JP7479317B2 - Information processing device, method and program - Google Patents

Description

The present invention relates to an information processing device, method, and program for processing health-related information.

Conventional technology includes a technology (Patent Document 1) that calculates and displays a predicted healthy life expectancy for each individual from the results of a health checkup, and claims that by allowing a quantitative grasp of health status, it can increase motivation to prevent disease and manage one's health. In addition, a method (Patent Document 2) that applies recent machine learning technology to predict health age from health checkup data and physical condition information, and a program (Patent Document 3) that analyzes the factors that affect it have been proposed.

Non-patent documents 1 and 2 report technology that can estimate biological age (or epigenetic age) using information on the epigenome (DNA methylation) that controls gene function.

JP 2007-287184 A JP 2019-145057 A JP 2020-004126 A

Field, Adam E., Neil A. Robertson, Tina Wang, Aaron Havas, Trey Ideker, and Peter D. Adams. "DNA methylation clocks in aging: categories, causes, and consequences." Molecular cell 71, no. 6 (2018): 882-895. Horvath, Steve, and Kenneth Raj. "DNA methylation-based biomarkers and the epigenetic clock theory of aging." Nature Reviews Genetics 19, no. 6 (2018): 371.

However, these health programs are premised on using health checkup results such as blood pressure and blood sugar levels, and therefore have issues with issues such as differences in the content of health checkups, or missing health checkup results due to missed questionnaires or missed tests, making it difficult to respond when there are items that could not be measured. Also, even when biological age is estimated using information on the epigenome, there is an issue in that it is not possible to grasp the specific lifestyle habits that have an impact.

In view of the problems with the above-mentioned conventional technology, the present invention aims to provide an information processing device, method, and program that can efficiently utilize biometric information and lifestyle habit information to obtain health-related information.

To achieve the above object, the present invention is an information processing device that, as a first feature, executes a first process of calculating a biological age by applying a biological age prediction model to the evaluation value of each sample by referring to a database that associates the actual age, the evaluation value of each item of biometric information reflecting lifestyle habits, and the actual value of each item of lifestyle habits for each of a plurality of samples, and a second process of calculating the degree of influence of each item of lifestyle habits on the biological age by referring to the database.

The second feature is that the information processing device further executes a third process of referring to the database and calculating, from samples whose biological ages are determined to be young for each range of actual ages, a model value of the evaluation value of the bioinformation within a certain range of the actual ages, and a fourth process of outputting, for a specified sample, the lifestyle habit items that are determined to be effective in making the biological age of the sample younger by comparing the evaluation value of the sample with the model value of the evaluation value of the bioinformation corresponding to the actual age of the sample. The information processing device is also characterized by a method and program that are compatible with the information processing device.

According to the first feature, the biological age of each sample is calculated from the actual values of the lifestyle habits, and the impact of lifestyle habits on health can be estimated by taking this biological age and biometric information into consideration, so that health-related information can be obtained by efficiently utilizing biometric information and lifestyle habit information via biological age. According to the second feature, it is further possible to output lifestyle habit items that are determined to be effective in making the biological age younger.

FIG. 1 is a schematic diagram for explaining the epigenome. FIG. 1 is a schematic diagram for explaining the epigenome. FIG. 1 is a schematic diagram for explaining DNA methylation rate. FIG. 2 is a functional block diagram of an information processing device according to an embodiment. 11 is a flowchart illustrating an operation of an information processing device according to an embodiment. 2 is a diagram showing a schematic example of data handled by each unit of the information processing device; FIG. FIG. 10 is a diagram showing a schematic example of registered information in a database to be constructed. 13 is an example of coefficients of a prediction model. FIG. 13 is a diagram showing an example of recording model values for a database. FIG. 13 is a diagram showing an example of data for explaining the processing of a lifestyle habit effect analysis unit and a lifestyle habit advice generation unit (processing of steps S3 and S4). FIG. 11 is a diagram for explaining the process of a lifestyle habit influence level analysis unit, as a common example to the example of FIG. 10 etc. FIG. 1 is a diagram illustrating a hardware configuration of a typical computer.

Before describing each embodiment of the present invention, the epigenome (DNA methylation) and DNA methylation rate, which are known medical matters, will be briefly described with reference to Figures 1 to 3. As shown in the upper part of Figure 1, the genome, which is the base sequence of DNA (the sequence of four organic bases, adenine (A), guanine (G), thymine (T), and cytosine (C)), is a genetic characteristic that each individual is born with and can be analyzed as an immutable, innate characteristic. On the other hand, as shown in the lower part of Figure 1 and Figure 2, the epigenome, which controls the function of DNA, is a variable, acquired characteristic that shows changes dependent on the environment and time and can be analyzed.

While DNA (genome) remains unchanged in most cells of each individual, the epigenome, which changes depending on the developmental stage, aging, and the external environment, is known to be responsible for cellular diversity and determine the fate of cells. For example, iPS cells (induced pluripotent stem cells) are created by resetting the epigenome, and aging is associated with the loss, disturbance, or disorder of information in the epigenome.

Cytosine (C) in DNA can be methylated and is involved in gene control (switches). This is particularly common when C and G are consecutive, and the position that can be methylated is called a CG site or CpG site. There are about 26 million CpG sites in humans. In the mock examples in Figures 2 and 3, each CpG site is indicated as CpG _i (i=1,2,…,n,…), with a subscript (i=1,2,…,n,…) added as an identifier.

As mentioned above, the epigenome, including DNA methylation, is known to act as a gene switch. In a hypermethylated state, DNA is folded and genes cannot be read, whereas in a hypomethylated state, DNA can be accessed and genes can be read.

As described above, the epigenome (DNA methylation) contains information on the monitoring of the physical environment of each individual. As shown in the schematic example of Figure 3, the state of DNA methylation may differ for each individual cell (and the organ that this cell constitutes), but it is possible to obtain information on the methylation rate of each site in a large number of DNAs. This methylation rate is a value that reflects the physical environment of each individual in some way. In Figure 3, as shown on the left, in four DNAs in four cells, there are two CpG sites, CpG ₁ and CpG ₂ , and when there is a methylation state as shown in the figure, the methylation rate of each CpG site is calculated as shown on the right (CpG ₁ is 100% for four out of four, and CpG ₂ is 25% for one out of four).

Figure 4 is a functional block diagram of an information processing device according to one embodiment. The information processing device 100 includes an epigenome dataset input unit 11, an epigenome data analysis unit 12, a lifestyle habit data input unit 13, a correspondence storage unit 14, a database 15, a biological age calculation unit 21, a lifestyle habit effect analysis unit 22, and a lifestyle habit advice generation unit 23.

Note that data D1 to D5, which are attached to some of the arrows between the functional blocks in Figure 4, are data that are handled by inputting and outputting between the functional blocks, and data with the same reference numerals in Figures 6, 8, 10, etc., described below, are examples of the data D1 to D5.

Figure 5 is a flowchart of the operation of the information processing device 100 according to one embodiment, and as shown in the figure, the flowchart is composed of steps S1 to S4. As an overall overview of the flowchart, step S1 corresponds to a process of constructing the database 15 by storing information about a large number of samples in the database 15 by the epigenome dataset input unit 11, the epigenome data analysis unit 12, the lifestyle habit data input unit 13, and the correspondence storage unit 14, and steps S2 to S4 correspond to a process of analyzing the database 15 thus constructed by the biological age calculation unit 21, the lifestyle habit effect analysis unit 22, and the lifestyle habit advice generation unit 23 in this order, and extracting various information related to lifestyle habits from the database 15.

The following describes the details of the processing content of each unit of the information processing device 100 in FIG. 4 while explaining each step of the flowchart in FIG. 5. FIG. 6 shows schematic examples of data handled by each unit of the information processing device 100, divided into data examples D1 to D5, and will be referred to as appropriate in the following description.

<Step S1> In step S1, information about a large number of samples (a large number of subjects from whom health-related information (genome-related information and lifestyle information) is to be obtained) is obtained to construct a database 15, and then the process proceeds to step S2. This step S1 can be configured as the following steps 11 to 14. Note that steps 11, 12 and step 13 can be performed in any order, or may be performed in parallel. Step 14 can be performed after steps 11, 12 and step 13 have been completed.

(Step 11) The epigenome data dataset input unit 11 acquires genome-related information for each sample and outputs it to the epigenome data analysis unit 12.

The genome-related information (a schematic example is shown as data example D1 in Figure 6) is in FASTQ format, which is a typical output format of next-generation sequencers, or BAM format mapped to a reference sequence, or a format equivalent thereto, and includes user information including user ID, age (chronological age, chronological age, actual age, which is distinguished from "biological age" described below), sex, and race. The epigenome dataset input unit 11 receives this data, checks the format, and outputs it to the epigenome data analysis unit 12.

If the data input is in FASTQ format, redundant sequence information and data with low sequencing accuracy are removed by checking the quality values contained in the data, trimming the sequence reads, and then mapping them to the human genome reference sequence to create a BAM file.

(Step 12) The epigenome data analysis unit 12 analyzes the genome-related information of each sample obtained in step 11 to calculate the modification level at each epigenome site (data example D2 in Figure 6 shows a modification level value of 40% calculated at the site CpG ₃₄ (chr12: address of bases 3428031-3428032 on chromosome 12) of a certain sample). The modification level is then output to the correspondence storage unit 14.

Here, the epigenetic modification level for each region can be calculated along with the probability of significance by statistical testing. The tools for the series of processes in steps 11 and 12 are publicly available, so manual adjustments may also be made. This results in data on the level of epigenetic DNA modification, such as DNA methylation, in the genome sequence of each subject (as shown as a schematic example in Figure 3 above).

(Step 13) The lifestyle data input unit 13 acquires lifestyle data for each sample (each sample that is the same as the sample for which genome-related information has been received by the epigenome dataset input unit 11) and outputs this to the correspondence storage unit 14.

This lifestyle data can be obtained as each item predefined in relation to lifestyle and its actual value (actual value given as a quantitative value or qualitative value). For example, lifestyle data may be obtained including questionnaire data related to lifestyle based on a general medical interview (for example, data related to smoking, diet, exercise, mental stress, etc., as shown in the lower part of Figure 1 described above). Furthermore, for example, lifestyle data may be obtained including quantitative data obtained via a smartphone or wearable device. For example, in the case of data related to diet, the number of units for each category, such as staple food and fruit, can be obtained in a food exchange table where 1 unit is converted to 80 kcal. In the case of data related to exercise, the frequency of exercise, exercise intensity, exercise time, etc., over a certain period such as one week can be obtained.

In order to accommodate multiple types of formats for each lifestyle item, the lifestyle data input unit 13 may check to ensure that each item and its actual value in units of quantity, etc., can be registered in a unified format. (In other words, when actual values in different formats are obtained for common items between samples, they may be converted into actual values in a common format according to a predetermined rule, and then output from the lifestyle data input unit 13 to the correspondence storage unit 14.)

In FIG. 6, data examples D3 and D4 are schematic examples of the original data and output data of the lifestyle habit data acquired in this way, including data related to diet and exercise. That is, original data D3 is an example of input data to lifestyle habit data input unit 13, and output data D4 is an example of data output from lifestyle habit data input unit 13 to correspondence storage unit 14.

(Step 14) The DNA modification level data and lifestyle habit data obtained in steps 12 and 13 above are linked by the correspondence storage unit 14 to the corresponding sample ID and user information (at least age among age, sex, etc.), and then stored in the database 15, thereby constructing the database 15.

Fig. 7 shows a schematic example of registered information in the database 15 constructed in step S1 in the above-mentioned manner in the form of a table. In Fig. 7, part P1 shows an example of a table (a table showing the contents of registered information in the database 15) in which the age of each sample user and DNA modification level (epigenetic modification rate) data, such as the methylation rate at the DNA methylation site (each CpG region CpG _i (i=1, 2, ...)) and data on lifestyle habits (each diet item Diet_i (i=1, 2, ...) and each exercise item Exer_i (i=1, 2, ...)) are accumulated for each user ID (User001, 002, 003, etc.).

In addition, FIG. 7 also shows that database 15 receives biological age (B_Age) calculated by the biological age prediction model and stores it in association with the user ID as part P2, which represents information that is further added to database 15 in step S2, which will be described later. (Part P1 in FIG. 7 represents the information of database 15 that is constructed in step S1.)

<Step S2> In step S2, the biological age calculation unit 21 performs the following steps 21, 22, and 23, and then proceeds to step S3.

(Step 21) By referring to the registered information in the database 15 constructed in step S1, the biological age of each registered sample is calculated by applying the specified biological age prediction model.

In the biological age prediction model, the modification rate of a specific site of the epigenome is analyzed, and biological age can be calculated as an index that represents the functional ability and degree of aging of biological tissues rather than chronological age. A model for calculating biological age is generally trained by a multiple linear regression model or a regression model (such as Elastic Net) with a regularization term added to suppress overlearning, using biological age obtained from various aging indices as the objective variable and epigenetic modification rate data (generally data on the modification level of modified sites) as the explanatory variable. Here, the epigenetic modification rate data may be the level of DNA methylation or DNA hydroxymethylation. (Note that the epigenetic data analysis unit 12 in step S1 can similarly calculate any type of epigenetic modification rate, not limited to DNA methylation rate.) For example, when DNA methylation rate is used as epigenetic modification rate data, biological age F can be calculated by a model such as the following formula (1) using multiple DNA methylation sites (CpG regions) that are correlated with age (biological age).

In formula (1), b _k (k=0,1,...n) is a coefficient in the prediction model, and CpG _k (k=1,2,...,n) is the methylation rate of the kth CpG region in the prediction model. If only some of the CpG regions k=1,2,...,n, rather than all of them, are correlated with biological age in the prediction model, the coefficient b _k for the methylation rate CpG _k of the uncorrelated CpG region may be set to b _k =0. If an arbitrary prediction model is used more generally, the methylation rate may be set to the modification level of the modified site.

As an example, if the target CpG region and its coefficients are specified as a model in which CpG regions 2, 4, 16, and 55 are meaningful, the above general formula (1) will specifically be expressed as the following formula (2), and each coefficient b _k (k=0, 2, 4, 16) will be given, for example, as shown in data example D5 in Figure 8 (and Figure 6). (Note that data example D5 in Figure 6 shows, in addition to the coefficients of the prediction model, a schematic example of the calculation of biological age using the prediction model, but this calculation is performed by the biological age calculation unit 21.)

Such models will reveal which CpG regions affect biological age. Various models have been proposed to predict biological age using epigenetic information, and any model can be specified in advance.

(Step 22) By the above step 21, for each sample in the database 15, information on the (real) age n=n(ID) and biological age F=F(ID) of the sample is obtained by linking it with its ID as follows:
(ID, age n, biological age F)

Therefore, in step 22, this information (ID, n, F) for each sample is used to determine the "BY group" described below, and the representative epigenetic modification rate for each BY group is calculated.

In other words, within the range of actual age plus or minus a (e.g. a=1) years, the group whose biological age is lower than chronological age and is the top t% will be called the BY group as the biologically young group, and for example, actual age of 45 years plus or minus a will be expressed as BY@45. Here, the value of t is specified in advance, and can be, for example, the probability Pr(n≦μ-σ) of being a group that is equal to or less than the mean value μ and the standard deviation σ (μ-σ) when age n follows a normal distribution.

According to the definition of this notation, for example, the BY group “BY@45” is a sample set that satisfies the following three conditions.
BY@45={ID|"45-a≦n(ID)≦45+a" and "F(ID)<n(ID)"
And "CDF(F(ID))≦t%"}

In the above, the first condition "45-a≦n(ID)≦45+a" means that the actual age n(ID) is in the range of 45±a years, and the second condition "F(ID)<n(ID)" means that the biological age F(ID) is smaller (younger) than the actual age n(ID). (Here, the value (scale) of biological age is predefined (in the model itself) so that being younger than the actual age value means being healthy.) Note that both actual age and biological age may be given as integers or real numbers.

The third condition, "CDF(F(ID))≦t%", is that in a "healthy group" around the actual age where "actual age is in the range of 45±a years" and "F<n" (biological age is younger than actual age), the value of the cumulative distribution function CDF of the value of F (biological age) is less than t%. (In other words, CDF(F(ID)) is the top percentile of the youngest biological age F(ID) of the sample with that ID within the "healthy group".)

In other words, the BY group "BY@45" is an example of a subset that is determined to be in exemplary health status taking biological age into consideration, extracted from a sample set (a sample set that satisfies the first condition) whose actual age is around 45 years old (determined to be the same as or close to 45 years old), and both the second and third conditions are imposed as conditions for extracting those that fall into the exemplary health status. As a variant, it is also possible to impose only either the second or third condition and obtain the BY group "BY@45".

In step 22, the BY group "BY@n" defined in this way is determined for various values of assumed actual age n (e.g., n=44, 45, 46, etc.), and a representative value of the epigenetic modification rate for each BY group "BY@n" is calculated. Any statistical indicator, such as the average or mode, may be used as the representative value. Here, as described above, the BY group "BY@n" is a group determined to be in exemplary health status among samples around actual age n, and therefore the epigenetic modification rate calculated as the representative value can be considered to reflect exemplary health status around actual age n.

(Step 23) The biological age calculated for each sample in step 21 above and the epigenetic modification rate (representative value) calculated for each BY group "BY@n" in step 22 are recorded and stored in database 15.

Here, biological age can be recorded in database 15 as (ID, personal information such as actual age, epigenetic modification rate, lifestyle, biological age) in a form further linked to the information (ID, personal information such as actual age, epigenetic modification rate, lifestyle) constructed in step S1. An example of such recording is shown in Figure 7 mentioned above, where information on biological age (B_Age) in part P2 is further recorded in addition to information on part P1 (actual age (Age), epigenetic modification rate, lifestyle).

The epigenetic modification rate (representative value) calculated for each BY group "BY@n" may also be separately recorded in the database 15. Figure 9 shows an example of such a record, which is an example of recording for each BY group "BY@n" with actual ages n=44, 45, 46 (years), etc., when using the biological age model of the above-mentioned formula (2) (a model in which the methylation rate of a specific CpG region correlates with biological age).

<Step S3> In step S3, by referring to the information (ID, actual age, DNA modification level, lifestyle, biological age) of all samples in the database 15, the lifestyle influence analysis unit 22 calculates the influence on biological age for each lifestyle item (each item such as each diet item Diet_i (i=1, 2, ...) and each exercise item Exer_i (i=1, 2, ...) as illustrated in Figure 7) that is considered to be correlated in the model used in the biological age calculation unit 21, and outputs this influence to the lifestyle advice generation unit 23 before proceeding to step S4.

In other words, lifestyle habits such as exercise and diet affect multiple epigenetic modifications, but the affected regions vary depending on the lifestyle item and its frequency and intensity. For example, when considering DNA methylation as an epigenetic modification, as shown in Figure 2 above as a schematic example, the impact of lifestyle items varies depending on the CpG region, and some of the CpG regions used in biological age prediction models are also affected.

The lifestyle habit influence analysis unit 22 can calculate this influence using the epigenetic modification rate (epigenetic modification rate used in the biological age prediction model) of all samples in the database 15 and the lifestyle habit data, as the partial correlation value r _Bk obtained by removing the influence between lifestyle habit items using existing statistical methods, using the following formula (3). By calculating as partial correlation, even if a plurality of different lifestyle habit items affect the CpG _k of interest, the confounding influence can be eliminated. (Note that if the correlation between independent variables is too strong, there may be cases where the estimated value of the partial correlation coefficient becomes unstable due to the problem of multicollinearity. For example, this may be the case when there are very similar types of lifestyle habit items (e.g., carbohydrate intake and calorie intake). To prevent instability in such cases, a setting may be set in advance to treat a plurality of very similar types of items as only one type of lifestyle habit item, or to use only one item, and data input may be accepted at the lifestyle habit data input unit 13.)

Here, S _ij is the cofactor in row i and column j of the variance-covariance matrix S of lifestyle item B and each epigenetic modification rate (the determinant created by removing the element in row i and column j is multiplied by (-1) ^i+j ) and can be calculated using the information of (lifestyle item, epigenetic modification rate) of all samples recorded in the database 15. The variance-covariance matrix S for calculating this cofactor Sij is as shown in the following formula (3A), and with regard to the subscripts representing its elements, x is the DNA methylation rate of CpG _k of interest as the target of calculation of the partial correlation r _Bk , y is one of the lifestyle items of interest as the target of calculation (lifestyle item B corresponding to the partial correlation r _Bk ), and z (z=z ₁ , z ₂ , ..., z _p ) is each of the p lifestyle items other than the item B. That is, the cofactor _S12 of the numerator in formula (3) is calculated from a determinant obtained by removing the element in row 1, column 2 of the matrix S (the element in the same row or column as _Sxy ), and similarly, the cofactors _S11 and _S22 of the denominator are calculated from determinants obtained by removing the element in row 1, column 1 of the matrix S (the element in the same row or column as _Sxx ) and the element in row 2, column 2 of the matrix S (the element in the same row or column as _Syy ), respectively. (In other words, the component in row i and column j of the cofactor matrix S~ of the matrix S in formula (3A) is _Sij used in formula (3). Note that since the matrix S is symmetric, the cofactor matrix S~ is also symmetric and it does not matter whether it is transposed or not.)

In Fig. 10, an example of the partial correlation _rBk calculated in this manner is shown as data example D6 as an example corresponding to the biological age prediction model of formula (2) and data example D5 also shown in Fig. 8 (i.e., an example targeting CpG regions 2, 4, 16, and 55 as DNA methylation among epigenetic modifications). Note that Fig. 10 is a diagram showing example data for explaining the processing (processing of steps S3 and S4) of the lifestyle habit effect analysis unit 22 and the lifestyle habit advice generation unit 23, and will be referred to as appropriate below.

The value of the influence r _Bk may be calculated by any method such as machine learning, other than calculating as partial correlation.

<Step S4> In step S4, for a sample designated as the subject for which advice information is to be generated (any of the samples recorded in database 15), the epigenetic modification rate and actual age of the subject's sample are compared with the epigenetic modification rate for each BY group "BY@n" as the model value recorded in database 15, and the lifestyle advice generation unit 23 outputs lifestyle items that are determined to be effective in making the subject's biological age younger as advice _information for promoting the health of the subject, by also taking into account the influence degree r Bk obtained from the lifestyle impact analysis unit 22, and then the flow of Figure 5 is terminated.

Specifically, the lifestyle habit impact analysis unit 22 can generate advice information in the following steps 31, 32, 33, and 34.

(Step 31) The influence degree r _Bk obtained as a partial correlation coefficient or the like is multiplied by the coefficient in the biological age prediction model of the corresponding epigenetic region to estimate the influence degree of each lifestyle item on biological age through the epigenetic region.

Figure 10 shows an example of this step 31, in which the influence _rBk is an element value of k rows and B columns by arranging the epigenetic modification rate item k as shown in data D6 and the lifestyle habit item B as column direction, and multiplying this by a column vector _bk such that the coefficient of the biological age prediction model as shown in data D5 is element k (the element in the kth row of the influence _rBk matrix is multiplied by _bk ) as a Hadamard product to obtain a matrix of influence " _rBk * _bk " as shown in data D7.

(Step 32) Furthermore, the difference between the epigenetic modification rate of the BY group “BY@m” corresponding to the subject’s actual age (say, m years old) (i.e., the model value at around m years old) is noted, and the influence of each lifestyle item on the epigenetic modification rate is calculated as a differential weight “Dw _k ” (absolute value of the difference).

Fig. 11 is a diagram for explaining the processing of the lifestyle habit influence analysis unit 22 as a common example with the examples of Fig. 10 etc., and data D8 shows an example of procedure 32. For example, suppose that subject User011 has an actual age of 45, and the methylation rate of the CpG region of the biological age compared with BY@45 in Fig. 9 is as shown in data example D8 in Fig. 11. Then, it is considered that the methylation rate of each CpG region can be brought close to the methylation rate (model value) of BY@45 (by improving the lifestyle of subject User011 in the future), and the absolute value of the difference is calculated as the difference weight "Dw _k ".

This differential weighting will be a different value depending on the subject. For example, the differential weighting for CpG2 in Figure 11 is 0.3 (= |0.9-0.6|) for User011 (methylation rate 0.6), as the difference from the model value of 0.9 in Figure 9 (and Figure 11), but is 0.2 (= |0.9-0.7|) for User012 (methylation rate 0.7), who is the same actual age of 45.

(Step 33) Using the influence degree "r _Bk *b _k " calculated in step 31 above and the differential weight "Dw _k " calculated in step 32, the influence degree EB of the lifestyle habit B is calculated by calculating the sum "Σ _k " of the epigenetic modification rate item k for each lifestyle item B for the product "r _Bk *b _k *Dw _k " as shown in the following formula (4). Here, the greater the differential weight of the epigenetic region, the more likely it is that improvement through lifestyle habits will be achieved, so the influence _{degree EB} is calculated as a weighted sum in this way. K is the total number of epigenetic modification rate items k in the prediction model.

Data example D9 in Fig. 11 is an example of calculating influence EB for subject User011 of data example D8 in the same figure as the sum of epigenetic modification rate item k (= _CpG2 , _CpG4 , _CpG16 , _CpG55 ) in the model for each lifestyle item _B (=Diet1, Diet2, Diet3, Exer1, ...) using formula (4). In this data example D9, for example, the influence for lifestyle item Diet1 of subject User011 is calculated as follows:
CpG ₂ item = (0.1) * (-5.0) * |0.9-0.6| = -0.15
CpG ₄ item = (-0.1) * (3.0) * |0.1-0.4| = -0.09
CpG ₁₆ item = (0.0) * (2.0) * |0.0-0.0| = 0.0
CpG ₅₅ entries = (0.2) * (-4.5) * |0.5-0.4| = -0.09
The total effect of the lifestyle habit item Diet1 of the subject User011 is -0.33.

(Step 34) The lifestyle habit B item for which the value of the influence degree _EB calculated in step 33 is determined to be large on the negative side is output as advice information that contributes to making the subject's biological age younger, i.e., contributes to improving the health condition. Note that, in regard to the output of the advice information, text or the like in a predetermined format corresponding to the degree of the negative value of the influence degree _EB and the lifestyle habit item B may be prepared in advance, and the advice information may be output in the form of this text information or the like.

In the case of data example D9 for subject User011 in FIG. 11, the sum of the values for Diet2 is -0.915, which indicates that it is the most effective. The value of this influence E _B represents the increase or decrease in biological age, so a negative value means that it slows down aging, and a positive value means that it accelerates aging. Here, since the lifestyle habit items are assumed to be in the positive direction, a positive value means that it slows down the lifestyle habit, and a negative value means that it increases it. In other words, in this example, a lifestyle habit that increases Diet2 is the most effective for this subject.

As described above, according to an embodiment of the present invention, since biological age is calculated using more fundamental genome-level information compared to conventional health programs, it is expected to be extremely effective in changing behavior to maintain and improve health. In addition, once data is registered from a sufficient number of user samples, lifestyle item parameters are accurately determined, so lifestyle data is not required for subsequent user samples. In other words, as long as there is genome-related information (information that reflects lifestyle habits as biometric information, such as genome modification rate) and actual age information, it is possible to present lifestyle habits to lower biological age, even without information on the type of lifestyle habits.

In other words, although it has been described that the subject for which advice information is generated by the lifestyle habit advice generation unit 23 in step S4 is specified from samples already registered in the database 15, it is also possible to specify a new unregistered subject. In this case, the actual age n of the new subject and DNA modification level data obtained by processing the genome-related information of the new subject by the epigenome data analysis unit 12 may be input to the lifestyle habit advice generation unit 23.

Below, we will explain various supplementary points, additions, alternatives, etc.

(1) According to an embodiment of the present invention, it is possible to generate advice information that encourages behavioral changes to maintain and improve health, which can contribute to Goal 3 of the United Nations-led Sustainable Development Goals (SDGs) to "Ensure healthy lives and promote well-being for all at all ages."

(2) In the database 15 constructed in step S1, DNA modification level data obtained from genome-related information is registered for each sample. However, instead of or in addition to this, other information having the same properties as the DNA modification level data, i.e., evaluation values of one or more optional items that reflect lifestyle habits as biological information, may be registered. For example, evaluation values of health checkup data such as BMI values, which are health index variables, blood test items (LDL cholesterol values), and some kind of biomarkers may be registered. Furthermore, with regard to lifestyle habits, in addition to items related to exercise and diet, actual values of items related to sleep, smoking, stress, etc. may be accepted by the lifestyle habits data input unit 13 and registered in the database 15, as shown in the schematic examples of Figures 1 and 2.

(3) In the embodiment of the present invention, biological age estimated from epigenetic information is used, which has the following significance. Biological age is an index that represents the functional capacity and degree of aging of biological tissue. Epigenetic age is one of biological ages, and is an age estimated by a mathematical algorithm based on the methylation state in the genome, and can be modeled by machine learning from methylation variables at hundreds to hundreds of thousands of locations. Since the DNA methylation rate is often used as the epigenome, it is also called DNA methylation age. The advantages of epigenetic age include that it is considered to be the most effective index of all biological ages, is strongly associated with mortality, serves as an index of health, and indicates the degree of aging progress for each tissue or organ. For example, it has been reported that cancer cells have advanced epigenetic ages, and that people with high epigenetic ages are more susceptible to viruses.

(4) FIG. 12 is a diagram showing an example of the hardware configuration of a general computer device 70. The information processing device 100 can be realized as one or more computer devices 70 having such a configuration. When the information processing device 100 is realized by two or more computer devices 70, information required for processing may be sent and received via a network. The computer device 70 includes a CPU (Central Processing Unit) 71 that executes predetermined instructions, a GPU (Graphics Processing Unit) 72 as a dedicated processor that executes some or all of the execution instructions of the CPU 71 in place of the CPU 71 or in cooperation with the CPU 71, a RAM 73 as a main storage device that provides a work area for the CPU 71 (and the GPU 72), a ROM 74 as an auxiliary storage device, a communication interface 75, a display 76, an input interface 77 that accepts user input via a mouse, a keyboard, a touch panel, etc., and a bus BS for transmitting and receiving data between them.

Each functional unit of the information processing device 100 can be realized by a CPU 71 and/or a GPU 72 that reads from a ROM 74 and executes a predetermined program corresponding to the function of each unit. Both the CPU 71 and the GPU 72 are a type of arithmetic device (processor). Here, when display-related processing is performed, the display 76 also operates in conjunction with the above, and when communication-related processing related to data transmission and reception is performed, the communication interface 75 also operates in conjunction with the above. The database 15 may be realized as a ROM 74 serving as an auxiliary storage device.

100...information processing device, 11...epigenetic data set input unit, 12...epigenetic data analysis unit, 13...lifestyle data input unit, 14...association storage unit, 15...database, 21...biological age calculation unit, 22...lifestyle effect analysis unit, 23...lifestyle advice generation unit

Claims (11)

a first process of calculating a biological age for each of a plurality of samples by applying a biological age prediction model to the evaluation value of each sample by referring to a database in which the actual age, the evaluation value of each item of biological information reflecting the lifestyle, and the actual value of each item of the lifestyle are associated with each other;
A second process of calculating an influence of each of the lifestyle items on the biological age by referring to the database ,
The information processing apparatus according to claim 1, wherein in the second process, the degree of influence is calculated using partial correlation between each item of the lifestyle habit and each item of the biological information . a first process of calculating a biological age for each of a plurality of samples by applying a biological age prediction model to the evaluation value of each sample by referring to a database in which the actual age, the evaluation value of each item of biological information reflecting the lifestyle, and the actual value of each item of the lifestyle are associated with each other;
A second process of calculating an influence of each of the lifestyle items on the biological age by referring to the database ,
a third process of calculating a model value of the evaluation value of the biometric information within a certain range of chronological ages from samples determined to have a young biological age within the certain range of chronological ages by referring to the database;
and a fourth process of outputting, for a specified sample, the lifestyle habit items that are determined to be effective in making the biological age of the sample younger by comparing an evaluation value of the biometric information of the sample with a model value of the evaluation value of the biometric information that corresponds to the actual age of the sample . In the second process, a partial correlation (rBk) between each item (B) of the lifestyle habit and each item (k) of the biological information is calculated;
In the fourth process, the comparison is made based on a difference between the evaluation value and the model value for each item (k) of biometric information in a designated sample;
3. The information processing device according to claim 2, further comprising: outputting the lifestyle habit item (B) determined to be effective in making the biological age of the sample younger, using the difference and the partial correlation (rBk). In the first process, when calculating the biological age by applying the biological age prediction model, the biological age is calculated as a weighted sum of the evaluation values of each item (k) of the biological information of each sample by a predetermined coefficient (bk);
The information processing device according to claim 3, characterized in that in the fourth process, a sum is calculated for each lifestyle item (B) by adding a value (rBk*bk*Dwk) obtained by multiplying the absolute value of the difference (Dwk) by the predetermined coefficient (bk) and the partial correlation (rBk) for each item (k) of the biological information, and a lifestyle item corresponding to a sum whose value is negative and has a larger absolute value is output as an item that is more effective in making the biological age of the sample younger. 5. The information processing device according to claim 1 , wherein at least a part of the items of biological information reflecting the lifestyle in the database includes an epigenetic modification rate. a first process of calculating a biological age for each of a plurality of samples by applying a biological age prediction model to the evaluation value of each sample by referring to a database in which the actual age, the evaluation value of each item of biological information reflecting the lifestyle, and the actual value of each item of the lifestyle are associated with each other;
A second process of calculating an influence of each of the lifestyle items on the biological age by referring to the database ,
2. An information processing apparatus comprising: a database including a health checkup data item among at least some of the items of biological information reflecting the lifestyle habits of the user; 7. The information processing apparatus according to claim 1 , wherein at least some of the lifestyle habit items in the database include items related to exercise, diet, smoking, sleep, or stress. A computer-implemented information processing method, comprising:
a first process of calculating a biological age for each of a plurality of samples by applying a biological age prediction model to the evaluation value of each sample by referring to a database in which the actual age, the evaluation value of each item of biological information reflecting the lifestyle, and the actual value of each item of the lifestyle are associated with each other;
A second process of calculating an influence of each of the lifestyle items on the biological age by referring to the database ,
The information processing method , wherein the second process calculates the degree of influence using partial correlation between each item of the lifestyle habit and each item of the biological information . A computer-implemented information processing method, comprising:
a first process of calculating a biological age for each of a plurality of samples by applying a biological age prediction model to the evaluation value of each sample by referring to a database in which the actual age, the evaluation value of each item of biological information reflecting the lifestyle, and the actual value of each item of the lifestyle are associated with each other;
A second process of calculating an influence of each of the lifestyle items on the biological age by referring to the database ,
a third process of calculating a model value of the evaluation value of the biological information within a certain range of chronological ages from samples determined to have a young biological age within the certain range of chronological ages by referring to the database;
and a fourth process of outputting, for a specified sample, the lifestyle habit items that are determined to be effective in making the biological age of the sample younger by comparing an evaluation value of the biometric information of the sample with a model value of the evaluation value of the biometric information that corresponds to the actual age of the sample. A computer-implemented information processing method, comprising:
a first process of calculating a biological age for each of a plurality of samples by applying a biological age prediction model to the evaluation value of each sample by referring to a database in which the actual age, the evaluation value of each item of biological information reflecting the lifestyle, and the actual value of each item of the lifestyle are associated with each other;
A second process of calculating an influence of each of the lifestyle items on the biological age by referring to the database ,
An information processing method, characterized in that at least some of the items of biological information reflecting the lifestyle habits in the database include items of health checkup data . 8. A program for causing a computer to function as the information processing device according to claim 1.

Images