JPH09305628A - Device for generating canonized data and its method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a canonized data generating device by which the use quantity of a storage area in a compound/reaction database system is greatly reduced. SOLUTION: Peculiar data concerning respective atoms received by an input means A and inter-atom connection pair data are given to a canocized data generating means B. Then, the canonized data generating means B generates canonized data based on these kinds of data. Canonized data generated in this way is an extremely short character, numerical and code string. Therefore, many kinds of canonizing data are preserved in a small storage area.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a processing apparatus and a processing method for performing information processing in the fields of chemistry and biochemistry, and particularly to uniquely specify a chemical structure of a compound from various data on each atom constituting the compound. And a canonicalized data creating method for creating canonicalized data.

[0002]

2. Description of the Related Art Conventionally, a compound database system containing compound information and a reaction database system containing compound reaction information have been developed. The compound database system contains compound information such as physical properties and actions of existing compounds, and the compound information is accessed with the structure of the compound as a key. By using this compound database, it is possible to efficiently refer to the properties and actions of compounds. In addition, the reaction database system stores reaction information of existing compounds, and the reaction information is accessed using the structure of the compound as a key. Using this reaction database system, researchers in synthetic chemistry can perform new synthesis of compounds,
Since similar reaction information can be efficiently referred to from the reaction information of existing compounds, it is essential in synthetic chemistry research.

In the compound database, for example,
There is a compound management system "MACCS" of MDL, Inc. in the United States. Further, the reaction database system is, for example, a comprehensive chemical information management system "ISIS" of MDL, Inc. in the United States.
There is also a reaction information management system "REACCS".

[0004]

By the way, the conventional compound / reaction database system has a function of displaying a structural diagram of a compound on a display together with information on the property and action of the compound. By displaying the structure diagram in this way, it is possible to construct a visually excellent system. However, since the structural diagram of this compound is stored in the system as image data (bitmap data) occupying a large storage area, a large-capacity storage device is required, which is a problem.

The present invention provides a canonicalized data creating apparatus and a canonicalized data creating method which can significantly reduce the amount of storage area used in a compound / reaction database system. With the goal.

[0006]

In order to solve the above-mentioned problems, the canonicalized data creating apparatus of the present invention has an input for accepting the input of the unique data for each atom constituting the compound and the bond pair data between the atoms. And a canonical data creating means for creating canonical data capable of uniquely identifying a chemical structure of a compound based on each data received by the input means. Then, the canonicalized data creating means classifies each atom into different classes for each equivalent atom, and assigns a different class number to each atom in each class. A second processing unit that gives each atom a canonicalization number uniquely corresponding to the structure of the compound based on the class number given to the atom, and the canonical process given to each atom by the second processing unit. Canonicalized data based on the digitization number Characterized in that it comprises a third processing unit that formed.

According to the canonical data creation device of the present invention having such a configuration, the unique data for each atom and the bond pair data between the atoms accepted by the input means are given to the canonical data creation means. To be Then, the canonicalized data creating means creates canonicalized data based on these data.

That is, in the canonicalized data creating means, first,
The processing of the first processing unit is executed, and each atom is classified into a different class for each equivalent atom based on the unique data for each atom and the bond pair data between the atoms. Then, a different class number is given to each atom for each class. Next, the processing of the second processing unit is executed, and the canonicalization number uniquely corresponding to the structure of the compound is assigned to each atom based on the class number given to each atom and the bond pair data between the atoms. give. Further, the process of the third processing unit is executed to create the canonicalized data based on the canonicalized number given to each atom and the unique data for each atom.

Here, the first processing unit gives each atom three types of attributes (a _i , b _ij , d _ij ) and determines that atoms having even one of these attributes are not equivalent. Utilizing what is possible, each atom is given a different class number for each equivalent atom, and among the three types of attributes (a _i , b _ij , d _ij ), a _i is the type of atom with the input number i Is a symbol, b _ij is the number of bonds whose kind symbol is j among the bonds adjacent to the atom of input number i, and d _ij is the number of bonds from the atom of input number i by the shortest path. The second processing unit assigns the canonicalization number 1 to the atom with the highest class number priority in the process of assigning the canonicalization number to each atom in ascending order from 1. , And thereafter, up to the canonicalization number n, the atoms which have already been given canonicalization numbers, and Among the atoms to which atoms that have not been given canonicalization numbers are bonded, select the atom with the smallest canonicalization number, and the atom that is bonded to that atom, and whose canonicalization number is still Among the atoms that are not given, the atom having the highest class number priority is given the canonicalization number n + 1, and the third processing unit gives each atom three types of attributes (P
_i , T _i , S _i ) are given and these attributes are arranged in a line to create the canonical data. Among the three types of attributes (P _i , T _i , S _i ), P _i is the canonicalization number i
Is the canonicalization number of the atom that has the smallest canonicalization number and is connected to the atom of, and T _i is the atom of canonicalization number i and the canonicalization number P _i.
Is a symbol for the type of bond with the atom, and S _i is the canonicalization number i
It is preferable that the symbol is the atom type symbol.

Further, the canonical data creation method of the present invention is
A method for creating canonical data that uniquely identifies the chemical structure of a compound, based on the unique data for each atom that constitutes the compound and the bond pair data between the atoms. Is classified into different classes for each equivalent atom, and the structure of the compound is determined based on the first step of giving each atom a different class number for each class and the class number given to each atom in the first step. The second step of giving each atom a canonicalization number uniquely corresponding to and the third step of creating the canonicalization data based on the canonicalization number given to each atom in the second step. And a step.

Here, the first step is to give three kinds of attributes (a _i , b _ij , di _ij ) to each atom, and to judge that atoms having any one of these attributes are not equivalent. , Each atom is given a different class number for each equivalent atom, and among the three types of attributes (a _i , b _ij , d _ij ), a _i is the type symbol of the atom with the input number i. And b _ij
Is the number of bonds whose kind symbol is j among the bonds adjacent to the atom with the input number i, and _dij is the number of paths that can pass from the atom with the input number i through j bonds by the shortest path. The second step is to give the atom having the highest class number priority in the process of giving each atom in ascending order from 1 and assigning the canonical number 1 thereafter. Up to, the atom with the canonicalization number already assigned and the atom with the canonicalization number not yet assigned has the smallest canonicalization number. Select an atom and assign the canonicalization number n + 1 to the atom that has the highest class number priority among the atoms that are bonded to that atom and have not yet been given a canonicalization number. In the third step, each atom has three types of attributes (P _i , T _i , S _i ) is given to create the canonical data by arranging these attributes in a line. Among the three types of attributes (P _i , T _i , S _i ), P _i is the canonicalization number. i is the canonicalization number of the atom that is bonded to the atom of i and has the smallest canonicalization number, and T _i is the type symbol of the bond between the atom of canonicalization number i and the atom of canonicalization number P _i . , S _i is preferably a symbol of the atom type with the canonicalization number i.

[0012]

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing a configuration of a canonical data creating apparatus 1 according to an embodiment of the present invention. As shown in FIG. 1, the canonical data creation device 1 includes an image memory 10 for storing image data 10a of a molecular structure diagram and a work memory 1 for temporarily storing symbol data 11a and the like.
1 and a main storage device 2 in which an operating system (OS) 21 and a canonical data creation program 22 are stored
0, binding table file 31 and compound information file 33
Is stored in the hard disk device 30.

Further, the canonical data preparation system 1 includes a display 40 for displaying a molecular structure diagram and a mouse 50 as a pointing device for accepting input of handwritten figures.
A keyboard 60 for receiving symbolic data such as chemical formulas, a printer 70 for outputting a molecular structure diagram, and a CPU 8 for controlling execution of the molecular structure diagram creation program 22.
It has 0 and. In addition to the mouse 50, the pointing device includes a tablet, a digitizer, a light pen, and the like.
It may be provided instead of 0.

The canonical data creation program 22 is a program that creates canonical data based on the unique data for each atom constituting the compound and the bond pair data between the atoms. This canonical data creation program 22
A main routine 100 for supervising the processing and a constituent atom classification routine (first processing unit) 101 for giving a class number to each atom constituting the compound are provided. Further, the canonicalization data creation program 22 is based on a canonicalization number assigning routine (second processing unit) 102 for giving a canonicalization number to each atom based on a class number, and a canonicalization number for each atom. And a canonicalized data creation routine (third processing unit) 103 for creating canonicalized data.

The hard disk device 30 is provided with a binding table file 31 capable of storing a plurality of binding tables 32. The bond table 32 records unique data for each atom constituting the compound and bond pair data between the atoms, and the canonical data creation program 22 can access these data via the bond table 32.

As shown in FIGS. 2 (a) and 2 (b),
The bond table 32 includes an atom table 32a in which unique data about each atom is recorded and an atom pair table 32b in which bond pair data between atoms are recorded. Specifically, in the atom table 32a, input numbers (also referred to as atom numbers), two-dimensional coordinates (X coordinates / Y coordinates) of atoms, element names (generally element symbols are used, but numbers such as atomic numbers). Column) is provided (see FIG. 2A), and the atom pair table 32b includes the bond atom pair data and the bond type (for example, single bond). Bonds are 1 and double bonds are 2), and a column for writing the structure (a column for distinguishing whether each molecule belongs to the cyclic portion or the chain portion of the molecular structure diagram) is provided (Fig. Two
(B)).

Here, the input number is a number for identifying each atom constituting the compound with a computer, and FIG.
Although it is a number in the example of (a), it may be a symbol. Also, the bond atom pair data is preferably represented as a combination of input numbers.

In order to create the canonicalized data,
Not all the data in the atom table 32a and the atom pair table 32b described above are required, but it is sufficient if the atomic number and element name are used as the unique data, and the bond atom pair data and the bond type are used as the bond pair data.

The hard disk device 30 also stores a compound information file 33 in which a list showing the relationship between compound numbers and canonical data (also called canonical data) corresponding to the compounds is recorded. As shown in FIG. 3, the compound information file 33 includes compound numbers C ₁ to
Canonical data for each compound of C ₇ and compound C ₁
~ C ₇ reference data for each compound (name, literature,
And physical properties) are recorded as a list corresponding to compound numbers C _{1 to} C ₇ . Therefore, if the compound information file 33 is accessed by using the compound numbers C _{1 to} C ₇ as keys, the canonicalized data and the reference data for each compound of the compound numbers C _{1 to} C ₇ can be read. Here, the canonicalized data is data composed of a plurality of symbols that uniquely specify the chemical structure of each compound.

The constituent atom classification routine 101 is the first
The canonicalization number assignment routine 102 corresponds to the second step, and the canonicalization data creation routine 103 corresponds to the third step.

Next, the outline of the operation of the canonical data creating device 1 will be described. As shown in FIG. 4, the operator operates the mouse 50 or the keyboard 60 to display the binding table 32 of the compound for which the molecular structure diagram is to be created in the binding table file 31.
Can be created within.

The mouse 50 is used for handwriting the molecular structure diagram of the compound on the display 40 using the mouse 50. The input numbers of the atoms determined in the input order are
It is written in the input number column of the combination table 32 created in the main storage device 20. Further, bond atom pair data indicating the bond relationship of each atom in the molecular structure diagram E ₁ is written in the bond atom pair column of the bond table 32. In this way, mouse 5
When the input is 0, the bond table 32 for identifying the compound is created from the handwritten molecular structure diagram E ₁ .

The input of the handwritten molecular structure diagram E ₁ by the mouse 50 will be described more specifically. First,
When the operator clicks the mouse 50, data on one of the atoms forming the bond atom pair is input. Next, when the operator moves the mouse 50 and clicks the mouse 50 again, data on one atom and the other atom forming a bond atom pair is input. Then, the click for inputting the data for the other atom is regarded as a click for designating one atom of the next bond atom pair when the next click is subsequently made.
In this way, one bond atom pair can be designated by two consecutive clicks. That is, the operator can input all the bonds constituting the compound by continuously designating the bond atom pairs while shifting the atoms one by one.

The atom number is written in the input number column of the atom table 32a in correspondence with the input data of the atom (for example, 1, 2, 3, ... In the order of atom input). Further, the input data regarding the bond relation of the atoms is written in the column of the bond atom pair of the atom pair table 32b. Further, when the operator inputs the element name of the atom, it is written in the element name column of the atom table 32a. Similarly, when the operator inputs the multiplicity of the bond connecting the bond atom pairs, it is written in the bond type column of the atom pair table 32b. The element name and multiplicity should be recognized as carbon and single bond, respectively, when it is not written. The molecular structure diagram and the bond table based on it are usually prepared by omitting hydrogen atoms.

In addition, the keyboard 60 is used to input a symbol string that specifies a binding table name corresponding to a predetermined compound using the keyboard 60. Based on the input symbol data 11a, the binding table name is input. The binding table 32 specified by is read from the binding table file 31.

Thus, the mouse 50 and the keyboard 6
0 constitutes input means A, and mouse 50 and keyboard 60
The binding table 32 is obtained by using either of the above. And
The canonicalized data creating program 22 which is the canonicalized data creating means B is executed, and the canonicalized data 34 is created based on each data of the connection table 32. The canonical data 34 thus created is written and stored in the compound information file 33 together with the compound reference data. The reason why the canonicalized data 34 is created and stored from the combination table 32 is that the storage area can be made smaller than that of the combination table 32 as it is. That is, the connection table 3 shown in FIG.
The canonical data 34 created based on 2 is “1% 1% 1
-2% 3% 5% N / 6% 7 / ", and the structure of the compound can be represented by extremely short letters, numbers, and symbol strings. The storage resources can be effectively used, and the device can be made compact and lightweight.

Further, the two-dimensional coordinate calculation processing is performed based on the bond atom pair data in the bond table 32,
Two-dimensional coordinate data of each atom can be obtained. From the two-dimensional coordinate data thus obtained, the aesthetically superior molecular structure diagram E ₂ is created. The created molecular structure diagram E ₂ is
It can be displayed on the display 40 or output from the printer 70.

The input from the keyboard 60 may be performed by directly writing the unique data and the join pair data into the join table created in the main storage device 20. In addition, image scanners and optical card readers (OC
A device for optically reading a figure or a character such as R) may be used as the input device of the present invention to accept the input of the connection table data.

Next, a canonicalized data creating method according to an embodiment of the present invention will be described. This method is
The canonical data creation device 1 is used. First, OS21
Under the control of, the main routine 100 of the canonical data creation program 22 is started.

As shown in the flow chart of FIG. 5, the main routine 100 first consists of the constituent atom classification routine 10.
1 is called to give a class number to each atom constituting the compound (S10). Next, the canonicalization number assignment routine 102 is called to assign a canonicalization number to each atom based on the class number assigned to each atom (S20). Furthermore, by calling the canonical data creation routine 103,
The canonicalization data is created based on the canonicalization number given to each atom (S30). The canonical data thus created is written and stored in the compound information file 33.

Next, the processing of the constituent atom classification routine 101 called in S10 will be described. This process
This is a process of classifying each atom constituting a compound into different classes for each equivalent atom and giving a class number corresponding to the class to which each atom belongs. For example, all 6 carbon atoms of benzene are equivalent, so they are all given the same class number. The 7 carbon atoms of toluene are represented by 5 different class numbers. That is, the two carbon atoms in the ortho position of toluene and the two carbon atoms in the meta position are equivalent and are given the same class number.

As shown in the flow chart of FIG. 6, first, based on the bond table 32, 3 atoms are formed for each atom constituting the compound.
Type attributes (a _i , b _ij , d _ij ) are given (S
101). Here, the attribute a _i is the type symbol (atom number in this example) of the atom of the input number i. Further, the attribute b _ij is the type symbol (in this example, the bond type is a single bond, a single bond, or a double bond) of the bonds adjacent to the atom with the input number i (that is, the bond having the atom with the input number i as one atom). The number of bonds (vector quantity) is j, where 2 is a heavy bond, 3 is a triple bond, and 4 is an aromatic bond. Furthermore, the attribute d _ij is the input number i
It is the number of paths (vector amount) that can be traveled from the atom of through the j bonds by the shortest path.

Next, the attributes (a _i , b _ij ,
(d _ij ) are arranged to form a numerical sequence (in this example, a 9-digit numerical sequence), the class numbers C _i ⁰ are given in ascending order of the numerical sequence, and each atom is classified into a plurality of classes (S102). The class number C _i ⁰ given here is the 0th-order class number, and is the 1st-order class number C in the loop processing after S103.
_i ¹ , secondary class numbers C _i ² , ... Are sequentially obtained.
If the number of classes is equal to the total number of atoms in this 0th order,
The processing may be terminated.

Next, the order n is set to 1 (S103). Then, the attribute V _ij ⁿ is given to each atom (S104). The attribute V _ij ⁿ is the number of atoms having a class number j in the order n−1 that is bonded to the atom with the input number i. Furthermore, the attributes (a _i , b _ij , d _ij , V _ij ⁿ ) are arranged for each atom, and the class numbers C _i ⁿ are given in the ascending order of the number sequence to classify each atom into a plurality of classes (S105). ). Then, check whether the number N _n of classes is equal to the number of classes N _(n-1) in the previous loop, the process is terminated if equal. Alternatively, it is checked whether the number N _{n of} classes is equal to the total number of atoms, and if they are equal, the process ends (S106). If they are not equal, 1 is added to n and the process returns to S104 (S107). In the processes of S102 and S105, the class numbers are given in ascending order of the number string, but the class numbers may be given in ascending order.

Next, 3,5-dimethyl-2,3,4,5
The step-by-step processing of the constituent atom classification routine 101 will be described in detail using tetrahydropyridine as an example.

First, the process of S101 is executed. When this processing is executed, the data as shown in FIGS. 7A and 7B has already been written in the binding table 32, and based on each data written in this binding table 32, the data is written in each atom. Three
The type attributes (a _i , b _ij , d _ij ) are given. Here, the input numbers recorded in the bond table 32 are arbitrary numbers given in the order in which each atom is handwritten and input, as shown in FIG.

The attribute a _i is obtained as follows. As described above, the attribute a _i is the type symbol of the atom with the input number i. Here, the element symbol of each atom is recorded in the bond table 32, and the type symbol (atomic number) of the atom can be obtained from these element symbols. Therefore, by reading the element symbol from the bond table 32, the attribute a _i corresponding to this element symbol can be obtained. As a result, a ₁ , a
_{2, a 4 ~a 8 = 6} ( atomic number of carbon), a ₃ = 7 (nitrogen atomic number) is obtained, respectively.

Further, the attribute b _ij is obtained as follows.
As described above, the attribute b _ij is the number of bonds whose kind symbol is j among the bonds adjacent to the atom of the input number i.
The bond type of each atom is recorded in the bond table 32, and the attribute b _{ij is} read by reading the bond type from the bond table 32.
Can be obtained. As a result, b _1j = (3,0,0,
0), b _2j = (1,1,0,0), b _3j = (1,1,
0,0), b _4j = (2,0,0,0), b _5j = (3
0,0,0), b _6j = (2,0,0,0), b _7j =
(1,0,0,0) and _b8j = (1,0,0,0) are obtained, respectively.

Specifically, the attribute b _ij is shown in FIGS. 9 (a) and 9 (b).
It is obtained using the reference table 12 shown in FIG. The reference table 12 is an array D (x, which indicates the bond relationship of two atoms.
y) and is created based on the bond atom pair and bond species data in the bond table 32. Here, x is the number of the first atom in the bond atom pair data, and y is the number of the second atom. For example, at the coordinate position of x = 2 and y = 3, the bond type 2 is circled. Is shown. That is, the bond type j is written in the array element indicated by the bond atom pair, and the reference table 12 is created.

The extraction of the attribute b _ij using this reference table 12 is performed as follows. First, of the array elements of the reference table 12, an array element that satisfies x = 1 or y = 1 (array element indicated by diagonal lines in FIG. 9A) is searched, and the data written in this array element ( Bond type) j is extracted. As a result, D (1,2) = 1, D (1,6)
= 1 and D (1,8) = 1 are obtained. Since the data j of the three array elements thus obtained are all 1, b ₁₁ = 3
Becomes Further, since array elements having data j of 2 or more cannot be obtained, b _{12 to} b ₁₄ = 0.

Next, of the array elements of the reference table 12, an array element satisfying X = 2 or Y = 2 (array element shown by hatching in FIG. 9B) is searched and written in this array element. Extracted data. As a result, D
(1,2) = 1 and D (2,3) = 2 are obtained. The data j of the array element thus obtained is 1 and 2, and each is one, so b ₂₁ = b ₂₂ = 1. Further, since array elements having data j of 3 or more cannot be obtained, b ₂₃ = b ₂₄ = 0.

By performing the same processing for i = 3 to 8, the attribute b _ij (i = 1 to 8, j = 1 to 1) shown in FIG.
4) is obtained.

Further, the attribute d _ij is obtained as follows. As described above, the attribute d _ij is the number of routes that can pass from the atom with the input number i through the j bonds by the shortest path.
That is, to explain based on the molecular structure diagram of FIG. 8, the paths that can pass from the atom of input number 1 through one bond are (input number 1 to input number 2), (input number 1 to input number 6) ,
There are a total of three (input number 1 to input number 8). In addition, the route that can pass from the atom of input number 1 through two bonds is
There are a total of two (input number 1 to input number 2 to input number 3) and (input number 1 to input number 6 to input number 5).

Further, the routes that can be traveled from the atom of input number 1 through the three bonds by the shortest path are (input number 1 to input number 2 to input number 3 to input number 4), (input number 1 to input number 1 to
Input number 6 to input number 5 to input number 4), (input number 1
-Input number 6-input number 5-input number 7) are three in total. Furthermore, the number of paths that can be traveled from the atom of input number 1 through the four bonds by the shortest path is 0. Looking at FIG. 8, there is a route that can be traveled from the atom of input number 1 through four bonds. For example, the route is (input number 1-input number 2-input number 3--input number 4-input number 5). However, as a route from the input number 1 (starting point) to the input number 5 (arriving point), there is a route that can be looped by two couplings via the input number 6. Therefore, the illustrated route is not the shortest route from the departure point to the arrival point. As a result of the above processing,
We obtain d _1j = (3,2,3,0).

By the same processing, d _2j = (2,3,
2,2), d _3j = (2,2,4,0), d _4j = (2
3,2,2), d _5j = (3,2,3,0), d _6j =
(2, 4, 2, 0), d _7j = (1, 2, 2, 3), d _8j
= (1,2,2,3) are obtained respectively.

Specifically, the attribute d _ij is also obtained by referring to the reference table 12 like the attribute b _ij . The extraction of the attribute d _ij with reference to the reference table 12 is i = 1, i =
It is performed in the order of 2, ... First, the attribute d _1j (i = 1) is extracted.

The extraction of the attribute d _1j (i = 1) is performed by arranging the array elements satisfying X = 1 or Y = 1 among the array elements of the reference table 12 (array elements shown by hatching in FIG. 11A).
Is searched to extract the array element in which the data is written. Then, 1 is set as the number of connection paths in the extracted array element.
Write. As a result, D (1,2), D (1,6),
The coupling path number 1 is written in D (1,8) (FIG. 11).
In (a), the number of bond paths is shown by enclosing it in a triangle.

Next, the subscripts S = (1,2), (1,6), (1,8) of each array element in which the coupling path number 1 is written are extracted. From these subscripts S, S = 2, 6, 8 is obtained by excluding 1 used in the extraction process of the previous stage. Based on S = 2,6,8 thus obtained, array elements satisfying X = 2,6,8 or Y = 2,6,8 (array elements shaded in FIG. 11B) ) Is searched, and the array element in which the data is written and the bond path number is not written is extracted. Then, 2 is written in the extracted array element as the number of coupling paths. As a result, D (2,3), D (5,
The number of coupling paths of 2 is written in 6).

Further, the subscripts S = (2,3), (5,6) of each array element in which the number of coupling paths 2 is written are extracted.
From these subscripts S, 2 and 6 used in the previous extraction process are removed, and S = 3 and 5 are obtained. Thus obtained S =
Based on 3, 5, array elements satisfying X = 3,5 or Y = 3,5 (array elements shaded in FIG. 12C) are searched, and data is written and the number of coupling paths is Extract unwritten array elements. Then, 3 is written as the number of coupling paths in the extracted array element. As a result, the coupling path number 3 is written in D (3,4), D (4,5) and D (5,7).

By the above processing, the number of coupling paths is written in all array elements. As a result, there are three array elements with the number of bond paths of 1, three array elements with the number of bond paths of two, three array elements with the number of bond paths of three, and zero array elements with the number of bond paths of four.
As a result, d _1j = (3,2,3,0) is obtained.

Next, the attribute d _2j (i = 2) is extracted.
The extraction of the attribute d _2j (i = 2) is performed by selecting an array element satisfying X = 2 or Y = 2 among the array elements of the reference table 12 (see FIG. 1).
3 (a), the array elements indicated by diagonal lines are searched,
Extract the array element in which the data was written. Then, 1 is written as the number of coupling paths in the extracted array element. As a result, the bond path number 1 is written in D (1, 2) and D (2, 3) (in FIG. 13A, the bond path number is enclosed by a triangle).

Next, the subscripts S = (1, 2), (2, 3) of each array element in which the coupling path number 1 is written are extracted. Except for 2 used in the previous extraction process from these subscripts S,
We obtain S = 1,3. S = 1,3 obtained in this way
Based on the above, an array element satisfying X = 1,3 or Y = 1,3 (array element shown by hatching in FIG. 13B) is searched, and data is written and the number of coupling paths is written. Extract missing array elements. Then, 2 is written in the extracted array element as the number of coupling paths. As a result, D
The coupling path number 2 is written in (1,6), D (1,8), and D (3,4).

Further, the subscripts S = (1,6), (1,8), (3,4) of each array element in which the number of coupling paths 2 is written.
To extract. From these subscripts S, S = 4, 6, and 8 are obtained by excluding 1 and 3 used in the previous extraction process. Based on S = 4,6,8 thus obtained, X = 4,6,
8 or array elements satisfying Y = 4, 6, 8 (FIG. 14 (c))
The array element indicated by the diagonal line in (1) is searched, and the array element to which the data is written and the number of bond paths is not written is extracted. Then, 3 is written as the number of coupling paths in the extracted array element. As a result, D (4,5), D
The number of coupling paths of 3 is written in (5, 6).

Furthermore, the subscripts S = (4,5), (5, 6) of each array element in which the coupling path number 3 is written are extracted. From these subscripts S, which were used in the previous extraction process 4,
With the exception of 6, we get S = 5,5 (ie S = 5 is applied twice). Based on S = 5,5 thus obtained, array elements satisfying X = 5 or Y = 5 (see FIG.
(D), the array elements indicated by diagonal lines are searched to extract array elements for which data is written and the number of coupling paths is not written. Then, 4 is written as the number of coupling paths in the extracted array element. As a result, D (5,
The number 4 of coupling paths is written in 7).

By the above processing, the number of coupling paths is written in all array elements. As a result, there are two array elements with the number of bond paths of 1, three array elements with the number of bond paths of two, two array elements with the number of bond paths of three, and two array elements with the number of bond paths of four.
As a result, d _2j = (2,3,2,2) is obtained.

Others of the attribute _dij (i = 3 to 8)
10 is repeated until the number of bond paths is written in all the array elements, the d shown in FIG.
_ij (i = 1 to 8, j = 1 to 4) are obtained. By the process of S101 described above, 3,5-dimethyl-2,
Three kinds of attributes ( _ai , _bij , _dij ) were given to each atom constituting 3,4,5-tetrahydropyridine.

Next, the process of S102 is executed. As described above, in S102, the attribute (a _i ,
b _ij and d _ij ) are arranged to form a 9-digit number sequence, and the class numbers C _i ⁰ are given in ascending order of the number sequence to classify each atom into a plurality of classes. The class number C _i ⁰ given here is the 0th-order class number of the atom with the input number i.

The process of S102 will be described in detail. The number sequence of the atom with the input number 1 is "630003230", and the number sequence of the atom with the input number 2 is "61100232".
2 ". In the following order," 711002240 ",
"620002322", "630003230",
"620002420", "610001223",
It becomes "610001223".

[0060] As a result, the numeric string of the atoms of input numbers 7 and 8 is minimized, the class number C ₇ to these atoms ⁰ = C
₈ ⁰ = 1 is given. Hereafter, the class number C ₂ ⁰ = 2 is given to the atom with the input number 2 in ascending order of the number string,
Class number C ₄ ⁰ = 3 is given to the atoms input number 4. Also, the class number C ₆ ⁰ = 4 is given to the atom of input number 6, class number C ₁ to atom input number 1 and 5
⁰ = C ₅ ⁰ = ₅ is given. Further, the class number C ₃ ⁰ = 6 is given to the atoms input number 3 (Fig. 15 (a)
reference). In this way, each atom is classified into 6 classes, and the number of classes N ₀ is 6.

Next, the processing of S103 is executed to set the dimension n to 1.

Further, the processing of S104 is executed. As described above, the attribute V _ij ^{1 (n = 1)} is given to each atom in S104. Here, the attribute V _ij ⁿ is the number of atoms having a class number j that is bonded to the atom with the input number i. That is,
Explaining based on the molecular structure diagram of FIG. 15 (a), the input numbers of atoms bonded to the atom of input number 1 are 2, 6 and 8, and the class numbers of these atoms are C ₂ ⁰ = 2, C ₆ ⁰
= 4 and C ₈ ⁰ = 1. As a result, 1 is written in the attribute V _1j ¹ for j = 1, 2, 4, and V _1j ¹ = (1,1,
0,1,0,0) is obtained.

The input numbers of the atoms bonded to the input number 2 are 1, 3 and the class numbers of these atoms are C ₁ ⁰ = 5 and C ₃ ⁰ = 6. As a result, j = 5,6
1 is written in the attribute V _2j ¹ of V _2j ¹ = (0,0,
0,0,1,1) is obtained. By similarly processing the atoms of input numbers 3 to 8, V _3j ¹ = (0,
^{1, 1,} 0, 0, 0), V _4j ¹ = (0, 0, 0, 0,
1, 1), V _5j ¹ = (1, 0, 1, 1, 0, 0), V _6j
¹ = (0,0,0,0,2,0), V _7j ¹ = (0,0,
0,0,1,0), V _8j ¹ = (0,0,0,0,1,
0) is obtained respectively.

Specifically, the attribute V _ij ⁿ is as shown in FIG.
It is obtained using the reference table 12 shown in (b). The extraction of the attribute V _ij ¹ using this reference table 12 is i =
1, i = 2, ... First, the attribute V _1j ¹
(I = 1) is extracted. The extraction of the attribute V _1j ¹ (i = 1) is performed by _selecting x = 1 or y from the array elements of the reference table 12.
= 1 is searched for an array element (array element indicated by diagonal lines in FIG. 9A), and the subscript S = (1, 2), (1, 6), (1 , 8). Excluding i = 1 from these subscripts S, S = 2
Get 6,8. Substituting the value of S thus obtained for the 0th-order class number C _i ⁰ , C ₂ ⁰ = 2, C ₆ ⁰ =
4, to obtain a C ₈ ⁰ = 1. Then, by writing ¹ to the attribute V _1j ¹ for j = 1, 2, 4, V _1j ¹ = (1,1,
0,1,0,0) is required.

Next, the attribute V _2j ¹ (i = 2) is extracted. The extraction of the attribute V _2j ¹ (i = 2) is performed in the reference table 12
Of the array elements of X = 2 or Y = 2 (array elements shaded in FIG. 9B) are searched, and the subscript S = (1,
2), (2, 3) are extracted. From these subscripts S i =
Except for 2, we get S = 1,3. Substituting the value of S thus obtained for the 0th-order class number C _i ⁰ , C ₁ ⁰ =
5, to obtain a C ₃ ⁰ = 6. Then, the attribute V _{2j with} j = 5 and 6
By writing ¹ to 1, V _2j ¹ = (0, 0, 0,
0,1,1) is required.

By performing similar processing for i = 3 to 8, the attribute V _ij ¹ (i = 1 to 8, j = 1 shown in FIG. 16 is obtained.
~ 6) is obtained.

Next, the processing of S105 is executed. As described above, in S105, the attribute (C _i ^n-1 ,
V _ij ⁿ ) are arranged and the class numbers C _i ⁿ are given in the ascending order of the number sequence to classify each atom into a plurality of classes.

Specifically, the number sequence of the atom with the input number 1 is "5110100", and the number sequence of the atom with the input number 2 is "2000011". In the following order, "6011
000 ”,“ 3000011 ”,“ 5101100 ”,
"4000020", "1000010", "1000"
010 ".

As a result, the number string of the atoms with the input numbers 7 and 8 becomes the minimum, and the class number C ₇ ¹ = C is assigned to these atoms.
₈ ¹ = 1 is given. Hereafter, the class number C ₂ ¹ = 2 is given to the atom with the input number 2 in ascending order of the number string,
Class number C ₄ ¹ = 3 is given to the atoms input number 4. Further, the atom with the input number 6 has the class number C ₆ ¹ = 4.
Is given, the class number C ₅ ¹ =
5 is given. Furthermore, the atom with the input number 1 is given the class number C ₁ ¹ = 6, and the atom with the input number 3 is given the class number C ₃ ¹ = 7. In this way, each atom is classified into seven classes, and the number of classes N ₁ is 7.

Next, the process of S106 is executed to check whether the number N _{n of} classes is equal to N _(n-1) . If they are equal, the process is terminated. Also, it is checked whether the number N _{n of} classes is equal to the total number of atoms, and if they are equal, the process is terminated. Here, the number N ₁ of classes is 7 and the number N _{0 of} classes is 6, so N ₁ and N ₀ are not equal. Moreover, since the total number of atoms is 8, the number of classes N ₁ is not equal to the total number of atoms. Thus, since they are not equal to each other, the process of S107 is executed to set n to 2.

Further, returning to S104, the attribute V is assigned to each atom.
give _ij ² . As a result, as shown in FIG. 17, V _1j ²
= (1,1,0,1,0,0,0), V _2j ² = (0,
0,0,0,0,1,1), V _3j ² = (0,1,1,
0,0,0,0), V _4j ² = (0,0,0,0,1,
0,1), V _5j ² = (1,0,1,1,0,0,0),
V _6j ² = (0,0,0,0,1,1,0), V _7j ² =
(0,0,0,0,1,0,0), V _8j ² = (0,0,
0,0,0,1,0) is obtained.

Then, the process of S105 is executed to give each atom a class number C _i ² . As a result, FIG.
As shown in (c), C ₁ ² = 7, C ₂ ² = 3, C ₃ ²
= 8, C ₄ ² = 4, C ₅ ² = 6, C ₆ ² = 5, C ₇ ² =
2, C ₈ ² = 1 is obtained. In this way each atom is 8
As a result of being classified into one class, the number N ₂ of classes becomes 8.
Since the number of classes N ₂ = 8 is equal to the total number of atoms, S106
The process is ended by the judgment of.

Next, using the flowchart of FIG.
Canonical numbering routine 1 called in S20 of FIG.
The process of 02 will be described. Here, the canonicalization number is
It is the number of each atom that is uniquely determined by the structure of the compound. That is, the input number given by handwriting the molecular structure diagram is an arbitrary number that changes depending on the input order. On the other hand, the canonicalized data 34 must be unique data that depends only on the structure of the compound. Therefore, it is difficult to directly create the unique canonicalized data 34 from an arbitrary input number. Therefore, in the canonical data creation program 22,
The input number is once converted into a canonicalization number, and the canonicalized data 34 is created based on this unique canonicalized number, so that the canonicalized data 34 can be created smoothly.

The process of the canonicalization number assigning routine 102 is as follows.
First, 1 is given to the variable k (S201). Next, the final class number C _i ^f obtained by the constituent atom classification routine 101
And find the canonicalization number k (where k =
1) is given (S202). When there are a plurality of atoms having the largest class number, an arbitrary atom is selected from these atoms and a canonicalization number k is given to this atom. Then, after the canonicalization numbers are given to all the atoms, the processing is ended (S203).

Next, 1 is added to the variable k (S204),
Atoms whose canonical numbers have been determined (hereinafter referred to as “determined atoms”) have no canonical numbers yet (hereinafter,
The determined atoms to which the undetermined atoms are bound are extracted (S205). Then, it is determined whether or not there are a plurality of extracted fixed atoms (S206), and when there are a plurality of extracted fixed atoms, the fixed atom with the smallest canonicalization number is selected from among these fixed atoms ( S207). Then, the undecided atom having the largest class number C _i ^f is extracted from the undecided atoms bonded to the selected decided atom, and the canonicalization number of this undecided atom is set to k (S208). In addition, class number C
^If there are multiple decided atoms with the largest _i ^f , select among these decided atoms.

When it is determined in S206 that the number of the decided atom is one, the undecided atom having the largest class number C _i ^f is selected from the undecided atoms bonded to the decided atom, A canonicalization number k is given to this undetermined atom (S20
9). After the processing of S208 and S209 is completed, S2
The process is returned to 03, and the loop of S203 to S209 is repeated until the canonicalization numbers are given to all the atoms.

Next, with respect to the process of the canonicalization number assigning routine 102, a specific example using 3,5-dimethyl-2,3,4,5-tetrahydropyridine will be described. First,
1 is given to the variable k in the process of S201, and then the process of S202 is performed. In the process of S202, the atom with the input number 3 is C ₃ ^f = 8, which is the maximum, so the atom with the input number 3 is given the canonicalization number k = 1. Next, the variable k is set to 2 in the process of S204.
Then, the atom of input number 3 is extracted as a determined atom in the process of S205.

Since there is only one determined atom thus extracted, the processing of S209 is performed next. Since the undetermined atoms bound to the atom with the input number 3 are the atoms with the input numbers 2 and 4, the atom with the largest class number C _i ^f is selected from these atoms. That is, the class number of the atom of input number 2 is C ₂ ^f = 3, and the class number of the atom of input number 4 is C ₄ ^f
= 4. Therefore, select the atom with input number 4,
The canonicalization number k = 2 is given to this atom.

Next, returning to the processing of S204, the variable k is set to 3
Then, in the process of S205, the atoms having the input numbers 3 and 4 are extracted as the determined atoms. Since there are a plurality of determined atoms thus extracted, the process of S207 is performed next, and the atom having the smallest canonicalization number is selected from the extracted determined atoms.
That is, the atom of input number 3 has a canonicalization number of 1, and the atom of input number 4 has a canonicalization number of 2. Therefore, the atom with the input number 3 is selected. Then, the process of S208 is performed,
The canonicalization number k = 3 is given to the atom of input number 2 which is bonded to the atom of input number 3.

Further, returning to the processing of S204, the variable k is set to 4, and the atoms of the input numbers 2 and 4 are extracted as the determined atoms in the processing of S205. Since there are a plurality of determined atoms thus extracted, the process of S207 is performed next, and the atom having the smallest canonicalization number is selected from the extracted determined atoms. That is, the atom of input number 2 has a canonicalization number of 3, and the atom of input number 4 has a canonicalization number of 2. Therefore, the atom with the input number 4 is selected. Then, the processing of S208 is performed, and the canonicalization number k = 4 is given to the atom of the input number 5 which is bonded to the atom of the input number 4.

By repeating the same process, the atom having the input number 1 is given the canonicalization number 5, and the atom having the input number 6 is given the canonicalization number 6. The atom with the input number 7 has the canonicalization number 7 and the atom with the input number 8 has the canonicalization number 8
Are given respectively.

After that, the process of S203 is performed, and since the canonicalization numbers of all the atoms have been obtained at this stage, the process ends. As a result, the canonicalization number as shown in FIG. 19 is obtained.

The constituent atom classification routine 101 and the canonicalization number assignment routine 102 of the present invention assign numbers to atoms (S
102, S105, S202, S208, S209) and atom selection (S202, S207, S20).
8, S209). Whether or not the numbers are assigned in ascending or descending order, and in selecting atoms, the one with the larger or the smaller number is given priority, the canonical data uniquely corresponding to the structure of the compound. To the extent that the task of creating is achieved, it is a free choice of the programmer (including the choice of using negative mathematics).

Therefore, in the present invention, "the class number has the highest priority" has the above-mentioned meaning, and does not necessarily mean that the one with the larger number is selected. However, in the process of assigning the canonicalization number, it is preferable to employ an algorithm in which numbers are assigned alternately from right to left, left, ...

Further, in the second processing unit or the second step of the present invention, the method of assigning the canonicalization number to each atom in ascending order from 1 is the arithmetic progression ({p + q (n-
1) | n = 1,2,3 ... n _max }),
"1" means the reference mathematics (the first term p of the arithmetic progression), and does not have to be 1 itself. Also, the tolerance q
Does not have to be 1.

Furthermore, in the second processing unit or the second step of the present invention, the canonicalization number is changed from the first term p (usually the total number of atoms) in descending order (negative tolerance q) to each atom. Can also be given to. In this case, next to the canonicalization number n, among the atoms to which atoms that have already been given canonicalization numbers and which have not yet been given canonicalization numbers are bonded, Select the atom with the highest canonicalization number and select the atom that has the highest class number priority among the atoms that are bonded to that atom and have not yet been given a canonicalization number. Canonicalization number n + q (since q is negative, it is actually a value like n−1). In the case of the descending order, the definition of the attribute P _i may be further amended to be the canonicalization number of the atom that has the largest canonicalization number and is connected to the atom of the canonicalization number i.

Next, using the flowchart of FIG. 20,
Canonicalized data creation routine 103 called in S30
Will be described. In this process, first, as shown in FIG. 21, the input number is replaced with the canonicalized number, and the connection table 32 is rewritten (S301). And this combination table 3
Based on 2, three types of data (P _i ,
T _i , S _i ) is calculated (S302). Here, P _i is the canonicalization number of the atom having the smallest number by being bonded to the atom having the canonicalization number i (i> 1). Further, T _i is a kind symbol of a bond between an atom having a canonicalization number i (i> 1) and an atom having a canonicalization number P _i (in this example, one single bond, a double bond =, and a triple bond is #, Aroma bond is%, etc.). Further, S _i is a kind symbol (element symbol in this example) of the atom having the canonicalization number i (i> 0).

Specifically, first, the element symbol of the atom of canonicalization number 1 is examined with reference to the atom table 32a. As a result, S ₁ = “N” is obtained. Next, the atom bonded to the atom having the canonicalization number 2 is examined with reference to the atom pair table 32b. As a result, canonicalization numbers 1, 4
To obtain the atom. Since the smallest canonicalization number is 1 among these atoms, P ₂ = 1. And canonicalization number 2
Since the bond between the atom of and the atom of canonical number 1 is a single bond,
T ₂ = “−”. Further, S ₂ = “C” is obtained by referring to the atom table 32a.

Next, the atom bonded to the atom of canonicalization number 3 is examined with reference to the atom pair table 32b. As a result, atoms having canonicalization numbers 1 and 5 are obtained. Since the smallest canonicalization number is 1 among these atoms, P ₃ = 1. Then, since the bond between the atom having the canonicalization number 3 and the atom having the canonicalization number 1 is a double bond, T ₃ = “=”. Further, by referring to the atom table 32a, S ₃ =
"C" is obtained. By performing the same process below,
P ₄ = 2, P ₅ = 3, P ₆ = 4, P ₇ = 4, P ₈ = 5,
_{T 4 ~T 8 = "-"} , S 4 ~S 8 = "C" is obtained.

Next, a bond atom pair not referred to when T _i is obtained in the process of S302 is extracted (S303).
This processing is performed with reference to the atom pair table 32b. As a result, the bond atom pair of the atom of canonicalization number 5 and the atom of canonicalization number 6 is extracted. Then, three types of data (R ¹ _j , R ² _j , H _j ) are obtained for the extracted bond atom pairs (S304). Here, R ¹ _j and R ² _j are the canonicalization numbers of the two atoms forming the bond. Also, H _j
Is the type symbol of the bond (in this example the same as T _i is used). R ¹ _j and R ² _j are R ¹ _j > R
The relationship of ² _j is satisfied. Further, with another bond atom pair (R ¹ _k , R ² _k ), the relation of R ¹ _j ≦ R ¹ _k is satisfied, or the relation of R ¹ _j = R ¹ _k and R ² _j <R ² _k is satisfied. Shall be satisfied.

By the above processing, the canonicalized tree structure data shown in FIG. 22 can be created.

Next, the data obtained in the processing of S302 and S304 are arranged in a line to create canonicalized data (S305). That is, a delimiter F different from the atom type symbol and the bond type symbol is defined, and S302 and S3 are defined.
The respective data obtained in the processing of 04 are arranged as follows. S ₁ , P ₂ , T ₂ , S ₂ , P ₃ , T ₃ , S ₃ , P ₄ , T
₄ , S ₄ , ..., P _N , T _N , S _N , F, R ¹¹ ₁ , H
^{_{1, R 2 1, F,}} R 1 2, H 2, R 2 2, ......, F, R
¹ _M , H _M , R ² _M , F where N is the total number of atoms and M is the total number of extracted bond atom pairs in S304.

The data string thus obtained is the canonical data uniquely corresponding to the structure of the compound. Specifically, if the delimiter F is "/" and the obtained data are arranged in a predetermined order, "N1 = C1 = C2-C3-C4-C4-C5-C / 5" is obtained.
-6 / "is obtained. Then, the canonical data is written and stored in the compound information file 33 (S306). After that, the process ends.

The present invention is not limited to the above-described embodiment, but may be modified as follows, for example, without departing from the spirit of the present invention.

(1) In the above embodiment, the data string including the atom type symbol S _i is used as the canonical data, but the atom type symbol with the highest frequency of occurrence (usually carbon C) is used. It may be excluded from the data string. That is, "N1-1 = 2-3-4-4-5- / 5-6 /" is obtained by omitting the carbon C symbol from the above canonicalized data. By thus shortening the data string, the amount of data written in the compound information file 33 can be reduced.

(2) In the canonicalization number assignment routine 102, the following process may be added when a plurality of undecided atoms having the largest class number C _i ^f are selected in the process of S209.

(A) When the undecided atom with the largest class number C _i ^f does not belong to the cyclic structure part, an arbitrary undecided atom is selected from a plurality of undecided atoms, and the canonical atom of this undecided atom is selected. Let the chemical number be k. After that, the process returns to S203.

(B) When the undecided atom with the largest class number C _i ^f belongs to the cyclic structure portion, the undecided atom selected in S209 (hereinafter referred to as the candidate atom) is bonded to these candidate atoms. The following vector quantity is defined for each candidate atom in the structure in which the bond with the determined atom is broken.

M _ik : the minimum number of bonds between the candidate atom i and the atom with the canonicalization number k. Priorities are set in advance for this attribute, the atom i with the highest priority is selected, and that atom is selected. Let the canonicalization number of k be k. After that, the process returns to S203.

Here, the selection criteria of atoms based on the attribute values of the atoms will be exemplified. First, the amount of scalar depends on its size. Regarding the vector amount, when the elements of the two vectors i and k are V _ij and V _kj , the size of the smallest j among the elements of V _ij ≠ V _kj is used as the criterion for determining the priority order. By using such a criterion, the attribute b
^{_{ij, d ij, V ij n}} , it is possible to prioritize the m _ij. Further, when the priority order is determined by a plurality of attributes, it is preferable to set the priority order among the attributes in advance and give priority to the determination with the attribute having the higher priority order.

When the canonical data of the C ₆₀ molecule shown in FIG. 23 (a) was obtained by the above canonical data creating method according to the present invention, the canonical data uniquely identifying the structure of the C ₆₀ molecule was obtained. The normalized data (FIG. 23 (b)) could be obtained in only 1.5 seconds. On the other hand, when the canonical data of the C ₆₀ molecule was obtained by the Morgan algorithm which did not go through the process of classifying the atoms into equivalent atoms, the canonical data of the C ₆₀ molecule was obtained. 5 to get
It took 50 seconds. Therefore, if the canonical data creating means and method according to the present invention are adopted in the present invention, the speed of information processing according to the present invention is significantly improved.

[0102]

As described in detail above, in the canonical data generation device of the present invention, the unique data for each atom and the bond pair data between atoms accepted by the input means are canonical data. Given to the means of creation. Then, the canonicalized data creating means creates canonicalized data based on these data. Further, according to the canonical data creation method of the present invention, the canonical data is created based on the unique data for each atom constituting the compound and the bond pair data between the atoms.

As described above, the canonicalized data created by the canonicalized data creating apparatus and the canonicalized data creating method of the present invention is a very short string of characters, numbers and symbols, and can be created in a small storage area. Normalized data can be saved. Therefore, if the canonical data creating device and the canonical data creating method of the present invention are used in a compound / reaction database system, the amount of storage area used in the compound / reaction database system can be significantly reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a canonical data creation device.

FIG. 2A is a diagram showing the contents of an atom table of a binding table. (B) is a figure which shows the content of the atom pair table of a connection table.

FIG. 3 is a diagram showing data contents of a compound information file.

FIG. 4 is a schematic diagram showing an outline of the operation of the canonical data creation device.

FIG. 5 is a flowchart showing an outline of processing of a main routine.

FIG. 6 is a flowchart showing an outline of processing of a constituent atom classification routine.

FIG. 7A is a diagram showing the contents of an atom table of a binding table. (B) is a figure which shows the content of the atom pair table of a connection table.

FIG. 8 is a diagram showing a relationship between each atom constituting 3,5-dimethyl-2,3,4,5-tetrahydropyridine and an input number.

9A and 9B are diagrams showing data contents of a reference table.

FIG. 10 is a diagram showing three types of attributes (a _i , b _ij , d _ij ) given to each atom constituting 3,5-dimethyl-2,3,4,5-tetrahydropyridine.

11A and 11B are views showing data contents of a reference table.

FIG. 12C is a diagram showing data contents of a reference table.

13A and 13B are diagrams showing data contents of a reference table.

14 (c) and (d) are views showing data contents of a reference table.

15 (a), (b) and (c) are 3,5-dimethyl-.
It is a figure which shows the relationship between each atom which comprises 2,3,4,5-tetrahydropyridine, and a class number.

FIG. 16 is an attribute V assigned to each atom constituting 3,5-dimethyl-2,3,4,5-tetrahydropyridine.
It is a figure which shows _ij ¹ .

FIG. 17 is an attribute V assigned to each atom constituting 3,5-dimethyl-2,3,4,5-tetrahydropyridine.
It is a figure which shows _ij ² .

FIG. 18 is a flowchart showing an outline of processing of a canonicalization number assignment routine.

FIG. 19 is a diagram showing a relationship between each atom constituting 3,5-dimethyl-2,3,4,5-tetrahydropyridine and a canonicalization number.

FIG. 20 is a flowchart showing an outline of processing of a canonicalized data creation routine.

FIG. 21 (a) is a diagram showing the contents of an atom table of a binding table. (B) is a figure which shows the content of the atom pair table of a connection table.

FIG. 22 is a diagram showing the data content of canonicalized tree structure data.

FIG. 23 (a) is a diagram showing a molecular structure of a C ₆₀ molecule. (B) is a diagram showing a canonical data C ₆₀ molecules.

[Explanation of symbols]

1 ... Canonicalized data creating device, 10 ... Image memory, 11 ...
Working memory, 20 ... Main storage device, 21 ... Operating system, 22 ... Canonicalized data creation program, 3
0 ... Hard disk device, 31 ... Binding table file, 32
... bond table, 33 ... compound information file, 34 ... canonical data, 40 ... display, 50 ... mouse, 60 ... keyboard, 70 ... printer, 80 ... CPU, 100 ... main routine, 101 ... constituent atom classification routine (first 1 processing unit), 102 ... Canonical number assigning routine (second processing unit), 103 ... Canonical data creation routine (3rd processing unit), A ... Input means, B ... Canonical data creation means.

Claims (4)

[Claims] 1. A chemical structure of a compound is uniquely defined based on input means for accepting input of unique data for each atom constituting a compound and bond pair data between atoms, and each data accepted by the input means. A canonical data creating device having canonical data creating means for creating canonically specified canonical data, wherein the canonical data creating means is a class for each atom of different equivalent atoms. A first processing unit that gives each atom a different class number for each class, and a unique structure of the compound based on the class number given to each atom in the first processing unit. A second processing unit that gives a corresponding canonicalization number to each atom, and a third processing unit that creates the canonicalization data based on the canonicalization number given to each atom by the second processing unit. And a processing unit. Canonical data creating apparatus. 2. The first processing unit gives each atom three kinds of attributes (a _i , b _ij , d _ij ),
Utilizing the fact that it is possible to determine that atoms having different even one of these attributes are not equivalent, each atom is given a different class number for each equivalent atom, and the three types of attributes (a _i , b _ij , d _ij ), a _i is the symbol of the atom of the input number i, and b _ij is the input number i
Is the number of bonds whose type symbol is j among the bonds adjacent to the atom, and d _ij is the number of routes that can pass from the atom with the input number i through j bonds by the shortest path, The second processing unit assigns the canonicalization number 1 to the atom having the highest priority of the class number in the process of assigning the canonicalization number to each atom in ascending order from 1, and thereafter, up to the canonicalization number n. Is given, the canonicalization number is the smallest among the atoms to which atoms that have already been given canonicalization numbers and which have not yet been given canonicalization numbers are bound. Select an atom, and give the canonicalization number n + 1 to the atom that has the highest priority of the class number among the atoms that are bonded to the atom and have not yet been given the canonicalization number. The third processing unit includes three types of attributes (P _i , T _i , S _i ),
The canonicalized data is created by arranging these attributes in a line. Among the three types of attributes (P _i , T _i , S _i ), P _i is an atom with a canonicalization number i. Is the canonicalization number of the atom that is bonded and has the smallest canonicalization number, T _i is the type symbol of the bond between the atom of canonicalization number i and the atom of canonicalization number P _i , and S _i is The canonical data creation device according to claim 1, wherein the canonical code is the atom type symbol of the canonicalization number i. 3. A canonical data creation method for creating canonical data capable of uniquely identifying the chemical structure of the compound based on the unique data for each atom constituting the compound and the bond pair data between the atoms. Where each atom is classified into a different class for each equivalent atom, and a first step of giving each atom a different class number for each class, and a class number given to each atom in the first step The second step of giving each atom a canonicalization number uniquely corresponding to the structure of the compound, and the canonical number given to each atom in the second step, And a third step of creating canonical data. 4. The first step is to give each atom three kinds of attributes (a _i , b _ij , d _ij ),
Utilizing the fact that it is possible to determine that atoms having different even one of these attributes are not equivalent, each atom is given a different class number for each equivalent atom, and the three types of attributes (a _i , b _ij , d _ij ), a _i is the symbol of the atom of the input number i, and b _ij is the input number i
Is the number of bonds whose type symbol is j among the bonds adjacent to the atom, and d _ij is the number of routes that can pass from the atom with the input number i through j bonds by the shortest path, In the second step, in the process of giving canonical numbers to atoms in ascending order from 1, the atom having the highest priority of the class number is given the canonical number 1, and thereafter the canonical numbers n are assigned. When attached, the atom that has already been given a canonicalization number and has the smallest canonicalization number among the atoms to which atoms that have not yet been given a canonicalization number are bound. And assigns the canonicalization number n + 1 to the atom that has the highest priority of the class number among the atoms that are bonded to the atom and have not yet been given the canonicalization number. In the third step, each atom has three kinds of attributes (P _i , T _i , S _i ),
The canonicalized data is created by arranging these attributes in a line. Among the three types of attributes (P _i , T _i , S _i ), P _i is an atom with a canonicalization number i. Is the canonicalization number of the atom that is bonded and has the smallest canonicalization number, T _i is the type symbol of the bond between the atom of canonicalization number i and the atom of canonicalization number P _i , and S _i is The canonical data creation method according to claim 3, wherein the atom type symbol of the canonicalization number i is used.