JPH0740272B2 - Recording / searching method for chemical reaction information - Google Patents

Description

Description: FIELD OF THE INVENTION The present invention relates to a method for recording and retrieving chemical reaction information, and more particularly to recording and retrieving a chemical reaction based on the recorded and stored chemical reaction information. Processing method and search method therefor.

[Technical Background of the Invention] In recent years, with the development of computers, various methods have been proposed and are being used for recording structural information of chemical substances, particularly organic compounds. The amount of organic compounds and organic chemical reactions that have been studied and elucidated to date has reached enormous amounts. However, it is possible to effectively utilize this known information to search for known chemical substances or chemical reactions in a short time. Furthermore, it is desired to find a method for synthesizing a novel substance having desired properties. To that end, as a form of expression of chemical substances and chemical reactions, a computer can process (that is, a computer can make logical judgments) instead of a conventional chemical structural formula that makes it easy for an engineer to grasp the structural characteristics. ) It is required to develop and utilize forms of expression.

As a method of recording chemical substances, WLN (Wiswesser Linear)
Notation) and other linear notation methods are typical, and details are given in, for example, WTWipke.
SRHeller, RJFeldmann, E.Hyde (Eds): “Computer
Representation and Manipulation of Chemical Inform
ation "(John Wiley and Sons, New York, 1974) [Wipke, Heller, Feldman, Hyde (eds.):" Computer representation and handling of chemical information "(John Willie and Sons)]. Join table (conn
ection table) is a list of, for example, the types of each atom in the structural formula of a chemical substance, the atoms of the partner that binds to it, the types of bonds, etc. There is an advantage that you can search in units.

Also, a method for recording information on the structural change (chemical reaction) of a chemical substance has been proposed, but a satisfactory expression method has not been known so far. For example,
As a method of recording information on chemical reactions, there is a method using a reaction code. Specifically, J.Valls, O.Scheiner: “C
hemical Information Systems ", E. Ash, E. Hyde (Eds),
(Ellis Horwood Limited, 1975) p.241-258 [bar,
Shiner: "Chemical Information System", Ash, Hyde, Ed. (Eris Hoawood Co.)], MA
Lobeck, Angew.Chem.Intern.Ed.Engl., 9 , 578 (1970)
[Lovec, Angevante Chemie International Edition in English] and HJZiegler, J.Chem.Inf.Comput.Sc
i., 19 , 141 (1979) [Ziegler, Journal of the
Chemical Information and Computational Science]. This method has a drawback in that when a new chemical reaction is found, it cannot be recorded because the viewpoint of expressing the chemical reaction is fixed. Further, since the structural information of the chemical substance and the information of its change are recorded in separate forms, there is a drawback that effective information retrieval cannot be performed.

In addition, a recording method devised from the standpoint of designing a synthetic route for a chemical substance is also known. For example, EJCorey, RD
Cramer, WJHowe, J.Am.Chem.Soc., 94 , 440 (1972) [Cori, Cramer, Hohe, Journal of American Chemical Society], I.Ugi, J.Bauer, J.Braud
t, J.Friedrich, J.Gasteiger, L.Jochum, W.Schubert, Ange
w.Chem.Intern.Ed.Engl., 18 , 111 (1979) [Ugi, Bauer, Blaut, Friedrich, Gasteiger, Jocham, Schubert, Angevante Chemie International Edition in English] There is. However, this method is not suitable for recording individual chemical reactions.

Heretofore, a desired chemical reaction has been searched based on the chemical reaction information (chemical reaction file) recorded and stored using the above various expression forms.
That is, a large number of chemical reactions and information related to the reactions are registered in advance in a computer, and a reaction including a specific starting material (reactant), product or partial structure can be efficiently detected in a short time. Has been tried. For example, a desired chemical reaction is detected by performing a screening according to search conditions such as a reaction name, a product, and a reaction condition from tens of thousands of chemical reaction files.

In this reaction search system, the input of a specific starting material, product, or partial structure involved in the reaction to the computer as the search condition and the output of the detected chemical reaction are the structural formulas that chemists are usually used to. Alternatively, it is desirable that the search can be performed in the form of a three-dimensional structure diagram, and the in-computer search based on the input conditions is preferably performed by an atom-atom tracking method using a bond table. Furthermore, the enormous amount of information can be registered and accumulated, and various chemical structures can be easily registered.
In addition, it is required to use an expression form that can be stored without occupying a large storage capacity, and to easily perform mutual conversion between the computer processable expression form and the input / output structural diagram.

It should be noted that the Applicant has applied the search condition to at least one starting material, at least one product, and / or at least one set of nodes involved in the reaction and a search condition including a condition regarding a bond between the nodes. We have already applied for a patent for a chemical reaction search processing method that is characterized by extracting information about a satisfied chemical reaction based on an imaginary transition structure and / or a binding table (Japanese Patent Application No. 60-221087).
issue).

SUMMARY OF THE INVENTION The present invention provides a processing method for searching for information on a specific chemical reaction based on chemical reaction information in a computer processable expression form and displaying and recording it in the form of a reaction formula. That is the purpose.

The present invention topologically superimposes the structure of a starting material and the structure of a product of a chemical reaction for producing at least one product from at least one starting material, and (1) is common to both the starting material and the product. Create a bond table that distinguishes between the bonds between the nodes, (2) the bonds between the nodes existing only in the starting material, and (3) the bonds between the nodes existing only in the generated material, and at least two of them adjacent to each other. Of a node clause and a bond between nodes existing only in the starting material connecting the nodes or between nodes existing only in the generated material and making it a binding table; A step of recording and storing the binding table of the chemical reaction and the binding table of the two-section subgraph obtained in (1) in a recording medium attached to a computer; In recording and search for basic reactions information including; step of searching a chemical reaction by performing processing on a computer and a connection table of the binding table and Section II subgraph of a chemical reaction.

According to the method of the present invention, a huge amount of chemical reaction information recorded and stored as a binding table derived from an imaginary transition structure is used, and it is suitable based on an input condition given as a partial reaction structure. By performing the processing, information regarding the target chemical reaction can be automatically detected.

In this specification, imaginary transition structures
tructure, hereinafter abbreviated as ITS) is a structural change of a substance involved in a chemical reaction due to (1) a bond existing only in a starting material, (2) a bond existing only in a product and (3) both. A two-dimensional or three-dimensional structure diagram (figure) that is classified into three types of bonds that exist in common
Say. This structural diagram is based on conventional structural formulas and three-dimensional structural diagrams of compounds, and is capable of showing chemical reactions in a form that is easily understood by engineers and easily understood.

In addition, a connection table can be derived from the imaginary transition structure, and the combination of the type of node in the chemical reaction, the partner node that binds to the node, and the connection between these nodes, which are distinguished by being classified into the above three types. Is a simple and clear list consisting of, and this binding table allows chemical reaction information to be stored and stored on a recording medium without requiring a large capacity. In particular,
The use of the binding table as the registration form facilitates the processing of information by a computer and facilitates the registration of chemical reactions, which facilitates the storage and management of such information.

In the above bond table, chemical reactions are basically represented by simple expressions of bonds between nodes (nodes) consisting of atoms, atomic groups, etc., and bonds between nodes in a reaction system are The above three types are represented separately.
For this reason, by performing simple graphic processing or arithmetic processing on the join table recorded and stored (registered) in the computer, paying attention to the distinction between the joins, information about the join change peculiar to the reaction, that is, "2 section part" Graph ”(2-
nodal subgraph) can be obtained.

In the present invention, a two-node subgraph representing a partial structure of a reaction is basically two adjacent nodes and one (2) a connection between the nodes which exists only in the starting material or (3 ) A subgraph that includes connections between nodes that exist only in the generated substance. Since this two-node subgraph has a simple structure, it can be easily derived and counted from the binding table of reactions by computer processing.

Further, according to the present invention, the reaction type peculiar to a chemical reaction is represented by a combination of a plurality of two-node subgraphs, specifically by the type and number of two-node subgraphs. Similar reactions have a common binomial subgraph combination. Since the reaction type is obtained in the form of a two-dimensional graphic following the imaginary transition structure of the reaction as a combination of the two-node subgraphs, the engineer can easily understand it visually. The two-node subgraph can be represented in the form of a connection table, and in this case, the two-node subgraph information can be represented very simply. Therefore,
It is easy to register the information in the computer and the registration does not require a large capacity. By registering the combination of the two-section subgraphs in advance in association with the reaction, it becomes possible to perform a simple reaction search through the subgraph.

The connection table obtained about the information about these two-node subgraphs can be recorded and stored in a computer, or can be recorded on a blank sheet of paper or displayed on a screen such as a cathode ray tube.

The search for a chemical reaction by a computer is performed by searching for a reaction that matches a search condition (clue, keyword) from a large number of registered reactions and screening it, but the computer processing is effectively performed in a short time. For this reason, setting search conditions is very important.
According to the present invention, since the search condition can be given in the form of a simple partial structure by combining the above-mentioned two-section subgraphs, the processing time is shorter than that of setting one complicated reaction structure, and the processing can be performed. No trouble will occur. In particular, a search condition can be partially digitized by performing a search based on the type and the number of the bisection subgraphs, and the computer processing is further facilitated. Therefore,
The method of the present invention is an effective method as a primary screening for reaction search.

Further, since information on chemical reactions can be efficiently searched based on the registered reaction information in a short time, the time required for collecting and investigating information in individual research of engineers is shortened, and It is possible to increase the density of information that can be obtained, and to carry out research efficiently.

In addition to these advantages, in the method of the present invention, partial structure search of chemical substances, structure-activity relationship, automatic structure determination of unknown compounds, and reaction mechanism when complex compounds are reacted under certain conditions and Prediction of reaction products (mechanis
tic evaluation of organic reaction) etc. can also be conducted together. In addition, the obtained information is used for structural analysis of chemical substances, molecular modeling, which is highly demanded by engineers involved in drug manufacturing, etc.
Heuristic analysis of organic sy
nthesis) and the like.

[Structure of the Invention] In the chemical reaction information recording / retrieval method of the present invention, a bond between nodes is composed of a bond existing only in a starting material, a bond existing only in a product, and a bond commonly existing in both. About the chemical reaction information represented as a bond table derived from the imaginary transition structure that is distinguished by the type,
An operation of extracting a two-node subgraph in which nodes involved in the reaction are connected by a specific connection is performed, and as a result, information about a change in the connection unique to the reaction is obtained as a connection table.

The method of the present invention will be described below by taking the reaction of hydrolyzing ethyl acetate with hydrochloric acid as an example.

This chemical reaction is represented by CH ₃ COOCH ₂ CH ₃ + H ₂ O + HCl → CH ₃ COOH + CH ₃ CH ₂ OH + HCl (Scheme 1). The imaginary transition structure (ITS) of the reaction can be expressed as follows, for example.

here, Represents the bonds that are commonly present in the starting material and the product, Represents a bond present only in the starting material, and It represents a bond that exists only in the product.

That is, the imaginary transition structure (ITS) is a two-dimensional or three-dimensional structure in which the structure of the starting material and the structure of the generated material are topologically overlapped and the bonds between the respective nodes are distinguished by the above three types i) to iii). The original structure. It should be noted that “topologically superimposing” specifically means that the nodes appearing in the structure of the starting material and the nodes appearing in the structure of the produced material are matched and these structures are combined into one.

In the above imaginary transition structure (ITS), the nodes of the substances involved in the chemical reaction may be represented by the atoms contained in the original system (starting substance group) and the producing system (producing substance group) as a unit. , Or a methyl group [above node (1),
(5)], a methylene group [the above-mentioned node (4),] or other functional group may be represented by an atomic group unit. When expressing a chemical reaction, some of the nodes appearing in the original system and the producing system may be omitted.

Also, the distinction between the three types of bonds is not limited to the display by the symbols such as i) to iii) above, and may be displayed by simple characters such as numbers (1, 2, 3) or the color (black). , Red, green) and the like, any means may be used as long as the user can judge by the five senses and computer processing is possible.

In the following, i) a common bond to the starting material and the product Is called a "colorless bond" or a constant bond (par-bond), and ii) a bond that exists only in the starting material. Is called "out-bond", and iii) the bond that exists only in the product Is called "in-bond", and out-bond and in-bond are collectively called "colored bond", and these three kinds of bonds appearing in the imaginary transition structure are collectively called "imaginary bond". 』I will call it.

Specifically, Table 1 shows the types of bonds appearing in the imaginary transition structure. In addition, in Table 1, the numerical value on the side means an index of bond entry and exit.

In Table 1, for example, symbols The bond represented by is a single in-bond and can be represented by a pair of numbers (0 + 1). Here, 0 means that there is no bond in the original system before the reaction, and +1 means that a single bond has occurred in the produced system after the reaction. Similarly, The bond represented by is a single out-bond and is represented by (1-1). There is a single bond in the original system before the reaction, but the single bond disappears in the production system after the reaction. It means that The bond represented by (2-1) is a double bond singly cleaved, It is written in.

Thus the type of bond is also a pair of numbers: (a, b)
[Where a is an integer that represents the bond multiplicity in the starting material, and b is an integer that represents the change in the bond multiplicity in the chemical reaction], and this can be represented by the “complex bond number” (complex bond number).
bond number) or “imaginary multipl”
icity). The comma (,) in the notation (a, b) may be omitted. According to this notation, even if the coupling multiplicity is two or more, it can be simply expressed. In addition, it is a particularly preferable method for recording and storing chemical reactions because it does not require much storage capacity and can be directly processed by a computer, and is an expression method that is preferably used when creating a binding table described below.

Alternatively, the chemical reaction is represented as a connection table that includes information about each node, the node of a binding partner that binds to the node, and the binding between the nodes.

The binding table for the hydrolysis reaction of the above ester is shown in Table 2. The connection table also includes information on the two-dimensional coordinates (xy coordinates) of each node.

As shown in Table 2, the binding table shows all nodes and their two-dimensional coordinates (node (1) for the original system (ethyl acetate, water, hydrochloric acid) and the productive system (acetic acid, ethanol, hydrochloric acid) involved in the reaction. Is a list in which all nodes connected to each node and the types of connections between these nodes are listed in order of node number.

Note that it is possible to create a binding table based on the imaginary transition structure of the reaction, and conversely, if the binding table contains the position information of each node, the imaginary transition structure can be created from the binding table. . In other words, it can be said that the imaginary transition structure and the binding table are in a one-sided relationship as registration and display form of chemical reaction information.

In the connection table, as described above, the input information regarding the position coordinates and the like of each node may be written together. If desired, additional information may be added to the bond table (and imaginary transition structure) regarding charge and stereochemistry of each node; various physical properties of chemical substances involved in the reaction; spectral information; and enthalpy and temperature of the reaction. Information about time, catalyst used, atmosphere, reaction phase, reaction yield, presence or absence of by-products, and the like may be described. Further, the binding table (and the imaginary transition structure) may be provided with a reaction name, a reaction number, etc. for each in order to facilitate the storage, management and retrieval of chemical reaction information. The binding table may be registered in the computer by recording and saving it in the main storage device in the computer, or by recording it through an appropriate recording medium (magnetic disk, optical disk, magnetic tape, etc.). Good.

Further, the registered connection table can be recorded on various recording materials such as plain paper by an appropriate recording device, or can be displayed (graphic display) on a color cathode ray tube connected to a computer or an electronic device.

For details of the imaginary transition structure and / or the binding table, Japanese Patent Application No. 60-177345 and Japanese Patent Application No. 60-180875 filed by the present applicant.
No. In addition, the imaginary transition structure including charges (DITS), the binding table, and the three-dimensional imaginary transition structure (IT
Details of SS) and the binding table are described in Japanese Patent Application Nos. 60-298603 and 60-298604 by the present applicant, respectively.

From the imaginary transition structure (ITS1) of the above reaction, a colored bond, that is, By extracting the part in which and are alternately connected, the reaction string [reaction string, reaction graph (graph
of reaction centers).

In other words, colorless from ITS1 Is obtained by deleting. However, if a colorless bond and a colored bond are combined (for example, For the purpose, the colorless bond shall be left without being deleted. For details of the reaction, refer to Japanese Patent Application No.
No. 60-185386.

In the present invention, in the above reaction graph, a colored bond A substructure (subgraph) including n adjacent nodes and the colored connection that are connected with each other, which includes only one reaction, is particularly referred to as an "n-nodal subg".
raph; where n is an integer of 2 or more). n
A node subgraph is a kind of reaction descriptor. For example, the reaction graph 1rc can be said to be a six-section subgraph. Further, the atom included in the n-node subgraph is referred to as "Ouchi atom", and the atom adjacent to the Onode atom and not included in the n-node subgraph is referred to as "Ogata atom".

A two-node subgraph (also called 2NSG) is basically a colored combination. 3 is a subgraph consisting of two adjacent nodes connected by and the colored connection. In addition, a colorless bond to one of these adjacent nodes One or more nodes (outer atoms) that are connected with each other and a subgraph including the bond are also included in the two-node subgraph.

As an example, as a two-node subgraph of the above reaction, from ITS1, two adjacent nodes connected by a colored bond, the colored bond, and optionally a colorless bond to one of the nodes. The following series of subgraphs are obtained by extracting the extraneous atoms and their bonds.

Here, the binomial subgraph is extracted for each node of interest, and when there is an extra atom adjacent to the node, the extra node is also included in the binode subgraph.

For comparison, a three-section partial graph of the above reaction is illustrated.

In the chemical reaction search described below, using the combination of the two-section subgraphs rather than the combination of the three-section subgraphs as the search condition simplifies each subgraph despite the large number of subgraphs. Therefore, the search process becomes easy.

However, the three-section subgraph , It is clear that it is a substitution reaction, while the corresponding two binomial subgraphs Combining cannot be said a priori, and therefore the possibility of substitution reaction cannot be confirmed even if the binomial subgraph suggests it.

In addition, the binomial subgraph is such that, after detecting a colored bond (a bond with b ≠ 0) from the bond table shown in Table 2, the node involved in the detected bond and, in some cases, a colorless bond to the node. It is obtained as a partial join table by deriving a node to join with (join with b = 0).

Another example is the reaction graph About reactions with Focusing on the join, as a bisection subgraph Is obtained.

When registering a binomial subgraph, it is preferable to enumerate and register the following types of binomial subgraphs and the number thereof in combination with Ogai atoms.

The obtained set of the two-node subgraph is recorded and stored (registered) in the computer. Further, it may be displayed or recorded via a suitable display recording means. At this time, the set of the two-node subgraph may be attached to the reaction ITS or the binding table,
Alternatively, 2NSG files may be created separately and independently and associated with the ITS or binding table of the reaction. Also, to facilitate the storage, management, and retrieval of information, the sets of two-section subgraphs may be numbered, coded, or otherwise common to each two-section subgraph set. The reaction names of may be registered together. The registration in the computer and the recording of the display can be performed by the same method as that described above.

Next, using information files about a number of chemical reactions registered in the computer in the form of a binding table and information files about a set of two-node subgraphs that simply describe reaction structures, The information is detected by the method described below.

First, the following two-node subgraph set is given as a search condition. Taking the reaction having the above reaction graph A as an example, a bisection partial graph If you try to search by Respectively, and these two-node subgraphs and their numbers are used as search conditions. This set of binomial subgraphs may be entered directly, or it may be automatically generated in the computer from the basic binomial subgraph.

The input is performed in the form of a structural formula based on the imaginary transition structure as described above, a connection table, or a combination thereof. Even if the input is made in the form of a structural formula, it can be automatically converted into a form of a computer-processable coupling table by computer processing based on this structural condition.

At this time, in order to increase the screening efficiency, a part of the node may be represented by an atomic group, and further simplified characters and symbols (HA = halogen, MT = metal) are used to represent the node and the bond species. Alternatively, conditions such as aromaticity and hellero ring may be added. If there are multiple possibilities for atom type or substituent position, so-called Markush expression (for example, X = Cl, OH) can be used. Conditions such as whether or not it can be added.

Input is performed using a character terminal such as an ordinary teletypewriter or a chemical typewriter devised to draw a chemical structural formula; and an image terminal (graphic terminal) using a keyboard, joystick, light pen, etc. You can The latter image terminal is easy to use because it can draw a two-dimensional figure directly on the graphic display. You can enter a question in any terminal, and if you want to set various types of search conditions, use the usual Boolean operations (AND, O
R, NOT) can be used.

Then, screening is performed based on the search conditions.
After searching for a reaction having the same number as or more than that number for each two-node subgraph (this is called a "hit"), only the reactions commonly detected are left. This hits the reaction with reaction graph A.

A bipartite subgraph that is a subgraph of the above bipartite subgraph Similarly, in the case of searching with, a set of two-node subgraphs is generated, and a reaction having each number of two-node subgraphs is hit with this as a search condition. Therefore, also in this case, the reaction having the reaction graph A is hit.

From a computer processing point of view, the search is preferably performed based on the connection table. In particular, it is preferable to search by atom by atom based on the connection table interactively so as to satisfy the search condition input in the form of structural formula on the graphic display.

In this way, a rough search (primary screening) can be performed using the binode subgraph. Screen the chemical reaction suitable for the purpose by repeating the same operation for the results of the primary screening, adding search conditions, or by giving other search conditions such as reaction graphs and ring descriptors. can do.

When the chemical reaction information file contains various data such as reaction conditions, presence or absence of by-products, yields, starting materials and physical properties of produced materials as additional information, it is represented by a reaction formula or the like. These additional information can be obtained as well as the target chemical reaction.

The information on the target chemical reaction obtained by the above reaction search can be displayed two-dimensionally or three-dimensionally as a reaction formula, an imaginary transition structure, or a binding table on a graphic display such as a CRT. At this time, regarding the starting material and the product, the optimized three-dimensional coordinate of the molecule is obtained by energy calculation (molecular force field calculation, etc.) based on the structural information obtained as the two-dimensional structure diagram. It is also possible to obtain a three-dimensional structure diagram (three-dimensional structure, space filling diagram, etc.) by performing an operation. In addition, a three-dimensional structural diagram viewed from an arbitrary angle can be enlarged or reduced and displayed on the display.

The obtained information can also be recorded (made into a hard copy) on a suitable recording material such as plain paper.
Further, it may be recorded and stored in an appropriate recording medium such as a magnetic disk.

The method of the present invention will be further described below with reference to examples.

[Example] The following Diels-Alder reactions (1) to (5): The imaginary transition structure of is expressed as follows.

From each ITS, the binomial subgraphs were extracted and counted, and a set of binomial subgraphs was obtained. The results are shown in Table 3 as a list.

In Table 3, the binomial subgraphs are enumerated by node.

On the other hand, the question formula: Given, the set of binomial subgraphs: There has occurred. Using this as a search condition, primary screening was performed on the binomial partial graph file, and a reaction having each 2NSG or more was hit. As a result,
All the reactions (1) to (5) were detected.

Similarly, a pair of two-node subgraphs was generated for each of the following query expressions, and the corresponding reaction was detected.

Claims (3)

[Claims] 1. A structure in which a starting material and a structure of a starting material in a chemical reaction for producing at least one starting material from at least one starting material are topologically overlapped with each other, and (1) common to both the starting material and the starting material. A binding table that distinguishes between the coupling between the nodes, (2) the coupling between the nodes existing only in the starting material, and (3) the coupling between the nodes existing only in the generated material, and at least two adjacent A step of extracting a two-node subgraph including a connection between nodes existing only in each node and a starting material connecting the nodes or a connection between nodes existing only in a generated material and using this as a connection table; A step of recording and storing the binding table of the chemical reaction and the binding table of the two-section subgraph obtained in the step in a recording medium attached to the computer; Recording and search for basic reactions information including; step of searching a chemical reaction by the performing processing on a computer and a connection table of the binding table and Section II subgraph of a chemical reaction. 2. A node in a connection table is represented by a number consecutive with its type, and a connection between nodes is represented by a pair of numbers (a, b).
[Wherein a is an integer representing the bond multiplicity in the starting material and b is an integer representing the change in the bond multiplicity in the chemical reaction]. Recording / search method. 3. The chemical reaction search comprises at least a bond between two nodes adjacent to each other and a bond existing only in one starting material connecting the nodes or a bond existing in only the product. The method for recording and retrieving chemical reaction information according to claim 1, which is performed by using one type of bisection subgraph and the number thereof.