Method, system, medium and electronic device for extracting reaction template of chemical reaction

Abstract

The present invention provides a reaction template extraction method, system, medium and electronic device for a chemical reaction, wherein the reaction template extraction method for a chemical reaction comprises: obtaining reactants and products of a chemical reaction; analyzing the atomic correspondence between the reactants and products, matching to obtain unchanged atomic mappings; performing estimated matching on unmatched atoms to obtain atomic balancing mappings; and extracting the reaction template of the chemical reaction from the atomic balancing mappings. The present invention utilizes computer technology to achieve fast and accurate atomic matching and balancing of chemical reactions, thereby enabling accurate reaction template extraction; it can also utilize the extracted reaction templates to match any compound and calculate its feasible chemical reactions, providing a strong foundation for both forward occurrence design and reverse synthesis design of chemical reactions.

Description

Technical Field

[0001] The invention belongs to the technical field of computer chemistry, and relates to a method for extracting chemical reaction templates, and in particular to a method, system, medium and electronic equipment for extracting reaction templates of chemical reactions. Background Art

[0002] A chemical reaction is a process in which molecules break into atoms and atoms rearrange and recombine to form new substances. The reaction is often accompanied by light emission, heat generation, color change, and precipitation. The basis for judging whether a reaction is a chemical reaction is whether the reaction generates new substances. According to the chemical bond theory, whether it is a chemical reaction can also be judged based on whether old bonds are broken and new bonds are generated during a change process.

[0003] Computational Chemistry is the science of using computers to study chemical reactions and material changes. Using computers as a technical means, it establishes theories and methods for resource-based and intelligent processing of chemical and chemical information. Computer chemistry revolves around the core of "chemical information" and starts with the basic concepts and methods of computer chemistry. It connects a series of seemingly scattered but intrinsically related contents such as experimental data, measurement signals, infrared spectra, molecular structures, chemical knowledge and the Internet. It shows the application of computer chemistry in data processing, signal analysis, structural analysis, organic synthesis, molecular design and many other aspects. Understanding substances, transforming substances, creating new substances, understanding reactions, controlling reaction processes and creating new reactions and new processes are the main subjects of computer chemistry research.

[0004] Chemical reactions in existing literature usually only record the chemical formulas of reactants and products, lacking the process of chemical reaction changes, making it difficult to directly provide intuitive results or template references for the study of chemical reactions. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a reaction template extraction method, system, medium and electronic equipment for chemical reactions, which are used to solve the problem that the prior art usually only records the chemical formulas of reactants and products and cannot provide a direct template reference for the study of chemical reactions.

[0006] To achieve the above-mentioned purpose and other related purposes, the present invention provides a method for extracting a reaction template of a chemical reaction, comprising: obtaining reactants and products of a chemical reaction; analyzing the atomic correspondence between the reactants and products, and matching to obtain unchanged atomic mappings; performing estimated matching on unmatched atoms to obtain atomic balancing mappings; and extracting the reaction template of the chemical reaction from the atomic balancing mappings.

[0007] In one embodiment of the present invention, an implementation process of analyzing the atomic correspondence between the reactants and the products includes: calculating the hierarchical hash of each atom in the reactants to obtain the hierarchical hash of each reactant atom; calculating the hierarchical hash of each atom in the products to obtain the hierarchical hash of each product atom; the calculation method of the hierarchical hash includes: using the hash function value of the atomic type of each atom in the compound as the 0th layer hash of the atom:

[0008]

[0009] Where H represents the hierarchical hash set of compounds, H i represents a hierarchical hash set of the i-th atom in a compound, represents the zero-level hash of the i-th atom, Hash() is an arbitrary hash function, AtomElement(V i ) represents the atomic type of the i-th atom; obtain the 0-th level hash value set of the compound according to the 0-th level hash of each atom in the compound; calculate and obtain the temporary j-th level hash value set L of the compound:

[0010] L={L1,L2,…,L k ,…,L |V|}

[0011]

[0012]

[0013] Among them, L k represents the temporary j-th level hash of atom k, |V| represents the number of atoms in the atom set V of the compound; for each adjacent atom l of each atom k in the atom set of the compound, the bond type BondType(E k,l ) and the j-1th level hash of atom l After adding set M, take the j-1th level hash of set M and atom k The new hash is used as the temporary j-th level hash value L of atom k k ; Obtain a temporary j-th level hash value set L of the compound according to the temporary j-th level hash value of each atom in the compound; If the temporary j-th level hash value set L of the compound is the same as the previous level hash value set of the compound, i.e., the j-1 level hash value set, then the hash value set of the compound has been calculated; otherwise, the temporary j-th level hash value set L of the compound is used as the j-th level hash value set, and the calculation of the next level hash value set, i.e., the j+1 level hash value set, is performed.

[0014] In one embodiment of the present invention, an implementation process of matching to obtain unchanged atomic mapping includes: searching for reactant atoms and product atoms having the same hierarchical hash, and matching reactant atoms and product atoms having the same hierarchical hash to the unchanged atomic mapping.

[0015] In one embodiment of the present invention, the estimated matching of unmatched atoms to obtain an atomic balancing mapping includes: the unmatched atoms include unmatched reactant atoms and unmatched product atoms; for each unmatched reactant atom, all unmatched product atoms are traversed and searched respectively, and the matching scores of each unmatched product atom and each unmatched reactant atom are calculated; the unmatched reactant atoms and unmatched product atoms with the highest matching scores are estimated to be matched as an atomic balancing mapping; and the above steps are repeated until the atomic balancing mapping is completed for all unmatched reactant atoms.

[0016] In one embodiment of the present invention, the estimated matching of unmatched atoms to obtain an implementation process of an atomic balancing mapping also includes: when the atomic balancing mapping of all the unmatched reactant atoms is completed, but the unmatched product atoms still remain, at least one special reactant is added; for each product atom that has not completed the atomic balancing mapping, the atoms of the special reactant are traversed and searched respectively, and the matching scores of the atoms of each special reactant and the product atoms of each uncompleted atomic balancing mapping are calculated; the product atom with the highest matching score and the atom of the special reactant are estimated to be matched as an atomic balancing mapping; and the above steps are repeated until the atomic balancing mapping is completed for all product atoms that have not completed the atomic balancing mapping.

[0017] In one embodiment of the present invention, the estimated matching of unmatched atoms to obtain an implementation process of atomic balancing mapping also includes: when the atomic balancing mapping of all unmatched product atoms is completed, but the unmatched reactant atoms still remain, at least one special product is added; for each reactant atom that has not completed the atomic balancing mapping, the atoms of the special product are traversed and searched respectively, and the matching scores of the atoms of each special product and the reactant atoms of each uncompleted atomic balancing mapping are calculated; the reactant atom with the highest matching score and the atom of the special product are estimated to be matched as an atomic balancing mapping; and the above steps are repeated until the atomic balancing mapping is completed for all reactant atoms that have not completed the atomic balancing mapping.

[0018] In one embodiment of the present invention, an implementation process of extracting the reaction template of the chemical reaction from the atom balancing map includes: deleting atoms without bonding or hybridization changes from the atom balancing map to obtain the minimum information describing the chemical reaction changes, that is, the reaction template of the chemical reaction; the reaction template includes the attribute information and atomic correspondence of each atom in the template; the attribute information of the atom includes the element information, hybridization type and bonding type of the atom.

[0019] The present invention also provides a reaction template extraction system for a chemical reaction, the reaction template extraction system for a chemical reaction comprising: an input module, for acquiring reactants and products of a chemical reaction; an analysis module, connected to the input module, for analyzing the atomic correspondence between the reactants and products; a first matching module, connected to the analysis module, for matching to obtain unchanged atomic mapping; a second matching module, connected to the first matching module, for performing estimated matching on unmatched atoms to obtain an atomic balancing mapping; and a template extraction module, connected to the second matching module, for extracting the reaction template of the chemical reaction from the atomic balancing mapping.

[0020] The present invention also provides an electronic device, comprising: a memory storing a computer program; a processor communicating with the memory and implementing the steps of the reaction template extraction method of the chemical reaction when calling and executing the computer program.

[0021] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the reaction template extraction method for the chemical reaction are implemented.

[0022] As described above, the reaction template extraction method, system, medium and electronic device of the chemical reaction described in the present invention have the following beneficial effects:

[0023] The present invention utilizes computer technology to achieve fast and accurate chemical reaction atom matching and balancing, thereby enabling accurate reaction template extraction; it can also utilize the extracted reaction template to match any compound and calculate its feasible chemical reaction, thus providing a strong foundation for both forward occurrence design and reverse synthesis design of chemical reactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 Shown is a schematic diagram of an implementation flow of a method for extracting a reaction template for a chemical reaction described in an embodiment of the present invention.

[0025] Figure 2 Shown is a matching illustration of the atomic correspondence described in an embodiment of the present invention.

[0026] Figure 3Shown is a schematic diagram of the template extraction process of the Suzuki coupling reaction described in an embodiment of the present invention.

[0027] Figure 4A Shown is an exemplary flowchart of the atom matching search process described in an embodiment of the present invention.

[0028] Figure 4B Shown is a schematic diagram of an exemplary flow chart of the trim search process according to an embodiment of the present invention.

[0029] Figure 5 Shown is a schematic diagram of an implementation structure of a reaction template extraction system for a chemical reaction described in an embodiment of the present invention.

[0030] Figure 6 Shown is a schematic diagram of an implementation structure of an electronic device according to an embodiment of the present invention.

[0031] Component number description

[0032] 500 Chemical Reaction Template Extraction System

[0033] 510 Input Module

[0034] 520 Analysis Module

[0035] 530 First matching module

[0036] 540 Second matching module

[0037] 550 Template Extraction Module

[0038] 600 Electronic equipment

[0039] 610 Memory

[0040] 620 processor

[0041] Steps S101 to S104 DETAILED DESCRIPTION

[0042] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0043] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0044] See also Figure 1 , an embodiment of the present invention provides a method for extracting a reaction template of a chemical reaction, comprising:

[0045] S101, obtaining reactants and products of a chemical reaction;

[0046] S102, analyzing the atomic correspondence between the reactants and the products, and matching to obtain unchanged atomic mapping;

[0047] S103, performing estimated matching on unmatched atoms to obtain an atomic balancing mapping;

[0048] S104, extracting a reaction template of the chemical reaction from the atomic balancing map.

[0049] Among them, the reactants and products of chemical reactions generally come from chemical literature or verified chemical experiments. These literature data and experimental data are data sources for extracting reaction templates of chemical reactions, which can be collectively referred to as chemical reaction data. A major problem in processing chemical reaction data is that chemical reactions in literature usually only record the chemical formulas of reactants and products, but lack atomic correspondence. Although it is difficult to perform appropriate atomic mapping before a few reactions are experimentally elucidated to their mechanism, most reactions should have a unique and correct atomic mapping method. The present invention is designed using this principle to identify the changes of each atom in the chemical reaction before and after the reaction by atomic mapping.

[0050] In chemical literature, only the main reactants and products are usually given, and the reaction formula also needs to be balanced during the mapping process. In addition, it is also necessary to distinguish the roles of different reagents, such as distinguishing pure solvents, solvents or catalysts that provide atoms for the main products, etc., and removing reaction reagents that do not directly provide atoms. After obtaining the atomic mapping relationship, the reaction template can be extracted from the reaction to provide an available reaction for calculating the reverse synthesis of any compound. How to extract the reaction template and how to define a suitable reaction template are the technical problems solved by the present invention.

[0051] In one embodiment of the present invention, an implementation process of analyzing the atomic correspondence between the reactants and the products in step S102 includes:

[0052] Calculate the hierarchical hash of each atom in the reactant to obtain the hierarchical hash of each reactant atom;

[0053] Calculating the hierarchical hash of each atom in the product to obtain the hierarchical hash of each atom in the product;

[0054] The calculation method of the hierarchical hash includes:

[0055] The hash function value of the atomic type of each atom in the compound is used as the 0th level hash of the atom:

[0056]

[0057] Where H represents the hierarchical hash set of compounds, H i represents a hierarchical hash set of the i-th atom in a compound, represents the zero-level hash of the i-th atom, Hash() is an arbitrary hash function, AtomElement(V i ) represents the atomic type of the i-th atom;

[0058] Obtain a set of level 0 hash values ​​of the compound according to the level 0 hash of each atom in the compound;

[0059] Calculate the temporary j-th level hash value set L of the compound:

[0060] L={L1,L2,…,L k ,…,L |V|}

[0061]

[0062]

[0063] Among them, L k represents the temporary j-th level hash of atom k, |V| represents the number of atoms in the atom set V of the compound;

[0064] For each adjacent atom l of each atom k in the atom set of the compound, the bond type BondType(E k,l ) and the j-1th level hash of atom l After adding set M, take the j-1th level hash of set M and atom k The new hash is used as the temporary j-th level hash value L of atom k k ;

[0065] Obtain a temporary j-th level hash value set L of the compound according to the temporary j-th level hash value of each atom in the compound;

[0066] If the temporary j-th layer hash value set L of the compound is the same as the previous layer, i.e., the j-1-th layer hash value set of the compound, then the hash value set of the compound has been calculated; otherwise, the temporary j-th layer hash value set L of the compound is used as the j-th layer hash value set, and the next layer, i.e., the j+1-th layer hash value set, is calculated.

[0067] In one embodiment of the present invention, a specific implementation process of the layered hash calculation method includes:

[0068] First, the hash function value of the atomic type of each atom is used as the zero-level hash of the atom:

[0069]

[0070] Where H represents the hierarchical hash set of compounds, H i represents a hierarchical hash set of the i-th atom in a compound, represents the zero-level hash of the i-th atom, Hash() is an arbitrary hash function, AtomElement(V i ) represents the atomic type of the i-th atom; G(V,E) is the graph representation of the compound, where V is the vertex set (i.e., the atom set) and E is the edge set (i.e., the chemical bond set); the left arrow symbol ← means assigning the value or / and set on the right side of the arrow to the left side;

[0071] After obtaining the 0th level hash value set of the compound, the jth level hash value set of the compound is calculated by the following steps and formula:

[0072] Step 1: Reset temporary variables

[0073]

[0074] Among them, L is the set used to store temporary hash values ​​in this calculation, which is reset to an empty set in each calculation.

[0075] Step 2: For each atom k (k is one of the vertices in the vertex set V), substitute the hash value of the previous layer (the (j-1) layer) of atom k and the atoms bonded to atom k into the following formula to calculate the new hash value:

[0076]

[0077]

[0078] Where M is the set of hash values ​​and bond types of all atoms bonded to atom k (denoted as adjacent atoms), (k,l)∈E means that there is an edge between vertex k and vertex l, that is, there is a chemical bond between atom k and atom l; BondType(Ek,l ) represents the type of chemical bond between atom k and atom l; L k Represents the kth hash value in the temporary hash set.

[0079] The second step is to: for each adjacent atom l of each atom k in the vertex set, set its bond type BondType (E k,l ), and its (j-1)th layer hash value After adding set M, take the (j-1)th level hash value of set M and atom k The new hash value is used as the temporary j-th level hash value L of atom k k The obtained temporary set is recorded as L, L = {L1, L2, ..., L k ,…,L |V|}, where |V| represents the number of vertices in the vertex set V.

[0080] Step 3 is for comparison and assignment. If the temporary hash set L obtained is the same as the hash set of the previous layer (the (j-1) layer), it means that the hash set has been calculated. Otherwise, the hash set L is used as the j-th layer hash set, and the calculation of the next layer (the (j+1) layer) is performed. The reactants and products are all compounds. The present invention calculates hierarchical hashes for each atom in the compound. The i-th layer hash takes into account the environmental information of the i-th layer bonds around each atom. The similarity between the hierarchical hashes of two atoms is defined as the number of the highest layer where the two hashes are the same. This similarity can be used as a search priority. For example, each compound is represented by a hierarchical hash value set, that is, H = {H0, H1, H2, ...}, and the similarity / similarity is defined as the highest layer number with the same hash. For example, the highest equal number of layers in set H and set G is H3 = G3, and for any number of layers x above 3 (x is a positive integer greater than 3), there is H x ! =G x , then the similarity between set H and set G is defined as 3. Figure 2 A matching diagram of the atomic correspondence described in an embodiment of the present invention.

[0081] When analyzing the atomic correspondence between the reactants and products, it is necessary to search for single atoms, and the characteristics of chemical reactions can be used to remove many impossible matching methods or atomic correspondences. For example, atoms can never match atoms of different elements. In addition, atoms are less likely to match atoms with completely different bonding environments than atoms with the same environment. These factors, such as atomic elements and the environment in which atoms are located, can be reflected through hierarchical hash calculations of atoms, providing a basis for subsequent further atomic matching.

[0082] In one embodiment of the present invention, an implementation process of matching in step S102 to obtain unchanged atomic mapping includes: searching for reactant atoms and product atoms having the same hierarchical hash, and matching reactant atoms and product atoms having the same hierarchical hash to the unchanged atomic mapping.

[0083] Furthermore, an isomorphism search is performed based on the graph representation of the reactants and the graph representation of the products to obtain a subgraph isomorphism shared by both the reactants and the products, where the subgraph isomorphism is the unchanged atomic mapping.

[0084] In the process of atomic correspondence matching, if an exhaustive enumeration or search is performed, the search range is too large to be easy to implement. Most chemical reactions only change a small part of the reactants, which means that there is usually a large subgraph isomorphism between the reactants and the products. Therefore, in order to reduce the workload, the present invention first performs a search based on partial isomorphism, which matches as large a group of atoms as possible. This process only explores exactly the same subgraph, so the complexity is significantly reduced.

[0085] In one embodiment of the present invention, an implementation process of performing estimated matching on unmatched atoms in step S103 to obtain an atomic balancing mapping includes: the unmatched atoms include unmatched reactant atoms and unmatched product atoms; for each unmatched reactant atom, traverse and search all unmatched product atoms according to the similarity between the hierarchical hashes of the atoms, and calculate the matching score of each unmatched product atom and each unmatched reactant atom; the unmatched reactant atoms and unmatched product atoms with the highest matching scores are estimated to be matched as an atomic balancing mapping; and the steps of atomic balancing mapping are repeated until the atomic balancing mapping is completed for all unmatched reactant atoms.

[0086] Wherein, the matching score is a comprehensive score of multi-dimensional matching elements; the multi-dimensional matching elements include the matching atomic element type, the matching atomic orbital hybridization type, the matching chemical bond number, the matching chemical group and the matching bond breaking type; the scores of each matching element are preset, including; the matching atomic element type score, the matching atomic orbital hybridization type score, the matching chemical bond number score, the matching chemical group score and the matching bond breaking type score. The multi-dimensional matching element comprehensive score is the comprehensive value of the scores of each matching element. The comprehensive value can be the direct accumulation of the scores of each matching element, or the value obtained by the operation function of each matching element score.

[0087] Hybridization is a theoretical explanation of the process by which atoms form molecules, including sp (such as BeCl2), sp 2 (such as BF3), sp 3 (such as CH4), sp 3d (such as PCl5), sp 3 d 2 (such as SF6) hybridization, etc. Orbital hybridization theory refers to the atomic orbital hybridization theory. The electrons outside the nucleus of an atom are arranged on atomic orbitals of different energy levels, such as S orbitals, P orbitals, etc. When atoms form molecules, in order to enhance the bonding ability (to make the bonding energy the lowest and most stable), different types of atomic orbitals with similar energies in the same atom are recombined to form new atomic orbitals with different energy, shape and direction from the original orbitals (the energy of this orbital is lower than the energy before hybridization). For example, sp hybridization, sp 2 Hybridization, etc. This process of atomic orbital recombination is called atomic orbital hybridization, and the new atomic orbital formed is called a hybrid orbital. After the hybrid orbital is formed, it combines with other atoms to reduce the energy of the entire molecule and reach a stable state.

[0088] For example, referring to Table 1, an example preset table of matching scores of matching items is provided, through which each matching item is scored, and the final matching score is the sum of the scores of these matching items.

[0089] Table 1:

[0090]

[0091]

[0092] by Figure 2 For example, the basic matching items include:

[0093] 1) Atom matching: all atoms except 9, 10, 19, 20;

[0094] 2) Key matching: 11-12, 12-13, 13-14, 14-15, 14-16, 21-22, etc.

[0095] 3) Bond formation and breaking: 17-19, 17-10, 7-9, 7-10, 16-21, 6-21, 20-21.

[0096] For a set of input reactants and products, there are many different atomic matching methods. The present invention scores different matching methods so that the best match has the highest matching score. The matching factors of the scoring statistics include: the type of atomic elements matched, the type of atomic orbital hybridization matched, the number of chemical bonds matched, the type of chemical groups matched and the type of bond breaking. The main principle of the matching scoring is to minimize the number of bond modifications, specifically including: assigning a score to each match (atoms and chemical bonds, etc.), and then adding up the scores of all matches on the reactants and products.

[0097] After step S102 completes the isomorphic matching of some subgraphs, step S103 performs an atomic-level search for the remaining unmatched atoms. This process can achieve corresponding relationship matching for atoms that have undergone chemical changes. During the atomic-level search, the current highest matching score is continuously recorded, and the highest score of the unmatched part is estimated, and the expected maximum score in the future is calculated. If the estimated maximum score is still less than the known highest score, the unmatched part can be pruned, see Figure 4A shown.

[0098] In one embodiment of the present invention, the estimated matching of unmatched atoms in step S103 to obtain an implementation process of an atomic balancing mapping also includes: when the atomic balancing mapping of all unmatched reactant atoms is completed, but the unmatched product atoms still remain, at least one special reactant is added; for each product atom that has not completed the atomic balancing mapping, the atoms of the special reactant are traversed and searched separately, and the matching scores of the atoms of each special reactant and the product atoms of each uncompleted atomic balancing mapping are calculated, and the product atoms with the highest matching scores and the atoms of the special reactant are estimated to be matched as an atomic balancing mapping (i.e., the special reactant atom balancing step); the above-mentioned special reactant atom balancing step is repeated until the atomic balancing mapping is completed for all product atoms that have not completed the atomic balancing mapping. For example: See Figure 4B shown.

[0099] In one embodiment of the present invention, the estimated matching of unmatched atoms in step S103 to obtain an implementation process of an atomic balancing mapping also includes: when the atomic balancing mapping of all unmatched product atoms is completed, but the unmatched reactant atoms still remain, at least one special product is added; for each reactant atom that has not completed the atomic balancing mapping, the atoms of the special product are traversed and searched respectively, and the matching scores of the atoms of each special product and the reactant atoms of each uncompleted atomic balancing mapping are calculated, and the reactant atoms with the highest matching scores and the atoms of the special product are estimated to be matched as an atomic balancing mapping (special generation atomic balancing step); the above-mentioned special reactant atom balancing step is repeated until the atomic balancing mapping is completed for all reactant atoms that have not completed the atomic balancing mapping. For example: See Figure 4B Some reactants do not donate atoms, such as catalysts, solvents, etc. Identify which parts donate atoms and which parts do not.

[0100] The present invention is based on the above basic search pruning and heuristic methods to balance chemical reaction equations. If after the search is completed, it is not possible to find a suitable mapping for each atom of the reactant or product, then try to increase the equation balancing coefficient of a certain reactant or product to obtain the increase method that increases the score the most. Then repeat the process until the mapping is completely successful.

[0101] In one embodiment of the present invention, an implementation process of extracting the reaction template of the chemical reaction from the atom balancing map in step S104 includes: deleting atoms without bonding or hybridization changes from the atom balancing map to obtain the minimum information describing the chemical reaction changes, that is, the reaction template of the chemical reaction; the reaction template includes the attribute information and atomic correspondence of each atom in the template; the attribute information of the atom includes the element information, hybridization type and bonding type of the atom.

[0102] After the atomic correspondence of the reaction is determined in step S103, the reaction template can be extracted from these atomic balance maps. Figure 3 The extraction process of Suzuki coupling reaction is shown. Atoms without bonding or hybridization changes are removed from the mapped reaction and the final reaction map is obtained. The map contains the minimum information to describe the change of this chemical reaction. Each atom in the reaction template is labeled with its atom type, which contains information about the element of the atom, the hybridization type and the bonding type. The atom correspondence is also saved in the reaction template. The reaction template can be regarded as a compact representation of the corresponding chemical reaction. Using the reaction template, potential reaction sites on any compound can be effectively found and applied to these sites so that the precursors of the retrosynthesis of the chemical reaction step can be found using the reaction represented by the reaction template.

[0103] The present application can achieve fast and accurate chemical reaction atom matching and balancing, thereby enabling accurate reaction template extraction; and using the extracted reaction template to match any compound, its feasible chemical reaction can be calculated.

[0104] The protection scope of the reaction template extraction method for the chemical reaction described in the present invention is not limited to the execution order of the steps listed in this embodiment. All solutions implemented by adding, reducing or replacing steps in the prior art based on the principles of the present invention are included in the protection scope of the present invention.

[0105] The present invention also provides a reaction template extraction system for a chemical reaction, and the reaction template extraction system for a chemical reaction can implement the reaction template extraction method for a chemical reaction described in the present invention. However, the implementation device of the reaction template extraction method for a chemical reaction described in the present invention includes but is not limited to the structure of the reaction template extraction system for a chemical reaction listed in this embodiment. All structural deformations and replacements of the prior art made according to the principles of the present invention are included in the protection scope of the present invention.

[0106] See also Figure 5As shown, an embodiment of the present invention further provides a reaction template extraction system for a chemical reaction. The reaction template extraction system 500 for a chemical reaction includes: an input module 510 , an analysis module 520 , a first matching module 530 , a second matching module 540 , and a template extraction module 550 .

[0107] The input module 510 obtains the reactants and products of a chemical reaction. Among them, the reactants and products of the chemical reaction are generally derived from chemical literature or verified chemical experiments. These literature data and experimental data are data sources for extracting reaction templates of chemical reactions, which can be collectively referred to as chemical reaction data. A major problem in processing chemical reaction data is that chemical reactions in the literature usually only record the chemical formulas of reactants and products, but lack atomic correspondence. Although it is difficult to perform appropriate atomic mapping before the mechanism of a few reactions is experimentally clarified, most reactions should have a unique and correct atomic mapping method. The present invention is designed using this principle to identify the changes of each atom in the chemical reaction before and after the reaction by atomic mapping.

[0108] The analysis module 520 is connected to the input module 510 to analyze the atomic correspondence between the reactants and the products. The analysis module 520 calculates the hierarchical hash of each atom in the reactants to obtain the hierarchical hash of each reactant atom; the analysis module 520 calculates the hierarchical hash of each atom in the products to obtain the hierarchical hash of each product atom; the calculation method of the hierarchical hash includes: the analysis module 520 uses the hash function value of the atomic type of each atom in the compound as the 0th layer hash of the atom: Where H represents the hierarchical hash set of compounds, H i represents a hierarchical hash set of the i-th atom in a compound, represents the zero-level hash of the i-th atom, Hash() is an arbitrary hash function, AtomElement(V i ) represents the atomic type of the i-th atom; obtain the 0-th level hash value set of the compound according to the 0-th level hash of each atom in the compound; calculate and obtain the temporary j-th level hash value set L of the compound:

[0109] L={L1,L2,…,L k ,…,L |V|}

[0110]

[0111]

[0112] Among them, L krepresents the temporary j-th level hash of atom k, |V| represents the number of atoms in the atom set V of the compound;

[0113] The analysis module 520 converts the bond type BondType (E k,l ) and the j-1th level hash of atom l After adding set M, take the j-1th level hash of set M and atom k The new hash is used as the temporary j-th level hash value L of atom k k .

[0114] The analysis module 520 obtains a temporary j-th level hash value set L of the compound according to the temporary j-th level hash value of each atom in the compound.

[0115] If the analysis module 520 determines that the temporary j-th layer hash value set L of the compound is the same as the previous layer, i.e., the j-1-th layer hash value set of the compound, the analysis module 520 has completed the calculation of the hash value set of the compound; otherwise, the analysis module 520 uses the temporary j-th layer hash value set L of the compound as the j-th layer hash value set and calculates the next layer, i.e., the j+1-th layer hash value set.

[0116] The first matching module 530 is connected to the analysis module 520 to match and obtain unchanged atomic mappings. The first matching module 530 searches for reactant atoms and product atoms with the same hierarchical hash and matches the reactant atoms and product atoms with the same hierarchical hash as the unchanged atomic mappings.

[0117] The second matching module 540 is connected to the first matching module 530, and performs estimated matching on unmatched atoms to obtain an atomic balancing mapping.

[0118] Furthermore, the unmatched atoms include unmatched reactant atoms and unmatched product atoms. The second matching module 540 traverses and searches all unmatched product atoms for each unmatched reactant atom, calculates the matching score of each unmatched product atom and each unmatched reactant atom; estimates the matching of the unmatched reactant atom and the unmatched product atom with the highest matching score as an atomic balancing mapping; repeats the atomic balancing mapping steps until the atomic balancing mapping is completed for all unmatched reactant atoms. Among them, the matching score is a comprehensive score of multi-dimensional matching elements; the multi-dimensional matching elements include matching atomic element types, matching atomic orbital hybridization types, matching chemical bond numbers, matching chemical groups and matching bond breaking types; each matching element score is preset, including; matching atomic element type score, matching atomic orbital hybridization type score, matching chemical bond number score, matching chemical group score and matching bond breaking type score. The multi-dimensional matching element comprehensive score is the comprehensive value of each matching element score. The comprehensive value may be a direct accumulation of scores of the matching elements, or a value obtained by using a calculation function to calculate the scores of the matching elements.

[0119] In one embodiment of the present invention, when the atomic balancing mapping of all the unmatched reactant atoms is completed, but the unmatched product atoms still remain, the second matching module 540 adds at least one special reactant; the second matching module 540 traverses and searches for the atoms of the special reactant for each product atom that has not completed the atomic balancing mapping, calculates the matching scores of the atoms of each special reactant and the product atoms of each uncompleted atomic balancing mapping, and estimates that the product atom with the highest matching score and the atom of the special reactant are matched as an atomic balancing mapping (i.e., the special reactant atom balancing step); the second matching module 540 repeats the above-mentioned special reactant atom balancing step until the atomic balancing mapping is completed for all product atoms that have not completed the atomic balancing mapping.

[0120] In one embodiment of the present invention, when the atomic balancing mapping of all the unmatched product atoms is completed, but the unmatched reactant atoms still remain, the second matching module 540 adds at least one special product; the second matching module 540 traverses and searches for the atoms of the special product for each reactant atom that has not completed the atomic balancing mapping, calculates the matching scores of the atoms of each special product and the reactant atoms of each uncompleted atomic balancing mapping, and estimates the matching of the reactant atom with the highest matching score and the atom of the special product as an atomic balancing mapping (special generation atom balancing step); the second matching module 540 repeats the above-mentioned special reactant atom balancing step until the atomic balancing mapping is completed for all reactant atoms that have not completed the atomic balancing mapping.

[0121] The template extraction module 550 is connected to the second matching module 540, and extracts the reaction template of the chemical reaction from the atom balancing map. The template extraction module 550 deletes atoms without bonding or hybridization changes from the atom balancing map to obtain the minimum information describing the chemical reaction changes, that is, the reaction template of the chemical reaction; the reaction template includes the attribute information and atomic correspondence of each atom in the template; the attribute information of the atom includes the element information, hybridization type and bonding type of the atom.

[0122] It should be noted that it should be understood that the division of the various modules of the above system is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or physically separated. Moreover, these modules can be implemented in the form of software calling through processing elements, or in the form of hardware, or some modules can be implemented in the form of software calling through processing elements, and some modules can be implemented in the form of hardware.

[0123] See also Figure 6 As shown, an embodiment of the present invention further provides an electronic device, the electronic device 600 comprising: a memory 610 and a processor 620. The memory 610 stores a computer program. The processor 620 is connected to the memory 610 for communication, and when the computer program is called and executed, the steps of the reaction template extraction method for the chemical reaction of the present invention are implemented.

[0124] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the reaction template extraction method for a chemical reaction are implemented.

[0125] The computer programs described in this application can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network.

[0126] In summary, the present invention effectively overcomes various shortcomings of the prior art and has high industrial utilization value.

[0127] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for extracting a reaction template of a chemical reaction, characterized in that: include: Obtain the reactants and products of a chemical reaction; Analyzing the atomic correspondence between the reactants and the products, and matching to obtain unchanged atomic mapping; Perform estimated matching on unmatched atoms to obtain atomic balancing mapping; An implementation process of performing estimated matching on unmatched atoms to obtain atomic balancing mapping includes: The unmatched atoms include unmatched reactant atoms and unmatched product atoms; For each unmatched reactant atom, traverse and search all unmatched product atoms respectively, and calculate the matching score between each unmatched product atom and each unmatched reactant atom; The unmatched reactant atoms and unmatched product atoms with the highest matching scores are estimated to be matched as an atomic balancing map; Repeat the above steps until the atomic balancing mapping is completed for all unmatched reactant atoms; extracting a reaction template for the chemical reaction from the atomic balance map; One implementation process of analyzing the atomic correspondence between the reactants and products includes: Calculate the hierarchical hash of each atom in the reactant to obtain the hierarchical hash of each reactant atom; Calculating the hierarchical hash of each atom in the product to obtain the hierarchical hash of each atom in the product; The calculation method of the hierarchical hash includes: The hash function value of the atomic type of each atom in the compound is used as the 0th level hash of the atom: Where H represents the hierarchical hash set of compounds, H i represents a hierarchical hash set of the i-th atom in a compound, represents the zero-level hash of the i-th atom, Hash() is an arbitrary hash function, AtomElement(V i ) represents the atomic type of the i-th atom; Obtain a set of level 0 hash values ​​of the compound according to the level 0 hash of each atom in the compound; Calculate the temporary j-th level hash value set L of the compound: L={L1,L2,…,L k ,…,L |V| } Among them, L k represents the temporary j-th level hash of atom k, |V| represents the number of atoms in the atom set V of the compound; For each adjacent atom l of each atom k in the atom set of the compound, the bond type BondType(E k,l ) and the j-1th level hash of atom l After adding set M, take the j-1th level hash of set M and atom k The new hash is used as the temporary j-th level hash value L of atom k k ; Obtain a temporary j-th level hash value set L of the compound according to the temporary j-th level hash value of each atom in the compound; If the temporary j-th layer hash value set L of the compound is the same as the previous layer, i.e., the j-1-th layer hash value set of the compound, then the hash value set of the compound has been calculated; otherwise, the temporary j-th layer hash value set L of the compound is used as the j-th layer hash value set, and the next layer, i.e., the j+1-th layer hash value set, is calculated.

2. The method for extracting a reaction template for a chemical reaction according to claim 1, characterized in that: An implementation process of matching to obtain unchanged atomic mapping includes: Reactant atoms and product atoms having the same hierarchical hash are searched for, and reactant atoms and product atoms having the same hierarchical hash are matched to the unchanged atom mapping.

3. The method for extracting a reaction template for a chemical reaction according to claim 1, characterized in that: An implementation process of performing estimated matching on unmatched atoms to obtain atomic balancing mapping also includes: When all the unmatched reactant atoms have completed the atomic balancing mapping, but the unmatched product atoms still remain, adding at least one special reactant; For each of the generated atoms of the incomplete atomic balancing mapping, traverse and search for the atoms of the special reactants respectively, and calculate the matching scores between the atoms of each of the special reactants and the generated atoms of each of the incomplete atomic balancing mappings; The product atoms with the highest matching scores and the atoms of the special reactants are estimated to match into an atom balancing map; Repeat the above steps until the atomic balancing mapping is completed for all generated atoms that have not completed the atomic balancing mapping.

4. The method for extracting a reaction template for a chemical reaction according to claim 1, characterized in that: An implementation process of performing estimated matching on unmatched atoms to obtain atomic balancing mapping also includes: When all the unmatched product atoms complete the atomic balancing mapping, but the unmatched reactant atoms still remain, adding at least one special product; For each reactant atom of each unfinished atomic balancing mapping, traverse and search for the atoms of the special product respectively, and calculate the matching score between the atom of each special product and the reactant atom of each unfinished atomic balancing mapping; The reactant atoms with the highest matching scores and the atoms of the special products are estimated to match into an atomic balancing map; Repeat the above steps until the atomic balancing mapping is completed for all reactant atoms that have not completed the atomic balancing mapping.

5. The method for extracting reaction templates for chemical reactions according to claim 1, characterized in that: An implementation process of extracting the reaction template of the chemical reaction from the atomic balance map includes: Atoms without bonding or hybridization changes are deleted from the atomic balancing map to obtain the minimum information describing the chemical reaction changes, namely, the reaction template of the chemical reaction; the reaction template includes the attribute information and atomic correspondence of each atom in the template; the attribute information of the atom includes the element information, hybridization type and bonding type of the atom.

6. A reaction template extraction system for a chemical reaction, characterized in that: The reaction template extraction system of the chemical reaction comprises: Input module, obtains reactants and products of a chemical reaction; An analysis module is connected to the input module and analyzes the atomic correspondence between the reactants and the products. An implementation process of analyzing the atomic correspondence between the reactants and the products includes: calculating the hierarchical hash of each atom in the reactants to obtain the hierarchical hash of each reactant atom; calculating the hierarchical hash of each atom in the products to obtain the hierarchical hash of each product atom; the calculation method of the hierarchical hash includes: using the hash function value of the atomic type of each atom in the compound as the 0th layer hash of the atom: Where H represents the hierarchical hash set of compounds, H i represents a hierarchical hash set of the i-th atom in a compound, represents the zero-level hash of the i-th atom, Hash() is an arbitrary hash function, AtomElement(V i ) represents the atomic type of the i-th atom; Obtain a set of level 0 hash values ​​of the compound according to the level 0 hash of each atom in the compound; Calculate the temporary j-th level hash value set L of the compound: L={L1,L2,…,L k ,…,L |V| } Among them, L k represents the temporary j-th level hash of atom k, |V| represents the number of atoms in the atom set V of the compound; For each adjacent atom l of each atom k in the atom set of the compound, the bond type BondType(E k,l ) and the j-1th level hash of atom l After adding set M, take the j-1th level hash of set M and atom k The new hash is used as the temporary j-th level hash value L of atom k k ; Obtain a temporary j-th level hash value set L of the compound according to the temporary j-th level hash value of each atom in the compound; If the temporary j-th layer hash value set L of the compound is the same as the previous layer, i.e., the j-1-th layer hash value set of the compound, the hash value set of the compound has been calculated; otherwise, the temporary j-th layer hash value set L of the compound is used as the j-th layer hash value set, and the next layer, i.e., the j+1-th layer hash value set, is calculated; A first matching module, connected to the analysis module, matching to obtain unchanged atomic mapping; A second matching module is connected to the first matching module, and performs an estimated match on unmatched atoms to obtain an atomic balancing map. An implementation process of performing an estimated match on unmatched atoms to obtain an atomic balancing map includes: the unmatched atoms include unmatched reactant atoms and unmatched product atoms; for each unmatched reactant atom, traverse and search all unmatched product atoms respectively, and calculate the matching score of each unmatched product atom and each unmatched reactant atom; the unmatched reactant atom and the unmatched product atom with the highest matching score are estimated to be matched as an atomic balancing map; and the above process is repeated until the atomic balancing map is completed for all unmatched reactant atoms; A template extraction module is connected to the second matching module and extracts a reaction template of the chemical reaction from the atomic balancing map.

7. An electronic device, characterized in that: The electronic device comprises: A memory storing a computer program; A processor is communicatively connected to the memory, and implements the steps of the method according to any one of claims 1 to 5 when calling and executing the computer program.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images