阅读说明：本技术 一种用于药物虚拟筛选的分子三维相似度的打分方法 (Molecular three-dimensional similarity scoring method for virtual drug screening ) 是由 严鑫 李瑞麟 卢峰 于 2021-07-22 设计创作，主要内容包括：本发明涉及计算机辅助药物研发技术领域,具体涉及一种用于药物虚拟筛选的分子三维相似度的打分方法,该方法包括步骤一、获取用于相似性比较的两分子的特征参数；步骤二、训练深度学习模型；以及步骤三、深度学习模型的外部验证。本发明通过训练深度神经网络,选取多维度相似性作为特征参数,经深度神经网络的变换,给出最终的综合打分即AUC值。该打分方法在药物筛选准确率上有显著的改进,能够尽可能多地将具有潜在生物活性的分子从含有大量分子的数据库中筛选出来,而不漏掉任何潜在活性分子,从而解决了假阳性或假阴性的问题,同时保持着筛选高通量的计算速度。故,本发明在药物虚拟筛选方面具有非常广阔的应用前景。(The invention relates to the technical field of computer-aided drug research and development, in particular to a molecular three-dimensional similarity scoring method for drug virtual screening, which comprises the following steps of firstly, obtaining characteristic parameters of two molecules for similarity comparison; step two, training a deep learning model; and step three, external verification of the deep learning model. According to the method, a deep neural network is trained, multi-dimensional similarity is selected as a characteristic parameter, and a final comprehensive score, namely an AUC value, is given through transformation of the deep neural network. The scoring method has significant improvement in drug screening accuracy, and can screen molecules with potential biological activity from a database containing a large number of molecules as much as possible without leaking any potential active molecules, thereby solving the problem of false positive or false negative and maintaining the high-throughput calculation speed of screening. Therefore, the invention has very wide application prospect in the aspect of virtual screening of the drugs.)

1. A molecule three-dimensional similarity scoring method for virtual drug screening is characterized in that: the method comprises the following steps:

step one, obtaining various characteristic parameters of two molecules for similarity comparison

Respectively reading topological structure and three-dimensional structure information of two molecules for similarity comparison, and calculating to obtain various characteristic parameters, wherein the characteristic parameters comprise: difference in the number of atoms between the two molecules (F1); number of rotatable chemical bonds of two molecules (F2); volume difference of two molecules (F3); similarity of shape of the two molecules (F4); similarity of two molecular hydrogen bond acceptors (F5); similarity of two molecular hydrogen bond donors (F6); similarity of two molecular aromatic rings (F7); similarity of hydrophobic centers of the two molecules (F8); similarity of positively charged groups of the two molecules (F9); and similarity of the negatively charged groups of the two molecules (F10); wherein:

the calculation mode of F1 is to read in the respective topological structure information of two molecules, then to take the absolute value of the difference of the total number of atoms of the two molecules;

the calculation mode of F2 is that on the basis of the calculation mode of F1, whether each chemical bond is a rotatable bond is judged to obtain the total number of the rotatable bonds of two molecules, and then the absolute value of the difference value of the total number of the rotatable bonds of the two molecules is taken;

the calculation method of F3 is based on the calculation method of F1, and the Van der Waals radius of the atoms is obtained according to the types of the atoms in two molecules, each atom is represented by a Gaussian sphere, the radius of the Gaussian sphere is the same as the Van der Waals radius of the atom, the position coordinate of the Gaussian sphere is the same as the coordinate of the atom, and the atom coordinate is from the input three-dimensional structure of the molecule; calculating the superposed volume of Gaussian ball groups in each molecule, wherein the ijth Gaussian ball group comprises a Gaussian ball corresponding to the ith atom and a Gaussian ball corresponding to the jth atom, and the superposed volume of the ijth Gaussian ball group is v_ij(ii) a Calculate the volume of overlap of each of the two molecules by itself asN is the total number of atoms in the molecule; then, taking the absolute value of the difference value of the superposed volumes of the two molecules;

the calculation method of F4 is to calculate the intermolecular superposition volume of two molecules under multiple superposition conditions based on the calculation method of F3Wherein v is_ijIs the congruent volume of the ith atom in the first molecule and the jth atom in the second molecule, N being the total number of atoms in the first molecule and M being the total number of atoms in the second molecule, the maximum of which is selectedAs the maximum intermolecular volume; calculating the similarity of the shapes of the two moleculesWherein V_AIs the self-stacking volume of the first molecule, V_BIs the self-stacking volume of the second molecule;

the calculation mode of F5 is to find out the position of a hydrogen bond receptor in two molecules on the basis of the calculation mode of F1; calculating the volume of overlap of the respective hydrogen bond acceptors in the two moleculesWherein F_ijIs the congruent volume between the ith hydrogen bond acceptor and the jth hydrogen bond acceptor; calculating the overlapping volume of intermolecular hydrogen bond acceptor of two molecules under various overlapping conditionsWherein F_ijThe maximum of the volume of overlap between the ith hydrogen bond acceptor in the first molecule and the jth hydrogen bond acceptor in the second molecule, N being the total number of hydrogen bond acceptors in the first molecule and M being the total number of hydrogen bond acceptors in the second molecule, is selected(ii) a stacking volume that is the largest intermolecular hydrogen bond acceptor; calculating the similarity of hydrogen bond receptors of two moleculesWherein P is_AIs the self-stacking volume of the hydrogen bond acceptor in the first molecule, P_BIs the self-stacking volume of the hydrogen bond acceptor in the second molecule;

f6 was calculated in the same way as F5, only the hydrogen bond acceptor in both molecules needs to be replaced by a hydrogen bond donor;

f7 was calculated in the same manner as F5, except that the hydrogen bond acceptor was replaced with an aromatic ring;

f8 was calculated in the same way as F5, only the hydrogen bond acceptor needs to be replaced by a hydrophobic center;

f9 was calculated in the same way as F5, only the hydrogen bond acceptor needs to be replaced with a positively charged group;

f10 was calculated in the same way as F5, only the hydrogen bond acceptor needs to be replaced by a negatively charged group;

step two, training a deep learning model

Using a DUD-E dataset with 102 biological target information in the dataset, each target having a corresponding active subset and Decoy subset, the data for each target is processed as follows:

selecting crystal structure molecules in the active molecule set of each target point as template molecules, respectively calculating F1-F10 characteristic parameters of every two molecules in the template molecules and other molecules according to the calculation mode in the step one and all the molecules in the Decoy subset, so that each target point is calculated to obtain a set of characteristic parameter data;

modeling by using a deep learning method, taking the characteristic parameter data of each target point obtained by calculation as input data, taking the activity of the molecule as a target function of two categories, and optimizing the model in such a way that the error of the prediction of the activity of the molecule of all the target points is minimized, so that the average value of AUC values is maximized; after the training is finished, a final deep learning model is obtained;

step three, external verification of deep learning model

Verifying the generalization ability of the deep learning model by using an MUV data set, and selecting 10 biological target point information in the MUV data set, wherein each target point has a corresponding active subset and a Decoy subset; selecting crystal structure molecules in the active molecule set of each target point as template molecules, respectively calculating F1-F10 characteristic parameters of every two molecules in the template molecules and other molecules according to the calculation mode in the step one, with other molecules in the active molecule set of the target point and all molecules in the Decoy molecule set; and inputting the characteristic parameters into the trained deep learning model, and calculating to obtain an AUC value of each target point virtual screening.

2. The scoring method for three-dimensional similarity of molecules for virtual screening of drugs according to claim 1, characterized in that: in the first step, the three-dimensional structure information includes the total number of atoms in the molecule, the total number of chemical bonds, the type of each atom and the coordinate value thereof.

3. The scoring method for three-dimensional similarity of molecules for virtual screening of drugs according to claim 1, characterized in that: and in the second step, calculating the virtually screened AUC value of each target point by adopting a 5-fold similar cross validation mode.

Technical Field

The invention relates to the technical field of computer-aided drug research and development, in particular to a molecular three-dimensional similarity scoring method for drug virtual screening.

Background

The research and development of the medicine has the characteristics of large investment, high risk and long period, generally, one medicine research and development period is more than 10 years, the research and development investment is in hundreds of millions of dollars, and the trend of year-by-year increase is presented. Drug screening is a key link of drug discovery, and high-throughput drug virtual screening can greatly reduce screening time and cost, and has important significance in accelerating drug research and development.

In the virtual screening of drugs, a molecular screening and sorting method based on molecular three-dimensional similarity is commonly used at present. Such scoring methods usually include similarity of molecular shapes and similarity of pharmacophores (usually, a score is given by selecting similarity between a database molecule and a template molecule or a pharmacophore model), and a comprehensive score of similarity is formed by a simple weighting function, and the effectiveness of the scoring function determines the effectiveness and the calculation speed of screening. However, a significant problem with this type of scoring method is that it is not accurate enough to allow for higher false positives or false negatives for virtual drug screening.

In recent years, deep learning techniques based on artificial neural networks have made major breakthroughs in the fields of unmanned driving, speech recognition, image recognition and the like, and have also made important progress in the biomedical field, and diagnostic applications for diseases such as skin cancer, congenital cataract, infantile autism and the like based on deep learning have been developed. With the development of technology, the pharmaceutical field also begins to focus on the utilization of deep learning technology to accelerate the development of drugs so as to reduce the development cost. Research shows that the deep learning technology has more advantages in the aspects of optimizing a synthetic route, predicting a drug target and virtually screening compared with the traditional machine learning method.

Therefore, the application of deep learning technology in virtual drug screening is urgently needed to be researched to solve the problem of high false positive or false negative in the existing virtual drug screening process.

Disclosure of Invention

The invention aims to provide a molecular three-dimensional similarity scoring method for virtual screening of drugs aiming at the defects in the prior art, so as to solve the problem of poor accuracy of the existing virtual screening of drugs and simultaneously keep the high-throughput computing speed of screening.

The purpose of the invention is realized by the following technical scheme:

the molecular three-dimensional similarity scoring method for the virtual drug screening is provided, and comprises the following steps: the method comprises the following steps:

step one, obtaining various characteristic parameters of two molecules for similarity comparison

Respectively reading topological structure and three-dimensional structure information of two molecules for similarity comparison, and calculating to obtain various characteristic parameters, wherein the characteristic parameters comprise: difference in the number of atoms between the two molecules (F1); number of rotatable chemical bonds of two molecules (F2); volume difference of two molecules (F3); similarity of shape of the two molecules (F4); similarity of two molecular hydrogen bond acceptors (F5); similarity of two molecular hydrogen bond donors (F6); similarity of two molecular aromatic rings (F7); similarity of hydrophobic centers of the two molecules (F8); similarity of positively charged groups of the two molecules (F9); and similarity of the negatively charged groups of the two molecules (F10); wherein:

the calculation mode of F1 is to read in the respective topological structure information of two molecules, then to take the absolute value of the difference of the total number of atoms of the two molecules;

the calculation mode of F2 is that on the basis of the calculation mode of F1, whether each chemical bond is a rotatable bond is judged to obtain the total number of the rotatable bonds of two molecules, and then the absolute value of the difference value of the total number of the rotatable bonds of the two molecules is taken;

the calculation method of F3 is based on the calculation method of F1, and the Van der Waals radius of the atoms is obtained according to the types of the atoms in two molecules, each atom is represented by a Gaussian sphere, the radius of the Gaussian sphere is the same as the Van der Waals radius of the atom, the position coordinate of the Gaussian sphere is the same as the coordinate of the atom, and the atom coordinate is from the input three-dimensional structure of the molecule; calculating the superposition volume of the Gaussian ball groups in each molecule, wherein the ijth Gaussian ball group comprises the Gaussian ball corresponding to the ith atom and the jth atom pairThe corresponding Gauss sphere, the overlapping volume of the ijth Gauss sphere group is v_ij(ii) a Calculate the volume of overlap of each of the two molecules by itself asN is the total number of atoms in the molecule; then, taking the absolute value of the difference value of the superposed volumes of the two molecules;

the calculation method of F4 is to calculate the intermolecular superposition volume of two molecules under multiple superposition conditions based on the calculation method of F3Wherein v is_ijIs the congruent volume of the ith atom in the first molecule and the jth atom in the second molecule, N being the total number of atoms in the first molecule and M being the total number of atoms in the second molecule, the maximum of which is selectedAs the maximum intermolecular volume; calculating the similarity of the shapes of the two moleculesWherein V_AIs the self-stacking volume of the first molecule, V_BIs the self-stacking volume of the second molecule;

the calculation mode of F5 is to find out the position of a hydrogen bond receptor in two molecules on the basis of the calculation mode of F1; calculating the volume of overlap of the respective hydrogen bond acceptors in the two moleculesWherein F_ijIs the congruent volume between the ith hydrogen bond acceptor and the jth hydrogen bond acceptor; calculating the overlapping volume of intermolecular hydrogen bond acceptor of two molecules under various overlapping conditionsWherein F_ijIs the volume of overlap of the ith hydrogen bond acceptor in the first molecule with the jth hydrogen bond acceptor in the second molecule, and N is the hydrogen bond acceptor in the first moleculeM is the total number of hydrogen bond acceptors in the second molecule, the maximum of which is selected(ii) a stacking volume that is the largest intermolecular hydrogen bond acceptor; calculating the similarity of hydrogen bond receptors of two moleculesWherein P is_AIs the self-stacking volume of the hydrogen bond acceptor in the first molecule, P_BIs the self-stacking volume of the hydrogen bond acceptor in the second molecule;

f6 was calculated in the same way as F5, only the hydrogen bond acceptor in both molecules needs to be replaced by a hydrogen bond donor;

f7 was calculated in the same manner as F5, except that the hydrogen bond acceptor was replaced with an aromatic ring;

f8 was calculated in the same way as F5, only the hydrogen bond acceptor needs to be replaced by a hydrophobic center;

f9 was calculated in the same way as F5, only the hydrogen bond acceptor needs to be replaced with a positively charged group;

f10 was calculated in the same way as F5, only the hydrogen bond acceptor needs to be replaced by a negatively charged group;

step two, training a deep learning model

Using a DUD-E dataset with 102 biological target information in the dataset, each target having a corresponding active subset and Decoy subset, the data for each target is processed as follows:

selecting crystal structure molecules in the active molecule set of each target point as template molecules, respectively calculating F1-F10 characteristic parameters of every two molecules in the template molecules and other molecules according to the calculation mode in the step one and all the molecules in the Decoy subset, so that each target point is calculated to obtain a set of characteristic parameter data;

modeling by using a deep learning method, taking the characteristic parameter data of each target point obtained by calculation as input data, taking the activity of the molecule as a target function of two categories, and optimizing the model in such a way that the error of the prediction of the activity of the molecule of all the target points is minimized, so that the average value of AUC values is maximized; after the training is finished, a final deep learning model is obtained;

step three, external verification of deep learning model

Verifying the generalization ability of the deep learning model by using an MUV data set, and selecting 10 biological target point information in the MUV data set, wherein each target point has a corresponding active subset and a Decoy subset; selecting crystal structure molecules in the active molecule set of each target point as template molecules, respectively calculating F1-F10 characteristic parameters of every two molecules in the template molecules and other molecules according to the calculation mode in the step one, with other molecules in the active molecule set of the target point and all molecules in the Decoy molecule set; and inputting the characteristic parameters into the trained deep learning model, and calculating to obtain an AUC value of each target point virtual screening.

In the above technical solution, in the first step, the three-dimensional structure information includes a total number of atoms in the molecule, a total number of chemical bonds, a type of each atom, and a coordinate value thereof.

In the above technical scheme, in the second step, a 5-fold cross validation like mode is adopted to calculate the virtually screened AUC value for each target point.

The invention has the beneficial effects that:

the invention relates to a molecular three-dimensional similarity scoring method for virtual drug screening, which comprises the following steps of obtaining characteristic parameters of two molecules for similarity comparison, wherein the characteristic parameters mainly comprise atom number difference, rotatable chemical bond number, volume difference, shape similarity, similarity of hydrogen bond receptors, similarity of hydrogen bond donors, similarity of aromatic rings, similarity of hydrophobic centers, similarity of positive electric groups and negative electric groups and the like of the two molecules; step two, training a deep learning model; and step three, external verification of the deep learning model, and verification of the generalization ability of the model by adopting an MUV data set. According to the invention, the deep neural network is trained, the multi-dimensional similarity selected in the step one is taken as a characteristic parameter, and the final comprehensive score, namely the AUC value is given through the transformation of the deep neural network. The AUC value evaluation index of the virtual drug screening is a common standard for evaluating the accuracy of the screening method, the value range of the AUC value is between 0 and 1, and the closer the value is to 1, the more accurate the screening method is indicated. Experiments prove that compared with the prior art, the method has obvious improvement on the accuracy of medicament screening, and simultaneously keeps the calculation speed of high screening flux. Therefore, the scoring method of the invention can screen molecules with potential biological activity from a database containing a large amount of molecules as much as possible, and the more accurate the screening method is, the easier the molecules with potential activity can be found, so that any potential active molecules can be avoided as much as possible, thereby solving the problem of false positive or false negative. Therefore, the invention has very wide application prospect in the aspect of virtual screening of the drugs.

Detailed Description

The present invention will be described in further detail with reference to specific examples, but the present invention is not limited thereto.

Taking target ADA17 in the DUD-E dataset as an example, the DUD-E dataset has 102 biological target information, and each target has a corresponding active component subset and a corresponding Decoy component subset. Among these, the data set for the ADA17 target contained 1,341 active molecules and 35,900 Decoy subsets. Selecting active molecules in the crystal structure as template molecules (hereinafter referred to as 'molecules A') and first molecules in the active molecule set (hereinafter referred to as 'molecules B'), and carrying out the following operations on data:

step one, obtaining various characteristic parameters of two molecules for similarity comparison:

reading the topological structure and three-dimensional structure information (including total number of atoms in the molecule, total number of chemical bonds, type of each atom, coordinate value of each atom and the like) of the molecule A and the molecule B, and respectively calculating to obtain characteristic parameters for similarity comparison according to the following steps:

step 1, taking the absolute value of the difference of the total numbers of atoms of the molecules A and B to obtain a first characteristic parameter F1.

And 2, judging whether each chemical bond in the molecules A and B is a rotatable bond, respectively obtaining the total number of the rotatable bonds of the molecules A and B, and taking the absolute value of the difference of the total number of the molecules A and B to obtain a second characteristic parameter F2.

Step 3, obtaining the Van der Waals radius of each atom in the molecules A and B according to the type of each atom, wherein each atom is represented by a Gaussian sphere, the radius of each Gaussian sphere is the same as the Van der Waals radius of each atom, the position coordinate of each Gaussian sphere is the same as the coordinate of each atom, and the atom coordinate is from the input three-dimensional structure of each molecule; calculating the superposed volume of a group of Gauss spheres (hereinafter referred to as Gauss sphere group) corresponding to any two atoms in the molecule A, wherein the ijth Gauss sphere group comprises a Gauss sphere corresponding to the ith atom and a Gauss sphere corresponding to the jth atom in the molecule A, and the superposed volume of the ijth Gauss sphere group is v_ij(ii) a Calculate the overlap volume of molecule A itself asN is the total number of atoms in molecule A; the same method was used to calculate the overlap volume of molecule B itself asM is the total number of atoms in molecule B; the absolute value of the difference between the volumes of the overlapping molecules A and B themselves was then taken to give F3.

Step 4, calculating the intermolecular superposition volume of the molecules A and B under various superposition conditionsWherein v is_ijIs the volume of overlap of the ith atom in molecule A and the jth atom in molecule B, N is the total number of atoms in molecule A, M is the total number of atoms in molecule B, and the maximum value is selectedAs the maximum intermolecular volume; calculating the similarity of the shapes of the two moleculesWherein V_AIs the self-stacking volume of molecule A, V_BIs the self-stacking volume of molecule B (i.e. calculated in step 4).

Step 5, finding out the positions of all hydrogen bond receptors in the molecules A and B; calculation of the overlap volume of Hydrogen bond acceptors in molecule AWherein F_ijThe volume of the i-th hydrogen bond acceptor and the j-th hydrogen bond acceptor in the molecule A are superposed, and N is the total number of the hydrogen bond acceptors in the molecule A; the same procedure was used to calculate the overlap volume of hydrogen bond acceptors in molecule BWherein F_ijIs the superposed volume between the ith hydrogen bond acceptor and the jth hydrogen bond acceptor in the molecule B, and M is the total number of the hydrogen bond acceptors in the molecule B; calculating the overlapping volume of intermolecular hydrogen bond acceptor of the molecules A and B under various overlapping conditionsWherein F_ijThe maximum value of the volume of the i-th hydrogen bond acceptor in the molecule A and the j-th hydrogen bond acceptor in the molecule B is selected as the superposition volume of the i-th hydrogen bond acceptor in the molecule A and the j-th hydrogen bond acceptor in the molecule B, N is the total number of the hydrogen bond acceptors in the molecule A, and M is the total number of the hydrogen bond acceptors in the molecule BAs the stacking volume of hydrogen bond acceptors between molecules a and B; calculating the similarity of hydrogen bond receptors of the molecules A and BWherein P is_AIs the self-stacking volume of the hydrogen bond acceptor in molecule A, P_BIs the self-stacking volume of the hydrogen bond acceptor in molecule B.

Step 6, consistent with the calculation mode of step 5, the similarity F6 of the hydrogen bond acceptor of the molecules A and B can be obtained only by replacing the hydrogen bond donor with the hydrogen bond acceptor.

Step 7, in accordance with the calculation method of step 5, the aromatic ring similarity F7 of the molecules a and B can be obtained only by replacing the hydrogen bond donor with an aromatic ring.

Step 8, the calculation mode is consistent with that of step 5, and the similarity F8 of the hydrophobic centers of the molecules A and B can be obtained only by replacing the hydrogen bond donor with the hydrophobic center.

Step 9, consistent with the calculation mode of step 5, the similarity F9 of the positive groups of the molecules A and B can be obtained only by replacing the hydrogen bond donor with the positive group.

Step 10, consistent with the calculation mode of step 5, the similarity of the electronegative groups of the molecules A and B, F10, can be obtained only by replacing the hydrogen bond donor with the electronegative group.

Thus, 10 characteristic parameters F1-F10 of the molecule A and the first molecule in the active molecule set, namely the molecule B, are obtained.

Step two, training a deep learning model:

also taking target ADA17 in the DUD-E dataset as an example, the characteristic parameters of molecule A and the second molecule in the active molecule set (hereinafter referred to as "molecule C") were calculated in the same manner as in step one, and 10 corresponding characteristic parameters were obtained.

By analogy, the third, fourth and fifth molecules (N is a natural number) in the molecule a and the active molecule set are calculated respectively until all active molecules in the active molecule set are calculated, and the number of characteristic parameters is 10 × 1341-13,410.

Next, in the same calculation manner as in the first step, the characteristic parameters of each molecule in the molecule a and Decoy molecule set were calculated, and 10 × 35900 to 359,000 characteristic parameters were obtained. All characteristic parameters of the target ADA17 are calculated up to this point.

Then, the same calculation as above is used to calculate the feature parameter sets for the other 101 targets in the DUD-E dataset. To this end, 102 targets in the DUD-E dataset each had a set of characteristic parameter data.

And finally, taking the 102 sets of characteristic parameter data as input characteristic data of the deep learning model, taking the activity of the molecules as a target function of two categories, and optimizing the model in such a direction that the error of prediction of the activity of the molecules of all the target points is minimized, so that the average value of the AUC value is maximized. In the training process, a 5-fold cross validation similar mode is adopted to calculate the virtually screened AUC value (see table 1) of each target point, the average AUCaver of all 102 target point AUC values is taken, and the model optimization direction is to enable the AUCaver value to be maximum. And after the training is finished, a final deep learning model can be obtained.

TABLE 1 AUC values calculated for DUD-E data sets using conventional methods and the method of this example

Step three, external verification of the deep learning model:

and verifying the generalization ability of the deep learning model by adopting the MUV data set. Taking target 466 in the MUV dataset as an example, the 466 target dataset contains 31 active molecule sets and 15000 Decoy molecule sets. Active molecules in the crystal structure are selected as template molecules (hereinafter referred to as molecules A). And (5) obtaining characteristic parameters of the molecule A and each molecule in the active molecule set and the Decoy molecule set respectively by adopting a calculation mode in the step two, and obtaining a data set of 10 (31+15000) ═ 150310 characteristic parameters in total. Inputting the characteristic parameters into the trained deep learning model, and calculating the AUC value of the target virtual screening (see Table 2).

TABLE 2 AUC values calculated for MUV data sets using the conventional method and the method of this example

The AUC value evaluation index of the virtual drug screening is a common standard for evaluating the accuracy of the screening method, the value range of the AUC value is between 0 and 1, and the closer the value is to 1, the more accurate the screening method is indicated. The experimental results shown in tables 1 and 2 demonstrate that the present invention provides a significant improvement in drug screening accuracy over the prior art while maintaining the computational speed of screening high throughput. By adopting the scoring method, molecules with potential biological activity can be screened from a database containing a large number of molecules as much as possible, and the more accurate the screening method is, the easier the molecules with potential activity can be found, so that any potential active molecules can be avoided as much as possible, and the problem of false positive or false negative is solved.

The above-mentioned embodiments are only for convenience of description, and are not intended to limit the present invention in any way, and those skilled in the art will understand that the technical features of the present invention can be modified or changed by other equivalent embodiments without departing from the scope of the present invention.