodiment of the information providing device 10 of the first embodiment.

<Learning stage>
The learning stage according to the second embodiment will be described with reference to FIG. For convenience of explanation, in order to distinguish it from the first neural network 11 of the first embodiment, it is referred to as the first neural network 11a in the second embodiment. In FIG. 4, the learning stage (S100a) includes three steps S101a, S102a and S103a.

(Step S101a) As the input information to be input to the first neural network 11a, the biochemical data of the information provider is further prepared in addition to the miRNA attribute of the information provider and the multidimensional miRNA information. Biochemical data include, for example, data on the results of biochemical tests on human specimens (eg, blood, urine, stool, etc.), captured images of human skin, ultrasonic images of humans, X-ray images of humans, and humans. Examples include data on the results of doctors' diagnoses. The biochemical data of the information provider may be the data of the test result obtained when the information provider undergoes a health examination such as a human dock.
Next, the miRNA attribute of the information provider, the multidimensional miRNA information, and the biochemical data are input to the first neural network 11a.

(Step S102a) As a result of inputting the miRNA attribute of the information provider, the multidimensional miRNA information, and the biochemical data into the first neural network 11a, a Q value is output from the first neural network 11a.

In the learning stage (S100a), step S103a is the same as step S103 in the learning stage (S100) of Example 1 described above.

<Information provision stage>
The information providing stage according to the second embodiment will be described. Examples of the information providing stage according to the second embodiment include Examples 2-1 and 2-2 of the information providing stages shown below.

<Example of information provision stage 2-1>
Example 2-1 of the information providing stage is the same as the information providing stage according to the first embodiment shown in FIG. 3 described above. However, in Example 2-1 of the information providing stage, in FIG. 3, instead of the first neural network 11, the first neural network 11a that has been learned by the learning stage according to the second embodiment described above is used. Other than this point, the information provision stage is the same as that of the first embodiment shown in FIG. 3, and the description thereof will be omitted.

<Example of information provision stage 2-2>
Example 2-2 of the information providing stage according to the second embodiment will be described with reference to FIG. Example 2-2 of the information providing stage provides the user's miRNA score from the user's miRNA information and biochemical data by using the first neural network 11a learned by the learning stage according to the above-described second embodiment. It is a stage.

In FIG. 5, Example 2-2 (S200a) of the information providing stage includes three steps S201a, S202a and S203a. In Example 2-2 (S200a) of the information providing stage, the information providing device 10 includes a first neural network 11a that has been learned by the learning stage according to the second embodiment described above. That is, the information providing device 10 compresses the number of dimensions of the miRNA information to a smaller number of dimensions based on the miRNA attribute of the information provider, the multidimensional miRNA information, and the biochemical data, and outputs the Q value. A trained first neural network 11a is provided.

(Step S201a) The information providing device 10 receives the user's miRNA information and biochemical data from the terminal device 20. The information output unit 12 inputs the user's miRNA information and biochemical data received from the terminal device 20 into the first neural network 11a.

(Step S202a) As a result of inputting the user's miRNA information and biochemical data into the first neural network 11a, a Q value is output from the first neural network 11a.

(Step S203a) The information output unit 12 converts the Q value output from the first neural network 11a into a miRNA score.

The information providing device 10 transmits the user's miRNA score to which the Q value output from the first neural network 11a has been converted to the terminal device 20. The terminal device 20 records the user's miRNA score received from the information providing device 10, and displays the user's miRNA score on the display screen in response to the information browsing operation by the user. As a result, the user can view his / her miRNA score on the terminal device 20. The miRNA score can concisely present the good or bad balance of the miRNA profile.

The above is the description of the embodiment of the information providing device 10 according to the first embodiment.

According to the first embodiment described above, there is an effect that the state of balance of the user's miRNA profile can be simply presented.

[Second Embodiment]
FIG. 6 is a schematic view showing an information providing system according to the second embodiment. The information providing system 1b shown in FIG. 6 includes an information providing device 10b and a terminal device 20. The information providing device 10b and the terminal device 20 may send and receive information online by communication, or may input and output information offline. Hereinafter, as an example, the information providing device 10b and the terminal device 20 will be described as transmitting and receiving information online by communication. The information providing device 10b is configured as a device for providing the miRNA score of the individual and the eating information recommended to the individual by using the miRNA information of the individual. The recommended dietary information for an individual is information indicating a diet for improving the balance of the individual's miRNA profile.

The terminal device 20 transmits information such as the user's miRNA information to the information providing device 10b. The information providing device 10b includes a first neural network 11, a second neural network 13, and an information output unit 12b. The information output unit 12b of the information providing device 10b uses the first neural network 11 to output information such as the user's miRNA score from the information such as the user's miRNA information received from the terminal device 20. The first neural network 11 according to the second embodiment may be the first neural network 11 according to the first embodiment described above, or the second embodiment according to the first embodiment described above. The first neural network 11a according to the above.

The information output unit 12b further uses the second neural network 13 to inform the user from the Q value output by using the first neural network 11 from the information such as the user's miRNA information received from the terminal device 20. Output recommended eating information. The information providing device 10b transmits information such as a user's miRNA score and eating information recommended to the user to the terminal device 20.

The information providing device 10b has the same configuration as the information providing device 10 of the first embodiment, and the functions of each part of the information providing device 10b are realized by the CPU included in the information providing device 10b executing a computer program. To.

Next, an example of the information providing device 10b according to the second embodiment will be described with reference to FIGS. 7 and 8. FIG. 7 is a diagram showing a learning method of a neural network according to an embodiment of the information providing device 10b of the second embodiment. FIG. 8 is a diagram showing an information providing method according to an embodiment of the information providing device 10b of the second embodiment.

<Learning stage>
In the second embodiment, the learning of two neural networks, the first neural network 11 and the second neural network 13, is performed. The learning method of the first neural network 11 is the above-described first embodiment (learning stage (S100) of the first neural network 11 shown in FIG. 2) or the second embodiment (first shown in FIG. 4). Since it is the same as the learning stage (S100a) of the neural network 11a), the description thereof will be omitted.

The learning stage of the second neural network 13 according to the second embodiment will be described with reference to FIG. 7. This learning stage is a stage of learning the second neural network 13 used in the information providing stage described later. The learning stage of the second neural network 13 may be performed by the information providing device 10b, or may be performed by a computer different from the information providing device 10b.

In FIG. 7, the learning stage (S100b) includes three steps S101b, S102b and S103b. In the learning stage (S100b), the Q value output from the first neural network 11 as a result of inputting the miRNA attribute of the information provider and the multidimensional miRNA information into the first neural network 11 and the constant of the information provider. The second neural network 13 is learned so as to output the ideal eating information based on the eating intake information of the period.
The information provider's food intake information for a certain period of time is information on the meals taken by the information provider for a certain period of time. In addition, the information of the specific food component contained in the meal taken by the information provider for a certain period may be included in the eating and intake information, and the information of the specific food component may be learned by the second neural network 13. The specific food ingredient may be a food ingredient that improves the balance of the miRNA profile.
Ideal eating information is information indicating a diet for improving the balance of miRNA profiles.

(Step S101b) Q output from the first neural network 11 as a result of inputting the miRNA attribute of the information provider and the multidimensional miRNA information into the first neural network 11 as the input information to be input to the second neural network 13. Prepare the value and the food intake information of the information provider for a certain period of time. Next, the Q value and the food intake information of the information provider for a certain period of time are input to the second neural network 13.

(Step S102b) The ideal eating information is output from the second neural network 13 as a result of inputting the Q value and the eating intake information of the information provider for a certain period into the second neural network 13.

(Step S103b) Learning of the second neural network 13. As a neural network learning method, for example, reinforcement learning may be performed. Reinforcement learning does not require correct answer data like supervised learning, so even if it takes a long time to actually acquire the correct answer data for the ideal eating information output of the second neural network 13, it is short-term. You can learn between them.

<Information provision stage>
In the information providing stage of the second embodiment, the user's miRNA score and the eating information recommended to the user are provided by using the first neural network 11 and the second neural network 13 that have been learned by the learning stage described above. It is a stage. At this information providing stage, the information providing device 10b includes a first neural network 11 and a second neural network 13 that have been learned by the learning stage described above. That is, the information providing device 10b is learned to output the Q value by compressing the number of dimensions of the miRNA information to a smaller number of dimensions based on the miRNA attribute of the information provider and the multidimensional miRNA information. The first neural network 11 and the second neural network 13 learned to output ideal eating information based on the Q value output from the first neural network 11 and the eating and intake information of the information provider for a certain period of time. And.

The method of providing the user's miRNA score using the first neural network 11 is the above-described first embodiment (information providing step (S200) by the first neural network 11 shown in FIG. 3) or the embodiment. Since it is the same as Example 2 (Example 2-1 or Example 2-2 at the information providing stage), the description thereof will be omitted.

The information providing stage by the second neural network 13 according to the second embodiment will be described with reference to FIG. In FIG. 8, the information providing step (S200b) includes two steps S201b and S202b.

(Step S201b) The information output unit 12 is in the information providing stage of the first embodiment (S200 in FIG. 3) or the information providing stage of the second embodiment (Example 2-1 or Example 2 of the information providing stage) described above. In the same manner as in -2), the Q value output from the first neural network 11 or 11a is input to the second neural network 13.

(Step S202b) As a result of inputting the Q value to the second neural network 13, the eating information recommended to the user is output from the second neural network 13. The information providing device 10 includes the information providing stage of the first embodiment (S200 in FIG. 3) or the information providing stage of the second embodiment (Example 2-1 or Example 2-2 of the information providing stage) described above. Similarly, the miRNA score of the user to which the Q value output from the first neural network 11 or 11a is converted and the eating information recommended to the user output from the second neural network 13 are transmitted to the terminal device 20. The terminal device 20 records the user's miRNA score received from the information providing device 10 and the eating information recommended to the user, and displays the user's miRNA score and the eating information recommended to the user according to the information browsing operation by the user. Display on the screen. As a result, the user can view his / her miRNA score and his / her recommended eating information on the terminal device 20. The miRNA score can concisely present the good or bad balance of the miRNA profile. In addition, the eating information recommended by the user can provide information indicating a meal for improving the balance of the user's miRNA profile.

The above is the description of the embodiment of the information providing device 10b according to the second embodiment.

According to the second embodiment described above, it is possible to concisely present the state of balance of the user's miRNA profile and to present information indicating a diet for improving the balance of the user's miRNA profile. The effect of being able to do it is obtained.

Although the present invention has been described above using the embodiments, it goes without saying that the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various changes or improvements can be made to the above embodiments. Further, it is clear from the description of the scope of claims that such a modified or improved form may be included in the technical scope of the present invention.

1,1b Information providing system 10,10b Information providing device 11, 11a First neural network 12, 12b Information output unit 13 Second neural network 20 Terminal device

Claims (7)

A first neural network learned to output a Q value by compressing the number of dimensions of miRNA information to a smaller number based on the information provider's miRNA attributes and multidimensional miRNA information.
It is provided with an information output unit that outputs a miRNA score indicating the good or bad balance of the user's miRNA profile corresponding to the Q value using the first neural network from the user's miRNA information .
The information output unit derives the miRNA score from the Q value output from the first neural network by a predefined function.
Information providing device. The miRNA attribute includes the race of the informant.
The information providing device according to claim 1 . The miRNA attribute includes the age and gender of the informant.
The information providing device according to claim 2 . In the first neural network, the learning is performed by further adding the biochemical data of the information provider.
The information providing device according to any one of claims 1 to 3 . The information output unit outputs the user's miRNA score corresponding to the Q value from the user's miRNA information and biochemical data using the first neural network.
The information providing device according to claim 4 . A second neural network trained to output ideal eating information based on the Q value output from the first neural network and the eating / intake information of the information provider for a certain period of time is further provided.
The information output unit outputs the eating information recommended to the user by using the second neural network from the Q value output from the miRNA information of the user by using the first neural network.
The information providing device according to claim 1 . By the information providing device according to any one of claims 1 to 6 .
An information providing method that provides a user's miRNA score.

Images