Chemical structure generating device, storage medium and chemical structure generating method

Abstract

The chemical structure generating device of the present invention comprises a generating unit and a control unit. The generating unit generates a product list consisting of one or more compounds based on a chemical reaction list and a reactant list consisting of one or more compounds. The control unit gives the product list to the generating unit as a new reactant list, updates a database having at least one of the reactant list and the product list, and causes the generating unit to generate a new product list based on the new reactant list and the chemical reaction list.

Description

Technical Field

[0001] The invention relates to a chemical structure generating device, a storage medium and a chemical structure generating method. Background Art

[0002] I am interested in methods for designing molecules with useful properties through machine learning using computer simulations. As an example, first, in the learning phase, a learning model is generated by machine learning such as deep learning using a learning dataset, and then in the execution phase, the learned model generated in the learning phase is used to predict physical property values ​​and search for molecules with desired physical properties.

[0003] Here, for example, a large amount of learning data (training data) is required for machine learning such as deep learning, so it is important to efficiently generate high-quality learning data. From this point of view, a chemical structure generation device that automatically generates a molecular structure suitable for use as learning data for machine learning is important.

[0004] Here, the number of possible molecular structures relative to the number of atoms increases exponentially as the number of atoms increases, resulting in a combinatorial explosion, so it is actually impossible to comprehensively generate candidate molecular structures. In addition, if the number of atoms increases, the ratio of the number of molecules that can be used, that is, the number of molecules that actually exist stably and can be used for commercial purposes to the number of possible molecular structures is sharply reduced. On the basis of machine learning, the use of learning data containing many molecular structures that cannot be used for commercial purposes is extremely inefficient in terms of time and economy. Therefore, it is desirable to be able to efficiently generate only chemical structure generation devices that can be used for commercial and valuable molecules.

[0005] Patent Document 1: International Publication No. 9501606

[0006] Patent Document 2: U.S. Patent Application Publication No. 2018 / 096100

[0007] Patent Document 3: Japanese Patent Application Publication No. 2001-058962

[0008] Patent Document 4: International Publication No. 9736252 Summary of the invention

[0009] An object of the present invention is to efficiently generate a molecular structure suitable for use as learning data for machine learning.

[0010] In order to solve the above problems and achieve the purpose, the chemical structure generation device of the present invention has a generation unit and a control unit. The generation unit generates a product list consisting of one or more compounds based on a chemical reaction list and a reactant list consisting of one or more compounds. The control unit gives the product list as a new reactant list to the generation unit, updates a database having at least one of the reactant list and the product list, and causes the generation unit to generate a new product list based on the new reactant list and the chemical reaction list.

[0011] In addition, the storage medium of the present invention stores a chemical structure generation program, which is a program that enables a computer to perform the following processing: generating a product list consisting of one or more compounds based on a chemical reaction list and a reactant list consisting of one or more compounds; giving the product list as a new reactant list; updating a database having at least one of a reactant list and a product list; and generating a new product list based on a new reactant list and a chemical reaction list.

[0012] In addition, the chemical structure generating method of the present invention is a chemical structure generating method executed by a chemical structure generating device, wherein, based on a chemical reaction list and a reactant list composed of one or more compounds, a product list composed of one or more compounds is generated; the product list is given as a new reactant list; a database having at least one of the reactant list and the product list is updated; and a new product list is generated based on the new reactant list and the chemical reaction list.

[0013] According to the present invention, it is possible to efficiently generate a molecular structure suitable for use as learning data for machine learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a diagram showing an overview of a chemical structure generating device according to an embodiment of the present invention.

[0015] Figure 2 This is a diagram showing an example of a chemical reaction list in the chemical structure generating device according to the first embodiment.

[0016] Figure 3 This is a diagram showing an example of a prohibited reaction list in the chemical structure generating device according to the first embodiment.

[0017] Figure 4 This is a flowchart showing the flow of processing performed by the chemical structure generation device according to the first embodiment.

[0018] Figure 5 This is a diagram for explaining the processing performed by the chemical structure generation device according to the first embodiment.

[0019] Figure 6 Yes Figure 4 A flowchart illustrating the processing flow of step S200 in more detail.

[0020] Figure 7 This is a diagram for explaining the processing performed by the chemical structure generation device according to the first embodiment.

[0021] Figure 8 This is a flowchart showing the flow of processing performed by the chemical structure generation device according to the second embodiment.

[0022] Fig. 9 Yes Figure 8 A flowchart illustrating the processing flow of step S200A in more detail. DETAILED DESCRIPTION

[0023] Hereinafter, a chemical structure generating device according to an embodiment of the present invention will be described with reference to the drawings.

[0024] (First Embodiment)

[0025] Figure 1 1 is a diagram showing an overview of a chemical structure generation device 1 according to an embodiment of the present invention. The chemical structure generation device 1 includes a processing circuit 100, a storage unit 120, a database 110, an input device 130, and a display 140. The processing circuit 100 has a generation function 100a and a control function 100b. The database 110 has a list of reactants of chemical reactions, i.e., a reactant list 110a, a list of products of chemical reactions, i.e., a product list 110b, a chemical reaction list 110c, a prohibited reaction list 110d, and a numerical information list 110e. The chemical structure generation device 1 is a device for generating a chemical structure as a reactant list 110a or a product list 110b, which is a basis for learning data for machine learning.

[0026] The processing circuit 100 is a processor (e.g., CPU (Central Processing Unit), GPU (Graphical Processing Unit), ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), etc.) that realizes functions such as the generation function 100a and the control function 100b by reading out a program from the storage unit 120 and executing it. That is, the processing functions such as the generation function 100a and the control function 100b are stored in the storage unit 120 in the form of a program that can be executed by the processor. Thus, the processing circuit 100 in the state where the program has been read out has functions such as the generation function 100a and the control function 100b. That is, the generation function 100a and the control function 100b are examples of a generation unit and a control unit, respectively. In addition, the details of the processing of each function of the processing circuit 100 will be described later.

[0027] The database 110 is a variety of databases that are referenced by the processing circuit 100 in the process of generating a new compound by the chemical structure generating device 1, and is composed of, for example, a reactant list 110a, a product list 110b, a chemical reaction list 110c, a prohibited reaction list 110d, a numerical information list 110e, etc. These databases are stored in the storage unit 120 as needed.

[0028] The chemical reaction list 110c, the prohibited reaction list 110d, and the numerical information list 110e are data used when the processing circuit 100 performs a process of generating a chemical structure. In contrast, the reactant list 110a or the product list 110b is data that becomes the basis of learning data for subsequent machine learning, and generating these lists is one of the purposes of the process performed by the chemical structure generating device 1.

[0029] The reactant list 110a is a list of reactants, etc., which are the basis of the chemical reaction, and is composed of compounds, etc. representing the reactants. More precisely, the reactant list 110a is composed of one or more elements, which are elements that represent operations that do not select simple substances, compounds, or reactants, or elements that represent intramolecular reactions. Here, the chemical structure generation device 1 of the embodiment updates the reactant list in sequence starting from the initial compound. Therefore, the reactant list 110a is composed of the initial state (0th generation (p=0)) reactant list X0, the first generation reactant list X1, the second generation reactant list X2, ..., the pth generation reactant list X3, ..., the pth generation reactant list X4, ..., the pth generation reactant list X5, ..., the pth generation reactant list X6, ..., the pth generation reactant list X7, ..., the pth generation reactant list X8, ..., the pth generation reactant list X9, ..., the pth generation reactant list X10, ..., the pth generation reactant list X11, ..., the pth generation reactant list X12, ..., the pth generation reactant list X13, ..., the pth generation reactant list X14, ..., the pth generation reactant list X15, ..., the pth generation reactant list X16, ..., the pth generation reactant list X17, ..., the pth generation reactant list X18, ..., the pth generation reactant list X29, ..., the pth generation reactant list X3, ..., the pth generation reactant list X4, ..., the pth generation reactant list X5, ..., the pth generation reactant list X6, ..., the pth generation reactant list X7, ..., the pth generation reactant list X8, ..., the pth generation reactant list X9, ..., the pth generation reactant list X19, ..., the pth generation reactant list X11, ..., the pth generation reactant list X11, ..., the pth generation reactant p Here, the k+1th generation reactant list X k+1 Can completely contain the list of reactants of generation k X k, the list of reactants for the k+1th generation X k+1 It is also possible to not completely include the list of reactants of the kth generation X k .

[0030] In addition, the list of reactants for each generation X k , can also be used as any constituent element, except for the compound x representing the reactant k1 …x kn In addition, it also contains special compounds x that represent special operations null 、x intra Etc. These special compounds are described later. Without including these special compounds as elements, the kth generation reactant list X k The element is the compound x representing the kth generation reactant k1 、x k2 …、x kn In contrast, in the case of including these special compounds as elements, the kth generation reactant list X k For example, the element x k1 、x k2 …、x kn 、x null 、x intra .

[0031] The product list 110b is a product list of a chemical reaction, and is composed of compounds representing products. More specifically, the product list 110b is composed of one or more elements, which are elements that represent operations that do not select simple substances, compounds, or reactants, or elements that represent intramolecular reactions. Here, as in the case of the reactant list 110a, the product list 110b is composed of the 0th generation (p=0) product list Y0, the first generation product list Y1, the second generation product list Y2, ..., the pth generation product list Y p constitute.

[0032] In addition, in some embodiments, the k+1 generation reactant list X k+1 The elements of and the k-th generation generated list Y k The elements of are equal.

[0033] The chemical reaction list 110c is a list of possible chemical reactions, which is a list of partial structures of reactants and a list of partial structures of products generated by chemical reactions using the reactants. In addition, the chemical reaction list 110c includes information related to the cost of the chemical reaction as needed.

[0034] exist Figure 2 , an example of a chemical reaction list 110 c is shown when there are two or one reactants and one corresponding product.

[0035] exist Figure 2 The first line of , records the situation where the structure "X1-CO-X2" is generated based on the partial structure "X1-COOH" and the partial structure "X2-H". In such a case, the list {{"X1-CO-X2"}, {"X1-COOH", "X2-H"} of the partial structure list of the reactants and the list of the partial structure list of the products generated by the chemical reaction {"X1-CO-X2"} that establish a corresponding relationship is an example of the chemical reaction list 110c.

[0036] In addition, Figure 2 The second line of the table records the situation where the structure "X1-CHOH-X2" is generated based on the partial structure "X1-CO-X2". In such a case, a list {{"X1-CHOH-X2"}, {"X1-CO-X2"}} of the partial structure list of the reactants and the partial structure list of the products generated by the chemical reaction ("X1-CHOH-X2") is established with a corresponding relationship, which is an example of a chemical reaction list 110c. In addition, this chemical reaction can also be represented as {{"X1-CHOH-X2"}, {"X1-CO-X2", "H2"}} by considering H2 as a partial structure.

[0037] In addition, Figure 2 The third line of, for example, records the case where the structure "X1-CHX3-X2" is generated from the partial structure "X1-CHOH-X2" and the partial structure "X3-H". In such a case, a list {{"X1-CHX3-X2"}, {"X1-CHOH-X2", "X3-H"}} in which a correspondence is established between the partial structure list of the reactants {"X1-CHOH-X2", "X3-H"} and the partial structure list of the product generated by the chemical reaction ("X1-CHX3-X2") is an example of a chemical reaction list 110c.

[0038] In addition, Figure 2 In the embodiment, the case where there are two or one reactants and one corresponding product is described, but the embodiment is not limited to this, and the reactants may be three or more. For example, in the case where structure D is generated from partial structure A, partial structure B, and partial structure C, a list {{D}, {A, B, C}} in which a corresponding relationship is established between the partial structure list {A, B, C} of the reactants and the partial structure list {D} of the product generated by the chemical reaction is an example of the chemical reaction list 110c.

[0039] In addition, in the above case, the case where there is one product is described, but the embodiment is not limited to this, and there may be two or more products. Figure 2 Consider the reverse reaction of the first row of the chemical reaction, where the structure "X1-COOH" and the structure "X2-H" are generated from the partial structure "X1-CO-X2". In such a case, a list {{"X1-COOH", "X2-H"}, {"X1-CO-X2"}}, which is a corresponding list of the partial structure list of the reactants {"X1-CO-X2"} and the partial structure list of the products generated by the chemical reaction {"X1-COOH", "X2-H"}, is an example of the chemical reaction list 110c.

[0040] If the above contents are expressed using a formula, for example, according to the partial structure y k and part of the structure y l Generate a new structure y kl ', the list {{y kl '},{y k ,y l}} is a chemical reaction list 110c. In addition, generally speaking, according to n partial structures y q1 ,y q2 …y qn Generate new m structures y q1q2 … qn；1 ,y q1q2 ...qn; 2, ..., y q1q2 ...qn;m, the list {{y q1q2 ...qn; 1, y q1q2 ...qn; 2, ..., y q1q2…qn；m},{y q1 ,y q2 …y qn}} is a chemical reaction list 110c.

[0041] Furthermore, in some embodiments, the chemical reaction list 110 c is different from the reactant list 110 a and the product list 110 b in that the elements are not updated in each generation and the elements are fixed in each generation.

[0042] The prohibited reaction list 110d is a list of prohibited chemical reactions, in which a partial structure list of reactants and a partial structure list of products generated by chemical reactions using the reactants are associated with each other, together with information indicating that the reactions are prohibited chemical reactions.

[0043] For example, according to the partial structure y k and part of the structure y l Generate a new structure y kl', when the generated new structure has no commercial value and is not suitable as a generated product, the list {{y kl '},{y k ,y l}} is a prohibited reaction list 110d. In addition, generally speaking, according to n partial structures y q1 ,y q2 …y qn Generate new m structures y q1q2…qn；1 ,y q1q2…qn；2 , …, y q1q2…qn；m In the case of a product that is not suitable as a product due to lack of commercial value, etc., the list {{y q1q2…qn；1 ,y q1q2…qn；2 , …, y q1q2…qn；m},{y q1 ,y q2 …y qn}} is a prohibited reaction list 110d. In addition, whether or not there is commercial value is determined for each field based on viewpoints such as cost, stability, and toxicity.

[0044] exist Figure 3 , an example of the prohibited reaction list 110d is shown. Figure 3 This is a diagram for explaining the processing related to the prohibited reaction list 110d.

[0045] exist Figure 3 The above paragraph describes the case where the structure "X1-CO-X2" is generated from the partial structure "X1-COOH" and the partial structure "X2-H". Since there is no particular reason for limiting the generation of such a structure, the generation of this structure is allowed.

[0046] On the other hand, Figure 3 The lower part of the description records the mechanism "X1-COOH" and the partial structure "X2-CO-CH2-CO-X3" x -CO-CHX1-CO-X3". When it is judged that the generated structure is chemically unstable, for example, when it has no commercial value in a certain field, the generation of such a structure is prohibited, and the generation of the structure is prohibited (Deny). That is, the list {{"X2-CO-CHX1-CO-X3"}, {"X1-COOH", "X2-CO-CH2-CO-X3"}} is the prohibited reaction list 110d.

[0047] The numerical information list 110e is a numerical information list indicating the cost and the like of each compound in the reactant list 110a. In the second embodiment, the numerical information list 110e will be described in detail.

[0048] The storage unit 120 is a storage area such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, or a hard disk, and stores various programs executed by the processing circuit 100 , the execution results of the various programs, the database 110 , and the like.

[0049] The input device 130 is a device for a user who uses the chemical structure generating device 1 to perform various operations. The input device 130 is composed of, for example, a mouse, a keyboard, a touch panel, or hardware keys.

[0050] The display 140 displays various information. For example, the display 140 displays the processing results of the CPU, a GUI (Graphical User Interface) for accepting various operations from the user, etc. The display 140 is composed of, for example, a liquid crystal display, an organic EL (Electro Luminescence) display, or a cathode ray tube display. In addition, in a method such as a touch panel, the input device 130 and the display 140 may also be integrally formed.

[0051] In addition, the embodiment is not limited to the above-mentioned example. According to the embodiment, the storage unit 120, the input device 130, the display 140, etc. may not be necessary components of the chemical structure generating device 1, that is, the compound generating device 1 may not include the storage unit 120, the input device 130, the display 140, etc. In addition, as another example, the storage unit 120, the input device 130, the display 140, etc. may be configured outside the compound generating device 1, for example, via a network, and exchange data with the processing circuit 100.

[0052] Next, the background of the embodiment will be described.

[0053] I am interested in methods for designing molecules with useful properties through machine learning using computer simulations. As an example, first, in the learning phase, a learning model is generated by machine learning such as deep learning using a learning dataset, and then in the execution phase, the learned model generated in the learning phase is used to predict physical property values ​​and search for molecules with desired physical properties.

[0054] Here, for example, a large amount of learning data is required for machine learning such as deep learning, so it is important to efficiently generate high-quality learning data. From this point of view, a chemical structure generation device that automatically generates a molecular structure suitable for use as learning data for machine learning is important.

[0055] Here, the number of possible molecular structures relative to the number of atoms increases exponentially as the number of atoms increases, resulting in a combinatorial explosion, so it is actually impossible to comprehensively generate candidate molecular structures. In addition, if the number of atoms increases, the ratio of the number of molecules that can be used, that is, the number of molecules that actually exist stably and can be used for commercial purposes to the number of possible molecular structures is sharply reduced. On the basis of machine learning, the use of learning data containing many molecular structures that cannot be used for commercial purposes is extremely inefficient in terms of time and economy. Therefore, it is desirable to be able to efficiently generate only chemical structure generation devices that can be used for commercial and valuable molecules.

[0056] In view of such background, the processing circuit 100 of the chemical structure generating device 1 of the embodiment has a generating function 100a and a control function 100b. The processing circuit 100 generates a list of products consisting of one or more compounds, namely a product list 110b, based on a list of reactants consisting of one or more compounds, namely a reactant list 110a and a list of chemical reactions, namely a chemical reaction list 110c, through the generating function 100a. The processing circuit 100 gives the product list 110b as a new reactant list to the program involved in the generating function 100a through the control function 100b, updates the database 110 having at least one of the reactant list 110a and the product list 110b, and causes the program involved in the generating function 100a to generate a new product list based on the new reactant list and the chemical reaction list 110c.

[0057] Thus, the chemical structure generation device 1 of the embodiment can efficiently generate high-quality learning data suitable as learning data for machine learning, and can improve the efficiency of molecular design by using a learned model generated using the learning data.

[0058] use Figures 4 to 7 The details of such processing will be described.

[0059] Figure 4 This is a flowchart showing the flow of processing performed by the chemical structure generation device according to the first embodiment.

[0060] First, the processing circuit 100 starts generating a compound list by generating function 100a (step S100). The processing circuit 100 sets p=0 in the initial state by generating function 100a, increases p based on the 0th generation reactant list X0, and generates the pth generation reactant list X100. p as a list of the compounds.

[0061] Next, the processing circuit 100 generates a p-th generation reactant list X based on the reactant list 110a by generating a function 100a. p and chemical reaction list 110c to generate the pth generation product list Y p As product list 110b. (Step S200). That is, the processing circuit 100 generates a product list 110b consisting of one or more compounds based on the reactant list consisting of one or more compounds in the reactant list 110a and the chemical reaction list 110c by generating function 100a. Figure 6 The detailed processing of step S200 is described again in detail.

[0062] Next, the processing circuit 100 uses the control function 100b to list the p-th generation product Y p As the new reactant list (the p+1th generation reactant list X p+1 ) is given to the program involved in the generation function 100a (step S300). That is, the processing circuit 100 controls the function 100b to generate the product list Y generated in step S200. p As a new list of reactants X p+1 Given to the program involved in generating function 100a.

[0063] In addition, while executing such processing, the processing circuit 100 updates the database 110 having at least one of the reactant list 110a and the product list 110b through the control function 100b. Here, updating the database 110 means an operation of changing the elements of the database 110 by adding new elements to the database 110, for example, an operation of inputting a part of the elements in the product list 110b as new elements of the new reactant list 110a to the database 110. For example, the processing circuit 100 updates the reactant list 110a in step S300 through the control function 100b.

[0064] In addition, at least a portion of the database 110 updated in this way is used as learning data in machine learning, for example. As an example, the processing circuit 100 generates learning data in machine learning based on the updated reactant list 110a through a learning data generation function not shown in the figure. As an example, the processing circuit 100 can also use the learning data generation function not shown in the figure to establish a corresponding relationship between each element of the updated reactant list 110a and the physical property value of the element, thereby creating supervised data in which a chemical structure and the physical property value of the chemical structure are established with a corresponding relationship, and the supervised data is used as learning data in machine learning. By using such learning data to perform machine learning to generate a learning completed model, it is possible to predict the physical property values ​​of unknown chemical structures, thereby, for example, it is possible to explore chemical structures with desired physical property values. In addition, as another example, the processing circuit 100 can also create unsupervised data based on the updated reactant list 110a through a learning data generation function not shown in the figure, and the unsupervised data is used as learning data in machine learning.

[0065] Next, the processing circuit 100 increases the value of p by 1 through the control function 100b (step S400). max In the case of (“Yes” in step S500), the process is completed. That is, when the number of iterations p exceeds the preset threshold value p max In the case of p, the processing circuit 100 having the control function 100b determines that the termination condition of the generated list creation process is satisfied and terminates the process. On the other hand, when the value of p is less than p max ) ("No" in step S500), the processing circuit 100 repeatedly performs the processing from step S200 to step S400. That is, the processing circuit 100 controls the function 100b to cause the program involved in the generation function 100a to generate a new product list based on the new reactant list in step S300 and the chemical reaction list in the chemical reaction list 110c. In addition, with respect to the termination condition of the product list creation process, the implementation is not limited to the structure that ends when the number of repeated generations p exceeds a constant number of repeated times as described above, and the processing circuit 100 may also have other termination conditions such as when the number of elements in the generated product list reaches a constant value to terminate the product list creation process.

[0066] exist Figure 5 , an example of a compound generated sequentially through such treatment is shown. Figure 5 This is a diagram for explaining the processing performed by the chemical structure generation device according to the first embodiment.

[0067] Compound 10a and compound 10b represent examples of compounds included in the 0th generation reactant list X0 when p=0 in the initial state. When p=0, in step S200, the processing circuit 100 generates compound 11 as the 0th generation product list Y0 based on compound 10a and compound 10b through the generation function 100a. When p=0, in step S300, the processing circuit 100 gives the 0th generation product list Y0 as the first generation reactant list X1 to the program involved in the generation function 100a through the control function 100b. That is, compound 11 is included in the first generation reactant list X1. In addition, the processing circuit 100 having the control function 100b stores these product lists generated by the generation function 100a in the storage unit 120. Then, in step S400, the processing circuit 100 increases the value of p by 1 through the control function 100b, and p=1.

[0068] Next, when p=1, in step S200, the processing circuit 100 generates compound 12a based on compound 10a and compound 11, and generates compound 12b based on compound 11 as the first generation product list Y1 through the generation function 100a. When p=1, in step S300, the processing circuit 100 gives the first generation product list Y1 as the second generation reactant list X2 to the program involved in the generation function 100a through the control function 100b. That is, compound 12a and compound 12b are included in the second generation reactant list X1. Next, in step S400, the processing circuit 100 increases the value of p by 1 through the control function 100b, and p=2.

[0069] Similarly, the processing circuit 100 generates compound 13a based on compound 10b and compound 12b, and generates compound 14 based on compound 10a and compound 13a, by generating function 100a. In this way, the chemical structure generating device 1 can efficiently generate various compounds in sequence. Figure 5 In the example, the chemical reaction of generating compound 13b from compound 12a and compound 11 corresponds to the chemical reaction included in the prohibited reaction list 110d. Therefore, in this case, the processing circuit 100 does not add the generated compound 13b to the product list.

[0070] Next, use Figure 6 ,right Figure 4 The details of the processing of step S200 are described below. Figure 6 The flowchart is Figure 4 A flowchart illustrating the processing flow of step S200 in more detail.

[0071] In addition, as mentioned above, the reactant list X for each generation k, can also be used as any constituent element, except for the compound x representing the reactant k1 …x kn In addition, it also contains special compounds x that represent special operations null 、x intra Etc. These special compounds will be described later. Here, first, the reactant list X of each generation is k The case where these special compounds are not contained is explained.

[0072] In order to avoid complicating the description, Figure 6 In the description, a case is described in which the chemical reaction used for generating the product list is a chemical reaction that generates one product from two reactants.

[0073] First, in step S210, the processing circuit 100 starts the pth generation of the generated list Y in the generated list 110b by generating function 100a. p Here, the processing circuit 100 generates a function 100a based on the 0th generation reactant list X0, the first generation reactant list X1, the second generation reactant list X2, ... the pth generation reactant list X p , start the list of generated products Y for each p generation p Immediately after the processing circuit 100 starts generating the list Y by generating function 100a p After the generation of p When the process reaches the p0th generation, the generated list Y0, Y1...Y p0 The union of is the data of learning data used for machine learning. In other words, the processing circuit 100 generates the generated product list Y0, Y1...Y by the generation function 100a. p0 , generating data to be used as learning data for machine learning.

[0074] First, in step S220, the processing circuit 100 generates a function 100a from the 0th generation reactant list X0, the first generation reactant list X1, the second generation reactant list X2, ... the pth generation reactant list X p Select an element from any list of reactants and extract it as the first element x m .

[0075] For example, as the list of reactants for generation 0 X0 = {x 01 , x 02}、First generation reactant list X1={x 11 , x 12}, consider x 01 = "CH3OH", x 02= "CH4", x 11 = "CH3COOH", x 12 = "C2H5COOH", p = 1. In this case, for example, in step S220, the processing circuit 100 selects an element from the 0th generation reactant list X0 by generating function 100a, and extracts element x 02 = "CH4" as the first element x m In addition, as another example, in step S220, the processing circuit 100 selects an element from the first generation reactant list X1 by generating function 100a, and extracts element x 12 = "C2H5COOH" as the first element x m .

[0076] Next, in step S230, the processing circuit 100 generates a function 100a from the p-th generation reactant list X p Select an element and extract it as the second element x n For example, in the case of p=1, in step S230, the processing circuit 100 selects an element from the first generation reactant list X1 by generating function 100a, and extracts element x 11 = "CH3COOH" as the second element x m Thus, in step S220 and step S230 , the processing circuit 100 extracts one or more compounds from the reactant list by generating the function 100 a .

[0077] In addition, the processing circuit 100 generates a function 100a to process various first elements x m and the second element x n , and repeat the processing of steps S220 to S280.

[0078] In addition, in this iterative process, the processing circuit 100 may process all first elements x m and the second element x n Steps S220 to S280 can be performed on the combination of x. On the contrary, only the first element x can be processed. m and the second element x n Steps S220 to S280 are performed on a portion of the possible selected combinations.

[0079] Next, in step S240 and step S250 , the processing circuit 100 selects a partial structure for chemical reaction from the selected elements by generating the function 100 a .

[0080] In step S240, the processing circuit 100 generates the first element x extracted in step S220 by the generation function 100a.m Extract the first part of the structure y m For example, when the first element x extracted in step S220 m In the case of "C2H5COOH", the structure matches the structure of "R-COOH", so in step S240, the processing circuit 100 extracts the first partial structure y by generating function 100a. m As y m = "COOH". In addition, as another example of partial structure extraction, when the first element x extracted in step S220 m In the case of "C2H5COOH", the structure matches the structure of the form "R-H", so in step S240, the processing circuit 100 extracts the first partial structure y by generating function 100a. m As y m =“H”.

[0081] In addition, similarly, in step S250, the processing circuit 100 generates the second element x extracted in step S230 by the generating function 100a. n Extract the second part structure y n .

[0082] In addition, the processing circuit 100 generates a function 100a to process various first partial structures y m And the second part structure y n , and repeat the processing of steps S240 to S280.

[0083] Next, the processing circuit 100 reads the chemical reaction list 110c from the storage unit 120 by generating the function 100a, and determines whether the structure extracted in step S240 and step S250 is included in the chemical reaction recorded in the chemical reaction list 110c. m and the second part structure y n If the chemical reaction list is inconsistent (“No” in step S260), the new first partial structure y m And the second part structure y n The processing of steps S240 to S280 is repeated. m and the second part structure y n If the chemical reaction list is consistent ("Yes" in step S260), the process proceeds to step S265. m is "COOH", the second part of the structure y nWhen the chemical reaction is "H", these partial structures match the chemical reaction list {"X1-CO-X2", {"X1-COOH", "X2-H"}. Therefore, in this case, the processing circuit 100 determines that a new compound can be generated through the chemical reaction through the generation function 100a, and the process proceeds to step S265.

[0084] In this way, the processing circuit 100 extracts chemical reactions in which one or more compounds extracted in steps S220 and S230 serve as reactants from the chemical reaction list 110 c by using the generation function 100 a .

[0085] Next, in step S265, the processing circuit 100 reads the prohibited response list 110d from the storage unit 120 through the generating function 100a, and determines whether the first partial structure y extracted in step S240 is m and the second partial structure y extracted in step S250 n Is it consistent with the prohibition rule? m and the second part structure y n If the prohibition rule is consistent (“Yes” in step S265), the processing circuit 100 does not perform new processing on the combination of the partial structures, and the new first partial structure y m And the second part structure y n , and repeat the process of step S240 to step S280. m is "-COOH", the second part structure y n In the case of "-CO-CH2-CO-", these partial structures match the prohibited reaction list {"X2-CO-CHX1-CO-X3", {"X1-COOH", "X2-CO-CHX1-CO-X3"}. Therefore, in such a case, the processing circuit 100 determines that the new compound generated by the chemical reaction has no commercial value through the generation function 100a, and the processing circuit 100 does not perform new processing for the combination of partial structures, and for the new first partial structure y m And the second part structure y n , and repeat the processing of steps S240 to S280. m and the second part structure y n If the result does not match the prohibition rule ("No" in step S265), the process proceeds to step S270, and the obtained product is added to the product list Y. p .

[0086] In this way, the processing circuit 100 generates the product list Y by generating the function 100a and also based on the prohibition rule.p In addition, in the above example, the prohibition rule is defined on the basis of chemical reactions, that is, the prohibition rule is a rule that specifies chemical reactions to be excluded from the chemical reaction list in step S260, but the embodiment is not limited to this. That is, the prohibition rule may also be defined on the basis of products and reactants. In other words, the prohibition rule may also be a rule that specifies chemical reactions to be excluded from the product list Y. p Excluded creations, and from creation list Y p In addition, the embodiment is not limited to the above example, and the processing circuit 100 may specify the product list Y by generating function 100a based on the rule that specifies the chemical reactions to be excluded from the chemical reaction list, i.e., the prohibition rule, instead of the rule that specifies the chemical reactions to be excluded from the chemical reaction list, i.e., the permission rule. p .

[0087] Next, in step S270, the processing circuit 100 generates a product y based on the chemical reaction list by using the generation function 100a. mn For example, in the first element x m is "C2H5COOH", the second element x n is "CH4", the first part of the structure y m is "-COOH", the second part structure y n When the chemical reaction list is {"X1-CO-X2", {"X1-COOH", "X2-H"}}, the processing circuit 100 generates the product y by the generation function 100a. mn It is "C2H5-CO-CH3".

[0088] Next, in step S275, the processing circuit 100 determines the product y generated in step S270 by generating function 100a. mn , which becomes the generated list Y p The candidate compound is included in the previously generated product list, that is, whether it is a compound that has been generated in the previous process. The processing circuit 100 determines that the product y generated in step S270 by the generation function 100a mn If the product y is included in the previously generated product list, that is, if it is determined that the product y is a compound that has been generated in the previous process ("Yes" in step S275), the processing circuit 100 generates a new product y for the product y. mn Do not include it in the p-th generation generated list Y p , for the new first part structure y m And the second part structure y n , repeat the processing of steps S240 to S280.

[0089] On the other hand, the processing circuit 100 determines that the product y generated in step S270 is mn If the compound y is not a compound that has been generated in the previous process ("No" in step S275), the process proceeds to step S280, and the processing circuit 100 processes the product y mn Append to the pth generation generation list Y p In this way, the processing circuit 100 generates the product y of the chemical reaction extracted in step S260 by the generating function 100a. nm The product generated in step S270 is added to the product Y composed of one or more compounds. p .

[0090] The processing circuit 100 generates a function 100a for a given first element x. m and the second element x n , change the first part structure y m And the second part structure y n , and for various first part structures y m And the second part structure y n , and the processing of steps S240 to S280 is repeated. In addition, the processing circuit 100 generates the first element x by the generating function 100a. m and the second element x n The processing of step S220 to step S280 is repeated.

[0091] In addition, the embodiment is not limited to the above-mentioned example.

[0092] exist Figure 6 In the embodiment, the chemical reaction for generating a compound is a chemical reaction that generates a product from two reactants (the dimension of the reaction is 2), but the embodiment is not limited to this. As an example, the chemical reaction that can also be used for generating a compound is a chemical reaction that generates a product from three reactants (the dimension of the reaction is 3), and it can also be a chemical reaction that generates two products from one reactant, and it can also be a chemical reaction that generates a product different from the reactant from one reactant.

[0093] For example, the case where the chemical reaction for generating a compound is a chemical reaction that generates one product from three reactants is described. In this case, instead of steps S220 to S230, the processing circuit 100 generates the reactant list X0 to X1 from the 0th to pth generation by the generation function 100a. p Extract the first element x from any list of reactants m1, from the list of reactants X0 to X1 from generation 0 to p p Extract the second element x from any reactant list m2 , and from the pth generation reactant list X p Extract the third element x m3 Alternatively, instead of steps S240 to S250, the processing circuit 100 generates a function 100a from the first element x m1 Extract the first part of the structure y m1 , from the second element x m2 Extract the second part structure y m2 , and from the third element x m3 Extract the third part structure y m3 The processing circuit 100 generates a function 100a to generate a first element x. m1 , the second element x m2 , the third element x m3 Combination of various first part structures y m1 、The second part structure y m2 、The third part structure y m3 The same process is performed for various first elements x m1 , the second element x m2 , the third element x m3 These treatments are performed in combination.

[0094] In addition, for example, the case where the chemical reaction for generating the compound is a chemical reaction that generates two products from one reactant is described. In this case, instead of steps S220 to S230, the processing circuit 100 generates a function 100a from the p-th generation reactant list X p Extract element x m Alternatively, instead of steps S240 to S250, the processing circuit 100 generates a function 100a from the element x. m Extract part of the structure y m In step S260, the processing circuit 100 obtains the partial structure y from the chemical reaction list 110c by generating the function 100a. m The matched chemical reaction is checked for prohibition rules in step S265, and then in step S270, two products y are generated based on the chemical reaction list. m：1 ,y m；2 In step S275, it is determined whether each of these products is a compound that has already been generated. If it is not a compound that has already been generated, then in step S280, the product is added to the p-th generation product list Y. p .

[0095] In the case where the chemical reaction for generating a compound is a chemical reaction that generates r products from q reactants, these processes can be easily generalized.

[0096] In addition, the chemical reaction for generating a compound may be a chemical reaction that generates one product from one reactant, such as a chemical reaction that generates a product "R-Br" from a reactant "R-OH". This chemical reaction can be represented by, for example, {{"R-Br"}, {"R-OH"}}.

[0097] In addition, in the embodiment, the processing circuit 100 having the generation function 100a is configured to generate the reactant lists X0 to X1 from the 0th generation to the pth generation in step S220. p Extract the first element x m , in step S230, from the pth generation reactant list X p Extract the second element x n For example, the processing circuit 100 having the generation function 100a may also select from the 0th to pth generation reactant lists X0 to X1 in step S230. p Extract the second element x n .

[0098] Then, in Figure 6 In the reactant list X p Contains special compounds x null The case of an element is described below. Special compound x null It is a special element that indicates the operation of not selecting a compound. When a special compound x is selected null In the case of , it is treated as a non-selective compound.

[0099] For example, consider the reactant list X p It is composed of n elements x1, x2, ... x as a common compound. n And special compounds x null A list of n+1 elements is constructed, and three elements are extracted from the list with repetitions allowed. k 、x l 、x m So extract three special compounds x null In the case of compounds other than k 、x l 、x m These three dimensional reactions. In addition, when extracting x null , and as x k 、x l So extract 2 x nullIn the case of a compound other than the above, as a compound is not selected for processing, the processing circuit 100 extracts the reactant as x by generating function 100a. k 、x l These two-dimensional reactions. In addition, when extracting x twice null , and as x k Extract an x ​​like this null In the case of compounds other than k This is a one-dimensional reaction. Like this, the reactant list X p Suppose there are n compounds x1, x2…x n And special compounds x null The list of n+1 elements is then repeated and d elements are extracted, so that the processing circuit 100 can count the number of reactants from 1 to d without omission through the generation function 100a, that is, the chemical reactions whose dimension is less than d. That is, the reactant list X p Contains operations that do not select compounds as elements x null , the processing circuit 100 generates a function 100a from the list X of the reaction parts p Extract element x null Instead of extracting the compound, it is possible to extract the element x by null The compounds obtained by the chemical reaction involving the reactants with a relatively small number of reactions are added to the product list Y p .

[0100] In addition, the reactant list X p The operation of changing the structure of a certain compound to form a compound different from the original compound may also be included as an element. Here, as examples of "the operation of changing the structure of a certain compound to form a compound different from the original compound", operations such as intramolecular reaction, isomerization, and functional group substitution are considered.

[0101] For such operations, use Figure 7 , taking the case of intramolecular reaction as an example. Figure 7 This is a diagram for explaining the processing performed by the chemical structure generation device according to the first embodiment.

[0102] First, we will explain the case of a normal chemical reaction that is not an intramolecular reaction. Consider the reactant list X p The compound x0 contained is Figure 7 Here, in the case of a common chemical reaction which is not an intramolecular reaction, such as Figure 7As shown in the middle section, two molecules of compound 15 are aggregated to generate compound 16. In this case, the reactants of the chemical reaction are represented as {"compound 15", "compound 15"}, the product of the chemical reaction is represented as "compound 16", and the chemical reaction is represented as {{"compound 16"}, {{"compound 15", "compound 15"}.

[0103] Next, consider the case involving intramolecular reactions. In the case of intramolecular reactions, such as Figure 7 As shown in the lower part of the figure, a compound 17 is generated by an intramolecular reaction in one molecule of compound 15. In this case, the reactant of the chemical reaction can be represented as "compound 15", the product of the chemical reaction can be represented as "compound 17", and the chemical reaction can be represented as {"compound 17", "compound 15"}, but the processing circuit 100 having the generating function 100a can also be represented as follows: intra As included in the reactant list X p That is, the processing circuit 100 having the generating function 100a uses the reactant of the reaction with the operation x that performs the intramolecular reaction. intra The corresponding special elements are represented by {“Compound 15”, x intra}, the product of the chemical reaction is represented as "Compound 17", and the chemical reaction is represented as {{"Compound 17"}, {"Compound 15", x intra}}, then proceed to Figure 6 The processing circuit 100 having the generating function 100a performs the intramolecular reaction operation x by intra As included in the reactant list X p The element is processed to add the compound obtained by intramolecular reaction of any compound in the reactant list to the product list Y p As shown in this example, by treating operations on molecules as elements included in a set of molecule candidates, the processing circuit 100 having the generation function 100a can systematically generate molecule candidates obtained by operations such as intramolecular reactions.

[0104] In addition, in step S270, the processing circuit 100 generates a product y by generating the function 100a. nm It is usually a certain chemical structure (compound), but the implementation is not limited thereto. For example, in step S270, the processing circuit 100 can also generate a special compound x through the generation function 100a. null 、x intra etc. as the product y nm For example, when all elements selected as reactants are xnull In the case of, the processing circuit 100 generates x in step S270 by generating function 100a. null As the product. In addition, all the elements selected as reactants are x intra In the case of, the processing circuit 100 generates x in step S270 by generating function 100a. intra In addition, when only a specific compound is selected as a reactant, there is no need to perform Figure 6 Step S240, step S250, step S260, step S265, step S275 and other processing are performed.

[0105] As described above, according to the chemical structure generating device 1 of the first embodiment, the processing circuit 100 generates a product list consisting of one or more compounds based on the chemical reaction list and the reactant list consisting of one or more compounds through the generating function 100a, gives the product list as a new reactant list through the control function 100b, and causes the generating unit to generate a new product list based on the new reactant list and the chemical reaction list. Thus, it is possible to efficiently generate a molecular structure suitable for use as learning data for machine learning, a molecular structure with commercial value, and a molecular structure that can actually be manufactured.

[0106] (Second Embodiment)

[0107] In the second embodiment, as a method for more efficiently generating a molecular structure with commercial value, a chemical structure generating apparatus is described which more efficiently generates a molecular structure with commercial value by assigning numerical information to each of candidate compounds and performing control.

[0108] In the second embodiment, the chemical structure generating device 1 assigns numerical information to each of the candidate compounds based on the numerical information list 110e and performs control based on the numerical information. Here, the numerical information assigned to each compound is, for example, information indicating the cost required to generate the compound.

[0109] Typically, this numerical information is assigned to each compound.

[0110] For example, consider the p-th generation reactant list X p From compounds x1, x2, ...x n Composition, generating compounds x1, x2, ..., x n The costs are w1, w2, ..., w n In this case, the numerical information list W corresponding to the p-th generation reactant list p The elements of are w1, w2, …w n .

[0111] In addition, the processing circuit 100 having the generation function 100a can also integrate the reactant list 110a and the numerical information list 110e to generate a compound list X' in which the numerical information is set as an attribute for each compound. p ={x1(w1), x2(w2), ...x n (w n )}.

[0112] In addition, the processing circuit 100 calculates the numerical information for each compound by the generation function 100a. Specifically, the processing circuit 100 having the generation function 100a gives the known numerical information as the numerical information to the compound whose numerical information is known, and for the numerical information of the other compounds, based on the synthesis path of the compound, the numerical information is recursively defined according to the numerical information of the reactant of the compound until the compound whose numerical information is known is reached.

[0113] use Figure 8 as well as Fig. 9 , the process of such processing is described in detail. Figure 8 as well as Fig. 9 This is a flowchart showing the flow of processing performed by the chemical structure generation device according to the second embodiment. Figure 8 It is a flowchart showing the overall flow of the processing. Fig. 9 Yes Figure 8 The process of step S200A is described in detail below. Figure 8 as well as Fig. 9 In the present invention, for the processing that has been described in the first embodiment, repeated description is omitted.

[0114] First, the processing circuit 100 starts generating a compound list by generating function 100a (step S100A). The processing circuit 100 sets p=0 in the initial state by generating function 100a, increases p based on the 0th generation reactant list X0 and the numerical information list W0 for the reactant, and generates the pth generation reactant list X0. p and the pth generation reactant list X p A list of numerical information corresponding to each reactant W p As a list of each compound and a list of numerical information (a list of cost information), as an example, the processing circuit 100 having the generation function 100a initially generates a reactant list X0 = {x1, x2, ... x n} and a list of information such as numerical information (cost information) W0 = {w1, w2, ..., w n}, and based on this information, initially generate a list of reactants X0'={x1(w1), x2(w2), ..., x n (w n )} and start the generation of compound list.

[0115] Next, the processing circuit 100 generates a p-th generation reactant list X based on the reactant list 110a by generating a function 100a. p and the chemical reaction list 110c as a list of chemical reactions, to generate the p-th generation product list Y p As the generated list 110b. In conjunction with this process, the processing circuit 100 also calculates the p-th generation generated list Y by generating function 100a. p The numerical information list W corresponding to each generated p That is, the processing circuit 100 having the generation function 100a generates the p-th generation product list Y in the form of establishing a correspondence between the numerical information and each product. p (Step S200A). Reuse Fig. 9 The detailed processing of step S200A will be described.

[0116] Next, the processing circuit 100 uses the control function 100b to list the p-th generation product Y p And a list of numerical information W corresponding to each of the products p As new reactant lists (the p+1 generation reactant list X p+1 ) and the new numerical information list W p+1 The processing circuit 100 having the control function 100b thus creates a p-th generation product list Y in which the numerical information is associated with each product. p As the new reactant list (the p+1th generation reactant list X p+1 ) is given to the procedures involved in generating function 100a.

[0117] Next, the processing circuit 100 increases the value of p by 1 through the control function 100b (step S400). Next, the processing circuit 100 determines the new reactant list X through the control function 100b. p In the second embodiment, the processing circuit 100 having the generation function 100a performs the use of the numerical information list W p The control of inhibits the generated compounds, so as p increases, the new reactant list X p The number of elements naturally decreases. Therefore, the processing circuit 100 with the control function 100b can use the new reactant list X pIf p is an empty set, set it as the end point of the processing related to p.

[0118] That is, in the new reactant list X p If the set is empty ("Yes" in step S500A), the process ends and a new reactant list X is created. p If it is not an empty set (No in step S500A), the process returns to step S200A. By performing such a stop condition of the process, for example, a compound having a cost less than a given threshold and having a high commercial value can be efficiently generated.

[0119] Next, use Fig. 9 , the details of the processing of step S200A are described. Fig. 9 In the table W corresponding to each reactant, p The following description will be given of a case where information on the cost of a product for each reactant is displayed.

[0120] In addition, Figure 6 Similarly, in Fig. 9 In the description, the chemical reaction used to generate the product list is a chemical reaction in which one product is generated from two reactants. However, as described above, the embodiment is also applicable when the number of reactants or products is different from this.

[0121] First, in step S210A, the processing circuit 100 starts the pth generation of the generated list Y in the generated list 110b by generating the function 100a. p Immediately after the processing circuit 100 starts generating the list Y by generating function 100a p After the generation of p Is an empty set.

[0122] Similar to the first embodiment, in step S220A, the processing circuit 100 generates a function 100a from the 0th generation reactant list X0, the first generation reactant list X1, the second generation reactant list X2, ... the pth generation reactant list X p Select an element from any list of reactants and extract it as the first element x m .

[0123] Next, in step S230A, the processing circuit 100 generates a function 100a from the p-th generation reactant list X p Select an element and extract it as the second element x n Thus, in step S220A and step S230A, the processing circuit 100 extracts one or more compounds from the reactant list by generating the function 100a.

[0124] In addition, the processing circuit 100 generates a function 100a to process various first elements x m and the second element x n , repeat the processing of steps S220A to S280A.

[0125] Next, in step S231, the processing circuit 100 generates a function 100a for the first element x selected in step S220A. m and the second element x selected in step S230A n The cost of the reactants (the sum of the raw material costs) is calculated for the reaction. Here, the cost of the product is the sum of the cost of the reactants (the sum of the raw material costs) and the cost involved in the chemical reaction. However, when the sum of the raw material costs exceeds the allowable cost, the cost of the product exceeds the maximum allowable cost without considering the cost involved in the chemical reaction. Therefore, the processing circuit 100 having the generation function 100a performs a trimming process in step S232 before entering the loop of the partial structure.

[0126] That is, in the cost of the reactants, for example, the first element x selected in step S220A m The cost of m and the second element x selected in step S230A n The cost of n of and w m +w n If the cost exceeds the allowable cost (No in step S232), the processing circuit 100 having the generation function 100a determines that the first element x selected in steps S220A and S230A is m and the second element x n Not a proper element, for the new first element x m and the second element x n The processing of step S220A to step S280A is repeated. On the other hand, in the cost of the reactant, for example, the first element x selected in step S220A m The cost of m and the second element x selected in step S230A n The cost of n of and w m +w n If the allowable cost is not exceeded (YES in step S232 ), the process proceeds to step S240 .

[0127] Next, in step S240 and step S250, the processing circuit 100 selects a partial structure for chemical reaction from the selected elements by generating function 100a. For example, in step S240, the processing circuit 100 generates a partial structure for chemical reaction from the first element x extracted in step S220A by generating function 100a. m Extract the first part of the structure y m In addition, similarly, in step S250, the processing circuit 100 generates the second element x extracted in step S230A by the generating function 100a. n Extract the second part structure y n .

[0128] In addition, the processing circuit 100 generates a function 100a for various first partial structures y m And the second part structure y n , repeat the processing of steps S240 to S280A.

[0129] Next, the processing circuit 100 reads the chemical reaction list 110c from the storage unit 120 by generating the function 100a, and determines whether the structure extracted in step S240 and step S250 is included in the chemical reaction recorded in the chemical reaction list 110c. m and the second part structure y n If the chemical reaction list is inconsistent (“No” in step S260), the new first partial structure y m And the second part structure y n The processing of steps S240 to S280A is repeated. m and the second part structure y n If it matches the chemical reaction list (YES in step S260 ), the process proceeds to step S270 .

[0130] In this way, the processing circuit 100 extracts chemical reactions in which one or more compounds extracted in steps S220A and S230A serve as reactants from the chemical reaction list 110c by using the generation function 100a.

[0131] Next, in step S270, the processing circuit 100 generates a product y based on the chemical reaction list by using the generation function 100a. mn .

[0132] Next, in step S275, the processing circuit 100 determines the product y generated in step S270 by generating function 100a. mn , which becomes the generated list Y pThe candidate compound is included in the previously generated product list, that is, whether it is a compound that has been generated in the previous process. The processing circuit 100 determines that the product y generated in step S270 by the generation function 100a mn If the product y is included in the previously generated product list, that is, if it is determined that the product y is a compound that has been generated in the previous process ("Yes" in step S275), the processing circuit 100 generates the product y. mn , not included in the p-th generation generated list Y p , for the new first part structure y m And the second part structure y n , repeat the processing of steps S240 to S280.

[0133] On the other hand, the processing circuit 100 determines that the product y generated in step S270 is mn If it is not a compound that has been generated in the previous process (NO in step S275), the process proceeds to step S271A.

[0134] In step S271A, the processing circuit 100 having the generation function 100a processes the product y generated in step S270. mn , generate cost w mn Here, the generation cost w mn is the cost of the reactants and w m +w n The cost of chemical reactions mn That is, the processing circuit 100 having the generation function 100a uses the sum of the costs of the reactants that become the raw materials of the candidate compounds. m +w n and the candidate compound y generated by the reactants mn Chemical reaction {y mn ,{x m , x n The costs involved mn The sum of generates and becomes the generated list Y p Each candidate compound y mn Here, the processing circuit 100 with the generation function 100a obtains the cost z of the chemical reaction from the chemical reaction list 110c. mn The chemical structure generating device 1 associates the chemical reaction and the cost of the chemical reaction with the chemical reaction list 110 c and stores the chemical reaction in the storage unit 120 .

[0135] Next, in step S272, the processing circuit 100 determines the product y calculated in step S271A by generating function 100a. mm The generation cost w mn Is it the maximum allowable cost c max Below. In the generation cost w mn Not the maximum allowable cost c max In the following case ("No" in step S272), the product y mn Not appended to the generated list Y p , for the new first part structure y m And the second part structure y n , and repeat the process of step S240 to step S280A. mn is the maximum allowable cost c max In the following case (YES in step S272), the process proceeds to step S280A.

[0136] In this way, the processing circuit 100 having the generation function 100a generates the generated list Y p The candidate compound y mn When the compound y is equal to each candidate compound y, in step S271A, mn The corresponding numerical information is the generation cost w mn , and based on the calculated numerical information, i.e., the generation cost w mn In step S272, it is determined whether the candidate compound y mn Included in spawn list Y p .

[0137] That is, the information representing the cost of the reactants is combined with the reactant list X p The processing circuit 100 having the generation function 100a establishes a corresponding relationship with each of the reactants included, and in step S271A, the processing circuit 100 having the generation function 100a calculates the product list Y based on the information indicating the cost of the reactants. p The candidate compound y mn The cost of mn , and in step S272 the calculated value becomes the generated list Y p The candidate compound y mn The cost of mn and the pre-set threshold, i.e., the maximum allowable cost c max For comparison, at cost w mn Above the threshold, i.e. the maximum allowable cost c max In the case of ("No" in step S272), compound y will not become a candidate mn Included in spawn list Y p.

[0138] In step S280A, the processing circuit 100 generates the product y mn Append to the pth generation generation list Y p , and the product y mn The generation cost w mn Append to the pth generation value information list W p In this way, the processing circuit 100 adds to the product Y composed of one or more compounds by generating the function 100a. p .

[0139] The processing circuit 100 generates a function 100a to generate a first element x m and the second element x n Change the first part structure y m And the second part structure y n , and for various first part structures y m And the second part structure y n , and the processing of steps S240 to S280A is repeated. In addition, the processing circuit 100 generates the function 100a for various first elements x m and the second element x n The processing of step S220A to step S280A is repeated.

[0140] The embodiment is not limited to the above-described examples.

[0141] In the embodiment, the numerical information list W corresponding to each compound is p The example in which the information of the cost of the compound is shown is described, but the embodiment is not limited to this. For example, the numerical information list W corresponding to each compound p It is a score specified for each compound. As an example, the processing circuit 100 having the generating function 100a determines the comprehensive commercial usefulness of each compound based not only on the price of the raw material alone but also on various factors such as the technical difficulty of the processing, the ease of supply of the raw material, and the market demand, and specifies a score for each compound, and compounds with higher scores or exceeding a constant threshold are preferentially added to the next generation of reactant lists. Here, the score specified for each compound may also be a score such that a higher score indicates a more preferred compound, or conversely, a score such that a lower score indicates a more preferred compound. In addition, as a numerical information list W corresponding to each compound p For example, information related to the physical properties of the compound, such as molecular weight, vapor pressure, boiling point, melting point, dipole moment, oil-water partition coefficient (such as logP), etc. can also be used.

[0142] In the embodiment, the processing circuit 100 having the generation function 100a generates the generated product list Y p When a candidate compound is selected, numerical information corresponding to each candidate compound is calculated, and based on the calculated numerical information, it is determined whether the candidate compound is included in the product list Y. p However, the embodiment is not limited to this. For example, the processing circuit 100 having the generation function 100a may generate the generated list Y p When a candidate compound is selected, information other than numerical information corresponding to each candidate compound is also calculated, and based on the calculated information, it is determined whether the candidate compound is included in the product list Y. p Here, as examples of information that is not numerical information, for example, information on whether a specific substituent is contained, information on the classification of hazardous materials during a public period, and the like are cited. In such a case, for example, the processing circuit 100 having the generation function 100a causes the compound that is a candidate for the product list to contain a specific substituent, or selects a compound from the product list Y. p Exclude compounds that are classified into specific dangerous goods categories.

[0143] for Fig. 9 The first element x in the flowchart m and the second element x n The iterative processing involved is for the numerical information list W of each compound p Regardless of the content, the first element x is m and the second element x n The processing circuit 100 having the generation function 100a may also select the first element x according to the priority of the numerical information of each compound, for example, in the order of cost from low to high and the order of score from high to low. m and the second element x n Exploration, generating a list of generated items Y p Alternative y mn .

[0144] In the embodiments, the cost of a compound is not limited to information indicating a simple price of the compound, and may include various costs other than price, such as difficulty of safety management, ease of supply of raw materials, and stability.

[0145] In addition, the numerical information in the embodiment is not limited to numerical information in the form of a scalar, and may be numerical information in the form of a vector or a tensor as necessary.

[0146] In addition, in the embodiment, the processing circuit 100 sequentially generates the p-th generation product list Y based on the 0-th generation reactant list X0. p That is, in the embodiment, a reactant list Y is generated based on an initial list such as the 0th generation reactant list X0. p However, the embodiment is not limited thereto, and a plurality of reactant lists corresponding to each of the plurality of initial lists may be generated based on the plurality of initial lists.

[0147] For example, consider the case where a certain compound (the compound itself is not necessarily the calculation object of the present compound generating device) requires more than two raw material groups (for example, group A, group B, group C...), and a compound list for each raw material group (a compound list for group A, a compound list for group B, a compound list for group C...) is generated within the allowable cost.

[0148] At this time, for example, the processing circuit 100 generates a function 100a based on the 0th generation reactant list X associated with the compound of group A. 0，A Generate a list Y of compounds of generation p related to compounds of group A p，A , based on the list of 0th generation reactants X associated with compounds of group B 0，B Generate a list Y of compounds of generation p related to compounds of group B p，B , based on the list of 0th generation reactants X associated with compounds of group C 0，C Generate a list Y of compounds of generation p related to compounds of group C p，c .

[0149] First, as a first example, consider a case where an allowable cost is specified for each of the raw material groups. In this case, the processing circuit 100 generates a function 100a to set the allowable cost specified for each of the raw material groups to Fig. 9 The maximum allowable cost c in step S272 max , and execute the processing described above. For example, consider the case where a list of compounds of two raw material groups, group A and group B, is generated, and the overall allowable cost c is given, and the ratio of the allowable cost of group A to the allowable cost of group B is set to 8:2. In this case, the processing circuit 100 is set to allow the cost of group A to be c. max =0.8c, for group B, the allowable cost is c max =0.2c, and perform various treatments.

[0150] Next, as a second example, consider a case where the allowable cost is not set for each of the raw material groups, but for the raw material group as a whole. In this case, the processing circuit 100 generates the function 100a to perform each process on each of the raw material groups in parallel or alternately, and then Fig. 9 In the processing of step S272 of , the information on the allowable cost of the compound list generated in the other raw material groups is referred to, and it is determined whether the production cost of the product is less than the maximum allowable cost. As an example, in the processing of step S272 in the process of generating the compound list of group A, if the sum of the production cost of the product currently being generated and the minimum value of the production cost in the compound list of group B exceeds the overall allowable cost, the processing circuit 100 does not add the product to the product list through the generation function 100a.

[0151] By performing the above-described processing, the processing circuit 100 can simultaneously and in parallel generate compound lists related to two or more raw material groups.

[0152] As described above, in the second embodiment, the processing circuit 100 assigns numerical information to each candidate compound and controls the generation of chemical structures, thereby more efficiently generating molecular structures with commercial value suitable for use as learning data for machine learning.

[0153] As described above, according to the embodiment of the present invention, it is possible to efficiently generate a molecular structure suitable for use as learning data for machine learning.

[0154] In addition, the embodiments of the present invention are presented as examples only, and various substitutions, changes, etc. can be made.

[0155] Regarding the above-mentioned embodiment, the following additional notes are disclosed as optional features of the invention.

[0156] (Note)

[0157] A data processing system comprises: a database having at least one of a reactant list consisting of one or more compounds and a product list consisting of one or more compounds, and a chemical structure generating device.

[0158] The chemical structure generating device has:

[0159] a generating unit, which generates the product list based on the reactant list and the chemical reaction list; and

[0160] The control unit provides the product list as a new reactant list to the generation unit, updates the database, and causes the generation unit to generate a new product list based on the new reactant list and the chemical reaction list.

[0161] Explanation of reference numerals: 100…processing circuit; 100a…generation function; 100b…control function; 110…database; 110a…reactant list; 110b…product list; 110c…chemical reaction list; 110d…prohibited reaction list; 110e…numerical information list; 120…storage unit; 130…input device; 140…display.

Claims

1. A chemical structure generating device, comprising: a generating unit that generates a product list consisting of one or more compounds based on the chemical reaction list and the reactant list consisting of one or more compounds; and a control unit, giving the product list as a new reactant list to the generating unit, updating a database having at least one of the reactant list and the product list, and causing the generating unit to generate a new product list based on the new reactant list and the chemical reaction list, The information is associated with each of the above reactants. The generating unit further calculates information corresponding to each of the candidate compounds when generating the candidate compounds for the product list, and determines whether to include the candidate compounds in the product list based on the calculated information. The above information corresponding to each of the above reactants is numerical information. The numerical information associated with each of the reactants is information indicating the cost of the product of each of the reactants. The generation unit generates numerical information corresponding to each of the candidate compounds using a sum of a cost of the reactant as a raw material of the candidate compound and a cost associated with a chemical reaction for generating the candidate compound from the reactant.

2. The chemical structure generating device according to claim 1, wherein: The generation unit extracts one or more compounds from the reactant list, extracts chemical reactions in which the one or more compounds serve as reactants from the chemical reaction list, and adds products of the chemical reactions to the product list consisting of the one or more compounds.

3. The chemical structure generating device according to claim 1 or 2, wherein: The generation unit further generates the product list based on a prohibition rule, wherein the prohibition rule is a rule that specifies products excluded from the generated product list or chemical reactions excluded from the chemical reaction list.

4. The chemical structure generating device according to claim 1 or 2, wherein: The generating unit determines whether the candidate compound for the product list is included in a previously generated product list, and if it is determined that the candidate compound is included in the previously generated product list, does not include the candidate compound in the product list.

5. The chemical structure generating device according to claim 1 or 2, wherein: The numerical information is associated with each of the above reactants. The generating unit generates candidates for the generated object list according to a priority level based on the numerical information.

6. The chemical structure generating device according to claim 1, wherein: The information representing the cost of the reactants is associated with each of the reactants. The generating unit calculates the cost of the candidate compound for the product list based on the information indicating the cost of the reactant, and compares the calculated cost of the candidate compound for the product list with a preset threshold value. When the cost exceeds the threshold value, the candidate compound is not included in the product list.

7. The chemical structure generating device according to claim 1 or 2, wherein: The above list of reactants includes operations that do not select compounds as elements, The generation unit adds, to the product list, compounds obtained by chemical reactions involving reactants having a smaller number than when the element is not extracted, by extracting the element from the reactant list instead of extracting the compound from the reactant list.

8. The chemical structure generating device according to claim 1 or 2, wherein: The above list of reactants includes, as elements, operations for transforming the structure of a certain compound to form a compound different from the above compound.

9. The chemical structure generating device according to claim 1 or 2, wherein: The generating unit adds a compound obtained by performing an intramolecular reaction on any one of the compounds in the reactant list to the product list by treating an operation of performing an intramolecular reaction as an element included in the reactant list.

10. The chemical structure generating device according to claim 1 or 2, wherein: The control unit stores the product list generated by the generation unit in the storage unit as the database.

11. The chemical structure generating device according to claim 1, wherein: At least a portion of the updated database is used as learning data in machine learning.

12. A chemical structure generating device, comprising: a generating unit that generates a product list consisting of one or more compounds based on the chemical reaction list and the reactant list consisting of one or more compounds; and a control unit, giving the product list as a new reactant list to the generating unit, updating a database having at least one of the reactant list and the product list, and causing the generating unit to generate a new product list based on the new reactant list and the chemical reaction list, The information is associated with each of the above reactants. The generating unit further calculates information corresponding to each of the candidate compounds when generating the candidate compounds for the product list, and determines whether to include the candidate compounds in the product list based on the calculated information. The information representing the cost of the reactants is associated with each of the reactants. The generating unit calculates the cost of the candidate compound for the product list based on the information indicating the cost of the reactant, and compares the calculated cost of the candidate compound for the product list with a preset threshold value. When the cost exceeds the threshold value, the candidate compound is not included in the product list.

13. The chemical structure generating device according to claim 12, wherein: The generation unit extracts one or more compounds from the reactant list, extracts chemical reactions in which the one or more compounds serve as reactants from the chemical reaction list, and adds products of the chemical reactions to the product list consisting of the one or more compounds.

14. The chemical structure generating device according to claim 12 or 13, wherein: The generation unit further generates the product list based on a prohibition rule, wherein the prohibition rule is a rule that specifies products excluded from the generated product list or chemical reactions excluded from the chemical reaction list.

15. The chemical structure generating device according to claim 12 or 13, wherein: The generating unit determines whether the candidate compound for the product list is included in a previously generated product list, and if it is determined that the candidate compound is included in the previously generated product list, does not include the candidate compound in the product list.

16. The chemical structure generating device according to claim 12, wherein: The above information corresponding to each of the above reactants is numerical information.

17. The chemical structure generating device according to claim 12 or 13, wherein: The numerical information is associated with each of the above reactants. The generating unit generates candidates for the generated object list according to a priority level based on the numerical information.

18. The chemical structure generating device according to claim 12 or 13, wherein: The above list of reactants includes operations that do not select compounds as elements, The generation unit adds, to the product list, compounds obtained by chemical reactions involving reactants having a smaller number than when the element is not extracted, by extracting the element from the reactant list instead of extracting the compound from the reactant list.

19. The chemical structure generating device according to claim 12 or 13, wherein: The above list of reactants includes, as elements, operations for transforming the structure of a certain compound to form a compound different from the above compound.

20. The chemical structure generating device according to claim 12 or 13, wherein: The generating unit adds a compound obtained by performing an intramolecular reaction on any one of the compounds in the reactant list to the product list by treating an operation of performing an intramolecular reaction as an element included in the reactant list.

21. The chemical structure generating device according to claim 12 or 13, wherein: The control unit stores the product list generated by the generation unit in the storage unit as the database.

22. The chemical structure generating device according to claim 12, wherein: At least a portion of the updated database is used as learning data in machine learning.

23. A storage medium storing a chemical structure generation program, wherein the chemical structure generation program causes a computer to perform the following processing: A process of generating a product list consisting of one or more compounds based on a chemical reaction list and a reactant list consisting of one or more compounds; Treat the above product list as a new reactant list; Performing a process of updating a database having at least one of the reactant list and the product list; and generating a new product list based on the new reactant list and the chemical reaction list, The information is associated with each of the above reactants. In the process of generating the product list, when generating candidate compounds for the product list, information corresponding to each of the candidate compounds is calculated, and whether to include the candidate compounds in the product list is determined based on the calculated information. The above information corresponding to each of the above reactants is numerical information. The numerical information associated with each of the reactants is information indicating the cost of the product of each of the reactants. In the process of generating the above-mentioned product list, numerical information corresponding to each compound among the above-mentioned candidate compounds is generated using the sum of the cost of the above-mentioned reactants as raw materials of the above-mentioned candidate compounds and the cost involved in the chemical reaction of generating the above-mentioned candidate compounds from the above-mentioned reactants.

24. A storage medium storing a chemical structure generation program, wherein the chemical structure generation program causes a computer to perform the following processing: A process of generating a product list consisting of one or more compounds based on a chemical reaction list and a reactant list consisting of one or more compounds; Treat the above product list as a new reactant list; Performing a process of updating a database having at least one of the reactant list and the product list; and generating a new product list based on the new reactant list and the chemical reaction list, The information is associated with each of the above reactants. In the process of generating the product list, when generating candidate compounds for the product list, information corresponding to each of the candidate compounds is calculated, and whether to include the candidate compounds in the product list is determined based on the calculated information. The information representing the cost of the reactants is associated with each of the reactants. In the process of generating the above-mentioned product list, the cost of the compound that is a candidate for the above-mentioned product list is calculated based on the information representing the cost of the above-mentioned reactants, and the calculated cost of the compound that is a candidate for the above-mentioned product list is compared with a predetermined threshold value. If the above-mentioned cost exceeds the above-mentioned threshold value, the compound that is a candidate for the above-mentioned product list is not included in the above-mentioned product list.

25. A chemical structure generation method, which is a chemical structure generation method performed by a chemical structure generation device, wherein: Based on a chemical reaction list and a reactant list consisting of one or more compounds, a product list consisting of one or more compounds is generated; Give the above product list as a new reactant list; updating a database having at least one of the reactant list and the product list; as well as Generate a new product list based on the new reactant list and the chemical reaction list. The information is associated with each of the above reactants. The process of generating the product list includes: when generating candidate compounds for the product list, further calculating information corresponding to each of the candidate compounds, and determining whether to include the candidate compounds in the product list based on the calculated information, The above information corresponding to each of the above reactants is numerical information. The numerical information associated with each of the reactants is information indicating the cost of the product of each of the reactants. The process of generating the above-mentioned product list includes: using the sum of the cost of the above-mentioned reactants that serve as raw materials for the above-mentioned candidate compounds and the cost involved in the chemical reaction that generates the above-mentioned candidate compounds based on the above-mentioned reactants, to generate numerical information corresponding to each compound among the above-mentioned candidate compounds.

26. A chemical structure generation method, which is a chemical structure generation method performed by a chemical structure generation device, wherein: Based on a chemical reaction list and a reactant list consisting of one or more compounds, a product list consisting of one or more compounds is generated; Give the above product list as a new reactant list; updating a database having at least one of the reactant list and the product list; as well as Generate a new product list based on the new reactant list and the chemical reaction list. The information is associated with each of the above reactants. The process of generating the product list includes: when generating candidate compounds for the product list, further calculating information corresponding to each of the candidate compounds, and determining whether to include the candidate compounds in the product list based on the calculated information, The information representing the cost of the reactants is associated with each of the reactants. The process of generating the above-mentioned product list includes: calculating the cost of the compound that is a candidate for the above-mentioned product list based on the information representing the cost of the above-mentioned reactants, and comparing the calculated cost of the compound that is a candidate for the above-mentioned product list with a predetermined threshold value. If the above-mentioned cost exceeds the above-mentioned threshold value, the compound that is a candidate for the above-mentioned product list is not included in the above-mentioned product list.

Images