A data preprocessing system for cleaning small molecule compounds and a method thereof

Abstract

The present application provides a kind of data preprocessing method for cleaning small molecule compounds, the data preprocessing method includes: S1 text preprocessing step, the S1 text preprocessing step includes: the original SMILES text of the small molecule compound is preprocessed into the standardized SMILES text of the small molecule compound;S2 chemical graph formatting step, the S2 chemical graph formatting step includes: each text element of the standardized SMILES text of the small molecule compound of the S1 is formatted and split, obtains the chemical graph information of the small molecule compound.The present application also provides a kind of data preprocessing system for cleaning small molecule compounds.The present application realizes the cleaning, deduplication and standardization of global data set, provides a kind of efficient, fast, accurate end-to-end small molecule compound cleaning integration method.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of medicine and artificial intelligence, and particularly relates to a data preprocessing system for cleaning small molecule compounds and a method thereof. BACKGROUND

[0002] The traditional method is based on the method of chemical informatics to realize compound standardization to obtain a cleaning and integration method of small molecule compounds.

[0003] However, with the advent of the big data era, high efficiency, precision and fast calculation speed are required. The algorithm efficiency of the traditional method based on chemical informatics is low, which cannot meet the needs of the big data era, and the data standards of various open source algorithms are not unified.

[0004] Specifically, there are many sources of SMILES compound information (such as open source databases such as Chembl and PubChem), lack of unified standardization operation, and cannot well distinguish clean and unclean data for duplicate checking.

[0005] In addition, at present, based on rules, there are some cleaning and deduplication process methods. In the process, only the database is built, and there is no actual application of downstream (such as machine learning and deep learning). Non-standard or repeated structures can still be encountered when using the method.

[0006] In addition, the mathematical graph of SMILES converted for graph neural network currently lacks standardization, and the algorithms called from individual open source frameworks lack unified standards.

[0007] Based on the above, the technical solutions for solving the above technical problems are provided. SUMMARY

[0008] The first object of the application is to provide an efficient, fast and accurate end-to-end small molecule compound cleaning and integration method.

[0009] The second object of the application is to obtain an efficient, fast and accurate end-to-end small molecule compound cleaning and integration system.

[0010] The first aspect of the application provides a data preprocessing method for cleaning small molecule compounds, which comprises:

[0011] S1, a text preprocessing step, comprising: according to a predetermined text processing rule, preprocessing the original SMILES text of the small molecule compound into a standardized SMILES text of the small molecule compound;

[0012] S2 chemical graph formatting step, the S2 chemical graph formatting step comprising: according to a predetermined text processing rule, format splitting each text element of the standardized SMILES text of the small molecule compound of the S1, obtaining a digital graph structure of chemical information of the small molecule compound.

[0013] In one preferred embodiment of the present application, further comprising an S3 step, wherein the digital graph structure of chemical information of the small molecule compound of the S2 is used for construction of an artificial intelligence model.

[0014] In one preferred embodiment of the present application,

[0015] In the S1 text preprocessing step, when the original SMILES text of the small molecule compound is preprocessed into the standardized SMILES text of the small molecule compound, the predetermined text processing rule comprises:

[0016] S1-1 step: optional structure standardization, wherein the data of the small molecule compound is processed into an original SMILES text;

[0017] S1-2 step: if the original SMILES text includes heavy metal components and organic compound components, removing the heavy metal components in the original SMILES text and retaining the organic compound components;

[0018] S1-3 step: if the original SMILES text includes multimer components, removing the multimer components in the original SMILES text and retaining the longest component;

[0019] S1-4 step: if the original SMILES text includes charges, adding or subtracting hydrogen atoms in the original SMILES text to achieve charge removal;

[0020] S1-5 step: removing special SMILES text information;

[0021] S1-6 step: outputting a standardized sequence to obtain the standardized SMILES text of the small molecule compound.

[0022] In one preferred embodiment of the present application, when each text element of the standardized SMILES text of the small molecule compound of the S1 is format split in the S2 chemical graph formatting step, the predetermined text processing rule comprises:

[0023] S2-1 step: splitting the standardized SMILES text of the small molecule compound of the S1 into each core text element to obtain text elements of the small molecule compound;

[0024] S2-2 step: text processing identification is performed on the property of the text element of the small molecule compound in the S2-1 step, simplified chemical information is identified and completed, and a chemical information graph of the small molecule compound is obtained;

[0025] S2-3 step: according to the chemical information graph of the small molecule compound in the S2-2 step, a coordinate system with atomic elements as nodes is established, and a digital coordinate system of the chemical information graph of the small molecule compound is constructed;

[0026] S2-4 step: according to the digital coordinate system of the chemical information graph of the small molecule compound in the S2-3 step, the element properties of the nodes and edges are added, and a digital graph structure of the chemical information of the small molecule compound is obtained.

[0027] In one preferred embodiment of the present application, S2-5 step: if necessary, the hydrogen atom information of the digital graph structure of the chemical information is completed.

[0028] In one specific embodiment, it further includes S2-6 step: the digital graph structure of the chemical information of the small molecule compound is completely output.

[0029] The second aspect of the present application provides a data preprocessing system for cleaning small molecule compounds, which is suitable for the data preprocessing method according to any one of claims 1-5, and is characterized in that it comprises:

[0030] S1 text preprocessing unit, the S1 text preprocessing unit is arranged to include: according to a predetermined text processing rule, the original SMILES data of the small molecule compound is preprocessed into the standardized SMILES text of the small molecule compound;

[0031] S2 chemical graph formatting unit, the S2 chemical graph formatting unit is arranged to include: according to a predetermined text processing rule, each text element of the standardized SMILES text of the small molecule compound in the S1 is format split to obtain a digital graph structure of the chemical information of the small molecule compound.

[0032] In one preferred embodiment of the present application, it further includes S3 unit, which is arranged to use the digital graph structure of the chemical information of the small molecule compound in the S2 for constructing an artificial intelligence model.

[0033] In one preferred embodiment of the present application, when the original SMILES text of the small molecule compound is preprocessed into the standardized SMILES text of the small molecule compound in the S1 text preprocessing unit, the predetermined text processing rule includes:

[0034] The S1-1 unit is configured to: optional structure standardization, wherein data of the small molecule compound is processed into original SMILES text;

[0035] The S1-2 unit is configured to: if the original SMILES text includes heavy metal components and organic compound components, removing the heavy metal components and retaining the organic compound components in the original SMILES text;

[0036] The S1-3 unit is configured to: if the original SMILES text includes multimer components, removing the multimer components and retaining the longest component in the original SMILES text;

[0037] The S1-4 unit is configured to: if the original SMILES text includes charges, adding or subtracting hydrogen atoms in the original SMILES text to achieve de-charging;

[0038] The S1-5 unit is configured to: removing special SMILES text information;

[0039] The S1-6 unit is configured to: outputting the standardized sequence to obtain the standardized SMILES text of the small molecule compound.

[0040] In a preferred embodiment of the present application, when the S2 chemical graph formatting unit formats each text element of the standardized SMILES text of the small molecule compound of S1, the predetermined text processing rule includes:

[0041] The S2-1 unit is configured to: splitting the standardized SMILES text of the small molecule compound of S1 into each core text element to obtain the text element of the small molecule compound;

[0042] The S2-2 unit is configured to: performing text processing identification on the nature of the text element of the small molecule compound of the S2-1 unit, identifying and completing the simplified chemical information to obtain the chemical information graph of the small molecule compound;

[0043] The S2-3 unit is configured to: establishing a coordinate system with atomic elements as nodes according to the chemical information graph of the small molecule compound of the S2-2 unit, and constructing a digital coordinate system of the chemical information graph of the small molecule compound;

[0044] The S2-4 unit is configured to: adding element attributes of nodes and edges according to the digital coordinate system of the chemical information graph of the small molecule compound of the S2-3 unit to obtain the digital graph structure of the chemical information of the small molecule compound.

[0045] In one preferred embodiment of the present application, the S2-5 unit is configured to, if necessary, complete the hydrogen atom information of the digital graph structure of the chemical information.

[0046] In one specific embodiment, the S2-6 unit is further included: the digital graph structure of the chemical information of the small molecule compound is output completely.

[0047] The third aspect of the present application provides an electronic device, comprising: a memory, a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to realize the data preprocessing method for cleaning small molecule compounds according to any one of the present application.

[0048] The present application can bring at least one of the following beneficial effects:

[0049] The method of the present application realizes a new method based on the combination of big data and natural language processing technology and part of chemical informatics, which can reduce the computing cost, and finally realize more accurate data preprocessing and more convenient downstream use. BRIEF DESCRIPTION OF DRAWINGS

[0050] The above-mentioned features, technical characteristics, advantages and their implementation methods will be further described in the following preferred embodiments in a clear and understandable manner, combined with the accompanying drawings.

[0051] Figure 1 is a data processing method flowchart (two independent but associated parts) in the present application;

[0052] Figure 2 is a workflow diagram in the present application;

[0053] Figure 3 is a schematic diagram of data variable conversion in the present application. DETAILED DESCRIPTION

[0054] The various aspects of the present application will be further described in detail below.

[0055] Unless otherwise defined or specified, all professional and scientific terms used herein have the same meaning as familiar to those skilled in the art. In addition, any method and material similar or equivalent to those described can be applied to the method of the present application.

[0056] The following explains the terms.

[0057] Unless otherwise expressly specified and limited, the "or" in the present application includes the "and" relationship. The "and" corresponds to the Boolean logical operator "AND", the "or" corresponds to the Boolean logical operator "OR", and the "AND" is a subset of "OR".

[0058] It is to be understood that, although the terms "first", "second", etc. can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. Thus, a first element could be termed a second element without departing from the teachings of the present inventive concept.

[0059] In the present invention, the terms "comprising", "containing" or "including" mean that various elements can be used in the mixture or composition of the present invention together. Therefore, the terms "consisting essentially of and "consisting of are included in the terms "comprising", "containing" or "including".

[0060] Unless otherwise defined and limited, the terms "connected", "coupled", "connected", of the present invention should be interpreted broadly, for example, can be fixedly connected, can be connected through an intermediate medium, can be internal connection of two elements or interaction relationship of two elements. For those skilled in the art, the specific meaning of the above terms in this application can be understood according to the specific circumstances.

[0061] For example, if an element (or component) is referred to as being "on", "coupled with" or "connected with" another element, it can be directly formed on, coupled with or connected with the other element, or one or more intermediate elements can be provided between them. In contrast, if the expressions "directly on", "directly coupled with" and "directly connected with" are used herein, it means that there is no intermediate element. Other words used to describe the relationship between elements should be similarly interpreted, such as "between" and "directly between", "attached" and "directly attached", "adjacent" and "directly adjacent", etc.

[0062] In addition, it should be noted that the words "front", "back", "left", "right", "up" and "down" used in the following description refer to the directions in the drawings. The words "inner" and "outer" are used to refer to the direction towards or away from the geometric center of a particular component. It can be understood that these terms are used herein to describe the relationship of one element, layer or region with respect to another element, layer or region as shown in the drawings. In addition to the orientation described in the drawings, these terms should also include other orientations of the device.

[0063] Other aspects of the present invention will be apparent to those skilled in the art from the disclosure herein.

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, specific implementations of the present application will be described below with reference to the drawings. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained from these drawings without creative labor, and other embodiments can also be obtained.

[0065] It should also be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present application in a schematic manner, and only show the components related to the present application in the diagrams, not the number, shape and size of the components when actually implemented. The actual implementation of each component may be a random change in shape, number and proportion, and the layout pattern of the components may also be more complex. For example, the thickness of the elements in the drawings can be exaggerated for clarity.

[0066] Embodiments

[0067] In the present application, the inventors have conducted extensive and in-depth experiments and found that, based on the demand reference of artificial intelligence assisted drug design, a new process method is constructed to perform end-to-end small molecule compound SMILES sequence cleaning, deduplication, and mathematical graph standardization, to provide more accurate and efficient data preprocessing methods for downstream artificial intelligence models.

[0068] To achieve the above purpose, a first aspect of the present application provides a data preprocessing method for cleaning small molecule compounds, the data preprocessing method comprising:

[0069] S1, a text preprocessing step, comprising: according to a predetermined text processing rule, preprocessing the original SMILES text of the small molecule compound into a standardized SMILES text of the small molecule compound;

[0070] S2, a chemical graph formatting step, comprising: according to a predetermined text processing rule, format splitting each text element of the standardized SMILES text of the small molecule compound of S1 to obtain a digital graph structure of chemical information of the small molecule compound.

[0071] In a preferred embodiment of the present application, S3 is further included, wherein the digital graph structure of chemical information of the small molecule compound of S2 is used for constructing an artificial intelligence model.

[0072] Exemplarily but not limitatively, the final display result is in Python list format, which can be saved in Python pickle format for downstream deep learning training.

[0073] In one preferred embodiment of the present application, in the step of S1 text preprocessing, when the original SMILES text of the small molecule compound is preprocessed into the standardized SMILES text of the small molecule compound, the predetermined text processing rule comprises:

[0074] S1-1 step: optional structure standardization, wherein the data of the small molecule compound is processed into the original SMILES text;

[0075] S1-2 step: if the original SMILES text comprises heavy metal components and organic compound components, the heavy metal components are removed from the original SMILES text, and the organic compound components are retained;

[0076] S1-3 step: if the original SMILES text comprises multimer components, the multimer components are removed from the original SMILES text, and the longest component is retained;

[0077] S1-4 step: if the original SMILES text comprises charges, hydrogen atoms are added or subtracted from the original SMILES text to achieve charge removal;

[0078] S1-5 step: remove special SMILES text information;

[0079] S1-6 step: output the standardized sequence to obtain the standardized SMILES text of the small molecule compound.

[0080] More specifically, the following describes each part of the S1 step in combination with the accompanying drawings. The following description is exemplary and not limiting, and therefore those skilled in the art can arbitrarily combine each of the following steps, which all fall within the scope of the present application.

[0081] S1-1 step: optional structure standardization, wherein the data of the small molecule compound is processed into the original SMILES text;

[0082] In one specific embodiment, the original data of the small molecule compound is input, then subjected to chemical structure standardization processing, and finally processed into the original SMILES text (usually in text format).

[0083] Specifically, when chemical structure standardization is performed, the predetermined text processing rule (S1-1 part) is used for text collation.

[0084] Specifically, the predetermined text processing rule (S1-1 part) includes but is not limited to:

[0085] The text of the original data is modified into S1-1-1 standard text by number rules.

[0086] S1-2 step: if the original SMILES text includes heavy metal components and organic compound components, then remove the heavy metal components from the original SMILES text, and keep the organic compound components.

[0087] The process of reorganization will split the SMILES sequence components using text rules and then calculate the longest chain. The S1-1-3 standard text is reorganized by the longest chain of the SMILES sequence. Exemplarily but not limitatively, the S1-1-3 standard text is, for example, the SMILES sequence shown in Figure 3

[0088] S1-2 step: if the original SMILES text includes heavy metal components and organic compound components, then remove the heavy metal components from the original SMILES text, and keep the organic compound components.

[0089] Specifically, the S1-2 step is used to remove the heavy metal part in the SMILES text.

[0090] More specifically, this part is operated using text processing rules (S1-2 part). Among them, the heavy metal that needs to be removed is defined as the atom that does not exist covalent bond.

[0091] Exemplarily but not limitatively, the SMILES representation text of part of the heavy metal atom is: “[Li]”, “[Ca]”, “[Na+]” and the like SMILES text elements of atoms.

[0092] S1-3 step: if the original SMILES text includes multimer components, then remove the multimer components from the original SMILES text, and keep the longest component;

[0093] Specifically, the purpose of the S1-3 step is to remove the multimer in the SMILES text and keep the longest sequence.

[0094] More specifically, in the text, it will be divided according to the “.” separator.

[0095] S1-4 step: if the original SMILES text includes charge, then add or subtract hydrogen atoms in the original SMILES text to achieve de-charging;

[0096] Specifically, the purpose of the S1-4 step is to zero the charge component in the SMILES text. More specifically, this process can be understood as a text processing rule (S1-4 part).

[0097] More specifically, the special components in the covalent bond are modified. For example: “[O-]” is modified to “O”.

[0098] S1-5 step: remove special SMILES text information; ​

[0099] The purpose of this step is to remove special marks or special atoms in the SMILES text.

[0100] More specifically, this process can be understood as a text processing rule (S1-5 part).

[0101] Exemplarily but not limitatively, the modified text is, for example: ”[1*]”, “*”, ”[2H]”.

[0102] S1-6 step: output the normalized sequence to obtain the normalized SMILES text of the small molecule compound.

[0103] In a preferred embodiment of the present application, when the S2 chemical graph formatting step formats and splits each text element of the normalized SMILES text of the small molecule compound of S1, the predetermined text processing rule comprises:

[0104] S2-1 step: splitting the normalized SMILES text of the small molecule compound of S1 into each core text element to obtain the text element of the small molecule compound.

[0105] S2-2 step: text processing identification of the nature of the text element of the small molecule compound of the S2-1 step, identifying and completing the simplified chemical information to obtain the chemical information graph of the small molecule compound.

[0106] S2-3 step: according to the chemical information graph of the small molecule compound of the S2-2 step, establishing a coordinate system with atomic elements as nodes, and constructing a digital coordinate system of the chemical information graph of the small molecule compound.

[0107] S2-4 step: according to the digital coordinate system of the chemical information graph of the small molecule compound of the S2-3 step, adding the element properties of the nodes and edges to obtain the digital graph structure of the chemical information of the small molecule compound.

[0108] More specifically, the S2 step is explained as follows in combination with the drawings. The following explanation is exemplary but not limitative, and therefore those skilled in the art can arbitrarily combine each of the following steps, which all belong to the scope of the present application.

[0109] S2-1 step: splitting the normalized SMILES text of the small molecule compound of S1 into each core text element to obtain the text element of the small molecule compound.

[0110] The purpose of S2-1 step is to split the normalized SMILES sequence into each key text element (tokenization).

[0111] In particular, the text elements include a chemical bond mark, an atom mark, a chiral mark, an organic compound ring mark, or a combination thereof.

[0112] S2-2 step: performing text processing identification on the properties of the text elements of the small molecule compound in the S2-1 step, identifying and completing the simplified chemical information to obtain a chemical information graph of the small molecule compound.

[0113] The purpose of S2-2 is to complete the missing elements by a text processing rule algorithm. SMILES usually hides part of the information, and this step will restore the hidden information to the default information.

[0114] By way of example but not limitation, the ‘—’ element is completed as a mark element of a single bond compound covalent bond.

[0115] S2-3 step: establishing a coordinate system with atomic elements as nodes according to the chemical information graph of the small molecule compound in the S2-2 step, and constructing a digital coordinate system of the chemical information graph of the small molecule compound;

[0116] The purpose of the S2-3 step is to separately mark the nodes and edges with coordinates by splitting the order of the elements.

[0117] By way of example but not limitation, the node element is an atom, and the element of the edge is a bond. The coordinate marking of 0, …, N is performed in the order of the input standardized SMILES sequence.

[0118] S2-4 step: adding element attributes of nodes and edges according to the digital coordinate system of the chemical information graph of the small molecule compound in the S2-3 step, to obtain a digital graph structure of the chemical information of the small molecule compound.

[0119] The purpose of the S2-4 step is to integrate the information of nodes and edges as an initial mathematical graph by the coordinate system of step S2-3, and to construct a graph.

[0120] By way of example but not limitation, the construction of the graph will take the coordinates of each node as a node list data structure. At the same time, the bond information of the compound completed by step 2 (-, =, #, : and other elements) is matched to create the edges of the mathematical graph.

[0121] Optionally, other marked elements can also be used to specially mark nodes or edges as attributes in the mathematical graph.

[0122] Exemplarily but not limitatively, the special marks include but are not limited to: chiral atom marks (@, @ @, / , \), atomic number (by regular query), single double triple bond (see step 4 information), aromaticity (identified by rules), whether in the ring of the compound (identified by regular expression), and other attributes.

[0123] In one preferred embodiment of the present application, it further comprises S2-5 step: if necessary, the hydrogen atom information of the digital graph structure of the chemical information is completed.

[0124] Exemplarily but not limitatively, hydrogen atoms can be added to the mathematical graph. The completion method is completed based on the rules of atomic attributes, and the related attribute information is completed.

[0125] In one specific embodiment, it further comprises S2-6 step: the digital graph structure of the chemical information of the small molecule compound is completely output.

[0126] Exemplarily but not limitatively, the output is as shown in the chemical structure diagram. Figure 3

[0127] Specifically, see Figure 1 , which shows one preferred specific embodiment of the present application.

[0128] The concept of the preferred embodiment is as follows: the method is divided into two parts: text preprocessing and text to mathematical graph.

[0129] Text preprocessing includes:

[0130] 1. Structure standardization

[0131] 2. Remove metal components from structure text, retain organic compound components

[0132] 3. Remove multimers from structure text, retain the longest component

[0133] 4. Add and subtract hydrogen atoms to the structure text to realize charge removal

[0134] 5. Remove special SMILES text information

[0135] 6. Output the standardized sequence

[0136] Text to graph includes:

[0137] 1. Split the SMILES sequence into core elements

[0138] 2. Text processing identifies the nature of the text element, identifies and completes the simplified chemical information

[0139] 3. Create an atomic element node coordinate system to build a mathematical graph ​

[0140] 4. Add node and edge element properties

[0141] 5. Complete hydrogen atom information

[0142] Output the complete compound graph.

[0143] More specifically, the S1 flow is the upper half of the flow, and the data output by this flow can be saved or converted. The following is a detailed description:

[0144] 1. The original SMILES data. The data format is text. The SMILES sequence is a text representation of a small molecule compound, such as Figure 3 as shown in the case.

[0145] 2. Chemical structure standardization, using text processing rules to organize the text. Through numerical rules, the original text is modified into the standard text in the method. At the same time, using regular method to split all SMILES main components to reorganize SMILES text into standard text. The reorganization process will use text rules to split SMILES sequence components and then calculate the longest chain. Through the longest chain, the SMILES sequence text is reorganized.

[0146] 3. Remove the multimer in the SMILES text, and keep the longest sequence. The text will be divided according to the “.” separator.

[0147] 4. Remove the heavy metal part in the SMILES text. This part is operated by using text processing rules. Heavy metals are defined as atoms that do not exist in covalent bonds. In the example, the SMILES text elements of some heavy metal atoms are: “[Li]”, “[Ca]”, “[Na+]”, etc.

[0148] 5. Zero the charge component in the SMILES text. This method uses text processing rules. The special component in the covalent bond is modified by rules. For example: “[O-]” is modified to “O”

[0149] 6. Remove special markers and special atoms in the SMILES text, which also uses text processing rules. Modified text examples: “[1*]”, “*”, “[2H]”.

[0150] 7. Output the standardized SMILES sequence

[0151] Figure 1 The S2 flow is the lower half of the flow, and the input is the SMILES sequence, and the output is the mathematical graph format variable:

[0152] 1. Tokenization of the standardized SMILES sequence into each token. Tokens include: bond token, atom token, chirality token, ring token of organic compounds

[0153] 2. Missing token completion by text processing rule algorithm. SMILES usually hides some information, this step will restore the hidden information to the default information. For example: complete the ‘—’ token as the token of single bond covalent bond of compound.

[0154] 3. Coordinate tokenization of nodes and edges by the order of tokenization. In the example, node token is atom, and edge token is bond. The coordinate tokenization of 0, …, N is performed by the order of the input standardized SMILES sequence.

[0155] 4. Integration of node and edge information by the coordinate system of step 3 as the initial mathematical graph.

[0156] The construction of the graph will use the coordinate of each node as a node list data structure. At the same time, the left and right node matching of the compound bond information completed in step 2 (-, =, #, : and other tokens) is used to create the edge of the mathematical graph.

[0157] 5. Special tokenization of nodes or edges by other tokenized tokens as attributes in the mathematical graph. In the example, special tokens include but are not limited to: chirality atom token (@, @@, / , \), atomic number (by rule query), single, double and triple bond (see step 4 information), aromaticity (identified by rule), whether in the ring of the compound (identified by regular expression), etc.

[0158] 6. (Optional) Hydrogen atom information completion. In the example, hydrogen atoms can be added to the mathematical graph. The completion method is based on the rules of atomic attributes, and the related attribute information is completed.

[0159] 7. Output of the chemical structure graph, which is the final display of Figure 3 .

[0160] In one preferred embodiment of the present application, it further comprises an S3 step, wherein the digital graph structure of the chemical information of the small molecule compound of S2 is used for construction of an artificial intelligence model.

[0161] Referring to Figure 2 , the workflow of the S3 step is shown. The S3 step comprises:

[0162] S3-1 obtains an original drug data set;

[0163] S3-2 data preprocessing (SMILES cleaning);

[0164] S3-3 Workflow of machine learning, deep learning

[0165] S3-4: Artificial intelligence model.

[0166] The following S3 steps are exemplarily illustrated:

[0167] Example 1:

[0168] 1. Input SMILES sequence dataset

[0169] 2. Respectively, each sequence is subjected to the S1 process shown in the figure. According to the parameters, it is determined whether to standardize the optional step. Figure 1

[0170] 3. Parallel computing is arranged through machine resource allocation to improve computing efficiency.

[0171] 4. The cleaned SMILES dataset is output and stored for other purposes. The storage method is SQL type database or table format such as csv, excel, etc.

[0172] Example 2:

[0173] 1. Input SMILES sequence dataset

[0174] 2. Respectively, each sequence is subjected to the S1 process shown in the figure. Figure 1

[0175] 3. Parallel computing is arranged through machine resource allocation to improve computing efficiency.

[0176] 4. The cleaned SMILES dataset is output

[0177] 5. Respectively, each cleaned SMILES sequence is subjected to the S2 process shown in the figure. Figure 1

[0178] 6. Parallel computing is arranged through machine resource allocation to improve computing efficiency.

[0179] 7. All compound graph data variables are output. The entire dataset is presented in python list format, and each mathematical graph has node list variable and edge list variable shown in the figure. Figure 3

[0180] 8. The last two steps shown in the figure are used to save the data for machine learning and deep learning training, and the saving method is python pickle format. Figure 2 For example, the entire process in some instances is as follows:

[0181] Figure 3

[0182] ​​​​​​1. Input raw SMILES format data from a source

[0183] 2. Perform S1 flow as shown in Figure 1

[0184] 3. Perform S2 flow as shown in Figure 1

[0185] In particular, the final display result is in Python list format, which can be saved in Python pickle format for downstream deep learning training.

[0186] In summary, compared with the original SMILES sequence text, the method realizes global data set cleaning, deduplication and standardization. The samples with conflicts and different original data are standardized for downstream analysis.

[0187] Compared with the traditional ETL data processing method, the method realizes the conversion from the original data to the data that can be used for training, and standardizes the entire workflow from the original data to the training data set to the data model training.

[0188] The second aspect of the application provides a data preprocessing system for cleaning small molecule compounds, which is suitable for the data preprocessing method described in the application, and comprises:

[0189] The S1 text preprocessing unit is configured to include: according to a predetermined text processing rule, preprocessing the original SMILES data of the small molecule compound into the standardized SMILES text of the small molecule compound.

[0190] The S2 chemical graph formatting unit is configured to include: according to a predetermined text processing rule, format splitting each text element of the standardized SMILES text of the small molecule compound of S1 to obtain a digital graph structure of chemical information of the small molecule compound.

[0191] In a preferred embodiment of the application, the S3 unit is configured, wherein the digital graph structure of chemical information of the small molecule compound of S2 is used for construction of an artificial intelligence model.

[0192] In a preferred embodiment of the application, when the original SMILES text of the small molecule compound is preprocessed into the standardized SMILES text of the small molecule compound in the S1 text preprocessing unit, the predetermined text processing rule comprises:

[0193] ​​The S1-1 unit is configured to: optional structure standardization, wherein data of the small molecule compound is processed into original SMILES text;

[0194] The S1-2 unit is configured to: if the original SMILES text includes heavy metal components and organic compound components, removing the heavy metal components and retaining the organic compound components in the original SMILES text;

[0195] The S1-3 unit is configured to: if the original SMILES text includes multimer components, removing the multimer components and retaining the longest component in the original SMILES text;

[0196] The S1-4 unit is configured to: if the original SMILES text includes charges, adding or subtracting hydrogen atoms in the original SMILES text to achieve de-charging;

[0197] The S1-5 unit is configured to: removing special SMILES text information;

[0198] The S1-6 unit is configured to: outputting the standardized sequence to obtain the standardized SMILES text of the small molecule compound.

[0199] In a preferred embodiment of the present application, when the S2 chemical graph formatting unit formats each text element of the standardized SMILES text of the small molecule compound of S1, the predetermined text processing rule includes:

[0200] The S2-1 unit is configured to: splitting the standardized SMILES text of the small molecule compound of S1 into each core text element to obtain the text element of the small molecule compound;

[0201] The S2-2 unit is configured to: performing text processing identification on the nature of the text element of the small molecule compound of the S2-1 unit, identifying and completing the simplified chemical information to obtain the chemical information graph of the small molecule compound;

[0202] The S2-3 unit is configured to: establishing a coordinate system with atomic elements as nodes according to the chemical information graph of the small molecule compound of the S2-2 unit, and constructing a digital coordinate system of the chemical information graph of the small molecule compound;

[0203] The S2-4 unit is configured to: adding element attributes of nodes and edges according to the digital coordinate system of the chemical information graph of the small molecule compound of the S2-3 unit to obtain the digital graph structure of the chemical information of the small molecule compound.

[0204] In one preferred embodiment of the present application, the S2-5 unit is configured to, if necessary, complete the hydrogen atom information of the digital graph structure of the chemical information.

[0205] In one embodiment, the S2-6 unit is further included: output the digital graph structure of the chemical information of the small molecule compound completely.

[0206] The third aspect of the present application provides an electronic device, comprising: a memory, a processor; wherein the memory is configured to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the data preprocessing method for cleaning small molecule compounds according to any one of the present application.

[0207] Based on the present application, those skilled in the art should understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, an apparatus and / or a method can be implemented using any number and combination of the aspects set forth herein. In addition, this apparatus and / or method can be implemented using other structures and / or functionality in addition to or instead of one or more of the aspects set forth herein.

[0208] Those skilled in the art know that, in addition to implementing the system provided by the present application and its various devices, modules, units in a pure computer readable program code manner, the system provided by the present application and its various devices, modules, units can also be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers and embedded microcontrollers, etc. by logically programming the method steps to achieve the same functions. Therefore, the system provided by the present application and its various devices, modules, units can be considered as a hardware component, and the devices, modules, units included therein for implementing various functions can also be considered as structures within the hardware component; the devices, modules, units for implementing various functions can also be considered as both software modules implementing methods and structures within hardware components.

[0209] It should be noted that the above embodiments can be freely combined as needed. The above is only the preferred embodiment of the present application, and it should be pointed out that for those skilled in the art, without departing from the principle of the present application, a number of improvements and refinements can also be made, which should be considered as the protection scope of the present application.

[0210] All documents mentioned in the present application are cited as references in the present application, just as each document is cited as a reference. In addition, it should be understood that those skilled in the art can make various modifications or modifications to the present application after reading the above description of the present application, and these equivalent forms also fall within the scope defined by the claims attached to the present application.

Claims

1. A data preprocessing method for cleaning small molecule compounds, characterized in that, The data preprocessing method comprises: S1, a text preprocessing step, comprising: according to a predetermined text processing rule, preprocessing the original SMILES text of the small molecule compound into the standardized SMILES text of the small molecule compound; wherein the predetermined text processing rule comprises: S1-1, an optional structure standardization step, wherein the data of the small molecule compound is processed into the original SMILES text; S1-2, if the original SMILES text includes heavy metal components and organic compound components, removing the heavy metal components in the original SMILES text and retaining the organic compound components; S1-3, if the original SMILES text includes multimer components, removing the multimer components in the original SMILES text and retaining the longest component; S1-4, if the original SMILES text includes charges, adding or subtracting hydrogen atoms in the original SMILES text to achieve charge removal; S1-5, removing special SMILES text information; S1-6, outputting the standardized sequence to obtain the standardized SMILES text of the small molecule compound; S2, a chemical graph formatting step, comprising: according to a predetermined text processing rule, formatting and splitting each text element of the standardized SMILES text of the small molecule compound of S1 to obtain a digital graph structure of chemical information of the small molecule compound; wherein the predetermined text processing rule comprises: S2-1, splitting the standardized SMILES text of the small molecule compound of S1 into each core text element to obtain the text element of the small molecule compound; S2-2, text processing and identifying the properties of the text element of the small molecule compound of S2-1 to identify and complete the simplified chemical information to obtain the chemical information graph of the small molecule compound; S2-3, establishing a coordinate system with atomic elements as nodes according to the chemical information graph of the small molecule compound of S2-2 to construct a digital coordinate system of the chemical information graph of the small molecule compound; S2-4, adding node and edge element attributes according to the digital coordinate system of the chemical information graph of the small molecule compound of S2-3 to obtain the digital graph structure of the chemical information of the small molecule compound.

2. The data preprocessing method for cleaning small molecule compounds according to claim 1, further comprising: S3, wherein the digital graph structure of the chemical information of the small molecule compound of S2 is used for constructing an artificial intelligence model.

3. The data preprocessing method for cleaning small molecule compounds according to claim 1, wherein, Further comprising: S2-5, if necessary, completing the hydrogen atom information of the digital graph structure of the chemical information.

4. A data pre-processing system for cleaning small molecule compounds, adapted to the data pre-processing method according to any one of claims 1 to 3, characterized in that, Further comprising: S1 text preprocessing unit, the S1 text preprocessing unit is set to include: according to the predetermined text processing rules, the original SMILES data of the small molecule compound is preprocessed into the standardized SMILES text of the small molecule compound; S2 chemical graph formatting unit, the S2 chemical graph formatting unit is set to include: according to the predetermined text processing rules, each text element of the standardized SMILES text of the small molecule compound of the S1 is format split, and the digital graph structure of the chemical information of the small molecule compound is obtained.

5. The data preprocessing system for cleaning small molecule compounds of claim 4, further comprising an S3 unit configured to use the digital graph structure of the chemical information of the small molecule compound of the S2 for construction of an artificial intelligence model.

6. The data preprocessing system for cleaning small molecule compounds of claim 4, wherein the predetermined text processing rules for preprocessing the original SMILES text of the small molecule compound into the standardized SMILES text of the small molecule compound in the S1 text preprocessing unit include: S1-1 unit configured to optional structure standardization, wherein the data of the small molecule compound is processed into the original SMILES text; S1-2 unit configured to remove heavy metal components from the original SMILES text if the original SMILES text includes heavy metal components and organic compound components, and retain the organic compound components; S1-3 unit configured to remove multimer components from the original SMILES text if the original SMILES text includes multimer components, and retain the longest component; S1-4 unit configured to add or subtract hydrogen atoms from the original SMILES text to achieve decharging if the original SMILES text includes charge; S1-5 unit configured to remove special SMILES text information; S1-6 unit configured to output a standardized sequence to obtain the standardized SMILES text of the small molecule compound.

7. The data preprocessing system for cleaning small molecule compounds of claim 4, wherein the predetermined text processing rules for format splitting each text element of the standardized SMILES text of the small molecule compound of the S1 in the S2 chemical graph formatting unit include: S2-1 unit configured to split the standardized SMILES text of the small molecule compound of the S1 into text elements of each core to obtain text elements of the small molecule compound; S2-2 unit configured to perform text processing identification on the properties of the text elements of the small molecule compound of the S2-1 unit, identify and complete simplified chemical information to obtain a chemical information graph of the small molecule compound; ​ ​ ​ S2-3 unit: according to the chemical information map of the small molecule compound of the S2-2 unit, a coordinate system with atomic elements as nodes is established, and a digital coordinate system of the chemical information map of the small molecule compound is constructed; S2-4 unit: according to the digital coordinate system of the chemical information map of the small molecule compound of the S2-3 unit, the element attributes of the nodes and edges are added, and the digital graph structure of the chemical information of the small molecule compound is obtained.

8. The data pre-processing system for cleaning small molecule compounds of claim 7, wherein, Also includes: S2-5 unit is set to: if necessary, the hydrogen atom information of the digital graph structure of the chemical information is completed.

9. An electronic device, comprising: Including: Memory, processor; wherein, the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to realize the data preprocessing method for cleaning small molecule compounds as claimed in any one of claims 1-3.

Images