Method for processing information, information processing device, and program

Abstract

To provide a method for processing information that enables the prediction of enzyme reactions or metabolic products even for unknown enzyme genes.SOLUTION: A computer generates a list of combinations of multiple enzyme reactions corresponding to multiple substrates and estimated metabolic products presumed to be generated by each enzyme reaction. The computer acquires information on enzyme genes possessed by an organism and information on metabolic products generated by enzyme reactions of the enzyme genes, and specifies combinations of enzyme reactions and the estimated metabolic products that match the acquired enzyme genes and metabolic products from the list.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an information processing method, an information processing device, and a program. [Background technology]

[0002] Enzyme genes possessed by microorganisms are genetic resources that may be capable of catalyzing new chemical reactions. In order to utilize microorganisms, it is necessary to understand what enzyme genes the microorganisms possess, and what reactions the enzyme genes possessed by the microorganisms catalyze and what metabolic products they produce. Patent Document 1 discloses a technology that can suitably search for literature related to metabolism, which synthesizes substances. According to the technology of Patent Document 1, it is possible to efficiently search for literature on enzymes, genes, and microorganisms related to a metabolic pathway in a living organism, based on information on the metabolic pathway (reactant compounds, reactant compounds, and enzymes). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-82097 Summary of the Invention [Problem to be solved by the invention]

[0004] In Patent Document 1, existing documents can be efficiently searched, but the search targets are limited to known enzyme genes and metabolites. Many of the enzyme genes possessed by microorganisms are unknown enzyme genes whose functions have not been elucidated, and many of the metabolic products produced by the enzymatic reactions of each enzyme gene are unknown metabolic products whose structures have not been elucidated, and such enzyme genes and metabolic products are not search targets. Therefore, it is difficult to decipher the functions and structures of unknown enzyme genes and metabolic products, and there is a problem that the enzyme genes possessed by microorganisms cannot be fully and effectively utilized at present.

[0005] An object of the present disclosure is to provide an information processing method and the like that is capable of estimating the enzymatic reaction and metabolic products even for unknown enzyme genes. [Means for solving the problem]

[0006] An information processing method according to one aspect of the present disclosure involves a computer executing a process of creating a list of combinations of multiple enzyme reactions corresponding to multiple substrates and estimated metabolic products estimated to be produced by each of the enzyme reactions, acquiring information about enzyme genes possessed by an organism and information about the metabolic products produced by the enzyme reactions caused by the enzyme genes, and identifying from the list a combination of enzyme reaction and estimated metabolic product that matches the acquired enzyme gene and metabolic product. Effect of the Invention

[0007] In one aspect of the present disclosure, the enzymatic reaction and metabolic products of an unknown enzyme gene can also be estimated. [Brief description of the drawings]

[0008] [Figure 1] 1 is a block diagram showing an example of the configuration of an information processing device; [Diagram 2] FIG. 1 is an explanatory diagram showing an example of a record layout of an enzyme reaction-estimated metabolite DB. [Diagram 3] 13 is a flowchart showing an example of a procedure for creating an enzyme reaction-estimated metabolite DB. [Figure 4] 13 is a flowchart showing an example of a procedure for predicting pairs of enzyme genes and predicted metabolites. [Diagram 5] FIG. 1 is an explanatory diagram of the process of predicting pairs of enzyme genes and predicted metabolites. [Figure 6] FIG. 11 is an explanatory diagram of a verification process. [Figure 7] FIG. 11 is an explanatory diagram of a verification process. [Figure 8] FIG. 2 is an explanatory diagram showing an example of the configuration of a learning model. [Figure 9]13 is a flowchart showing an example of a procedure for predicting a pair of an enzyme gene and a predicted metabolite according to the second embodiment. [Figure 10] FIG. 13 is an explanatory diagram of a verification result of the process of the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] The information processing method, information processing device, and program of the present disclosure will be described below with reference to the drawings showing the embodiments. In the following embodiments, a configuration for estimating an enzyme reaction and a metabolic product resulting from an enzyme reaction of an enzyme possessed by a microorganism will be described as an example. However, the configuration of this embodiment is not limited to a microorganism, and can be similarly applied to various enzymes possessed by animals or plants. That is, the configuration can be such that an enzyme reaction and a metabolic product resulting from an enzyme reaction of an enzyme possessed by an animal or plant can be estimated.

[0010] (Embodiment 1) An information processing device that estimates the enzyme reaction and the metabolic products resulting from the enzyme reaction of an enzyme gene possessed by a microorganism will be described. Fig. 1 is a block diagram showing an example of the configuration of the information processing device. The information processing device 10 is capable of various information processing and sending and receiving information, and is configured, for example, as a personal computer, a server computer, a workstation, or a tablet terminal.

[0011] The information processing device 10 includes a control unit 11, a storage unit 12, a communication unit 13, an input unit 14, a display unit 15, a reading unit 16, etc., and these units are connected to each other via a bus. The control unit 11 has one or more processors such as a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), or a GPU (Graphics Processing Unit). The control unit 11 appropriately executes a program 12P stored in the storage unit 12 to perform various information processing and control processing to be performed by the information processing device 10. When the control unit 11 has multiple processors, each process may be executed by the same processor, or each process may be executed by a different processor.

[0012] The storage unit 12 includes a RAM (Random Access Memory), a flash memory, a hard disk, an SSD (Solid State Drive), etc. The storage unit 12 prestores a program 12P (program product, computer program) executed by the control unit 11 and various data required for executing the program 12P, etc. The storage unit 12 also temporarily stores data, etc. generated when the control unit 11 executes the program 12P. The storage unit 12 also stores an enzyme reaction-estimated metabolite DB 12a, which will be described later. The enzyme reaction-estimated metabolite DB 12a may be stored in another storage device connected to the information processing device 10, or may be stored in another storage device with which the information processing device 10 can communicate.

[0013] The communication unit 13 is a communication module for connecting to a network by wired communication or wireless communication, and transmits and receives information to and from other devices via the network. The network may be the Internet or a public communication line, or may be a LAN (Local Area Network) constructed in a facility where the information processing device 10 is installed. The input unit 14 accepts an operation input by a user and sends a control signal corresponding to the operation content to the control unit 11. The display unit 15 is a liquid crystal display or an organic EL display, etc., and displays various information according to instructions from the control unit 11. The input unit 14 and the display unit 15 may be a touch panel integrally configured. In addition, the input unit 14 and the display unit 15 are not essential, and the information processing device 10 may be configured to accept operations through a connected computer, or to output information to be displayed to an external display device.

[0014] The reading unit 16 reads information stored in a portable storage medium 10a such as a CD (Compact Disc), a DVD (Digital Versatile Disc), a USB (Universal Serial Bus) memory, an SD card, a micro SD card, or a Compact Flash (registered trademark). The program 12P (program product) and various data stored in the storage unit 12 may be read by the control unit 11 from the portable storage medium 10a via the reading unit 16 and stored in the storage unit 12. The program 12P and various data may be downloaded by the control unit 11 from another device via the communication unit 13 and stored in the storage unit 12, or may be written in the storage unit 12 during the manufacturing stage of the information processing device 10.

[0015] The information processing device 10 is not limited to a single computer, but may be a multi-computer including multiple computers, or may be a virtual machine virtually constructed by software in one device. In addition, when the information processing device 10 is configured as a server computer, the information processing device 10 may be a local server installed in the facility where the information processing device 10 is installed, or may be a cloud server connected to the information processing device 10 via a network such as the Internet. In the following, the information processing device 10 will be described as being one computer. In addition, the program 12P may be executed on a single computer, or may be distributed and executed on multiple computers connected to each other via a network.

[0016] FIG. 2 is an explanatory diagram showing an example of a record layout of the enzyme reaction-estimated metabolite DB 12a. The enzyme reaction-estimated metabolite DB 12a is a database that stores combinations (pairs) of a plurality of enzyme reactions and metabolites (hereinafter referred to as estimated metabolites) estimated to be generated by each enzyme reaction. The enzyme reaction-estimated metabolite DB 12a shown in FIG. 2 includes a pair ID sequence, an enzyme reaction sequence, and an estimated metabolite sequence, and stores information on the enzyme reaction and information on the estimated metabolite in association with a pair ID assigned to each pair. The information on the enzyme reaction includes an EC number, SMARTS, name, etc. assigned to each enzyme reaction. The EC number is an enzyme number assigned by the Enzyme Commission of the International Union of Biochemistry and Molecular Biology according to the type of reaction catalyzed by each enzyme. SMARTS is a notation method for structure search that extends the SMILES notation for search queries. The SMILES (Simplified Molecular Input Line Entry System) notation is a display method in which the chemical structure of a molecule is expressed as a string of alphanumeric characters in ASCII code. The name of the enzyme reaction includes a systematic name and a common name. The information on the estimated metabolite includes the SMILES, mass spectrum, accurate mass, and name of the metabolite. The mass spectrum is a spectrum obtained by mass spectrometry of the metabolite, in which the horizontal axis indicates mass (mass-to-charge ratio m / z (m is the mass of the ion, z is the charge)) and the vertical axis indicates detection intensity (abundance of ions). For the mass spectrum here, for example, a product ion spectrum generated by fragmentation of a specific ion can be used. When the names of the enzyme reaction and the metabolite cannot be specified, no names may be stored, and provisional names may be assigned and stored. The enzyme reaction-estimated metabolite DB12a estimates and accumulates each pair of the enzyme reaction and the estimated metabolite by a pair creation process between the enzyme reaction and the estimated metabolite described later. The contents stored in the enzyme reaction-estimated metabolite DB12a are not limited to the example shown in FIG. 2, and other information on the enzyme reaction and other information on the estimated metabolite may be stored.For example, information on putative metabolites may include primary data obtained by a mass spectrometer, such as retention time, ontology of metabolites, partial structure information of metabolites, and secondary data obtained by secondarily analyzing these data, etc. Also, information on a plurality of substrates, information on enzyme reactions corresponding to each substrate, and information on putative metabolites corresponding to each enzyme reaction may be stored in association with each other.

[0017] The process of creating pairs of multiple enzyme reactions and estimated metabolites that may be generated by each enzyme reaction will be described below. Fig. 3 is a flow chart showing an example of the process of creating the enzyme reaction-estimated metabolite DB 12a. The following process is executed by the control unit 11 according to the program 12P stored in the storage unit 12 of the information processing device 10.

[0018] When creating the enzyme reaction-estimated metabolite DB 12a, the control unit 11 of the information processing device 10 acquires a substrate DB in which information on substrates that undergo a chemical structure change (enzymatic reaction) by an enzyme is stored (S11). The substrate DB may be, for example, HMDB (Human Metabolome Database) provided by The Human Metabolome Project. The substrate DB is not limited to HMDB, and may be KNApSAcK, a secondary metabolite database system managed by the Nara Institute of Science and Technology, PubChem managed by NCBI (National Center for Biotechnology Information), or the like. Such a substrate DB stores information on metabolites (compounds), such as the name, molecular formula, molecular weight, compositional formula, and SMILES, which is a character string representing the compositional formula (two-dimensional structure of a molecule). In addition to the above, the substrate DB may be, for example, KEGG (Kyoto Encyclopedia of Genes and Genomes) managed by the Kyoto University Bioinformatics Center. KEGG stores combinations of biological species and metabolites in the organisms, and stores the name, composition formula, SMILES, and the like of the metabolites as information on the metabolites. In this embodiment, a compound (metabolite) listed in the substrate DB is used as a substrate to generate an estimated metabolic product. The substrate DB may be obtained in advance from a predetermined server and stored in the storage unit 12, and in this case, the control unit 11 may read out the substrate DB from the storage unit 12.

[0019] Next, the control unit 11 acquires an enzyme reaction DB in which information on enzyme reactions is stored (S12). For example, the enzyme reaction DB may be BRENDA (The Comprehensive Enzyme Information System), an enzyme information database managed by the Institute of Biochemistry and Bioinformatics at the Braunschweig University of Technology in Germany. The enzyme reaction DB stores enzyme classification, EC number, name, function, structure, substrate, production reaction, product, and the like. In this embodiment, an estimated metabolite is created using an enzyme reaction included in the enzyme reaction DB. The enzyme reaction DB may also be acquired in advance from a predetermined server and stored in the storage unit 12, and in this case, the control unit 11 may read out the enzyme reaction DB from the storage unit 12.

[0020] The control unit 11 reads out information on one enzyme reaction from the enzyme reaction DB acquired in step S12 (S13). Then, the control unit 11 changes the EC number included in the read out information on the enzyme reaction so as to relax the substrate specificity in the enzyme reaction (S14). For example, the control unit 11 generates an EC number with relaxed substrate specificity by changing the number in the second division, the number in the third division, or the number in the fourth division, which indicates the classification result according to substrate specificity, in the EC number. Specifically, for example, if the read out enzyme reaction is a reaction of transferring an alkyl group, such as EC2.5.1.1, EC2.5.1.2, or EC2.5.1.3, the EC number is changed to EC2.5.1. EC2.5.1 indicates a reaction of transferring an alkyl group, and EC2.5.1.1, EC2.5.1.2, EC2.5.1.3, etc. each indicate a reaction of transferring an alkyl group to a different substrate. By changing EC2.5.1.1, EC2.5.1.2, EC2.5.1.3, etc. to EC2.5.1, the EC numbers can be generalized to include a reaction in which an alkyl group is transferred even to an unknown substrate if a nitrogen atom is present, thereby relaxing the substrate specificity. In addition, the substrate specificity can be relaxed by changing, for example, EC numbers such as EC1.1.1.X (X is a number), EC1.1.2.X, EC1.1.3.X, etc. to EC1.1. By relaxing the substrate specificity in the EC numbers in this way, it is possible to add a type of substrate that causes the enzyme reaction indicated by the EC number. By setting the above-mentioned EC number changing rules in advance and storing them in the storage unit 12, the control unit 11 can automatically perform the EC number changing process based on the EC number of the enzyme reaction read out from the enzyme reaction DB.

[0021] The control unit 11 converts the EC number changed in step S14 into a SMARTS notation representing a chemical structure and stores (registers) it in an enzyme reaction list (S15). The enzyme reaction list is prepared in, for example, the storage unit 12. The control unit 11 judges whether the above-mentioned processing has been completed for all enzyme reactions stored in the enzyme reaction DB (S16). If it is judged that the processing has not been completed (S16: NO), the control unit 11 returns to step S13 and executes the processing of steps S13 to S15 for unprocessed enzyme reactions. In this way, an enzyme reaction list is generated that includes the enzyme reaction read out from the enzyme reaction DB as well as the enzyme reaction with the relaxed substrate specificity of the enzyme reaction. The enzyme reactions registered in the enzyme reaction DB are known enzyme reactions, but the enzyme reactions indicated by the EC numbers with the relaxed substrate specificity as described above may include enzyme reactions with substrates not previously known (unknown enzyme reactions). Therefore, in this embodiment, the control unit 11 can generate an enzyme reaction list in which the SMARTS of unknown enzyme reactions is registered in addition to known enzyme reactions.

[0022] When the control unit 11 judges that the above-mentioned process has been completed for all the enzyme reactions stored in the enzyme reaction DB (S16: YES), it reads out one of the SMARTS of the enzyme reactions registered in the enzyme reaction list (S17). Then, the control unit 11 estimates a metabolite (estimated metabolite) produced by the enzyme reaction represented by the SMARTS based on the read SMARTS and the substrate DB read in step S11 (S18). Here, the control unit 11 simulates the enzyme reaction represented by the SMARTS read in step S17 using, for example, the RDKit, which is an open source library of Python, and identifies the estimated metabolite by the enzyme reaction. The inventor of the present application constructed a program that simulates the enzyme reaction using the RDKit on Python Ver. 3.7, and identified the estimated metabolite using this program. The control unit 11 obtains the SMILES that represents the composition formula of the estimated metabolite by simulating the enzyme reaction using the program.

[0023] Next, the control unit 11 predicts the mass spectrum and accurate mass of the estimated metabolite (S19). Here, the control unit 11 predicts the mass spectrum and calculates the accurate mass of the estimated metabolite from the SMILES of the estimated metabolite acquired in step S18 using, for example, the CFM-ID disclosed in the literature "Wang et al., Anal. Chem., 93, 11692-11700 (2021)". The control unit 11 pairs the information of the enzyme reaction read from the enzyme reaction list in step S17 and the information of the estimated metabolite estimated in step S18 and stores them in the enzyme reaction-estimated metabolite DB 12a (S20). Specifically, the control unit 11 issues a pair ID, and stores the EC number and SMARTS of the enzyme reaction, and the SMILES, mass spectrum, and accurate mass of the estimated metabolite in the enzyme reaction-estimated metabolite DB 12a in association with the issued pair ID. This makes it possible to create the enzyme reaction-estimated metabolite DB 12a as shown in FIG. 2.

[0024] The control unit 11 judges whether the above-mentioned processing has been completed for all enzyme reactions (SMARTS) registered in the enzyme reaction list (S21). If it is judged that the processing has not been completed (S21: NO), the control unit 11 returns to step S17 and executes the processing of steps S17 to S20 for the unprocessed enzyme reactions. If it is judged that the processing for all enzyme reactions has been completed (S21: YES), the control unit 11 ends the series of processing. As a result, for all enzyme reactions registered in the enzyme reaction list, estimated metabolites that may be produced are identified, and an enzyme reaction-estimated metabolite DB 12a is generated in which each enzyme reaction is associated with an estimated metabolite by the enzyme reaction. Since the enzyme reaction list registers not only enzyme reactions with known substrates but also enzyme reactions with unknown substrates, the enzyme reaction-estimated metabolite DB 12a registers unknown metabolites (hypothetical metabolites, hypothetical metabolites) by enzyme reactions with previously unknown substrates in addition to known metabolites by enzyme reactions with known substrates.

[0025] The inventors of the present application used approximately 556,000 compounds listed in a substrate DB such as HMDB as substrates, and set up 142 enzyme reactions, including unknown enzyme reactions, from enzyme reactions listed in an enzyme reaction DB such as BRENDA, performed simulations of the enzyme reactions, and created approximately 100.9 billion pairs of enzyme reactions and predicted metabolites.

[0026] The information processing device 10 of this embodiment uses the enzyme reaction-estimated metabolite DB 12a created by the above-mentioned process to perform a process of estimating pairs between an enzyme gene possessed by an arbitrary microorganism and a metabolite (estimated metabolite) that may be produced by an enzyme reaction caused by this enzyme gene. Note that, in this embodiment, based on the assumption that a metabolite detected only in a specific strain group is produced by an enzyme gene possessed only by the strain group, pairs between an enzyme gene present in strains (microorganisms) of multiple samples and a metabolite thought to have been obtained by a genetic reaction caused by this enzyme gene are estimated.

[0027] The following describes a process for estimating pairs of enzyme genes present in strains of multiple samples and estimated metabolites produced by these enzyme genes using the enzyme reaction-estimated metabolite DB 12a. Fig. 4 is a flow chart showing an example of a process procedure for estimating pairs of enzyme genes and estimated metabolites, and Fig. 5 is an explanatory diagram of the process for estimating pairs of enzyme genes and estimated metabolites. The following process is executed by the control unit 11 according to a program 12P stored in the storage unit 12 of the information processing device 10.

[0028] When performing the estimation process of pairs of enzyme genes and estimated metabolites, the user prepares a plurality of sample strains (microorganisms) and genome data of each strain, and metabolome data of the metabolites in each strain. The strains can be obtained, for example, from the RIKEN BioResource Research Center (RIKEN BRC), and the genome data of the strains can be obtained, for example, from Genome, a genome database managed by NCBI. Note that the sources of the strains and genome data are not limited to these. The metabolome data of the metabolites can be obtained, for example, by anaerobically culturing each strain in BHI (brain heart infusion) medium, extracting the metabolites in one layer, and performing mass analysis using a liquid chromatography quadrupole time-of-flight mass spectrometer. Note that the metabolome data includes at least the mass spectrum and accurate mass of the metabolites.

[0029] The control unit 11 of the information processing device 10 acquires genome data of the strain to be processed (S31). In this embodiment, since the user has acquired genome data of the sample strain in advance and stored it in the storage unit 12, the control unit 11 only needs to read out the genome data from the storage unit 12, but at this point, the control unit 11 may acquire it from a predetermined server. The control unit 11 also acquires metabolome data of metabolites detected in the strain to be processed (S32). Here, too, the user has acquired metabolome data of metabolites in the sample strain in advance and stored it in the storage unit 12, so the control unit 11 only needs to read out the metabolome data from the storage unit 12, but at this point, the control unit 11 may acquire it from another device that stores metabolome data.

[0030] Next, the control unit 11 executes pan-genome analysis on the genome data of the sample to be processed (S33). For example, the control unit 11 performs pan-genome analysis using BPGA disclosed in the literature "Chaudhari et al., Sci.Rep., 6, 27373 (2016)". Pan-genome analysis is a process of classifying each gene sequence into orthologs that are presumed to have the same function based on the similarity of the gene sequences possessed by each of the sample strains, determining whether each strain has each ortholog, and creating a list indicating the presence / absence of each ortholog for each strain. FIG. 5A shows an example of a list showing the results of pan-genome analysis, and indicates the presence / absence of orthologs A, B, C, etc. for each of the four samples 1 to 4. In FIG. 5A, the presence of orthologs A, B, C, etc. is indicated by a pentagon, and the absence of orthologs is indicated by a horizontal line. The example in Figure 5A shows that ortholog A is present only in strains of samples 1 and 2, ortholog B is present only in strains of samples 1 and 4, and ortholog C is present only in strains of samples 2 to 4. Pan-genome analysis may be performed using programs and methods other than BPGA.

[0031] Next, the control unit 11 identifies a gene sequence (ortholog) of interest from the pan-genome analysis result (S34). Here, in this embodiment, a gene sequence of interest is identified based on the assumption that a metabolite detected only in a specific strain group is produced by an enzyme gene possessed only by the strain group. FIG. 5B shows a list of metabolites in the strains of each sample shown in the metabolome data acquired in step S32, and indicates the presence or absence of each of the metabolites α, β, γ, etc. for each of the four samples 1 to 4. In FIG. 5B, for each of the metabolites α, β, γ, etc., the vertical axis shows each sample, and the horizontal axis shows the abundance (detection intensity) in each sample. The example of FIG. 5B shows that metabolite α was detected only in the strains of samples 1 and 2, metabolite β was detected in the strains of samples 1 to 4, and metabolite γ was detected in the strains of samples 1, 3 to 4. The control unit 11 judges whether or not there is a gene sequence (ortholog) present only in the strain in which each metabolic product was detected, and if it is judged that there is, specifies the gene sequence present only in this strain as a notable gene sequence. According to the list of orthologs shown in FIG. 5A and the list of metabolic products shown in FIG. 5B, metabolic product α was detected only in samples 1 and 2, and the gene sequence present only in samples 1 and 2 is ortholog A, so it can be presumed that metabolic product α was produced by the gene of ortholog A. In this case, the control unit 11 specifies ortholog A as a notable gene sequence, and specifies metabolic product α as a metabolic product produced by an enzyme gene having the gene sequence of the ortholog A. If the control unit 11 cannot identify a notable gene sequence, it may display a message indicating that the pair of the enzyme gene and the presumed metabolic product could not be presumed on the display unit 15, and end the process.

[0032] When a notable gene sequence can be identified, the control unit 11 estimates an EC number corresponding to the identified gene sequence (S35). For example, the control unit 11 estimates an EC number corresponding to the gene sequence identified in step S34 using DeepEC disclosed in the literature "Ryu et al., PNAS, 116, 13996-14001 (2019)". In DeepEC, an EC number is estimated using a learning model that learns the gene sequence of an enzyme gene that causes a certain chemical reaction and the EC number of the enzyme gene by deep learning. The estimation of the EC number may be performed using a program and method other than DeepEC.

[0033] By the above process, the EC number of the gene sequence of interest and metabolome data (specifically, mass spectrum and accurate mass) of the metabolite estimated to be produced by the enzyme expressed from this gene sequence were obtained. Then, the control unit 11 can screen for pairs that match (fit) the function of the enzyme gene, mass and structure of the metabolite by identifying pairs of enzyme reactions and estimated metabolites that match (fit) the EC number of the gene sequence of interest and the metabolite metabolome data from the contents stored in the enzyme reaction-estimated metabolite DB 12a. In this embodiment, about 100.9 billion enzyme reaction-estimated metabolite pairs are registered in the enzyme reaction-estimated metabolite DB 12a, and high-throughput screening can be realized to identify the enzyme reaction and estimated metabolite of interest from these pairs.

[0034] Specifically, the control unit 11 first narrows down the pairs of enzyme reaction-estimated metabolite registered in the enzyme reaction-estimated metabolite DB 12a based on the accurate mass included in the metabolite metabolome data acquired in step S32 (S36). Here, the control unit 11 extracts pairs of enzyme reaction-estimated metabolite registered in the enzyme reaction-estimated metabolite DB 12a in which the accurate masses of the metabolites match (match). The control unit 11 may extract not only pairs in which the accurate masses of the metabolites match, but also pairs in which the difference in the accurate masses is within a predetermined range. Next, the control unit 11 narrows down the pairs of enzyme reaction-estimated metabolite narrowed down in step S36 based on the EC number of the gene sequence of interest estimated in step S35 (S37). Here, the control unit 11 extracts pairs of enzyme reaction-estimated metabolite narrowed down based on the accurate masses of the metabolites in which the EC number of the enzyme reaction matches (matches). Note that the control unit 11 may extract not only pairs having the same EC number of the enzyme reaction, but also pairs having the same numbers in the first to third divisions of the EC number, for example.

[0035] Next, the control unit 11 specifies one of the pairs of enzyme reaction-estimated metabolite narrowed down in step S37, and reads out the mass spectrum of the estimated metabolite in the specified pair from the enzyme reaction-estimated metabolite DB 12a (S38). Then, the control unit 11 calculates the similarity between the read out mass spectrum and the mass spectrum included in the metabolome data of the metabolite acquired in step S32 (S39). The similarity can be, for example, a correlation coefficient or a cosine similarity, but other similarity can also be used. Note that the control unit 11 may extract a high-intensity characteristic spectrum (a spectrum of any mass) from the mass spectrum included in the metabolome data of the metabolite acquired in step S32, and calculate the similarity between the extracted mass spectrum and the mass spectrum of the estimated metabolite read out in step S38. The control unit 11 may also calculate the similarity between the high-intensity characteristic spectrum and the mass spectrum of the estimated metabolite by weighting the high-intensity characteristic spectrum to emphasize the similarity in the high-intensity characteristic spectrum. The control unit 11 stores the calculated similarity of the mass spectrum in the storage unit 12 in association with, for example, a pair ID. The control unit 11 judges whether the calculation process of the similarity of the mass spectrum of the metabolite is completed for all pairs of the enzyme reaction-estimated metabolite narrowed down in step S37 (S40), and if it is judged that the calculation process is not completed (S40: NO), the control unit 11 returns to step S38 and executes the processes of steps S38 to S39 for the unprocessed pairs. As a result, the similarity between the mass spectrum of the metabolite in the sample strain and each pair of the enzyme reaction-estimated metabolite narrowed down by the exact mass of the metabolite and the EC number of the enzyme gene is calculated.

[0036] When it is determined that the processing has been completed for all pairs (S40: YES), the control unit 11 identifies pairs with high similarity of the metabolite mass spectrum calculated for each pair in step S39 (S41). For example, the control unit 11 identifies a predetermined number of pairs (e.g., 3, 5, etc.) in descending order of high similarity of the mass spectrum. Then, the control unit 11 outputs a screening result screen displaying the identified pairs (S42). For example, the control unit 11 generates and displays a screening result screen displaying the EC number of the enzyme reaction in each identified pair and the structural formula of the metabolite. The control unit 11 may generate a screening result screen displaying only the pair with the highest similarity of the mass spectrum. The control unit 11 may also generate a screening result screen displaying the structural formula of the substrate used in the enzyme reaction, the enzyme reaction formula, etc. The control unit 11 may output and display the screening result screen on the display unit 15, may be configured to transmit the screening result screen from the communication unit 13 to a specified device via a network, or may be configured to transmit the screening result screen to a printer connected via a network or directly to print it.

[0037] The above-mentioned process makes it possible to estimate the enzyme gene possessed by the sample strain and the metabolic product that may be produced by the enzyme reaction by this enzyme gene. Since the pair of the enzyme reaction and the estimated metabolic product registered in the enzyme reaction-estimated metabolic product DB12a includes an enzyme reaction with an unknown substrate and an unknown metabolic product, even if the enzyme gene and metabolic product possessed by the sample strain are unknown, it is possible to screen the enzyme gene and the metabolic product by this enzyme gene. In addition, since a huge number of pairs of enzyme reactions and estimated metabolic products are registered in the enzyme reaction-estimated metabolic product DB12a, it is possible to realize high-throughput screening to identify the enzyme reaction and estimated metabolic product of interest from these pairs.

[0038] In order to confirm that the process shown in Figure 4 can predict an appropriate pair of enzyme reactions and predicted metabolic products, the inventors of the present application verified whether it is possible to search for the sphingolipid biosynthesis initiation enzyme "Serine palmitoyl-CoA transferase (Spt)" and its metabolic product "3-ketosphinganine (3KS)," which are known to be possessed only by bacteria of the Bacteroidetes phylum among the five bacterial phyla predominant in the human intestine. Figures 6 and 7 are explanatory diagrams of the verification process, and Figure 6A shows the enzyme reaction formula of the sphingolipid biosynthesis initiation enzyme.

[0039] The inventors of the present application first obtained genome data of 16 enterobacteria strains including 8 strains of bacteria belonging to the phylum Bacteroidetes, 5 strains of bacteria belonging to the phylum Firmicutes, 1 strain of bacteria belonging to the phylum Fusobacteria, 1 strain of bacteria belonging to the phylum Actinovacteria, and 1 strain of bacteria belonging to the phylum Proteobacteria, and performed non-targeted metabolome analysis on the 16 strains to obtain metabolome data. As a result, the control unit 11 of the information processing device 10 obtains genome data of the 16 enterobacteria strains and metabolome data of metabolites detected in the strains (S31, S32 in FIG. 4). As a result of the metabolome analysis here, experimental data was obtained in which a total of 6780 ions were detected from the 16 enterobacteria strains. FIG. 6B shows the detection intensity of the metabolite 3KS detected from each strain, and as known, the metabolite 3KS was detected only in 7 strains of bacteria belonging to the phylum Bacteroidetes other than Alistipes. In addition, as metabolome data of metabolite 3KS, an accurate mass of 300.290 and the mass spectrum shown in Figure 6C were obtained. Therefore, we decided to focus on the ion (metabolite 3KS) with an accurate mass of 300.290 and the mass spectrum shown in Figure 6C for verification.

[0040] The inventors of the present application also performed pan-genome analysis on the genome data of 16 enterobacteria strains using the information processing device 10, and created a list showing the presence / absence of 41053 genes (orthologues) for each of the 16 strains (S33 in FIG. 4). From the pan-genome analysis results, genes present only in 7 strains of Bacteroidetes bacteria other than Alistipes (genes not present in the other 9 strains) were searched for, and 248 genes were found to be relevant. The inventors of the present application estimated the EC numbers of the 248 genes by, for example, DeepEC using the information processing device 10 (S34, S35 in FIG. 4). Based on the estimated EC numbers of the 248 genes, proteins involved in DNA and protein modification were excluded, and the number was narrowed down to 109 genes that are likely to be enzyme genes. Therefore, it was decided to verify the 109 genes as genes of interest.

[0041] By the above process, the EC numbers of 109 genes of interest, the accurate mass 300.290 of the metabolite likely produced by the gene, and the mass spectrum shown in FIG. 6C were obtained. Therefore, the inventor of the present application used the information processing device 10 to narrow down the pairs that match (suit) the experimental data obtained as described above from the approximately 100.9 billion pairs stored in the enzyme reaction-estimated metabolite DB 12a. Specifically, the control unit 11 of the information processing device 10 narrowed down the pairs whose metabolites have an accurate mass of 300.290 from the stored contents of the enzyme reaction-estimated metabolite DB 12a (S36 in FIG. 4), and as a result, 216 pairs were obtained. In addition, the control unit 11 narrowed down the 216 pairs to pairs that match (suit) the EC numbers of the 109 genes (S37 in FIG. 4), and as a result, 198 pairs were obtained. In addition, since the EC numbers classify enzyme reactions (enzyme characteristics) according to the numbers in the first to third divisions, the EC numbers corresponding to 26 enzyme reactions can be identified by extracting the EC numbers based on the numbers in the first to third divisions from the EC numbers of the 109 enzyme genes. Therefore, it is possible to narrow down the pairs that match (suit) the 26 EC numbers, and in this case, 198 pairs were obtained.

[0042] The control unit 11 calculates the similarity between the mass spectrum of the metabolite of each of the narrowed down 198 pairs and the mass spectrum shown in FIG. 6C (S38 to S40 in FIG. 4). FIG. 7A is a graph in which the similarity of the mass spectrum calculated for each of the 198 pairs is arranged from the left in descending order. The control unit 11 identifies pairs with high mass spectral similarity (S41 in FIG. 4) and outputs a screening result screen displaying the identified pairs (S42 in FIG. 4). FIG. 7B shows enzyme reaction formulas of three pairs in descending order of mass spectral similarity, and the control unit 11 may generate a screening result screen displaying, for example, the three pairs shown in FIG. 7B. In the example shown in FIG. 7A and FIG. 7B, the pair with the highest mass spectral similarity was the pair of enzyme reaction in which 3KS (3-ketosphinganine) is produced by Spt (serine palmitoyl-CoA transferase) shown in FIG. 6A. That is, it was demonstrated that pairs of known enzyme reactions and estimated metabolites can be accurately estimated. Therefore, by using the technique disclosed in this embodiment, it is possible to estimate pairs of unknown enzyme reactions and estimated metabolic products with high accuracy, and it is possible to screen useful enzyme genes and metabolic products from unused microbial resources (genetic resources) with high throughput. This is expected to enable the enzyme genes possessed by microorganisms to be fully and effectively utilized.

[0043] In the above-described embodiment, the information processing device 10 is configured to narrow down the stored contents of the enzyme reaction-estimated metabolite DB12a by the exact mass of the metabolite, then narrow down the stored contents by the EC number of the gene, and calculate the similarity of the mass spectrum of the metabolite for the narrowed-down pairs. That is, the enzyme gene of the metabolite is predicted by narrowing down the genes present in the strain having a specific metabolite, but this configuration is not limited to this. For example, the metabolite produced by the enzyme gene may be predicted by narrowing down the specific metabolite present only in the strain having the specific gene. Specifically, the information processing device 10 may be configured to narrow down the stored contents of the enzyme reaction-estimated metabolite DB12a by the EC number of the gene, then narrow down the stored contents by the exact mass of the metabolite, and calculate the similarity of the mass spectrum of the metabolite for the narrowed-down pairs.

[0044] In the above-described embodiment, the information processing device 10 performs the process of creating the enzyme reaction-estimated metabolite DB 12a shown in FIG. 3 and the process of estimating the pair of the enzyme gene and the estimated metabolite shown in FIG. 4. However, these two processes may be executed by different devices. In addition, the present embodiment is not limited to the configuration in which the information processing device 10 locally performs the process of estimating the pair of the enzyme gene and the estimated metabolite. For example, a server may be provided that has a storage unit that stores the enzyme reaction-estimated metabolite DB 12a and performs the process of estimating the pair of the enzyme gene and the estimated metabolite. In this case, the information processing device 10 is configured to transmit the EC number of the gene sequence of interest and the metabolome data (accurate mass and mass spectrum) of the metabolite to the server, and to obtain the pair of the enzyme reaction and the metabolite estimated by the server. In this case, the server executes the processes of steps S36 to S41 in FIG. 4 and transmits the screening result screen to the information processing device 10. Even in such a configuration, the same process as in the present embodiment is possible and the same effect can be obtained.

[0045] The above-mentioned embodiment is a configuration for estimating the enzyme reaction (enzyme gene) and the metabolic products resulting from the enzyme reaction of an enzyme possessed by a microorganism, but is not limited to this configuration. The configuration of this embodiment can be similarly applied to various enzymes possessed by animals or plants (organisms), and it becomes possible to estimate the enzyme reaction (enzyme gene) and the metabolic products resulting from the enzyme reaction of various enzymes possessed by animals or plants.

[0046] In this embodiment, a matching pair of enzyme reaction-presumed metabolite is identified based on the EC number of the enzyme gene and the accurate mass and mass spectrum of the metabolite, but this is not limited to the configuration. For example, a matching pair of enzyme reaction-presumed metabolite may be identified based on the gene information of the metabolite instead of the metabolome data of the metabolite.

[0047] (Embodiment 2) In addition to the narrowing down of the first embodiment described above, an information processing device that estimates the enzyme reaction and the metabolic products resulting from the enzyme reaction for the enzyme gene possessed by the microorganism by narrowing down based on the retention time obtained by the analysis by the chromatographic mass spectrometer will be described. The information processing device of this embodiment has a configuration similar to that of the information processing device 10 of the first embodiment shown in FIG. 1, so the description of the configuration will be omitted. In addition to the configuration shown in FIG. 1, the information processing device 10 of this embodiment stores a learning model M (see FIG. 8) in the storage unit 12. The learning model M is a model that has learned training data by machine learning, for example, and is expected to be used as a program module that constitutes artificial intelligence software. The learning model M performs a predetermined calculation on an input value and outputs the calculation result, and the storage unit 12 stores information such as information on the layers included in the learning model M, information on the nodes that constitute each layer, and weights (coupling coefficients) between nodes as information that defines the learning model M. Instead of storing the learning model M in the storage unit 12, the information processing device 10 may access a server that stores the learning model M and read it out.

[0048] FIG. 8 is an explanatory diagram showing a configuration example of the learning model M. The learning model M shown in FIG. 8 is a model that inputs a molecular descriptor of a metabolite, performs a calculation to estimate a retention time (hereinafter referred to as an estimated retention time) obtained when the metabolite is analyzed using a liquid chromatograph quadrupole time-of-flight mass spectrometer based on the input data, and outputs the calculation result (estimated retention time). The molecular descriptor is a value that expresses the characteristics of the metabolite as a numerical value based on the chemical structure. The retention time indicates the time required for a component contained in a sample strain to be eluted from a separation part when the components are separated by liquid chromatography in a liquid chromatograph quadrupole time-of-flight mass spectrometer. Specifically, the time from when the sample is injected into a column (stationary phase) together with the mobile phase to when it is eluted from the column (more specifically, the time until the detection signal intensity of the sample eluted from the column reaches a peak) is called the retention time.

[0049] The learning model M is configured using, for example, a partial least squares regression (PLS) algorithm. Note that the learning model M may be configured using algorithms such as a support vector machine (SVM), a decision tree, a random forest, a convolutional neural network (CNN), a transformer, or a combination of multiple algorithms in addition to PLS.

[0050] The learning model M is generated by learning using training data in which a molecular descriptor for training is associated with a retention time (correct estimated retention time) obtained by performing an analysis of a metabolite indicated by the molecular descriptor using a liquid chromatograph quadrupole time-of-flight mass spectrometer. The learning model M learns to output the correct estimated retention time when a molecular descriptor included in the training data is input. Specifically, the learning model M performs a calculation based on the input molecular descriptor, estimates the estimated retention time of the metabolite indicated by the molecular descriptor, and obtains the estimation result. Then, the learning model M compares the estimated retention time, which is the estimation result, with the correct estimated retention time, and optimizes parameters such as weights (coupling coefficients) between nodes in the learning model M so that the two are close to each other. The parameter optimization method can be a steepest descent method, an error backpropagation method, or the like. As a result, a learning model M that outputs the estimated retention time of a metabolite when a molecular descriptor of the metabolite is input is obtained.

[0051] The estimated retention time output by the learning model M shown in FIG. 8 is a numerical value expressed by a continuous value, and the learning model M outputs any numerical value among the continuous values ​​as a so-called regression problem. However, it may be configured to be regarded as a classification problem and to discriminate any numerical value selectable for the estimated retention time. In this case, the learning model M is configured to discriminate the optimal numerical value from a plurality of numerical values ​​set in advance for the estimated retention time and output the discrimination result. The generation of training data and the learning of the learning model M may be performed by the information processing device 10 or by another learning device. When learning is performed by another learning device, the learning model M generated by the other learning device is downloaded from the learning device to the information processing device 10 via a network or a portable storage medium 10a and stored in the storage unit 12. The learning model M is not limited to the configuration shown in FIG. 8. For example, the learning model M may be configured to input other information representing the chemical structure of a metabolite, such as a structural formula, a rational formula, a composition formula, a SMILES notation, an InChI (International Chemical Identifier) ​​notation, a MOL file, or an SDF, in addition to or instead of a molecular descriptor.

[0052] FIG. 9 is a flow chart showing an example of a procedure for estimating pairs of enzyme genes and estimated metabolites in the second embodiment. The process shown in FIG. 9 is the process shown in FIG. 4, in which steps S51 to S54 are added between steps S39 and S40, and step S55 is added instead of step S41. Explanation of the same steps as in FIG. 4 is omitted. In this embodiment, the metabolome data of the metabolites in the sample strain includes the mass spectrum and accurate mass of the metabolites as well as the retention time. The retention time can be obtained by performing an analysis using a liquid chromatograph quadrupole time-of-flight mass spectrometer on the metabolites obtained by anaerobically culturing the sample strain in BHI medium and extracting the metabolites in a single layer. That is, in this embodiment, in step S32, the control unit 11 acquires the mass spectrum, accurate mass, and retention time of the metabolites as the metabolome data of the metabolites detected in the strain to be processed. In this embodiment, too, if metabolome data is stored in the memory unit 12, the control unit 11 can simply read the metabolome data from the memory unit 12, or can obtain the metabolome data from another device that has metabolome data stored at this point.

[0053] In the information processing device 10 of this embodiment, after processing step S39, the control unit 11 proceeds to step S51, and reads information on the pair for which the mass spectrum was read in step S38 from the enzyme reaction-estimated metabolite DB 12a (S51). For example, the control unit 11 reads out the SMILES of the estimated metabolite in the pair. Note that, if information expressing the chemical structure of the estimated metabolite other than the SMILES is stored in the enzyme reaction-estimated metabolite DB 12a, the control unit 11 may read out other information. Also, the control unit 11 may read out the SMILES together with the mass spectrum of the estimated metabolite in the identified pair in step S38, in which case it is not necessary to perform step S51.

[0054] The control unit 11 calculates molecular descriptors for the estimated metabolites of the pair from which the SMILES has been read out, using the PaDEL descriptor (Yap et al., 2010, J. Comp. Chem., 32, 1466-1474) (S52). The PaDEL descriptor is software for calculating molecular descriptors that represent the characteristics of a compound (metabolite) in numerical values ​​based on the chemical structure, and calculates the molecular descriptors of the compound from an SDF generated from the SMILES of the compound, for example. Thus, the control unit 11 generates an SDF from the read out SMILES, and calculates the molecular descriptors of the estimated metabolites of the pair from the generated SDF. The generation of an SDF from a SMILES can be performed, for example, using ChemDraw.

[0055] Next, the control unit 11 acquires the estimated retention time of the estimated metabolite of the pair from the molecular descriptor calculated in step S52 (S53). Here, the control unit 11 inputs the calculated molecular descriptor to the learning model M, and acquires the estimated retention time of the estimated metabolite indicated by the input molecular descriptor as an output value from the learning model M. The control unit 11 calculates the difference between the acquired estimated retention time and the retention time included in the metabolome data of the metabolite acquired in step S32 (S54). The control unit 11 stores the similarity of the mass spectrum calculated in step S39 and the difference in retention time calculated in step S54 in the storage unit 12, for example, in association with the pair ID.

[0056] The control unit 11 judges whether the calculation process of the similarity of the metabolite mass spectrum and the calculation process of the difference in retention time have been completed for all pairs (S40), and if it is judged that the calculation process has not been completed (S40: NO), the control unit 11 returns to step S38 and executes the processes of steps S38 to S39 and S51 to S54 for the unprocessed pairs. As a result, for each pair of enzyme reaction-estimated metabolite narrowed down by the exact mass of the metabolite and the EC number of the enzyme gene, and the sample strain, the similarity of the metabolite mass spectrum and the difference in retention time obtained when an analysis is performed using a liquid chromatograph quadrupole time-of-flight mass spectrometer are calculated.

[0057] When it is determined that the processing has been completed for all pairs (S40: YES), the control unit 11 identifies pairs in which the similarity of the metabolite mass spectrum calculated for each pair in step S39 is high and the difference in retention time calculated for each pair in step S54 is small (S55). For example, the control unit 11 extracts a predetermined number of pairs (e.g., three, five, etc.) in descending order of the similarity of the mass spectrum, and identifies the pair with the smallest difference in retention time among the extracted pairs. The control unit 11 may also extract a predetermined number of pairs in descending order of the difference in retention time, and identify the pair with the highest similarity of the mass spectrum among the extracted pairs. The control unit 11 may also extract a predetermined number of pairs in descending order of the similarity of the mass spectrum, and may also extract a predetermined number of pairs in descending order of the difference in retention time, and identify the pairs extracted in duplicate. It is presumed that the pairs identified in this way are the enzyme gene possessed by the strain of the sample and the metabolic product that may be produced by the enzymatic reaction caused by this enzyme gene. Thereafter, the control unit 11 performs the processing of step S42. Here, the control unit 11 may generate a screening result screen that displays pairs having high mass spectral similarity and small retention time differences, or may generate a screening result screen that displays a predetermined number of pairs having high mass spectral similarity and a predetermined number of pairs having small retention time differences.

[0058] FIG. 10 is an explanatory diagram of the verification result of the processing of the second embodiment. The graph shown in FIG. 10 shows the difference in retention time on the horizontal axis and the similarity of mass spectrum on the vertical axis, and is a graph in which the difference in retention time calculated between 3KS actually measured in the sample strain and the similarity of mass spectrum are plotted for each pair registered in the enzyme reaction-estimated metabolite DB12a. In FIG. 10, for 198 pairs narrowed down by the accurate mass of the metabolite and the EC number of the gene, the similarity between the mass spectrum of the metabolite of each pair and the mass spectrum of the metabolite detected in the sample strain, and the difference between the retention time estimated from the molecular descriptor of the metabolite of each pair (estimated value of retention time by liquid chromatograph quadrupole time-of-flight mass spectrometer) and the retention time of the metabolite detected in the sample strain are plotted. In the graph shown in FIG. 10, the two pairs surrounded by the dashed closed curve are the pair (1st) and the pair (2nd) with the highest similarity of the mass spectrum shown in the graph of FIG. 7A, respectively, and the 1st pair was the pair of the enzyme reaction in which 3KS is produced by Spt shown in FIG. 6A. As shown in the graph of FIG. 10, it was possible to separate the two pairs with high similarity of the mass spectrum from the other pairs. Therefore, in this embodiment, it was demonstrated that it is possible to more accurately estimate whether the sample strain is a pair of known enzyme reaction and putative metabolite by further narrowing down the pairs screened based on the similarity of the mass spectrum based on the difference in retention time.

[0059] The above-mentioned process can estimate the enzyme gene possessed by the sample strain and the metabolites that may be produced by the enzymatic reaction of the enzyme gene. In this embodiment, the metabolome analysis (measurement of retention time) is performed using a liquid chromatograph quadrupole time-of-flight mass spectrometer. However, the measurement of retention time may be performed using other analytical devices as long as the mass spectrometer is a chromatograph that can acquire retention time. For example, in addition to liquid chromatography, an analytical device using a separation method based on a different separation principle, such as gas chromatography or capillary electrophoresis, may be used. In addition to a quadrupole time-of-flight mass spectrometer, an analytical device using a mass analysis method based on a different principle, such as an orbitrap mass spectrometer or a triple quadrupole mass spectrometer, may be used.

[0060] In the above-described embodiment, the information processing device 10 calculates the similarity of the mass spectrum of the metabolite and the difference in retention time for pairs narrowed down by the accurate mass of the metabolite and the EC number of the gene for the stored contents of the enzyme reaction-estimated metabolite DB 12a, and performs the narrowing down by the similarity of the mass spectrum and the difference (approximation) of the retention time of the metabolite. Specifically, the configuration is such that pairs with high similarity of the mass spectrum and small difference in retention time are specified as the pair to be estimated (enzyme reaction and estimated metabolite), but is not limited to this configuration. For example, a configuration may be adopted in which pairs with high similarity of the mass spectrum are narrowed down to a predetermined number, and then the difference in retention time is calculated for the narrowed down pairs, and pairs with small difference in retention time (pairs with similar difference) are specified. Also, a configuration may be adopted in which pairs with small difference in retention time are narrowed down to a predetermined number, and then the similarity of the mass spectrum is calculated for the narrowed down pairs, and pairs with high similarity are specified. Furthermore, the combination may be narrowed down based on the similarity of the mass spectra and the difference in retention time separately, and the resulting pairs may be presented.

[0061] The matters described in each of the above-mentioned embodiments can be combined with each other. In addition, the independent claims and dependent claims described in the claims can be combined with each other in any and all combinations regardless of the citation format. Furthermore, the claims use a format in which a claim cites two or more other claims (multiple claim format), but this is not limited to this. A multiple claim (multi-multi claim) that cites at least one multiple claim may also be used.

[0062] The embodiments disclosed herein are illustrative in all respects and should not be considered as limiting. The scope of the present invention is defined by the claims, not by the above meaning, and is intended to include all modifications within the scope and meaning equivalent to the claims. [Explanation of symbols]

[0063] 10. Information processing device 11 Control section 12 Storage section 13. Communications Department 14 Input section 15 Display section 12a Enzyme reaction-estimated metabolite DB M Learning Model

Claims

1. creating a list of combinations of a plurality of enzyme reactions corresponding to a plurality of substrates and putative metabolic products putatively produced by each of the enzyme reactions; acquiring information on an enzyme gene possessed by an organism and information on a metabolic product produced by an enzymatic reaction mediated by the enzyme gene; A combination of an enzyme reaction and a predicted metabolic product that matches the obtained enzyme gene and metabolic product is identified from the list. An information processing method in which processing is performed by a computer.

2. changing the substrate specificity of the enzyme reaction to add a substrate type for the enzyme reaction; A list is created that includes combinations of the enzyme reaction and a putative metabolic product that is putatively produced by the enzyme reaction on the added substrate. The information processing method according to claim 1 , wherein the processing is executed by the computer.

3. A combination of an enzyme reaction and a predicted metabolic product that matches the enzyme gene and metabolic product is identified from the list based on an Enzyme Commission number (EC number) corresponding to the enzyme reaction predicted from the information on the enzyme gene.

3. The information processing method according to claim 1, wherein the processing is executed by the computer.

4. The information on the enzyme gene possessed by the organism includes genome data of the organism, Obtain genome data for multiple organisms, performing a pan-genome analysis on the acquired genome data to identify the presence or absence of multiple enzyme genes for each of the organisms; Identifying a gene of interest based on the presence or absence of the enzyme genes in the organisms; Deducing the EC number of the identified genes of interest; Based on the estimated EC number, a combination of an enzyme reaction and a putative metabolite that matches the enzyme gene and the metabolite is identified from the list. The information processing method according to claim 3 , wherein the processing is executed by the computer.

5. the information about the metabolite includes an accurate mass of the metabolite; A combination of an enzyme reaction and a putative metabolite that matches the enzyme gene and the metabolite is identified from the list based on the accurate mass of the metabolite.

3. The information processing method according to claim 1, wherein the processing is executed by the computer.

6. the information about the metabolite includes a mass spectrum of the metabolite; A combination of an enzyme reaction and a putative metabolic product that matches the enzyme gene and the metabolic product is identified from the list based on the mass spectrum of the metabolic product.

3. The information processing method according to claim 1, wherein the processing is executed by the computer.

7. Calculating a similarity between the mass spectrum of the metabolite and the mass spectrum of the putative metabolites in the list; Based on the calculated similarity, a combination of an enzyme reaction and a putative metabolite that matches the enzyme gene and the metabolite is identified from the list. The information processing method according to claim 6 , wherein the processing is executed by the computer.

8. The information about the metabolites includes retention times obtained by analysis using a chromatography-mass spectrometry device; inputting information representing the chemical structure of the estimated metabolite in the list into a learning model that has been trained to output an estimated retention time when information representing the chemical structure of the metabolite is input, thereby obtaining an estimated retention time of the estimated metabolite; Calculating the difference between the estimated retention time of the estimated metabolite and the retention time of the metabolite; Based on the calculated difference, a combination of an enzyme reaction and a putative metabolic product that matches the enzyme gene and the metabolic product is identified from the list.

3. The information processing method according to claim 1, wherein the processing is executed by the computer.

9. An information processing device having a control unit, The control unit is creating a list of combinations of a plurality of enzyme reactions corresponding to a plurality of substrates and putative metabolic products putatively produced by each of the enzyme reactions; acquiring information on an enzyme gene possessed by an organism and information on a metabolic product produced by an enzymatic reaction mediated by the enzyme gene; A combination of an enzyme reaction and a predicted metabolic product that matches the obtained enzyme gene and metabolic product is identified from the list. Information processing device.

10. storing a list of combinations of a plurality of enzyme reactions corresponding to a plurality of substrates and estimated metabolic products estimated to be produced by each of the enzyme reactions in a storage unit; acquiring information on an enzyme gene possessed by an organism and information on a metabolic product produced by an enzymatic reaction mediated by the enzyme gene; A combination of an enzyme reaction and a predicted metabolic product that matches the obtained enzyme gene and metabolic product is identified from the list. A program that causes a computer to carry out processing.

Images