JP7338315B2 - Chemical substance search device, chemical substance search method, and chemical substance search program - Google Patents

Description

The present invention relates to a chemical substance searching device, a chemical substance searching method, and a chemical substance searching program for searching for at least one of new chemical substances suitable for a certain use and new uses of existing chemical substances.

Conventionally, the search for new chemical substances suitable for certain uses or the search for new uses of chemical substances has involved a lot of trial-and-error searches through experiments, and the searches have required a lot of time and money.

As new chemicals were discovered, databases were built with property data for the new chemicals. The database can be searched for chemical substances corresponding to the property data stored in the database. In addition to that, the database utilizes the structural data of the property data stored in the database, and among the chemical substances represented by the property data stored in the database, other chemical substances that are similar in structure to a predetermined chemical substance. chemicals can be searched in the database.

For example, as a database related to inorganic compounds, there is an inorganic crystal structure database (ICSD) that stores names, molecular formulas, three-dimensional atomic coordinate values, crystallographic data, and the like of inorganic compounds.
In ICSD, for example, when searching for inorganic compounds similar to Li _1.865 CoP ₂ O ₇ [#261899 (14)], which is a positive electrode material for secondary batteries, for example, as shown in FIG. Compounds [ _Li1.73FeP2O7 ; #174531(14), _Li1.855Co ( _P2O7 ) _; #183602(14) _{, Li2Mn} ( _{P2O7 )} ; #419562( ₁₄ ) , and Li ₂ Mn(P ₂ O ₇ ); #253224(14)] are obtained as a result of the search. In FIG. 25, the number following # represents the number of the inorganic compound in the database, and "(14)" represents the number of the space group.
However, in this method, the inorganic compounds obtained as search results are all known as battery materials. Can't find a use for it. Also, naturally, it is impossible to find new inorganic compounds that are suitable for a certain application and are not stored in the database.

On the other hand, in a database of chemical substances, a method of automatically classifying chemical substances based on their partial structures has been proposed in order to improve the classification efficiency of chemical substances.
However, even with this method, although classification based on structural similarity is possible, it is not possible to find new uses for the classified chemical substances. Also, of course, it is not possible to find new chemicals that are suitable for a certain application and are not stored in the database.

JP 2006-113893 A

This case provides a chemical substance search device, a chemical substance search method, and a chemical substance search program that can search for at least one of new chemical substances suitable for a certain use and new uses of existing chemical substances. With the goal.

In one aspect, the chemical substance search apparatus of the present invention includes chemical substance data including properties of each of a plurality of chemical substances, and similarity data including similarities between two of the plurality of chemical substances. wherein there are two or more types of similarities, two or more types of similarity data, and each of the two or more types of similarities has at least one type of characteristic. for a use associated with at least one of said properties using said similarity data of a chemical substance database, each of said similarity data comprising one or more of two or more said similarities searching for chemicals that are suitable chemicals but not stored in the chemicals database; and properties of chemicals stored in the chemicals database but within the chemicals data for the chemicals. a search unit that searches for at least one of the search for the characteristic that does not exist.

In another aspect, the present chemical substance search method comprises:
the computer
A chemical substance database including chemical substance data including properties of each of a plurality of chemical substances and similarity data including similarities between two of the plurality of chemical substances, wherein the similar there are two or more types of similarity data, two or more types of similarity data, each of the two or more types of similarities is generated based on one or more types of characteristics, and each of the similarity data is a chemical suitable for use in relation to at least one of said properties, but in said chemical database Searching for chemicals that are not stored and/or searching for properties of chemicals that are stored in the chemical database but are not in the chemical data for the chemicals.

In another aspect, the subject chemical substance discovery program includes:
to the computer,
A chemical substance database including chemical substance data including properties of each of a plurality of chemical substances and similarity data including similarities between two of the plurality of chemical substances, wherein the similar there are two or more types of similarity data, two or more types of similarity data, each of the two or more types of similarities is generated based on one or more types of characteristics, and each of the similarity data is a chemical suitable for use in relation to at least one of said properties, but in said chemical database causing a search to occur for chemicals not stored and/or properties for chemicals stored in the chemical database but not in the chemical data for the chemical; .

In one aspect, it is possible to provide a chemical substance searching device capable of searching for at least one of a new chemical substance suitable for a certain use and a new use of an existing chemical substance.
Also, in one aspect, it is possible to provide a chemical substance search method capable of searching for at least one of new chemical substances suitable for a certain use and new uses of existing chemical substances.
Also, in one aspect, it is possible to provide a chemical substance search program capable of searching for at least one of new chemical substances suitable for a certain use and new uses of existing chemical substances.

FIG. 1 is a diagram showing an example of how acetic acid and methyl acetate are graphed. FIG. 2 is a diagram showing an example of a combination when the same elements in molecules A and B are combined to create nodes of a conflict graph. FIG. 3 is a diagram showing an example of edge creation rules in a conflict graph. FIG. 4 is a diagram showing an example of a conflict graph between molecule A and molecule B. FIG. FIG. 5 is a diagram showing an example of maximum independent sets in a graph. FIG. 6 is a diagram showing an example of the flow of finding the maximum common substructure between molecule A and molecule B by finding the maximum independent set of the conflict graph (solving the maximum independent set problem). FIG. 7 is an explanatory diagram for explaining an example of a method of searching for the maximum independent set in a graph having six nodes. FIG. 8 is an explanatory diagram for explaining an example of a method of searching for the maximum independent set in a graph having six nodes. FIG. 9 is a diagram showing an example of maximum independent sets in the conflict graph. FIG. 10 is a schematic diagram showing an example of the flow when performing the Weisfeiler-Lehman procedure and assigning WL labels to node atoms. FIG. 11 is a diagram showing a configuration example of the disclosed chemical substance search device. FIG. 12 is a diagram showing another configuration example of the disclosed chemical substance search device. FIG. 13 is a diagram showing another configuration example of the disclosed chemical substance search device. FIG. 14 is a diagram showing a functional configuration example as one embodiment of the disclosed chemical substance search device. FIG. 15 is a diagram showing a functional configuration example as another embodiment of the disclosed chemical substance search device. FIG. 16 is a flowchart of an example method for creating a chemical database. FIG. 17 is a flow chart of another example of a method of creating a chemical substance database. FIG. 18 is a flow chart of one example of the disclosed chemical substance search method. FIG. 19 is a flow chart of another example of the disclosed chemical substance search method. FIG. 20 is a flow chart combining the flow chart of FIG. 16 and the flow chart of FIG. FIG. 21 is a diagram visualizing an example of searching for a new battery material with a high discharge potential using the disclosed chemical substance search method. FIG. 22 is a flowchart of a specific example of the disclosed chemical substance search method. FIG. 23 is a diagram visualizing part of the chemical substance database created according to the flowchart of FIG. FIG. 24 is a diagram visualizing the results of searching for Li compounds according to the flowchart of FIG. FIG. 25 is an example of search results of Li compounds by ICSD.

(Chemical Substance Search Device, Chemical Substance Search Method, and Chemical Substance Search Program)
The chemical substance search device of this case has a search unit and, if necessary, a chemical substance database.
In the chemical substance search method of this case, a computer performs a search using similarity data in a chemical substance database.
The chemical substance search program of this case causes a computer to perform a search using similarity data in a chemical substance database.
The chemical substance searching device of the present case performs, for example, the chemical substance searching method of the present case.
The chemical substance search method of the present case is performed, for example, by the chemical substance search program of the present case.

The chemical database includes chemical data and similarity data.
The chemical substance data is data containing the properties of each chemical substance of a plurality of chemical substances.
Similarity data is data that includes the similarity of two chemicals of a plurality of chemicals.
There are two or more types of similarities in the chemical substance database.
There are two or more types of similarity data in the chemical substance database.
In the chemical database, each similarity of two or more similarities is generated based on one or more properties.
In the chemical database, each piece of similarity data includes one or more of two or more similarities.

The searching unit uses the similarity data of the chemical substance database to perform at least one of the following searches (1) and (2).
(1) searching for chemicals that are suitable for the application associated with at least one of the properties, but are not stored in the chemical database; (2) are chemical properties stored in the chemical database; but find properties that are not in the substance data for that chemical

According to the disclosed chemical substance search device, chemical substance search method, and chemical substance search program, it is possible to search for at least one of a new chemical substance suitable for a certain use and a new use of an existing chemical substance. "New" here does not mean new in the world, but refers to chemical substances that are not included in the chemical substance database used, or uses that are not included in the chemical substance data of the chemical substance database used. In terms of new means new.

<Chemical substance database>
The chemical substance database includes chemical substance data, similarity data, and, if necessary, other data.
Chemical substances stored in the chemical substance database are not particularly limited and can be appropriately selected according to the purpose. Examples include inorganic compounds, organic compounds, salts, polymers, proteins, DNA, complexes, and the like.
The chemical substance database may be configured by incorporating information from one or more existing chemical substance databases.
The existing chemical substance database is not particularly limited and can be appropriately selected according to the purpose. Crystal structure database (CRYSTMET), mass spectrum database (NIST17), mass spectrum database (Wiley Registry), CAS REGISTRY and the like.

The data structure of the chemical substance database is not particularly limited and can be appropriately selected according to the purpose. The chemical substance database may be a graph database or a relational database.
A graph database contains graphs. A graph includes a group of nodes (vertices) and a group of edges (branches) representing connectivity relationships between the nodes.
A relational database is a database designed based on a relational model (relational data model).
When the chemical substance database is a graph database, chemical substance data constitute nodes and similarity data constitute edges representing relationships between nodes.
A graph in a graph database is represented by G=(V, E), where V is the set of nodes and E is the set of edges. V is a finite set and E is a set of two-element sets chosen from V.

A chemical substance database is, for example, a so-called knowledge base that enables retrieval of knowledge and organizes the knowledge.

<<Chemical substance data>>
The chemical substance data is data containing the properties of each chemical substance of a plurality of chemical substances.
The properties of a chemical substance are, for example, information for specifying the chemical substance, and the properties of the chemical substance include, for example, chemical properties, physical properties, and the like.
Properties of chemical substances include, for example, name, molecular weight, compositional formula, molecular formula, structural formula, chemical formula, electronic formula, amino acid sequence, three-dimensional structure, crystal structure (crystal system, space group, point group), boiling point, melting point. , electrical conductivity, ionic conductivity, discharge potential, dielectric constant, thermal conductivity, specific heat, and the like.

A chemical property may also include derived properties that are derived and generated from a chemical property. Derived properties include, for example, derived properties obtained by translating at least one component of the properties of a chemical substance into a higher-level concept.
An example of a derived property is a derived property obtained by translating at least one constituent element of a composition formula, which is a property of a chemical substance, into a higher concept. For example, taking the chemical substance Li _1.865 CoP ₂ O ₇ as an example, in the composition formula Li _1.865 CoP ₂ O ₇ , the constituent element Co is a transition metal (T), which is one of its superordinate concepts. to produce the derivative characteristic Li _1.865 TP ₂ O ₇ . The derived property produced is one of the properties of the chemical Li _1.865 CoP ₂ O ₇ .

<<Similarity data>>
Similarity data is data that includes the similarity of two chemicals of a plurality of chemicals.
There are two or more types of similarities in the chemical substance database.
There are two or more types of similarity data in the chemical substance database.
In the chemical database, each similarity of two or more types of similarity is generated based on one or more properties.
In the chemical substance database, each piece of similarity data includes one or more types of two or more similarities, and may include two or more types.
In a chemical database, there need not be similarity data for every combination of two chemicals in every chemical.

The similarity is not particularly limited as long as it is generated based on one or more characteristics, and can be appropriately selected according to the content to be searched, the type of chemical substance database to be created, and the like.
The method for generating similarity is not particularly limited and can be appropriately selected according to the purpose. For example, the following methods (1) and (2) for evaluating the degree of structural similarity between compounds mentioned.
(1) Fingerprint method (2) A method of searching for common partial structures between compounds by solving the maximum independent set problem of the conflict graph with an Ising model equation and solving it with an annealing machine or the like (for example, the following non- See Patent Document X)
Non-Patent Literature X: Maritza Hernandez, Arman Zaribafiyan, Maliheh Aramon, Mohammad Naghibi "A Novel Graph-based Approach for Determining Molecular Similarity". arXiv: 1601.06693 (https://arxiv.org/pdf/1601.06693.pdf)
Note that the method of generating similarity is not limited to these methods. For example, similarity is generated by cosine similarity, correlation coefficient, correlation function, edit distance (also called Levenshtein distance), etc. related to similarity calculation. You may

The similarity may include, for example, a similarity score. In this case, for example, it is determined that the higher the score, the higher the similarity.
Regarding similarity data, for example, when the similarity of a certain property between two chemical substances is low, the chemical substance database stores the similarity including the similarity between the two chemical substances with the low similarity. It does not have to have data. Whether the similarity is high or low can be appropriately determined using, for example, a similarity score (score). In other words, when the similarity score for similarity is below the threshold value, there may be no similarity data including the similarity between two chemical substances with low similarity.

At least one of the two or more types of similarity data preferably includes two or more types of similarity. By doing so, it is possible to perform an evaluation that considers multiple similarities, making it easier to find chemical substances with high similarities.

At least one of the two or more types of similarity data preferably includes numerical difference data of at least one property between the two chemical entities. By doing so, given the difference between two chemicals with respect to that property, the search can find the chemical that has that property better.

At least one type of similarity data of the two or more types of similarity data preferably includes similarity generated based on derived characteristics.
Derived properties are properties that are not stored in the regular chemical database. Therefore, at least one type of similarity data among two or more types of similarity data includes similarity generated based on derived characteristics, so that the search is performed from a different perspective from the search in a known chemical substance database. be able to.
For example, consider the case where at least one of the two or more types of similarity data includes two or more similarities, and one of the similarities is generated based on derived characteristics. In this case, if the similarity data includes two or more types of similarities, the number of combinations of two chemical substances associated with the similarity data may become too small. However, if at least one of the two or more types of similarity is generated based on the derived property, the derived property tends to increase the number of combinations of two chemical substances associated with the similarity data than the original property. contribute to As a result, even when a plurality of similarities are considered for one similarity data in the search, it is possible to adjust the number of combinations of two chemical substances to be associated to be appropriate.

Here, as an example of a similarity (similarity score) calculation method, a method of calculating a similarity by solving a maximum independent set problem of a conflict graph to search for common substructures between compounds to be compared. explain.

When calculating the degree of structural similarity between compounds by solving the maximum independent set problem of the conflict graph, the compounds are graphed and handled. Here, graphing a compound means representing the structure of the compound using, for example, information on the types of atoms (elements) in the compound and information on the bonding state of each atom.
The structure of a compound can be expressed using, for example, the MOL format or the SDF (Structure Data File) format. Generally, the SDF format means that structural information of a plurality of compounds expressed in MOL format is collected in one file. Further, in the SDF format file, it is possible to handle additional information (for example, catalog number, CAS number, molecular weight, etc.) for each compound in addition to the MOL format structural information. The structures of these compounds are, for example, "atom 1 (name), atom 2 (name), element information of atom 1, element information of atom 2, bond order of atom 1 and atom 2" into one It can be graphed in CSV (Comma-Separated Value) format with rows.

In the following, a method of creating a conflict graph will be described with an example of creating a conflict graph between acetic acid (CH ₃ COOH) and methyl acetate (CH ₃ COOCH ₃ ).
First, acetic acid (hereinafter sometimes referred to as "molecule A") and methyl acetate (hereinafter sometimes referred to as "molecule B") are graphed as shown in FIG. In FIG. 1, the atoms forming acetic acid are indicated by A1, A2, A3, and A5, and the atoms forming methyl acetate are indicated by B1 to B5. In addition, in FIG. 1, A1, A2, B1, B2, and B4 represent carbon, A3, A5, B3, and B5 represent oxygen, single bonds are represented by thin solid lines, and double bonds are represented by thick solid lines. ing. In the example shown in FIG. 1, atoms other than hydrogen are selected and graphed, but when graphing a compound, all atoms including hydrogen may be selected and graphed.

Next, vertices (atoms) in graphed molecules A and B are combined to create vertices (nodes) of the conflict graph. At this time, for example, as shown in FIG. 2, it is preferable to combine the same elements in molecules A and B to form a conflict graph node. In the example shown in FIG. 2, a combination of A1, A2, B1, B2, and B4 representing carbon and a combination of A3, A5, B3, and B5 representing oxygen are nodes of the conflict graph. In this way, by using combinations of the same elements as nodes, it is possible to create a conflict graph with nodes that can be included in the maximum independent set. Fewer computer bits are required.
In the example of FIG. 2, the combination of carbon in molecule A and carbon in molecule B creates 6 nodes, and the combination of oxygen in molecule A and oxygen in molecule B creates 4 nodes. The number of nodes in the conflict graph created from B is ten.

Next, create edges (branches, sides) in the conflict graph. At this time, compare two nodes and create an edge between the two nodes if they consist of atoms with different states (e.g. atomic number, presence or absence of bonds, bond order, etc.) do. On the other hand, if two nodes are compared and consist of atoms in the same situation, no edge is created between the two nodes.

Now, with reference to FIG. 3, rules for creating edges in the conflict graph will be described.
First, in the example shown in FIG. 3, whether or not to create an edge between node [A1B1] and node [A2B2] will be described. As can be seen from the structure of molecule A graphed in FIG. 3, carbon A1 of molecule A contained in node [A1B1] and carbon A2 of molecule A contained in node [A2B2] are bonded to each other (single bond). ing. Similarly, carbon B1 of molecule B included in node [A1B1] and carbon B2 of molecule B included in node [A2B2] are bonded to each other (single bond). That is, the state of bonding between carbon A1 and carbon A2 and the state of bonding between carbon B1 and carbon B2 are the same.
Thus, in the example of FIG. 3, the states of carbon A1 and carbon A2 in molecule A are the same as the states of carbon B1 and carbon B2 in molecule B, and node [A1B1] and node [A2B2] are nodes composed of atoms in the same situation. Therefore, in the example shown in FIG. 3, no edge is created between node [A1B1] and node [A2B2].

Next, in the example shown in FIG. 3, whether or not to create an edge between node [A1B4] and node [A2B2] will be described. As can be seen from the structure of molecule A graphed in FIG. 3, carbon A1 of molecule A contained in node [A1B4] and carbon A2 of molecule A contained in node [A2B2] are bonded to each other (single bond). ing. On the other hand, as can be seen from the graphed structure of molecule B, oxygen B3 is sandwiched between carbon B4 of molecule B contained in node [A1B4] and carbon B2 of molecule B contained in node [A2B2]. , not directly coupled. That is, the state of bonding between carbon A1 and carbon A2 differs from the state of bonding between carbon B4 and carbon B2.
That is, in the example of FIG. 3, the states of carbon A1 and carbon A2 in molecule A are different from the states of carbon B4 and carbon B2 in molecule B, and node [A1B4] and node [A2B2] are different from each other. Nodes composed of the atoms of the situation become each other. Therefore, in the example shown in FIG. 3, an edge is created between node [A1B4] and node [A2B2].

Thus, the conflict graph creates edges between nodes if they are composed of atoms in different situations, and edges between nodes that are composed of atoms in the same situation. can be created based on the rule not to create edges.
FIG. 4 is a diagram showing an example of a conflict graph between molecule A and molecule B. FIG. As shown in FIG. 4, for example, in nodes [A2B2] and [A5B5], the bond between carbon A2 and oxygen A5 in molecule A and the bond between carbon B2 and carbon B5 in molecule B are the same. It has become. Therefore, since nodes [A2B2] and [A5B5] are composed of atoms in the same situation, an edge is created between nodes [A2B2] and [A5B5]. do not have.

Here, the edges of the conflict graph can be created, for example, based on the chemical structure data of two compounds for calculating the degree of structural similarity. For example, when chemical structure data of a compound is entered using an SDF format file, the edge of the conflict graph is created by performing calculations using a calculator such as a computer based on the information contained in the SDF format file. can (specify)

Next, a method of solving the maximum independent set problem of the created conflict graph in an example of the prior art as described in Non-Patent Document X will be described.
A maximum independent set (MIS) in a conflict graph means a set including the largest number of nodes having no edge between them among the sets of nodes forming the conflict graph. In other words, the maximum independent set in the conflict graph means the set having the maximum size (the number of nodes) among the sets formed by nodes having no edges between them.

FIG. 5 is a diagram showing an example of maximum independent sets in a graph. In FIG. 5, the nodes included in the set are denoted by the code "1", the nodes not included in the set are denoted by the code "0", and the points where edges exist between the nodes are indicated by solid lines. The nodes are connected by dotted lines where no edge exists. Here, as shown in FIG. 5, for simplification of explanation, a graph having six nodes will be explained as an example.
In the example shown in FIG. 5, there are three sets having the maximum number of nodes among the sets composed of nodes having no edge between them, and the number of nodes in these sets is three. That is, in the example shown in FIG. 5, the maximum independent sets in the graph are the three sets surrounded by the dashed-dotted line.

Here, as described above, the conflict graph creates an edge between nodes when the nodes are composed of atoms in different situations, and creates edges between the nodes when the nodes are composed of atoms in the same situation. It is created based on the rule of not creating edges between nodes. For this reason, in the conflict graph, among the sets composed of nodes without edges between the nodes, obtaining the maximum independent set, which is the set with the maximum number of nodes, is the partial structure common between the two molecules. It is synonymous with finding the maximum of In other words, finding the largest independent set in the conflict graph can identify the largest common substructure in two molecules.

That is, two molecules are graphed, a conflict graph is created based on the graphed molecular structure, and the maximum independent set in the conflict graph is obtained, thereby obtaining the maximum common substructure of the two molecules. can.
FIG. 6 shows an example of the flow of determining the maximum common substructure between molecule A (acetic acid) and molecule B (methyl acetate) by determining the maximum independent set of the conflict graph (solving the maximum independent set problem). indicates As shown in FIG. 6, the molecules A and B are graphed, the same elements are combined to form nodes, and edges are formed according to the states of the atoms forming the nodes to create a conflict graph. Then, by obtaining the maximum independent set in the created conflict graph, the maximum common partial structure between molecules A and B can be obtained.

Here, an example of a specific method for obtaining (searching for) the maximum independent set of the conflict graph will be described.
The maximally independent set search of the conflict graph can be done, for example, by using the Hamiltonian, whose minimization means searching for the maximally independent set. More specifically, for example, the Hamiltonian (H) represented by the following formula (1) can be used.

ここで、上記式（１）において、ｎは、コンフリクトグラフにおけるノードの数であり、ｂ_iは、ｉ番目のノードに対するバイアスを表す数値である。
さらに、ｗ_ijは、ｉ番目のノードとｊ番目のノードとの間にエッジが存在するときは、０ではない正の数であり、ｉ番目のノードとｊ番目のノードとの間にエッジが存在しないときは、０である。
また、ｘ_iは、ｉ番目のノードが０又は１であることを表すバイナリ変数であり、ｘ_jは、ｊ番目のノードが０又は１であることを表すバイナリ変数である。
なお、α及びβは、正の数である。

Here, in the above equation (1), n is the number of nodes in the conflict graph, and b _i is a numerical value representing the bias for the i-th node.
In addition, w _ij is a non-zero positive number if an edge exists between the i-th node and the j-th node, and if there is an edge between the i-th node and the j-th node. It is 0 when it does not exist.
Also, x _i is a binary variable that indicates that the i-th node is 0 or 1, and x _j is a binary variable that indicates that the j-th node is 0 or 1.
Note that α and β are positive numbers.

The relationship between the Hamiltonian represented by Equation (1) above and the search for the maximum independent set will be described in more detail. Equation (1) above is a Hamiltonian representing an Ising model equation in Quadratic Unconstrained Binary Optimization (QUBO) format.
In the above formula (1), if x _i is 1, it means that the i-th node is included in the set that is a candidate for the maximum independent set, and if it is 0, the i-th node is included in the maximum independent set is not included in the candidate set for Similarly, in the above equation (1), x _j means that if it is 1, the j-th node is included in the set that is a candidate for the maximum independent set, and if it is 0, the j-th node is It means that it is not included in the set that is a candidate for the maximum independent set.
For this reason, with respect to the above formula (1), under the constraint that there is no edge between nodes whose state is 1 (bit is set to 1), a combination in which as many nodes as possible have a state of 1 is searched for. Thus, the maximum independent set can be searched.

Here, each term in the above formula (1) will be explained.
One item on the right side of the above equation (1) (the term with a coefficient of -α) is that the more i when x _i is 1 (the greater the number of nodes included in the set that is a candidate for the maximum independent set), It is the term whose value becomes smaller. A smaller value in one item on the right side of the above equation (1) means a large negative number. That is, in the above equation (1), when the bits of many nodes become 1 due to the action of one item on the right side, the value of the Hamiltonian (H) decreases.

The two terms on the right side of the above equation (1) (the _term with a coefficient of β) are: This is a penalty term that increases in value. In other words, the two items on the right side of the above equation (1) are 0 when there is no edge between nodes whose bit is 1, and are positive numbers otherwise. That is, in the above equation (1), due to the action of the two items on the right side, if an edge exists between nodes whose bits are 1, the value of the Hamiltonian (H) increases.

As described above, the above formula (1) has a smaller value when the bits of many nodes are 1, and a larger value when there is an edge between nodes whose bits are 1. Therefore, the above formula ( It can be said that minimizing 1) means searching for the maximum independent set.

Here, the relationship between the Hamiltonian represented by the above equation (1) and the search for the maximum independent set will be described using an example with reference to the drawings.
Consider a case where a bit is set in each node as in the example shown in FIG. 7 in a graph having six nodes. In the example of FIG. 7, as in FIG. 5, solid lines connect the nodes where edges exist, and dotted lines connect the nodes where no edges exist.

In the example of FIG. 7, if b _i in the above equation (1) is 1 and w _ij is 1 when an edge exists between the i-th node and the j-th node, then the above equation (1) is It looks like this:

In this way, in the example of FIG. 7, if there is no edge between nodes whose bit is 1 (if there is no contradiction as an independent set), the two items on the right side are 0, and one item is The value becomes the Hamiltonian value as it is.

Next, consider a case where bits are set in each node as in the example shown in FIG. As in the example of FIG. 7, if b _i in the above equation (1) is 1 and w _ij is 1 when an edge exists between the i-th node and the j-th node, then the above equation (1) ) becomes:

Thus, in the example of FIG. 8, since there is an edge between the nodes whose bit is 1, the two terms on the right side are not 0, and the value of the Hamiltonian is the two terms on the right side. becomes the sum of Here, in the example shown in FIGS. 7 and 8, if α>5β, for example, −3α<−4α+5β, so the value of the Hamiltonian in the example of FIG. 7 is higher than the value of the Hamiltonian in the example of FIG. is smaller. The example of FIG. 7 is a set of nodes that is consistent as the maximum independent set, and it can be seen that the maximum independent set can be searched by searching for a combination of nodes that reduces the value of the Hamiltonian in equation (1) above.

Next, a method for calculating the structural similarity between molecules based on the searched maximum independent sets in an example of the prior art as described in Non-Patent Document X will be described.
The degree of structural similarity between molecules can be calculated using, for example, the following formula (2).

ここで、上記式（２）において、Ｓ（Ｇ_Ａ，Ｇ_Ｂ）は、グラフ化した第一の分子（例えば、分子Ａ）とグラフ化した第二の分子（例えば、分子Ｂ）との類似度を表し、０～１で表され、１に近づく程、類似度が高いことを意味する
また、Ｖ_Ａは、グラフ化した第一の分子におけるノード原子の総数を表し、Ｖ_ｃ ^Ａは、グラフ化した第一の分子におけるノード原子の内、コンフリクトグラフの最大独立集合に含まれるノード原子の数を表す。なお、ノード原子とは、グラフ化した分子における頂点の原子を意味する。
さらに、Ｖ_Ｂは、グラフ化した第二の分子におけるノード原子の総数を表し、Ｖ_ｃ ^Ｂは、グラフ化した第二の分子におけるノード原子の内、コンフリクトグラフの最大独立集合に含まれるノード原子の数を表す。
δは、０～１の数である。

Here, in the above formula (2), S(G _A , G _B ) is the similarity between the graphed first molecule (eg, molecule A) and the graphed second molecule (eg, molecule B). The degree is expressed from 0 to 1, and the closer to 1, the higher the degree of similarity. V _A represents the total number of node atoms in the graphed first molecule, and V _c ^A is It represents the number of node atoms included in the maximum independent set of the conflict graph among the node atoms in the graphed first molecule. A node atom means an atom at a vertex in a graphed molecule.
Furthermore, V _B represents the total number of node atoms in the graphed second molecule, and V _c ^B is the node atom included in the maximum independent set of the conflict graph among the node atoms in the graphed second molecule. represents the number of
δ is a number between 0 and 1;

Further, in the above formula (2), max {A, B} means selecting the larger value of A and B, and min {A, B} is the value of A and B. means to choose the smaller one.

Here, similarly to FIG. 1 and the like, acetic acid (molecule A) and methyl acetate (molecule B) will be used as examples to explain the similarity calculation method.
In the conflict graph shown in FIG. 9, the maximum independent set consists of four nodes, node [A1B1], node [A2B2], node [A3B3], and node [A5B5]. That is, in the example of FIG. 9, |V _A | is 4, |V _c ^A | is 4, |V _B | is 5, and |V _c ^B | Also, in this example, if δ is 0.5 and the first molecule and the second molecule are averaged (handled equally), the above equation (2) becomes as follows.

Thus, in the example of FIG. 9, the structural similarity between molecules is calculated to be 0.9 based on the above formula (2).

<Search part>
The searching unit uses the similarity data of the chemical substance database to perform at least one of the following searches (1) and (2).
(1) searching for chemicals that are suitable for the application associated with at least one of the properties, but are not stored in the chemical database; (2) are chemical properties stored in the chemical database; but find properties that are not in the substance data for that chemical

The search in the search unit is executed by, for example, making inferences based on condition settings (rules) related to similarity data.
Inference is performed, for example, by an inference engine. An inference engine is a program that makes inferences based on rules.

Rules are not particularly limited as long as they relate to similarity data, and can be appropriately selected according to the purpose. Examples of rules for performing the search of (1) or (2) above.
A method for creating a rule is not particularly limited, and can be appropriately selected according to the purpose.

In making inferences, various logics and inferences are used alone or in combination, depending on the rules. Such logic and reasoning include, for example, propositional logic, predicate logic, modal logic, temporal logic, fuzzy reasoning, inductive reasoning ( inductive reasoning), hypothetical reasoning, and the like.

For example, the inference engine is provided with rules and data to be analyzed. Here, the analysis target data is data required for the determination process of the selected rule, and specifically includes, for example, chemical substance data and similarity data. The inference engine derives results through rule-based inference.

The chemical substance search program disclosed in the present application can be created using various known program languages according to the configuration of the computer system to be used and the type and version of the operating system.

The chemical substance search program disclosed in this case may be recorded on a recording medium such as an internal hard disk or an external hard disk, or may be recorded on a recording medium such as a CD-ROM, DVD-ROM, MO disk, or USB memory. You can leave it.
Furthermore, when recording the chemical substance search program disclosed in this case on the above recording medium, use it directly or by installing it on a hard disk through the recording medium reading device of the computer system as necessary. can be done. Also, the chemical substance search program disclosed herein may be recorded in an external storage area (such as another computer) accessible from the computer system through an information communication network. In this case, the chemical substance search program disclosed in this case recorded in the external storage area can be used directly from the external storage area through the information communication network or by installing it on the hard disk if necessary.
In addition, the chemical substance search program disclosed in the present application may be divided for each arbitrary process and recorded in a plurality of recording media.

(recoding media)
The recording medium disclosed in this case records the chemical substance search program disclosed in this case.
The recording medium is computer readable.
The recording medium disclosed in the present application is not particularly limited, and can be appropriately selected according to the purpose. Examples include internal hard disks, external hard disks, CD-ROMs, DVD-ROMs, MO disks, and USB memories. be done.
Also, the recording medium disclosed herein may be a plurality of recording media in which the chemical substance search program disclosed herein is divided for each arbitrary process and recorded.
The recording medium may be transient or non-transitory.

Hereinafter, an example of the technology disclosed in the present application will be described in more detail using configuration examples of apparatuses, flowcharts, and the like.
FIG. 11 shows a hardware configuration example of the chemical substance search device disclosed in the present application.
In the chemical substance search device 10, for example, the control unit 11, the memory 12, the storage unit 13, the display unit 14, the input unit 15, the output unit 16, and the I/O interface unit 17 are connected via the system bus 18. there is

The control unit 11 performs operations (four arithmetic operations, comparison operations, annealing method operations, etc.), hardware and software operation control, and the like.
The controller 11 is not particularly limited and can be appropriately selected according to the purpose, and may be, for example, a CPU (Central Processing Unit).
The search unit in the chemical substance search device disclosed in this application can be realized by the control unit 11, for example.

The memory 12 is a memory such as RAM (Random Access Memory) and ROM (Read Only Memory). The RAM stores an OS (Operating System) and application programs read from the ROM and the storage unit 13, and functions as a main memory and a work area for the control unit 11. FIG.

The storage unit 13 is a device that stores various programs and data, such as a hard disk. The storage unit 13 stores programs executed by the control unit 11, data necessary for executing the programs, an OS, and the like. For example, a chemical substance database is stored in the storage unit 13 .
Further, the chemical substance search program disclosed herein is stored in the storage unit 13, loaded into the RAM (main memory) of the memory 12, and executed by the control unit 11, for example.

The display unit 14 is a display device such as a CRT monitor or liquid crystal panel.
The input unit 15 is an input device for various data, and includes, for example, a keyboard and a pointing device (such as a mouse).
The output unit 16 is an output device for various data, such as a printer.
The I/O interface unit 17 is an interface for connecting various external devices. The I/O interface unit 17 is, for example, a CD-ROM (Compact Disc Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), an MO disk (Magneto-Optical disk), a USB memory [USB (Universal Serial Bus ) flash drive].

FIG. 12 shows another hardware configuration example of the chemical substance searching device disclosed in the present application.
The chemical substance search device shown in FIG. 12 is an example of a cloud type chemical substance search device, and the control unit 11 is independent of the storage unit 13 and the like. In the example shown in FIG. 12, a computer 30 storing the storage unit 13 and the like and a computer 40 storing the control unit 11 are connected via the network interface units 19 and 20 .
The network interface units 19 and 20 are hardware that performs communication using the Internet.

FIG. 13 shows another hardware configuration example of the chemical substance search device disclosed in the present application.
The chemical substance search device shown in FIG. 13 is an example of a cloud type chemical substance search device, and the storage unit 13 is independent of the control unit 11 and the like. In the example shown in FIG. 13, a computer 30 storing the control unit 11 and the like and a computer 40 storing the storage unit 13 are connected via the network interface units 19 and 20 .

FIG. 14 shows a functional configuration example as one embodiment of the chemical substance search device disclosed in the present application.
A chemical substance searching device 1A shown in FIG.

FIG. 15 shows a functional configuration example as another embodiment of the chemical substance searching device disclosed in the present application.
A chemical substance searching device 1B shown in FIG. 15 has a searching unit 2 and a chemical substance database 3.

FIG. 16 shows a flow chart of an example of a method for creating a chemical substance database.
First, chemical substance data is extracted from another database (S1). The chemical substance data is data including the properties (chemical properties, physical properties) of each of a plurality of chemical substances. The extracted chemical substance data may be used in the file format of the other database as it is, or may be processed into another file format.
Next, based on the chemical substance data, two types of similarity data, namely first similarity data and second similarity data are created (S2). Similarity data is data that includes the similarity of two chemicals of a plurality of chemicals. Affinity is generated based on one or more characteristics. The first similarity data and the second similarity data are generated based on different characteristics.
As described above, a chemical substance database having chemical substance data and two types of similarity data is created.

FIG. 17 shows a flow chart of another example of a method of creating a chemical substance database. An example of this is where chemical data is processed to create derived properties.
First, chemical substance data is extracted from another database (S11). The chemical substance data is data including the properties (chemical properties, physical properties) of each of a plurality of chemical substances. The extracted chemical substance data may be used in the file format of the other database as it is, or may be processed into another file format.
Next, first similarity data is created based on the chemical substance data (S12). The first similarity data is data containing similarities between two of the plurality of chemical substances. Affinity is generated based on one or more characteristics.
Next, the chemical substance data is processed to create derived characteristics obtained by translating at least one component of the chemical substance characteristics into a superordinate concept (S13).
Next, the derived characteristics are used to create second similarity data (S14). The second similarity data is data containing similarities between two of the plurality of chemical substances. Affinity is generated based on derived characteristics.
As described above, a chemical substance database having chemical substance data and two types of similarity data is created.

FIG. 18 shows a flow chart of an example of the disclosed chemical substance search method.
The flowchart of FIG. 18 is a flow chart of a chemical substance search method for searching for new chemical substances that are applicable to a certain application but are not stored in the chemical substance database.
First, based on the similarity data in the chemical substance database (DB), a rule for searching for the new chemical substance is created (S21). Rules are set up so that searches can be made for chemicals that are suitable for use in relation to at least one of the properties, but are not stored in the chemical database.
Next, the rule is applied to a chemical database (DB) to search for chemicals that are suitable for the application associated with at least one of the properties, but are not stored in the chemical database ( S22). The search is performed, for example, by inference by an inference engine.
Then, it is determined whether or not the corresponding chemical substance is found on the chemical substance database (S23). If a corresponding chemical substance is found on the chemical substance database, a rule different from the already created rule is created (S21). On the other hand, if the corresponding chemical substance is not found on the chemical substance database, the chemical substance becomes a new chemical substance or a candidate for a use related to a certain characteristic. Note that one or more chemical substances may be discovered. When multiple chemicals are discovered, for example, the product of similarity scores contained in the similarity data may be used to rank the chemicals.

FIG. 19 shows a flowchart of another example of the disclosed chemical substance searching method.
The flow chart of FIG. 19 is a flow chart of a chemical substance search method for searching for properties that are properties of a chemical substance that are stored in the chemical substance database but that are not in the chemical substance data of the chemical substance.
First, based on the similarity data in the chemical substance database (DB), a rule for searching for the new chemical substance is created (S31). Rules are set up to allow searching for properties that are properties of chemicals that are stored in the chemical database, but that are not in the chemical data for that chemical.
Next, the rule is applied to the chemical substance database to search for properties of the chemical substance stored in the chemical substance database but not in the chemical substance data of the chemical substance (S32). The search is performed, for example, by inference by an inference engine.
Then, it is determined whether or not a chemical substance having the corresponding property is found on the chemical substance database (S33). If the corresponding chemical substance is not found on the chemical substance database, a rule different from the already created rule is created (S31). On the other hand, if a corresponding chemical substance is found on the chemical substance database, it is found that the chemical substance has a use related to a certain property. Note that one or more chemical substances may be discovered. When multiple chemicals are discovered, for example, the product of similarity scores contained in the similarity data may be used to rank the chemicals.

An example of a flow chart of a chemical substance search method when creating a chemical substance database in the disclosed chemical substance search method is, for example, a flow chart such as that shown in FIG. 20, which is a combination of the flow charts shown in FIGS.

An example of a method for searching for a new battery material with a high discharge potential using the disclosed method for searching for chemical substances will be described.
FIG. 21 is a diagram visualizing an example of searching for a new battery material with a high discharge potential using the disclosed chemical substance search method.
In FIG. 21, solid-line ellipses correspond to chemical substance data 121, 123, 124, 125, and 126 stored in the chemical substance database. Each chemical substance data includes characteristic data such as composition formula, crystal structure, discharge potential, etc. for a certain chemical substance. However, some chemical substance data 123 do not have discharge potential data. The discharge potential is one of the characteristics of battery materials, and the higher the discharge potential, the more useful the battery material.
The chemicals in chemical data 121, 124, 125, and 126 are known as battery materials. Chemicals in chemical data 123 are not known battery materials.
The chemical substance of the chemical substance data 121 and the chemical substance of the chemical substance data 124 have the same crystal structure A.
The chemical substance of the chemical substance data 123 and the chemical substance of the chemical substance data 126 have the same crystal structure C.
A chemical substance in the chemical substance data 125 has a crystal structure B.
Horizontal solid lines correspond to the first similarity data 101, 104, 105, 106, 107, and 108 stored in the chemical substance database.
Solid vertical lines correspond to second similarity data 111, 112, 113, 115, 116, and 117 stored in the chemical substance database.
The first similarity data and the second similarity data each include similarities of chemicals in two chemical data connected by a line.
The similarity included in the first similarity data is the similarity determined to be the same between the two chemical substances when the combination of elements is the same between the compositional formulas of the two chemical substances. . Similarity here includes the case where the compositional formula is exactly the same. (Ni) and (Co) in the ellipses mean that the chemical substance contains Ni and Co as constituent elements, respectively. The similarity in the first similarity data is scored as a similarity score, for example.
Similarity included in the second similarity data is defined as two chemicals being identical if their crystal structures are the same and the derived properties are the same between them. It is the similarity that is judged to exist. The derived characteristic here is a composition formula in which an element (for example, Ni, Co, etc.) included in the transition metal element is replaced with a transition metal element in the composition formula. The similarity in the second similarity data is scored as a similarity score, for example.
Further, a portion of the first similarity data and a portion of the second similarity data include numerical difference data indicating a numerical difference in discharge potential of the two chemicals. That is, the first similarity data 106 includes numerical difference data that the discharge potential of the chemical substance of the chemical substance data 125 is higher than the discharge potential of the chemical substance of the chemical substance data 124 . The second similarity data 113 includes numerical difference data that the discharge potential of the chemical substance of the chemical substance data 121 is higher than the discharge potential of the chemical substance of the chemical substance data 124 .

At least one of the following (1-1) and (2-1) searches are performed using the first similarity data and the second similarity data of the chemical substance database as described above.
(1-1) Search for chemical substances that are suitable as battery materials but are not stored in the chemical substance database (2-1) Properties of chemical substances stored in the chemical substance database properties that indicate utility) but are not in the substance data for the chemical

Regarding (1-1) above, the existence of similarity indicated by dashed line 102 is predicted from first similarity data 106 and second similarity data 113 . The similarity of dashed line 102 is the similarity in the first similarity data. From the first affinity data 107 and the second affinity data 115 it is expected that there is an affinity indicated by dashed line 103 . The similarity of dashed line 103 is the similarity in the first similarity data. Further, from the second similarity data 113 and 115, the existence of similarities indicated by dashed line 114 is expected. The similarity of dashed line 113 is the similarity in the second similarity data. The dashed arrow attached to the similarity 102 indicates that the chemical substance of the chemical substance data 125 [ This corresponds to the numerical difference data that the discharge potential of the crystal structure B(Co)] is higher, and is expected to indicate a numerical difference in the discharge potential. Further, the dashed arrow attached to the similarity 114 indicates that the discharge potential of the chemical substance [crystal structure A (Ni)] of the chemical substance data 121 is higher than the discharge potential of the chemical substance [crystal structure A (Co)] of the chemical substance data 124. Corresponding to the numerical difference data that the discharge potential is higher, it is an expectation that indicates the numerical difference of the discharge potential. Therefore, at the intersection of the expected similarities 102, 103, and 114, the chemistry of crystal structure A (Ni) and the chemistry of crystal structure B (Co) are similar to and higher than the discharge potentials of those chemistries. Chemicals of crystal structure B(Ni)122 can be found as chemicals expected to have high discharge potentials.

Regarding (1-2) above, from the first similarity data 106, the first similarity data 107, the second similarity data 113, the second similarity data 115, the similarity 102, and the similarity 103, The chemical substance [crystal structure C(Ni)] of the chemical substance data 123 is expected to be useful as a battery material.
Specifically, it is as follows. From the second similarity data 115, the chemical substance [crystal structure C(Ni)] of the chemical substance data 123 is the chemical substance [crystal structure C(Co)] of the chemical substance data 126, and the crystal structure and the above-mentioned derived property are judged to be similar. Then, from the first similarity data 106 and 107, the chemical substance [crystal structure C (Co)] of the chemical substance data 126 and the chemical substance [crystal structure A (Co)] of the chemical substance data 124 have the composition formula are judged to be similar. Further, from the second similarity data 113, the chemical substance [crystal structure A (Co)] of the chemical substance data 124 and the chemical substance [crystal structure A (Ni)] of the chemical substance data 121 have the crystal structure and derived characteristics are similar. Further, from the similarities 102 and 103, the chemical substance [crystal structure A (Ni)] of the chemical substance data 121 and the chemical substance [crystal structure C (Ni)] of the chemical substance data 123 have similar composition formulas. is determined to be The second similarity data 113 indicates that the discharge potential of the chemical substance [crystal structure A (Ni)] of the chemical substance data 121 is higher than the discharge potential of the chemical substance [crystal structure A (Co)] of the chemical substance data 124 . We have numerical difference data that the From these, the discharge potential of the chemical substance [crystal structure C(Ni)] of the chemical substance data 123 is higher than the discharge potential of the chemical substance [crystal structure C(Co)] of the chemical substance data 126, which is the battery material. is expected. That is, the chemical substance [crystal structure C(Ni)] of the chemical substance data 123 is expected to be useful as a battery material.

In addition, in the disclosed chemical substance search method, the search results may be visualized as shown in FIG. The visualization method is not particularly limited and can be appropriately selected depending on the purpose.

Further specific examples are shown below.
In the following specific example, Li compound data stored in an inorganic crystal structure database (ICSD) is used to create a chemical substance database in the disclosed chemical substance search method, and using the created chemical substance database, the following searches are performed. (1-A) and a method of searching (2-A).
・Search (1-A): Search for Li compounds that are not stored in the chemical substance database and are suitable for battery material applications related to discharge characteristics ・Search (2-A): Store in the chemical substance database Search for properties that are not in the chemical substance data of the Li compound

FIG. 22 shows a flow chart of the above specific example.
In the above specific example, first, Li compound data (Li compound data) is extracted from the inorganic crystal structure database (ICSD) (S41). The number of Li compounds corresponding to the extracted Li compound data is about 8,000. Each of the extracted Li compound data records characteristics such as compositional formula, crystal structure, and discharge potential. Data on the discharge potential is not recorded for all Li compounds.
Next, first similarity data is created using the extracted Li compound data (S42). The first similarity data is data containing the similarity of two Li compounds among the plurality of Li compounds. The similarity here is the similarity in which two Li compounds are determined to be the same when the combination of elements is the same between the compositional formulas of the two Li compounds. The similarity in the first similarity data is scored as a similarity score using a known method. Also, at least part of the first similarity data for two Li compounds whose Li compound data has discharge potential data includes discharge potential numerical difference data.
Next, the Li compound data is processed to generate derived characteristics obtained by translating at least one component of the characteristics of the chemical substance into a higher concept (S43). The derived characteristic here is a composition formula in which an element (for example, Ni, Co, etc.) included in the transition metal element is replaced with a transition metal element in the composition formula.
Next, second similarity data is created (S44). The second similarity data is data containing the similarity of two Li compounds among the plurality of Li compounds. Similarity here means that two Li compounds are judged to be identical if they have the same crystal structure and the aforementioned derived properties are the same between the two Li compounds. is similarity. The similarity in the second similarity data is scored as a similarity score using a known method. Also, at least part of the second similarity data for two Li compounds whose Li compound data has discharge potential data includes discharge potential numerical difference data.
As described above, a chemical substance database (DB) is created.

Next, based on the first similarity data and the second similarity data in the chemical substance database (DB), rules for searching for new Li compounds are created (S45). The rules are set so that at least one of the aforementioned search (1-A) and search (2-A) can be performed.
Next, the created rule is applied to the chemical substance database (DB), and the aforementioned searches (1-A) and (2-A) are performed (S46).
Then, as a result of searching (1-A), it is determined whether or not the searched Li compound is recorded in the chemical substance database (S47). If the corresponding Li compound is found on the chemical substance database, a rule different from the already created rule is created (S45). On the other hand, if the searched Li compound is not recorded in the chemical substance database as a result of the search (1-A), a new Li compound was found, so the search (1-A) ends.
Further, if the corresponding Li compound is not found on the chemical substance database as a result of the search (2-A), a rule different from the already created rule is created (S45). If the corresponding Li compound is found on the chemical substance database as a result of the search (2-A), the search is terminated.
When multiple Li compounds are discovered, for example, the product of similarity scores included in the similarity data may be used to rank those Li compounds.

An example of creating a chemical substance database according to the flowchart of FIG. 22 and performing the searches (1-A) and (2-A) will be described with reference to FIG.
Part of the chemical substance database created according to the flowchart of FIG. 22 is visualized as shown in FIG.
In FIG. 23, chemical substance data 251, 252, 253, 255, 256, 257, 258, and 259 exist as chemical substance data. Also, first similarity data 203, 204, 206, 207, 209, 210, 211, and 212 exist as first similarity data. Second similarity data 222, 223, 225, 226, 229, 230, 231, 232, 233, 234, and 235 exist as second similarity data. In FIG. 23, the number following # represents the number of the inorganic compound in the ICSD database, and "(14)" represents the space group number.
Chemical substance data 251 does not have discharge potential data.
The first similarity data 206 includes numerical difference data (arrows in FIG. 23) that the discharge potential of the Li compound of the chemical substance data 255 is higher than the discharge potential of the Li compound of the chemical substance data 256 .
The first similarity data 210 includes numerical difference data (arrows in FIG. 23) that the discharge potential of the Li compound of chemical substance data 257 is higher than the discharge potential of the Li compound of chemical substance data 258 .

The searches (1-A) and (2-B) described above are performed on the chemical substance database having such chemical substance data, first similarity data, and second similarity data.
Then, for example, it becomes as shown in FIG. In FIG. 24, similarities 201, 202, 205, and 208 corresponding to similarities in the first similarity data are generated with respect to FIG. Further, similarities 221, 224, 227, and 228 corresponding to similarities in the second similarity data are generated.
A new Li compound 254 (Li ₂ CoP ₂ O ₇ ) is then found, which was not stored in the chemical substance database, due to those similarities generated.
23, the discharge potential of the Li compound 254 is expected to be higher than the discharge potential of the Li compound of the chemical substance data 255 and the discharge potential of the Li compound of the chemical substance data 257.
That is, the Li compound 254 (Li ₂ CoP ₂ O ₇ ) useful as a battery material could be found.
Furthermore, the generated similarity predicts that the Li compounds of chemical substance data 251 may be useful as battery materials.

The following additional remarks are disclosed regarding the above embodiments.
(Appendix 1)
A chemical substance database including chemical substance data including properties of each of a plurality of chemical substances and similarity data including similarities between two of the plurality of chemical substances, wherein the similar there are two or more types of similarity data, two or more types of similarity data, each of the two or more types of similarities is generated based on one or more types of characteristics, and each of the similarity data is a chemical suitable for use in relation to at least one of said properties, but in said chemical database A search that searches for unstored chemicals and/or searches for properties of chemicals that are stored in the chemical database but that are not in the chemical data for the chemicals. A chemical substance searching device characterized by having a part.
(Appendix 2)
The chemical substance search device according to appendix 1, wherein at least one type of the two or more types of similarity data includes two or more types of similarity.
(Appendix 3)
3. The chemical substance search device according to appendix 1 or 2, wherein at least one type of the two or more types of similarity data includes numerical difference data of at least one characteristic between the two chemical substances.
(Appendix 4)
the chemical substance data includes, as the property, a derived property generated by deriving from the property of the chemical substance;
4. The chemical substance search device according to any one of Appendices 1 to 3, wherein at least one type of similarity data among the two or more types of similarity data includes similarity generated based on the derived characteristic.
(Appendix 5)
5. The chemical substance search device according to appendix 4, wherein the derived property is a derived property obtained by translating at least one component of the property of the chemical substance into a higher concept.
(Appendix 6)
6. The chemical substance search device according to any one of Appendices 1 to 5, having the chemical substance database.
(Appendix 7)
the computer
A chemical substance database including chemical substance data including properties of each of a plurality of chemical substances and similarity data including similarities between two of the plurality of chemical substances, wherein the similar there are two or more types of similarity data, two or more types of similarity data, each of the two or more types of similarities is generated based on one or more types of characteristics, and each of the similarity data is a chemical suitable for use in relation to at least one of said properties, but in said chemical database searching for unstored chemicals and/or searching for properties of chemicals stored in the chemical database but not in the chemical data for the chemical;
A chemical substance search method characterized by:
(Appendix 8)
The chemical substance search method according to appendix 7, wherein at least one type of the two or more types of similarity data includes two or more types of similarity.
(Appendix 9)
9. The chemical substance search method according to appendix 7 or 8, wherein at least one type of the two or more types of similarity data includes numerical difference data of at least one property between the two chemical substances.
(Appendix 10)
the chemical substance data includes, as the property, a derived property generated by deriving from the property of the chemical substance;
10. The chemical substance search method according to any one of Appendices 7 to 9, wherein at least one type of similarity data among the two or more types of similarity data includes similarity generated based on the derived characteristic.
(Appendix 11)
11. The chemical substance search method according to appendix 10, wherein the derived property is a derived property obtained by translating at least one component of the property of the chemical substance into a higher concept.
(Appendix 12)
to the computer,
A chemical substance database including chemical substance data including properties of each of a plurality of chemical substances and similarity data including similarities between two of the plurality of chemical substances, wherein the similar there are two or more types of similarity data, two or more types of similarity data, each of the two or more types of similarities is generated based on one or more types of characteristics, and each of the similarity data is a chemical suitable for use in relation to at least one of said properties, but in said chemical database causing a search to occur for chemicals not stored and/or properties for chemicals stored in the chemical database but not in the chemical data for the chemical; ,
A chemical substance search program characterized by:
(Appendix 13)
13. The chemical substance search program according to appendix 12, wherein at least one type of the two or more types of similarity data includes two or more types of similarity.
(Appendix 14)
14. The chemical substance search program according to appendix 12 or 13, wherein at least one type of the two or more types of similarity data includes numerical difference data of at least one property between two of the chemical substances.
(Appendix 15)
the chemical substance data includes, as the property, a derived property generated by deriving from the property of the chemical substance;
15. The chemical substance search program according to any one of Appendices 12 to 14, wherein at least one type of similarity data among the two or more types of similarity data includes similarity generated based on the derived characteristic.
(Appendix 16)
16. The chemical substance search program according to appendix 15, wherein the derived property is a derived property obtained by hyper-conceptualizing at least one component of the property of the chemical substance.

1A chemical substance search device 1B chemical substance search device 2 search unit 3 chemical substance database 10 chemical substance search device 11 control unit 12 memory 13 storage unit 14 display unit 15 input unit 16 output unit 17 I/O interface unit 18 system bus 19 network Interface unit 20 Network interface unit 30 Computer 40 Computer

Claims (8)

the similarity data of a chemical substance database including chemical substance data indicating properties of each of a plurality of chemical substances and similarity data indicating similarity between two of the plurality of chemical substances; using a first search, which is a search for chemicals used in applications associated with at least one of said properties, but not stored in said chemical database; a search unit that performs at least one search of a second search that is a search for a property of the chemical substance that is present but not included in the chemical substance data of the chemical substance ;
In the chemical substance database, there are two or more types of similarities, two or more types of similarity data, and each similarity data of the two or more types of similarity data is based on one or more types of chemical substance data wherein each of the similarity data indicates one or more of the two or more similarities,
wherein the chemical substance data indicates, as the property, a derived property obtained by translating at least one of the constituent elements of the property of the chemical substance into a higher concept;
at least one type of the similarity data among the two or more types of the similarity data includes similarity data generated based on the derived characteristic;
generating said similarity data for said property by searching for the largest independent set based on the molecular structures of two said chemical entities;
In the first search, searching for a chemical substance selected from the derived properties and having a high similarity based on the similarity data;
A chemical substance search device , wherein in the second search, the chemical substance is searched so that the similarity based on the similarity data is high . 2. The chemical substance search device according to claim 1, wherein at least one type of the two or more types of similarity data includes two or more types of similarity. 3. The chemical substance search device according to claim 1, wherein at least one of the two or more types of similarity data includes numerical difference data of at least one property between the two chemical substances. 4. The chemical substance search device according to any one of claims 1 to 3 , comprising the chemical substance database. 2. From claim 1, wherein the search unit generates the similarity data related to the properties by searching for a maximum independent set based on the molecular structures of the two chemical substances using the following formula (1) 5. The chemical substance search device according to any one of 4.

However, in the above formula (1),
The H is a Hamiltonian that means that minimizing the H means searching for the maximum independent set,
The n is the number of nodes in the conflict graph of the two chemical substances graphed, the first chemical substance and the second chemical substance,
In the conflict graph, a combination of each node atom that constitutes the graphed first chemical substance and each node atom that constitutes the graphed second chemical substance is set as the node, and a plurality of the nodes It is understood as a graph created based on the rule that an edge is created between the nodes that are not identical by comparing each other, and an edge is not created between the nodes that are identical by comparing the plurality of nodes. and
b. _i is a number representing the bias for the i-th node;
w _ij teeth,
a positive number other than 0 when an edge exists between the i-th node and the j-th node;
0 when no edge exists between the i-th node and the j-th node;
Said x _i is a binary variable representing whether the i-th node is 0 or 1,
Said x _j is a binary variable representing whether the j-th node is 0 or 1,
The α and the β are positive numbers. 6. The chemical substance searching device according to claim 5, wherein said similarity data relating to said characteristic is generated based on the searched maximum independent set using the following formula (2).

However, in the above formula (2),
said GA _represents said graphed first chemical;
said G _B represents said graphed second chemical;
The S (G _A , G _B ) represents the degree of similarity between the graphed first chemical substance and the graphed second chemical substance, and is represented by 0 to 1. The closer to 1, the more similar. means high degree,
said VA _represents the total number of said node atoms in said graphed first chemical;
The V _c ^A represents the number of the node atoms included in the maximum independent set of the conflict graph among the node atoms in the graphed first chemical substance,
said V _B represents the total number of said node atoms in said graphed second chemical;
The V _c ^B represents the number of the node atoms included in the maximum independent set of the conflict graph among the node atoms in the graphed second chemical substance,
The δ is a number of 0-1. the computer
the similarity data of a chemical substance database showing chemical substance data indicating characteristics of each of a plurality of chemical substances and similarity data including similarity between two of the plurality of chemical substances; using a first search, which is a search for chemicals used in applications associated with at least one of said properties, but not stored in said chemical database; performing at least one of a second search that is a search for a property of the chemical substance that is present but not in the chemical substance data of the chemical substance;
In the chemical substance database, there are two or more types of similarities, two or more types of similarity data, and each similarity data of the two or more types of similarity data is generated based on one or more types of the characteristics wherein each of the similarity data indicates one or more of the two or more similarities;
wherein the chemical substance data includes, as the property, a derived property obtained by translating at least one of the constituent elements of the property of the chemical substance into a higher concept;
at least one type of the similarity data among the two or more types of the similarity data includes similarity data generated based on the derived characteristic;
generating said similarity data for said property by searching for the largest independent set based on the molecular structures of said two said chemical entities;
In the first search, searching for a chemical substance selected from the derived properties and having a high similarity based on the similarity data;
A chemical substance search method , wherein in the second search, the chemical substance is searched so that the similarity based on the similarity data is high . to the computer,
the similarity data of a chemical substance database including chemical substance data indicating properties of each of a plurality of chemical substances and similarity data indicating similarity between two of the plurality of chemical substances; using a first search, which is a search for chemicals used in applications associated with at least one of said properties, but not stored in said chemical database; perform at least one search of a second search that is a search for a property of the chemical substance that is present but not in the chemical substance data of the chemical substance ;
In the chemical substance database, there are two or more types of similarities, two or more types of similarity data, and each similarity data of the two or more types of similarity data is generated based on one or more types of the characteristics wherein each of the similarity data indicates one or more of the two or more similarities;
wherein the chemical substance data includes, as the property, a derived property obtained by translating at least one of the constituent elements of the property of the chemical substance into a higher concept;
at least one type of the similarity data among the two or more types of the similarity data includes similarity data generated based on the derived characteristic;
generating said similarity data for said property by searching for the largest independent set based on the molecular structures of said two said chemical entities;
In the first search, searching for a chemical substance selected from the derived characteristics and having a high similarity based on the similarity data;
A chemical substance search program characterized in that, in the second search, the chemical substance is searched so as to increase the similarity based on the similarity data.

Images