A method for quantitative screening of new drug leads based on chemical space

Through chemical space-based methods, the Canonical SMILES encoding of traditional Chinese medicine and compounds is obtained, the FCFP_6 fingerprint is calculated and clustered, and a generalized linear model is constructed, which solves the problem of ignoring key structural features and inaccurate calculation of molecular similarity in the prior art, and realizes the problem of molecular similarity calculations being ignored and inaccurate. Accurate prediction of IC50 values ​​of traditional Chinese medicine ingredients and screening of new drug pilots.

CN118506921BActive Publication Date: 2024-12-06INST OF BASIC RES & CLINICAL MEDICINE CHINA ACAD OF CHINESE MEDICAL SCI
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410758008.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-13
Publication Date
2024-12-06
Estimated Expiration
2044-06-13

AI Technical Summary

Technical Problem

The prior art may ignore or overemphasize key structural features in biological activity in molecular similarity calculations, resulting in misleading results and inaccurate calculation results for large or complex compounds.

Method used

Methods for quantitative screening of new drug pilots based on chemical space. By obtaining the chemical composition data of traditional Chinese medicine and the Canonical SMILES encoding of compounds, FCFP_6 fingerprints are calculated and clustered, a generalized linear model is constructed, and IC50 values ​​are quantitatively predicted, and pilots for developing new drugs are screened out. .

Benefits of technology

The Canonical SMILES coded repetition check of the compound was determined to determine the set of chemical components without repeats, accurately divide molecules with similar structures into the chemical space, and build a GLM model to accurately predict the IC50 value of traditional Chinese medicine components, and screen out effective new drug pioneers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118506921B_ABST
    Figure CN118506921B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for quantitatively screening new drug leads based on chemical space, comprising: obtaining chemical component data of traditional Chinese medicine with specific efficacy through different databases, and obtaining Canonical SMILES coding of the chemical component data of traditional Chinese medicine; obtaining IC 50 The compound set with the highest value was combined as the background data; after mixing the TCM chemical component data and the compound set, the FCFP_6 fingerprints of all molecules were calculated and clustered according to the Tanimoto distance, and the background data and TCM chemical component data with similar structures were classified into the same chemical space; a generalized linear model was constructed, and based on the same chemical space, the IC values ​​of the chemical components of TCM were quantitatively predicted by the generalized linear model. 50 value; IC 50 The IC values ​​of Chinese herbal medicine components can be more accurately and quantitatively predicted. 50 value.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical fields of chemical informatics and computational biology, and in particular relates to a method for quantitatively screening new drug leads based on chemical space. Background Art

[0002] "Chemical space" was proposed by Christopher M. Dobson in 2004 and often refers to "multidimensional descriptor space", that is, a region defined by specifically selected descriptors and their restrictions. Chemical space characterizes molecular properties by calculating molecular descriptors, and uses these descriptor values ​​as nodes to form a multidimensional space. Then, by judging the distance between two points in the multidimensional space, the similarity of molecules is compared to infer the similarity of the properties of two compounds. Similar drug molecules usually have similar properties and active molecules. Similarity calculation has always been an important method for discovering lead compounds in new drug research and development.

[0003] A variety of methods have been developed to describe molecular shapes and determine shape similarities between small molecules. The most commonly used methods include atomic distance-based methods, surface-based methods (such as spherical harmonics and three-dimensional Zernike descriptors), and representation methods based on atom-centered Gaussian superposition. Among them, existing technologies have performed well in both retrospective and prospective virtual screening, but they also have certain disadvantages: first, molecular similarity calculations may ignore or overemphasize key structural features in certain biological activities, sometimes producing misleading results; second, for large or highly complex compounds, the results of molecular similarity calculations may not be accurate enough because these structural elements may not be fully described. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a method for quantitatively screening new drug leads based on chemical space to solve the problems existing in the above prior art.

[0005] To achieve the above object, the present invention provides a method for quantitatively screening new drug leads based on chemical space, comprising:

[0006] Based on different databases of the traditional Chinese medicine system, we obtain the chemical component data of traditional Chinese medicine with specific efficacy, and obtain the Canonical SMILES code of the chemical component data of traditional Chinese medicine;

[0007] Based on the ChEMBL database, 50 The compound set with the highest value is combined as background data, and the Canonical SMILES code of the compound is obtained;

[0008] After mixing the Chinese medicine chemical component data and the compound set, the FCFP_6 fingerprints of all molecules are calculated and clustering calculations are performed based on the Tanimoto distance, and the background data and Chinese medicine chemical component data with similar structures are classified into the same chemical space;

[0009] Construct a generalized linear model, and based on the same chemical space, quantitatively predict the IC values ​​of the chemical components of the traditional Chinese medicine by using the generalized linear model 50 value; the IC 50 The values ​​are compared with those of drugs already on the market to screen out lead compounds for developing new drugs.

[0010] Preferably, the database includes: a TCMSP database and a NPASS database.

[0011] Preferably, the biological activity of the background data corresponds to the specific efficacy.

[0012] Preferably, based on clustering calculation, molecules with similar structures are divided into the same or adjacent chemical spaces, wherein each chemical space includes: background data and Chinese medicine chemical component data.

[0013] Preferably, the process of constructing a generalized linear model includes:

[0014] By calculating chemical space, local sensitive hashing, and Consin similarity, the background data and Chinese medicine chemical component data with similar structures are classified into the same chemical space, and a minimum spanning tree diagram is drawn;

[0015] Based on the chemical component set of the minimum spanning tree diagram, the molecular descriptors were calculated using MOLD2 software to construct a generalized linear model.

[0016] Preferably, the molecular descriptors calculated using MOLD2 software include but are not limited to the physicochemical properties of the molecules.

[0017] Preferably, the method further comprises quantitatively predicting the IC 50 The values ​​were experimentally verified to ensure the effectiveness of the screened new drug leads.

[0018] In order to achieve the above technical objectives, the present invention also provides a computer storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method for quantitative screening of new drug leads based on chemical space is implemented.

[0019] Compared with the prior art, the present invention has the following advantages and technical effects:

[0020] The present invention provides a method for quantitatively screening new drug leads based on chemical space, comprising: firstly, obtaining chemical component data of traditional Chinese medicine with specific efficacy based on different databases of traditional Chinese medicine system, and obtaining Canonical SMILES coding of chemical component data of traditional Chinese medicine; obtaining chemical component data with IC based on ChEMBL database; 50 The compound set with the same value is combined as the background data, and the Canonical SMILES code of the compound is obtained; then, after the Chinese medicine chemical component data and the compound set are mixed, the FCFP_6 fingerprints of all molecules are calculated and clustering calculation is performed according to the Tanimoto distance, and the background data and Chinese medicine chemical component data with similar structures are classified into the same chemical space; finally, a generalized linear model is constructed, and based on the same chemical space, the IC values ​​of the Chinese medicine chemical components are quantitatively predicted by the generalized linear model. 50 value; the IC 50 The values ​​are compared with those of drugs already on the market to screen out lead compounds for developing new drugs.

[0021] The present invention checks the duplicates of the Chinese medicine ingredients by using the Canonical SMILES codes of the compounds, and can determine the set of non-repetitive chemical ingredients contained in a certain type of Chinese medicine. The present invention also constructs a compound with a certain activity and IC 50 The present invention divides all molecules into a certain number of chemical spaces. The closer the spatial distance of molecules with similar structures is, the more molecules with similar structures can be divided into the same chemical space. The present invention can more accurately predict the IC value of traditional Chinese medicine components through quantitative prediction of generalized linear model. 50 value. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0023] Figure 1 It is a schematic diagram of the framework of an embodiment of the present invention;

[0024] Figure 2 is a flow chart of a method according to an embodiment of the present invention;

[0025] Figure 3 Schematic diagram of the inhibitory effect of patchouli alcohol (B) according to an embodiment of the present invention on the growth of Plasmodium falciparum standard strain 3D7;

[0026] Figure 4 Schematic diagram of the inhibitory effect of menthol (E) on the growth of Plasmodium falciparum standard strain 3D7 according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0028] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0029] Embodiment 1

[0030] like Figure 1-2 As shown, this embodiment provides a method for quantitatively screening new drug leads based on chemical space, comprising:

[0031] S1, based on different databases of the traditional Chinese medicine system, obtain the chemical component data of traditional Chinese medicine with specific efficacy, and obtain the Canonical SMILES code of the chemical component data of traditional Chinese medicine;

[0032] Specifically, a compilation of Chinese medicines and their corresponding chemical components. For Chinese medicines with a certain efficacy, we searched the 2020 edition of the Chinese Pharmacopoeia and found Chinese medicinal materials in the traditional Chinese medicine system. Then, we used these Chinese medicines and their corresponding Latin names as keywords to search in the TCMSP and NPASS databases to obtain the names of chemical components. The Canonical SMILES code of the active ingredient was confirmed from PubChem (https: / / PubChem.ncbi.nlm.nih.gov).

[0033] Since there are many synonyms of chemical components, it is impossible to determine whether they are the same compound based on the name of the component alone. Therefore, the Canonical SMILES code of the compound will be obtained from the official website of PubChem, and then the Chinese medicine components will be checked for duplication. Finally, a set of non-duplicate chemical components contained in a certain type of Chinese medicine with different effects was determined.

[0034] S2, based on the ChEMBL database, obtained 50 The compound set with the highest value is combined as background data;

[0035] Specifically, the IC 50 We used compounds as background data. For a certain activity, such as antimalarial activity, we searched the ChEMBL database (https: / / www.ebi.ac.uk / chembl / ) and downloaded the results. We then searched the Canonical SMILES of the compounds in PubChem.

[0036] Since the SMILES codes obtained by different systems are inconsistent, for the next step of chemical space calculation, the Canonical SMILES code of PubChem is uniformly selected. 50 The chemical composition set of the values ​​is used as the background data set.

[0037] S3, after mixing the Chinese medicine chemical component data and the compound set, calculating the FCFP_6 fingerprints of all molecules and performing clustering calculations based on the Tanimoto distance, and classifying the background data and Chinese medicine chemical component data with similar structures into the same chemical space;

[0038] Specifically, after mixing the compounds obtained in the first two steps, the FCFP_6 fingerprints of all molecules are calculated and clustering calculation is performed according to the Tanimoto distance of the fingerprints: first, the first compound is randomly selected as the cluster center, and the compound farthest from the first compound is selected as the next cluster center; it is continuously iterated until there are a sufficient number of cluster centers; then the Tanimoto distances of other compounds and the cluster center compounds are calculated, and other compounds are assigned to the nearest cluster center according to the distance size, and all molecules are divided into a certain number of spaces, and the closer the spatial distance of molecules with similar structures is. This embodiment can divide molecules with similar structures into the same chemical space.

[0039] S4, constructing a generalized linear model, and quantitatively predicting the IC values ​​of the chemical components of the traditional Chinese medicine through the generalized linear model based on the same chemical space. 50 value; the IC 50 The values ​​are compared with those of drugs already on the market to screen out lead compounds for developing new drugs.

[0040] Specifically, for each set of molecules in the chemical space, the SMILES dataset is used as input to calculate the molecular fingerprint MHFP6, and then the following steps are performed: (I) Establish an LSH forest index; (II) Use cosine similarity to construct a c-approximate k-neighbor graph; (III) Calculate the minimum spanning tree (MST) of the c-approximate k-neighbor graph; (IV) Generate the layout of the MST. (V) Select adjacent molecular sets in the layout, and the set is required to contain background data and Chinese medicine chemical component data. Using the Mold2 software released by the US FDA, 777 compound molecular descriptors were calculated for each two-dimensional structure of each compound, and a GLM model was constructed to predict IC 50 Quantitative prediction of IC values ​​of chemical components of traditional Chinese medicine 50The value of the MST graph is compared with the drugs that have been marketed, and the lead compounds that may be developed into new drugs are screened out. This embodiment calculates the MST graph to make compounds with similar structures closer in space. This embodiment builds a machine learning GLM model based on this, and the prediction results will be more accurate, which can more accurately predict the IC value of traditional Chinese medicine ingredients. 50 value.

[0041] Embodiment 2

[0042] This embodiment takes antimalarial drug screening as an example, and the specific process includes:

[0043] S1, Compilation of Chinese medicines and their corresponding chemical components. We found 10 Chinese medicines with antimalarial effects from the Chinese Pharmacopoeia. We searched from TCMSP and NAPASS, and obtained the Canonical SMILES codes of the compounds from the official website of PubChem for duplicate checking. These Chinese medicines contain 555 chemical components.

[0044] S2, collected from the ChEMBL database 50 We used the keyword "anti-malaria" to search in the ChEMBL database and downloaded the results, a total of 7662 compounds. The Canonical SMILES code of the compound was obtained from the PubChem website.

[0045] S3, calculate the chemical space. After mixing the compounds obtained in the first two steps, calculate the FCFP_6 fingerprints of all molecules and perform clustering calculations based on the Tanimoto distance of the fingerprints, clustering them into 10 categories. Each chemical space includes background data and Chinese medicine chemical component data.

[0046] S4, Quantitative prediction of IC of chemical components of traditional Chinese medicine 50 For each data set, a minimum spanning tree (MST) diagram was drawn, with a total of 10 diagrams drawn. The chemical components of traditional Chinese medicine were identified in the diagram, and the pIC values ​​of the background data were 50 Then find the adjacent chemical component set, use MOLD2 to calculate the molecular descriptor, and then build a GLM model to quantitatively predict the IC value of the chemical components of traditional Chinese medicine. 50 Compared with the antimalarial drug artemisinin, which is already on the market, the chemical components with development potential were screened out. IC 50 The value is based on artemisinin IC 50 The prediction was based on 1.0 nM, and the two compounds that may have antimalarial activity are B and C.

[0047] In order to verify the accuracy of the method, a compound predicted to have antimalarial activity and a compound without antimalarial activity were selected for experimental verification. Due to the limitation of purchasing experimental raw materials, patchouli alcohol (B) and menthol (E) were selected for antimalarial experimental verification in this example. 50 The value is 35 nM, and the predicted IC of E 50 The value is 1590nM.

[0048] In this embodiment, the antimalarial experimental steps are as follows:

[0049] 1. Materials: bovine serum albumin Plasmodium falciparum, 0.16% saponin (Sigma, LOT: BCBR4223V), 0.1M EDTA (Sigma, LOT: BCBR3854V), 0.4M Tris-Base (Sigma, LOT: SLBR9609V), 0.8% Triton X-100 (Sigma, LOT: SLBR3411V), etc.

[0050] 2. Methods. Synchronize the ring stage Plasmodium falciparum. Smear and count the infection rate. Rationally match the synchronized infected red blood cells, healthy red blood cells and culture medium to obtain a Plasmodium in vitro culture system with a hematocrit of 2% and an infection rate of 1% (ring body, 2-4 hours of worm age). PMI 1640 plate (50μl / well). A1-A10 plate 25μl of drug (3 replicates for each drug), mix and perform multiple dilutions (Ax-Hx). Except A12-H12, each well is plated with 50μl of the Plasmodium culture medium prepared in step 2.3. A12-H12, 50μl of healthy red blood cells (hematocrit is 2% / culture medium). Culture for 72 hours under conventional in vitro culture conditions for Plasmodium falciparum (5% O2, 3% CO2, 92% N2, 37°C). Add 100μl of lysis solution to each well and place in a drawer away from light for 1 hour. Determination by fluorescence microplate reader: Ex = 485nm; Em = 525nm.

[0051] 3. Experimental results. In this experiment, the SYBR Green assay was used to determine the half inhibitory concentration (IC 50 ) values, and calculated by nonlinear regression using GraphPad Prism 8.0 software (values ​​are mean ± standard error), and a four-parameter logistic fitting curve was used. Each value in the curve is the mean ± standard deviation of three biological replicates. The results showed that compound B had a high concentration in the nanomolar range (IC 50=252.8nM) showed a certain inhibitory effect on the proliferation of parasites, indicating that although the compound has a certain degree of difference in efficacy with standard antimalarial drugs (such as chloroquine (about 2.360nM), mefloquine (about 3.962nM) and artemisinin (about 11.768nM)), it still has a strong inhibitory effect on malarial parasites. The inhibitory effect of patchouli alcohol (B) on the growth of Plasmodium falciparum standard strain 3D7, as shown in Figure 3 Compound E in the nanomolar concentration range (IC 50 =5.393e+149nM) had no significant inhibitory effect on the proliferation of parasites. Menthol (E) had an inhibitory effect on the growth of Plasmodium falciparum standard strain 3D7. Figure 4 shown.

[0052] The antimalarial experiment showed that the predicted results were basically consistent with the experimental results, indicating that the method of Example 2 is practical.

[0053] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for quantitatively screening new drug leads based on chemical space, characterized in that: The following steps are involved: Based on different databases of the traditional Chinese medicine system, we obtain the chemical component data of traditional Chinese medicine with specific efficacy, and obtain the Canonical SMILES code of the chemical component data of traditional Chinese medicine; Based on the ChEMBL database, 50 The compound set with the highest value is combined as background data, and the Canonical SMILES code of the compound is obtained; After mixing the Chinese medicine chemical component data and the compound set, the FCFP_6 fingerprints of all molecules are calculated and clustering calculations are performed based on the Tanimoto distance, and the background data and Chinese medicine chemical component data with similar structures are classified into the same chemical space; The process of calculating the FCFP_6 fingerprints of all molecules and clustering them according to Tanimoto distance includes: First, randomly select the first compound as the cluster center, and select the compound farthest from the first compound as the next cluster center; continue to iterate until there are enough cluster centers; then calculate the Tanimoto distance between other compounds and the cluster center compound, and assign other compounds to the nearest cluster center according to the distance, dividing all molecules into a certain number of spaces, and the closer the spatial distance of molecules with similar structures; Construct a generalized linear model, and based on the same chemical space, quantitatively predict the IC values ​​of the chemical components of the traditional Chinese medicine by using the generalized linear model 50 value; the IC 50 By comparing the value with the drugs on the market, the lead compounds for developing new drugs can be screened out; The process of building a generalized linear model involves: By calculating chemical space, local sensitive hashing, and Consin similarity, the background data and Chinese medicine chemical component data with similar structures are classified into the same chemical space, and a minimum spanning tree diagram is drawn; Based on the chemical component set of the minimum spanning tree diagram, the molecular descriptors are calculated using MOLD2 software to construct a generalized linear model; For each molecular set in the chemical space, the SMILES dataset is used as input to calculate the molecular fingerprint MHFP6, and then the following steps are performed: (I) establish the LSH forest index; (II) use cosine similarity to construct a c-approximate k-nearest neighbor graph; (III) calculate the minimum spanning tree MST of the c-approximate k-nearest neighbor graph; (IV) generate the layout of MST; (V) select the adjacent molecular sets in the layout, which must contain background data and Chinese medicine chemical component data.

2. The method for quantitative screening of new drug leads based on chemical space according to claim 1, characterized in that: The database includes: TCMSP database and NPASS database.

3. The method for quantitative screening of new drug leads based on chemical space according to claim 1, characterized in that: The biological activity of the background data has a corresponding relationship with the specific efficacy.

4. The method for quantitative screening of new drug leads based on chemical space according to claim 1, characterized in that: Based on clustering calculations, molecules with similar structures are divided into the same or adjacent chemical spaces, where each chemical space includes: background data and Chinese medicine chemical component data.

5. The method for quantitative screening of new drug leads based on chemical space according to claim 1, characterized in that: The molecular descriptors calculated using MOLD2 software include, but are not limited to, the physicochemical properties of the molecules.

6. The method for quantitative screening of new drug leads based on chemical space according to claim 1, characterized in that: It also includes IC for quantitative prediction 50 The values ​​were experimentally verified to ensure the effectiveness of the screened new drug leads.

7. A computer storage medium, characterized in that: The computer storage medium stores computer program instructions, which, when executed by a processor, implement the method for quantitatively screening new drug leads based on chemical space as described in any one of claims 1 to 6.

Citation Information

Patent Citations

  • Lead compound virtual screening method and device

    CN107862173A

  • Automatic drug design method and system, computing equipment and computer readable storage medium

    CN112116963A