Method for comparing similarity of molecular structural formula of compound based on hydrogen nuclear magnetic resonance spectrum data

Abstract

The embodiment of the present application discloses a compound molecular structure formula similarity comparison method based on nuclear magnetic hydrogen spectrum data, comprising: comparing the nuclear magnetic data of a target compound with the nuclear magnetic data of known compounds to obtain a nuclear magnetic data similarity score; comparing the nuclear magnetic image of the target compound with the nuclear magnetic image of the known compounds to obtain a nuclear magnetic image similarity score; weighting the nuclear magnetic data similarity score and the nuclear magnetic image similarity score to obtain a final similarity score; if the final similarity score is the highest and is higher than a preset threshold, the known compound corresponding to the final similarity score with the highest score is the same as the target compound; otherwise, the target compound is determined to be a new compound. The present application can reduce human intervention, reduce spectrum analysis time, effectively obtain similar chemical structure formulas for researchers to refer to, and improve efficiency by combining nuclear magnetic spectrum information and nuclear magnetic data information.

Description

Technical Field

[0001] This invention relates to the field of nuclear magnetic resonance (NMR) data processing, and in particular to a method for comparing the similarity of compound molecular structures based on NMR data. Background Technology

[0002] Nuclear magnetic resonance (NMR) is a crucial method for identifying compound structures and studying chemical kinetics, with wide applications in applied chemistry, polymer chemistry, and forestry chemical engineering. The structural formula of a compound directly influences its extranuclear chemical environment, thereby affecting the charge distribution of hydrogen or carbon nuclei, which manifests in the spectrum as chemical shifts, coupling peaks, and other phenomena. The interpretation of NMR spectra is highly dependent on the researcher's knowledge and experience; improving the efficiency and accuracy of structural analysis is a major challenge. With the rapid development of computer technology, utilizing computational models to assist in solving this problem is becoming increasingly important.

[0003] Compared to carbon (C) spectroscopy, proton (H) spectroscopy (THS) has significant advantages in structural analysis. Firstly, there are far more HHS spectra available, facilitating the establishment of essential databases for retrieval. Secondly, HHS sensitivity is several times higher than C spectroscopy, allowing for more effective information even with limited sample quantities. Nevertheless, research on computer-aided CHS analysis is quite mature and widely used in the identification and retrieval of unknown compounds, while research and application of HHS analysis lag significantly. The main reason is that CHS analysis typically only requires one parameter, chemical shift, which is intuitive, simple, and easy to output. HHS, on the other hand, involves multiple dimensions such as chemical shift, integral, and split peaks, making data processing difficult and hindering similarity comparisons by computers.

[0004] Generally, the interpretation of proton NMR spectra involves the following steps: 1. Determining if the spectrum meets the interpretation requirements; 2. Adjusting and labeling the spectrum parameters; 3. Differentiating impurity peaks, solvent peaks, etc.; 4. Analyzing the spectrum based on chemical shifts and signal peak integral areas; 5. Inferring possible structural formulas based on the spectrum analysis results; 6. Confirming the inferred structural formulas, ensuring that each functional group is represented in the spectrum. The first three steps are relatively straightforward and can be easily replaced, but they are prone to misinterpretation. The latter three steps can be quickly performed using database searches.

[0005] To address the challenges of digitizing proton NMR spectra, efficiently perform computer-aided analysis of proton NMR spectra, and enable rapid identification and confirmation of compounds, it is essential to propose a method for comparing the chemical structures of compounds that can directly analyze proton NMR spectra and avoid manual intervention. Summary of the Invention

[0006] The technical problem to be solved by this invention is to provide a method for comparing the similarity of compound molecular structures based on 1H NMR data, which is a method for comparing the chemical structures of compounds by analyzing 1H NMR spectra and avoiding human intervention.

[0007] To solve the above-mentioned technical problems, the objective of this invention is achieved through the following technical solution:

[0008] A method for comparing the similarity of compound molecular structures based on 1H NMR spectroscopy data includes:

[0009] Obtain NMR data of known compounds;

[0010] Based on the NMR data of the known compounds, generate corresponding NMR images of the known compounds;

[0011] Acquire NMR data of the target compound;

[0012] Based on the NMR data of the target compound, a corresponding NMR image of the target compound is generated;

[0013] The similarity score of the NMR data is obtained by comparing the NMR data of the target compound with that of known compounds.

[0014] The NMR image of the target compound is compared with the NMR images of known compounds to obtain an NMR image similarity score;

[0015] The final similarity score is obtained by weighting the similarity scores of the NMR data and the NMR image.

[0016] If the final similarity score is the highest and is higher than a preset threshold, then the known compound corresponding to the highest final similarity score is the same as the target compound; otherwise, the target compound is identified as a new compound.

[0017] Preferably, the acquisition of NMR data of known compounds includes:

[0018] Obtain known compound C i Raw NMR data;

[0019] The known compound C i The raw NMR data were preprocessed to obtain preprocessed NMR data of known compounds.

[0020] The preprocessed NMR data of the known compound is converted into a fixed number of data points, thus obtaining the NMR data D of the known compound. i ;

[0021] Where i is a positive integer between 1 and n.

[0022] Preferably, generating a corresponding NMR image of the known compound based on the known compound's NMR data includes:

[0023] Based on the known compound NMR data D i The corresponding known compound NMR image I is generated. i .

[0024] Preferably, the acquisition of NMR data of the target compound includes:

[0025] Obtain target compound C T Raw NMR data;

[0026] The target compound C T The raw NMR data is preprocessed to obtain preprocessed NMR data of the target compound.

[0027] The preprocessed NMR data of the target compound is converted into a fixed number of data points, thus obtaining the NMR data D of the target compound. T .

[0028] Preferably, generating a corresponding NMR image of the target compound based on the target compound's NMR data includes:

[0029] Based on the NMR data of the target compound T, D T The corresponding NMR image I of the target compound is generated. T .

[0030] Preferably, the target compound NMR data D T Compared with known compound NMR data D i A similarity comparison was performed to obtain a similarity score for the NMR data, including:

[0031] By using data similarity comparison methods, the NMR data of the target compound D T Compared with known compound NMR data D i A similarity comparison was performed to obtain the similarity score of the NMR data. in

[0032] Preferably, the step of comparing the NMR image of the target compound with the NMR images of known compounds to obtain an NMR image similarity score includes:

[0033] The NMR image of the target compound was compared using an image similarity comparison method. T NMR images of known compounds I i Similarity comparisons were performed to obtain NMR image similarity scores. in

[0034] Preferably, the step of weighting the similarity scores of the NMR data and the NMR image to obtain the final similarity score includes:

[0035] Calculate the known compound C i With target compound C T The final similarity score Where 0 < α < 1;

[0036] The final similarity score dataset is obtained as S = {S1, S2, ..., S...} n}

[0037] Preferably, the NMR images of the known compound and the target compound are in PNG format.

[0038] This invention also provides a system for comparing the similarity of compound molecular structures based on NMR data, comprising:

[0039] Known compound NMR data acquisition unit, used to acquire NMR data of known compounds;

[0040] A known compound NMR image generation unit is used to generate a corresponding known compound NMR image based on the known compound NMR data;

[0041] The target compound NMR data acquisition unit is used to acquire the target compound NMR data;

[0042] The target compound NMR image generation unit is used to generate a corresponding target compound NMR image based on the target compound NMR data;

[0043] The NMR data similarity score comparison unit is used to compare the NMR data of the target compound with the NMR data of known compounds to obtain the NMR data similarity score;

[0044] The NMR image similarity score comparison unit is used to compare the NMR image of the target compound with the NMR image of a known compound to obtain an NMR image similarity score.

[0045] The final similarity score processing unit is used to perform weighted processing based on the similarity scores of the NMR data and the similarity scores of the NMR image to obtain the final similarity score;

[0046] The judgment unit is used to determine whether the known compound corresponding to the highest final similarity score is the same as the target compound if the final similarity score is the highest and the final similarity score is higher than a preset threshold; otherwise, the target compound is identified as a new compound.

[0047] The advantages of this invention compared to the prior art are:

[0048] (1) This invention develops a novel method for digitizing hydrogen spectrum data, enabling automatic computer processing of nuclear magnetic resonance hydrogen spectrum data.

[0049] (2) By combining NMR spectrum information and NMR data information, this invention can reduce human intervention, reduce spectrum analysis time, effectively obtain similar chemical structural formulas for researchers to refer to, and improve efficiency.

[0050] (3) This invention has scalability by constructing and updating a dataset of known compound molecules.

[0051] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart illustrating the method for comparing the similarity of compound molecular structures based on 1H NMR spectroscopy data in this embodiment.

[0054] Figure 2 This is a graph showing the similarity calculation results between the target compound and known compounds in this embodiment. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0057] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0058] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0059] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for comparing the structural similarity of compounds based on 1H NMR data, provided in an embodiment of the present invention. This embodiment includes steps S1 to S8, specifically:

[0060] S1. Obtain NMR data for known compounds;

[0061] S2. Based on the known compound NMR data, generate the corresponding known compound NMR image;

[0062] S3. Obtain NMR data of the target compound;

[0063] S4. Based on the NMR data of the target compound, generate the corresponding NMR image of the target compound;

[0064] S5. Compare the NMR data of the target compound with the NMR data of known compounds to obtain the NMR data similarity score;

[0065] S6. Compare the NMR image of the target compound with the NMR images of known compounds to obtain the NMR image similarity score;

[0066] S7. The final similarity score is obtained by weighting the similarity scores of the NMR data and the NMR image.

[0067] S8. If the final similarity score is the highest and the final similarity score is higher than a preset threshold, then the known compound corresponding to the highest final similarity score is the same as the target compound; otherwise, the target compound is identified as a new compound.

[0068] In one embodiment, step S1, acquiring NMR data of a known compound, includes:

[0069] S11. Obtain known compound C i Raw NMR data;

[0070] S12, the known compound C i The raw NMR data were preprocessed to obtain preprocessed NMR data of known compounds.

[0071] S13. Convert the preprocessed NMR data of the known compound into a fixed number of data points, thus obtaining the NMR data D of the known compound. i ;

[0072] Where i is a positive integer between 1 and n.

[0073] In one embodiment, S2, generating the corresponding NMR image of the known compound based on the known compound's NMR data, includes:

[0074] Based on the known compound NMR data D i The corresponding known compound NMR image I is generated. i .

[0075] Specifically, the known set of compounds is C = {C1, C2, ..., C}. n}, obtain C for each known compound i The raw NMR data (where i is a positive integer between 1 and n) are processed through preprocessing techniques such as data completion, Fourier transform, phase correction, and baseline correction to transform the C... i The NMR data is converted into a fixed number of data points (default is 32786), which is the NMR data D of the known compound. i ;

[0076] Based on known compound NMR data D i This can be combined with other data, such as NMR peaks, for each known compound C i Generate NMR images of the corresponding known compounds I i And save the image in PNG format.

[0077] In one embodiment, S3, acquiring the NMR data of the target compound includes:

[0078] Obtain target compound C T Raw NMR data;

[0079] The target compound C T The raw NMR data is preprocessed to obtain preprocessed NMR data of the target compound.

[0080] The preprocessed NMR data of the target compound is converted into a fixed number of data points, thus obtaining the NMR data D of the target compound. T .

[0081] In one embodiment, S4, generating a corresponding NMR image of the target compound based on the target compound's NMR data, includes:

[0082] Based on the NMR data of the target compound T, D T The corresponding NMR image I of the target compound is generated. T .

[0083] Specifically, the raw NMR data of the target compound T is obtained. Preprocessing techniques such as data completion, Fourier transform, phase correction, and baseline correction are used to transform the NMR data of the target compound into a fixed number of data points (default 32786), which is the target compound NMR data D. T ;

[0084] Based on the NMR data of the target compound D T It can be combined with other data, such as NMR peaks, to generate NMR images of the target compound. T And save the image in PNG format.

[0085] In one embodiment, S5, the NMR data D of the target compound T Compared with known compound NMR data D i A similarity comparison was performed to obtain a similarity score for the NMR data, including:

[0086] By using data similarity comparison methods, the NMR data of the target compound D T Compared with known compound NMR data D i A similarity comparison was performed to obtain the similarity score of the NMR data. in

[0087] In this embodiment, the data similarity comparison method can be one of various existing conventional methods, such as the Pearson correlation coefficient comparison method.

[0088] In one embodiment, S6, comparing the NMR image of the target compound with the NMR images of known compounds to obtain an NMR image similarity score, includes:

[0089] The NMR image of the target compound was compared using an image similarity comparison method. T NMR images of known compounds I i Similarity comparisons were performed to obtain NMR image similarity scores. in

[0090] In this embodiment, the image similarity comparison method can employ various existing conventional methods, such as the SSIM algorithm.

[0091] In one embodiment, S7, the weighted processing based on the nuclear magnetic data similarity score and the nuclear magnetic image similarity score to obtain the final similarity score includes:

[0092] Calculate each known compound C in the compound set C i and the target compound C T of the final similarity score where 0 < α < 1;

[0093] Obtain the final similarity score dataset S = {S1, S2,..., S n}.

[0094] If all the scores in the final similarity score dataset S are less than the specified threshold TH (0 < TH < 1), it is considered that the target compound T cannot find a matching item in the known compound set C and is a new compound, then T is added to C for future research use; otherwise, it is considered that the compound in S that is higher than the threshold TH and has the highest score has the most similar chemical structural formula to the target compound T.

[0095] Next, taking the Figure 2 compound as an example, in combination with Figure 1 the specific implementation manner, the present invention will be further described in detail. The selected image similarity comparison algorithm is SSIM, the selected data similarity comparison algorithm is the Pearson correlation coefficient, the parameter α = 0.8, and TH = 0.8.

[0096] The method for comparing the chemical structural formulas of compounds based on nuclear magnetic data includes the following processes:

[0097] (1) Based on the nuclear magnetic data of known compounds, generate corresponding nuclear magnetic images and save them in the PNG format as the known compound dataset;

[0098] (2) Obtain the nuclear magnetic data of the target compound, and generate a nuclear magnetic image and save it in the PNG format;

[0099] (3) Based on the nuclear magnetic image of the target compound, calculate the nuclear magnetic image similarity score between this compound and other known compounds;

[0100] (4) Based on the nuclear magnetic data of the target compound, calculate the nuclear magnetic data similarity score between this compound and other known compounds;

[0101] (5) Based on the results of steps (2) and (3), calculate the final similarity score between the target compound and other compounds by weighting;

[0102] (vi) The compound with the highest final similarity score and a final similarity score higher than the threshold is considered to be the same as the target compound; otherwise, the target compound is considered to be a new compound.

[0103] The process specifically includes the following steps:

[0104] Step A: C23H19N3, C23H23N3O2S, and C26H29N3O2S are known compounds. Based on the NMR data of these three compounds, NMR images are generated, such as... Figure 2 As shown, this constitutes a dataset of known compounds.

[0105] The second process specifically includes the following steps:

[0106] Step B: Obtain NMR data for the target compound C24H19N3 and generate an NMR image, such as... Figure 2 As shown.

[0107] The third process specifically includes the following steps:

[0108] Step C: Based on the NMR image of the target compound C24H19N3, the image similarity score of C24H19N3 with three other known compounds is calculated using an image similarity comparison algorithm. The SSIM algorithm is used as the image similarity algorithm, and the structural_similarity function of the skimage package in Python is used to calculate the image similarity score. The image similarity score is 0.956 between C24H19N3 and C23H19N3, 0.932 between C24H19N3 and C23H23N3O2S, and 0.883 between C24H19N3 and C26H29N3O2S.

[0109] The fourth process specifically includes the following steps:

[0110] Step D: Based on the NMR data of the target compound C24H19N3, a data similarity comparison algorithm is used to calculate the data similarity scores of C24H19N3 with three other known compounds. The data similarity scores are: C24H19N3 with C23H19N3 = 0.305, C24H19N3 with C23H23N3O2S = 0.176, and C24H19N3 with C26H29N3O2S = 0.071.

[0111] The fifth process specifically includes the following steps:

[0112] Step E: Based on the results of steps three and four, calculate the final similarity scores of C24H19N3 with the other three known compounds. The final similarity score between C24H19N3 and C23H19N3 is 0.8×|0.956|+(1-0.8)×|0.305|=0.826. Similarly, the final similarity scores between C24H19N3 and C23H23N3O2S are 0.781, and between C24H19N3 and C26H29N3O2S are 0.721.

[0113] The sixth process specifically includes the following steps:

[0114] Step F: The final similarity scores of C24H19N3 and C23H19N3, and the final similarity scores of C24H19N3 and C23H23N3O2S are both higher than the specified threshold TH. However, since the final similarity scores of C24H19N3 and C23H19N3 are higher, the target compound C24H19N3 is determined to be most similar to the chemical structure of compound C23H19N3 in the known compound set.

[0115] In this embodiment, the known compound NMR image and the target compound NMR image can be in various formats. In one embodiment, the known compound NMR image and the target compound NMR image are in PNG format.

[0116] This embodiment also provides a similarity comparison system for compound molecular structures based on NMR data, corresponding to the above method, including:

[0117] Known compound NMR data acquisition unit, used to acquire NMR data of known compounds;

[0118] A known compound NMR image generation unit is used to generate a corresponding known compound NMR image based on the known compound NMR data;

[0119] The target compound NMR data acquisition unit is used to acquire the target compound NMR data;

[0120] The target compound NMR image generation unit is used to generate a corresponding target compound NMR image based on the target compound NMR data;

[0121] The NMR data similarity score comparison unit is used to compare the NMR data of the target compound with the NMR data of known compounds to obtain the NMR data similarity score;

[0122] The NMR image similarity score comparison unit is used to compare the NMR image of the target compound with the NMR image of a known compound to obtain an NMR image similarity score.

[0123] The final similarity score processing unit is used to perform weighted processing based on the similarity scores of the NMR data and the similarity scores of the NMR image to obtain the final similarity score;

[0124] The judgment unit is used to determine whether the known compound corresponding to the highest final similarity score is the same as the target compound if the final similarity score is the highest and the final similarity score is higher than a preset threshold; otherwise, the target compound is identified as a new compound.

[0125] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for comparing the similarity of compound molecular structures based on 1H NMR spectroscopy data, characterized in that, include: Obtain NMR data of known compounds; Based on the NMR data of the known compounds, generate corresponding NMR images of the known compounds; Acquire NMR data of the target compound; Based on the NMR data of the target compound, a corresponding NMR image of the target compound is generated; The similarity score of the NMR data is obtained by comparing the NMR data of the target compound with that of known compounds. The NMR image of the target compound is compared with the NMR images of known compounds to obtain an NMR image similarity score; The final similarity score is obtained by weighting the similarity scores of the NMR data and the NMR image. If the final similarity score is the highest and the final similarity score is higher than a preset threshold, then the known compound corresponding to the highest final similarity score is the same as the target compound. Conversely, the target compound is identified as a new compound. in, The acquisition of NMR data for known compounds includes: Obtain known compound C i Raw NMR data; The known compound C i The raw NMR data were preprocessed to obtain preprocessed NMR data of known compounds. The preprocessed NMR data of the known compound is converted into a fixed number of data points, thus obtaining the NMR data D of the known compound. i ; Where i is a positive integer between 1 and n; The step of generating corresponding NMR images of known compounds based on the known compound NMR data includes: Based on the known compound NMR data D i The corresponding known compound NMR image I is generated. i ; The acquisition of NMR data of the target compound includes: Obtain target compound C T Raw NMR data; The target compound C T The raw NMR data is preprocessed to obtain preprocessed NMR data of the target compound. The preprocessed NMR data of the target compound is converted into a fixed number of data points, thus obtaining the NMR data D of the target compound. T ; The step of generating a corresponding NMR image of the target compound based on the target compound's NMR data includes: Based on the NMR data of the target compound D T The corresponding NMR image I of the target compound is generated. T ; The target compound NMR data D T Compared with known compound NMR data D i A similarity comparison was performed to obtain a similarity score for the NMR data, including: By using data similarity comparison methods, the NMR data of the target compound D T Compared with known compound NMR data D i A similarity comparison was performed to obtain the similarity score of the NMR data. in The step of comparing the NMR image of the target compound with the NMR images of known compounds to obtain an NMR image similarity score includes: The NMR image of the target compound was compared using an image similarity comparison method. T NMR images of known compounds I i Similarity comparisons were performed to obtain NMR image similarity scores. in The step of weighting the similarity scores of the NMR data and the NMR image to obtain the final similarity score includes: Calculate the known compound C i With target compound C T The final similarity score Where 0 < α < 1; The final similarity score dataset is obtained as S = {S1, S2, ..., S...} n } 2. The method for comparing the similarity of compound molecular structures based on proton NMR spectroscopy data according to claim 1, characterized in that, The NMR images of the known compound and the target compound are in PNG format.

3. A system for comparing the similarity of compound molecular structures based on NMR data, utilizing the method for comparing the similarity of compound molecular structures based on proton NMR spectroscopy data as described in any one of claims 1-2, characterized in that, include: Known compound NMR data acquisition unit, used to acquire NMR data of known compounds; A known compound NMR image generation unit is used to generate a corresponding known compound NMR image based on the known compound NMR data; The target compound NMR data acquisition unit is used to acquire the target compound NMR data; The target compound NMR image generation unit is used to generate a corresponding target compound NMR image based on the target compound NMR data; The NMR data similarity score comparison unit is used to compare the NMR data of the target compound with the NMR data of known compounds to obtain the NMR data similarity score; The NMR image similarity score comparison unit is used to compare the NMR image of the target compound with the NMR image of a known compound to obtain an NMR image similarity score. The final similarity score processing unit is used to perform weighted processing based on the similarity scores of the NMR data and the similarity scores of the NMR image to obtain the final similarity score; The judgment unit is used to determine that if the final similarity score is the highest and the final similarity score is higher than a preset threshold, then the known compound corresponding to the highest final similarity score is the same as the target compound. Conversely, the target compound is identified as a new compound.

Images