Molecular docking method and apparatus based on coherent Ising machine

Abstract

The present invention relates to a molecular docking method and apparatus based on a coherent Ising machine, which belongs to the field of drug design technology and solves the problem that the conventional heuristic algorithm takes a long time but cannot obtain a standard solution. The method includes the steps of obtaining a ligand diagram by constructing a three-dimensional molecular structure diagram based on a selected drug molecule, obtaining a receptor diagram by constructing a pseudo-atom point diagram inside a receptor target based on a selected receptor molecule, constructing a ligand-receptor similarity diagram including a plurality of vertices and a plurality of edges between any two vertices based on the ligand diagram and the receptor diagram, and determining whether any one of the plurality of vertices of the ligand-receptor similarity diagram includes a point of the ligand diagram and a point of the receptor diagram, and each edge exists, and determining whether the selected drug molecule inhibits the activity of the selected receptor by constructing a pharmacophore model based on the ligand-receptor similarity diagram, and the pharmacophore model calculates the maximum weight group between the selected drug molecule and the selected receptor, thereby selecting a drug molecule compound. The pharmacophore model solves molecular simulation problems more quickly and accurately.

Description

[Technical field]

[0001] The present invention relates to the field of drug design technology, and in particular to a molecular docking method and apparatus based on a coherent Ising machine. [Background technology]

[0002] Traditional drug screening is a very expensive and resource-intensive process, generally costing billions of dollars, with a success rate of only over 10%. In recent years, with the development of powerful molecular modeling tools and the increasing number of analyzed structures of protein-small molecule complexes, structure-based drug design has become an indispensable tool in new drug development. The focus of molecular docking research is to simulate the molecular recognition process through computation. It aims to minimize the free energy of the entire system by simulating the optimal conformation between the protein and the ligand. It is a very important part in the early stage of drug screening, and molecular docking can accelerate the drug research and development process.

[0003] Conventional molecular docking software typically uses different algorithms to explore and sample conformational space and evaluate binding states using scoring functions. Commonly used software includes GOLD, VINA, etc.

[0004] Most of the traditional model selection methods in the pharmaceutical field adopt heuristic algorithms and systematic search algorithms, which take a long time and may not be able to calculate the optimal solution, and the false positive rate is high in the drug research and development process. The disadvantages are as follows:

[0005] Heuristic algorithms take a lot of time for iterations, may not find the optimal solution, and are computationally intensive.

[0006] When using a system search algorithm, one may be faced with the problem of combinatorial explosion, and conventional computers have relatively poor capabilities in dealing with such problems.

[0007] Conventional models require a large number of samples to obtain a lower energy conformation, which may not be the global optimum. The model used in this application can solve the problem more efficiently and quickly than conventional methods. Summary of the Invention [Problem to be solved by the invention]

[0008] In view of the above, the embodiments of the present application relate to a molecular docking method and apparatus based on a coherent Ising machine, which solves problems such as the fact that conventional heuristic algorithms take a long time but may not be able to obtain a standard solution. [Means for solving the problem]

[0009] In one aspect, an embodiment of the present application provides a molecular docking method based on a coherent Ising machine, which includes the steps of: obtaining a ligand diagram by selecting a drug molecule from a drug molecule library waiting for screening and constructing a three-dimensional molecular structure diagram based on the selected drug molecule; obtaining a receptor diagram by selecting a receptor from a receptor library and constructing a pseudo-atom point diagram inside a receptor target based on the selected receptor molecule; constructing a ligand-receptor similarity diagram based on the ligand diagram and the receptor diagram, the ligand-receptor similarity diagram includes a plurality of vertices and a plurality of edges, any one of the plurality of vertices includes a point in the ligand diagram and a point in the receptor diagram, each edge is an edge between any two vertices, and determining whether each edge exists; determining whether the selected drug molecule inhibits the activity of the selected receptor by constructing a pharmacophore model based on the ligand-receptor similarity diagram, and the pharmacophore model is used to screen a drug molecule compound by calculating a maximum weight group between the selected drug molecule and the selected receptor.

[0010] The beneficial effects of the above-mentioned technical solution are as follows: the method of solving molecular simulation problems using the pharmacophore model constructed based on the ligand-receptor similarity map in this application is faster and more accurate. The pharmacophore model can select drug molecule compounds by calculating the maximum weight group between drug molecules and receptors, and can determine whether the selected drug molecules inhibit the activity of the selected receptor.

[0011] According to a further refinement of the method, determining whether each edge exists further comprises determining whether an edge exists between any two vertices of the plurality of vertices according to a distance between the two vertices.

[0012] According to a further improvement of the method, determining whether an edge exists between any two of the plurality of vertices based on the distance between the any two vertices further comprises determining a first distance between a point of a ligand diagram in a first vertex and a point of a ligand diagram in a second vertex, determining a second distance between a point of a receptor diagram in the first vertex and a point of a receptor diagram in the second vertex, and determining that an edge exists between the any two vertices when the first distance and the second distance are both less than a predetermined distance.

[0013] In a further refinement of the method, the predetermined distance is in the range of 0.1 Angstroms to 2 Angstroms.

[0014] A further refinement of the method is that the maximum weight group between the selected drug molecule and the selected receptor corresponds to the atomic similarity between the selected drug molecule and the selected receptor.

[0015] A further refinement of the method comprises determining a maximum weight group between the selected drug molecule and the selected receptor according to the following formula: [ka] Here, w(i,j) is the weight of the vertex x(i,j) in the ligand-receptor similarity diagram and is determined by the type to which the atom / group corresponding to i,j belongs, w(i,j),(k,l) is the weight of the edge between x(i,j) and x(k,l) in the ligand-receptor similarity diagram, and the K1 term is the coefficient of the maximum weight group.

[0016] A further refinement of the method comprises determining the pharmacophore model according to the following formula: [ka] In addition, w (i,j) is the vertex x in the ligand-receptor similarity diagram. (i,j) is the weight of i, j, which is determined by the type of atom / group that corresponds to i, j, and w(i,j),(k,l) is the x in the ligand-receptor similarity diagram. (i,j) and x (k,l) where K1 is the weight of the edges of the vectors, K2 and K3 are the Lagrange coefficients, and are used to constrain the number of connections of the nodes and the number of solutions.

[0017] In a further refinement of the method, the selected receptor comprises a protein, a nucleic acid or a polysaccharide and the selected drug molecule comprises aspirin, oseltamivir or remdesivir.

[0018] In another aspect, embodiments of the present application provide a molecular docking apparatus based on a coherent Ising machine. The apparatus includes a ligand selection module used for selecting a drug molecule from a drug molecule library waiting for selection, a ligand diagram construction module used for obtaining a ligand diagram by constructing a three-dimensional molecular structure diagram based on the selected drug molecule, a receptor selection module used for selecting a receptor from a receptor library, a receptor diagram construction module for obtaining a receptor diagram by constructing a pseudo atomic point diagram inside a receptor target based on the selected receptor molecule, a ligand receptor similarity diagram construction module for constructing a ligand receptor similarity diagram based on the ligand diagram and the receptor diagram, the ligand receptor similarity diagram including a plurality of vertices and a plurality of edges, any one of the plurality of vertices including one point of the ligand diagram and one point of the receptor diagram, each edge being an edge between any two vertices, and determining whether each edge exists, and a pharmacophore model for determining whether the selected drug molecule inhibits the activity of the selected receptor by constructing a pharmacophore model based on the ligand receptor similarity diagram, the pharmacophore model for selecting a drug molecule compound by calculating a maximum weight group between the selected drug molecule and the selected receptor.

[0019] According to a further improvement of the method, the ligand-receptor similarity diagram construction module determines whether an edge exists between any two vertices among the plurality of vertices based on the distance between the any two vertices, and the ligand-receptor similarity diagram construction module further comprises a first distance determination module, a second distance determination module, and an edge determination module, wherein the first distance determination module is used for determining a first distance between a point of the ligand diagram among the first vertex and a point of the ligand diagram among the second vertex, the second distance determination module is used for determining a second distance between a point of the receptor diagram among the first vertex and a point of the receptor diagram among the second vertex, and the edge determination module is used for determining that an edge exists between the any two vertices when the first distance and the second distance are both smaller than a predetermined distance.

[0020] Compared with the prior art, the present application can achieve at least one of the following beneficial effects:

[0021] 1. The method of solving molecular simulation problems using the pharmacophore model constructed based on the ligand-receptor similarity map of the present application is faster and more accurate. The pharmacophore model can screen drug molecule compounds by calculating the maximum weight group between drug molecules and receptors, and can determine whether the selected drug molecule inhibits the activity of the selected receptor.

[0022] 2. The maximum weight group between the selected drug molecule and the selected receptor corresponds to the atomic similarity between the selected drug molecule and the selected receptor. The present application utilizes the properties of entanglement, overlapping and full connection of quantum computers to propose a better model for determining molecular binding patterns, and provides it for users to use by posting it on the Web.

[0023] 3. For the pharmaceutical field, this application is more convenient and rapid, and can provide more excellent potential pharmaceutical compounds.

[0024] In the present application, the above-mentioned technical solutions may be combined with each other to realize more preferred combinations. Additional features and advantages of the present application will be described hereinafter, and some of the advantages will be apparent from the description or may be learned by practice of the present application. The objectives and other advantages of the present application may be realized and obtained by means of the particular points pointed out in the description and drawings. [Brief description of the drawings]

[0025] The drawings are used only for the purpose of illustrating specific embodiments and are not to be construed as limitations on the present application. In all drawings, the same reference numerals represent the same elements. [Figure 1] FIG. 2 is a flow chart diagram of a molecular docking method based on a coherent Ising machine according to an embodiment of the present application. [Diagram 2] FIG. 1 shows two-dimensional molecular structure diagrams of selected drug molecules. [Diagram 3] FIG. 1 shows the construction of a three-dimensional molecular structure based on a selected drug molecule. [Figure 4] FIG. 2 is a schematic diagram showing pseudo atomic points inside a receptor target according to an embodiment of the present application. [Diagram 5] FIG. 1 is a pseudo-atom diagram of a protein target structure with pseudo-atom points according to an embodiment of the present application. [Figure 6] FIG. 1 is a schematic diagram of constructing the vertices of a ligand-receptor similarity map according to an embodiment of the present application. [Figure 7] FIG. 2 is a ligand-receptor analogy diagram according to an embodiment of the present application. [Figure 8] FIG. 2 is an overall flowchart of a molecular docking method based on a coherent Ising machine according to an embodiment of the present application. [Figure 9] FIG. 1 is a block diagram of a molecular docking apparatus based on a coherent Ising machine according to an embodiment of the present application. [Figure 10A] FIG. 2 is a schematic diagram of a ligand structure according to a first specific example of the present application. [Figure 10B] FIG. 2 is a ligand diagram according to a first specific example of the present application. [Figure 11A] FIG. 2 is a pseudo-atomic diagram according to a first specific example of the present application. [Figure 11B] FIG. 2 is a receptor diagram according to a first specific example of the present application. [Figure 12] FIG. 2 is a diagram showing the three-dimensional structure of MMP9 and its inhibitor after docking according to the first specific example of the present application. [Figure 13A] FIG. 2 is a schematic diagram of a ligand structure according to a second specific example of the present application. [Figure 13B] FIG. 2 is a ligand diagram according to a second specific example of the present application. [Figure 14A] FIG. 2 is a pseudo-atomic diagram according to a second specific example of the present application. [Figure 14B] FIG. 2 is a receptor diagram according to a second specific example of the present application. [Figure 15] FIG. 2 is a three-dimensional structure diagram of mouse thiamine pyrophosphate kinase and thiamine after docking according to a second specific example of the present application. [Figure 16A] FIG. 2 is a schematic diagram of the ligand structure of the third specific example of the present application. [Figure 16B] FIG. 2 is a ligand diagram of a third specific example of the present application. [Figure 17A] FIG. 2 is a pseudo-atomic diagram of a third specific example of the present application. [Figure 17B] FIG. 2 is a receptor diagram of a third specific example of the present application. [Figure 18] FIG. 2 is a three-dimensional structure diagram of the docking of IMP-1 metallo-β-lactamase from Pseudomonas aeruginosa with a biaryl succinic acid inhibitor according to a third specific example of the present application. [Figure 19] A hardware diagram of a quantum computer. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] Hereinafter, preferred embodiments of the present application will be described in detail with reference to the drawings. The drawings constitute a part of the present application, and are used to explain the principles of the present application together with the embodiments of the present application, and are not intended to limit the scope of the present application.

[0027] A specific embodiment of the present application provides a molecular docking method based on a coherent Ising machine. As shown in FIG. 1, the molecular docking method based on a coherent Ising machine includes steps S102, S104, S106, and S108. In step S102, a drug molecule is selected from a drug molecule library waiting for screening, and a three-dimensional molecular structure diagram is constructed based on the selected drug molecule to obtain a ligand diagram, the ligand diagram being circular and including a plurality of atoms and a plurality of edges in the drug molecule, each edge being an edge between any two atoms. In step S104, a receptor is selected from a receptor library, and a pseudo-atom point diagram inside the receptor target is constructed based on the selected receptor to obtain a receptor diagram, the receptor diagram being circular and including a plurality of atoms and a plurality of edges in the receptor molecule, each edge being an edge between any two atoms. In step S106, a ligand-receptor similarity diagram is constructed based on the ligand diagram and the receptor diagram, the ligand-receptor similarity diagram includes a plurality of vertices and a plurality of edges, and any one of the plurality of vertices includes a point in the ligand diagram and a point in the receptor diagram, and each edge is an edge between any two vertices, and whether each edge exists is determined. In step S108, a pharmacophore model is constructed based on the ligand-receptor similarity diagram to determine whether the selected drug molecule inhibits the activity of the selected receptor, and the pharmacophore model selects a drug molecule compound by calculating the maximum weight group between the selected drug molecule and the selected receptor. A receptor is a polymeric substance that can recognize and selectively bind to a certain type of ligand (signal molecule).

[0028] Compared with the prior art, the method of solving molecular simulation problems using the pharmacophore model constructed based on the ligand-receptor similarity map of the present embodiment is faster and more accurate, and the pharmacophore model can screen drug molecule compounds by calculating the maximum weight group between the drug molecule and the receptor, and can determine whether the selected drug molecule inhibits the activity of the selected receptor.

[0029] The present method of solving molecular simulation problems using pharmacophore models is faster and more accurate.

[0030] In the following, steps S102 to S108 of the molecular docking method based on the coherent Ising machine according to the embodiment of the present application will be described in detail as shown in FIGS.

[0031] In step S102, a drug molecule (see FIG. 2) is selected from the drug molecule library waiting for selection, and a three-dimensional molecular structure diagram (see FIG. 3) is constructed based on the selected drug molecule to obtain a ligand diagram. Specifically, a three-dimensional molecular structure diagram (drug molecule structure diagram) is constructed based on the selected drug molecule, and then a ligand diagram is generated based on the three-dimensional molecular structure diagram. As shown in FIG. 10B, FIG. 13B, and FIG. 16B, the ligand diagram is circular and includes multiple atoms and multiple edges in the drug molecule, and each edge is between any two atoms. The selected drug molecule includes aspirin, oseltamivir, remdesivir, etc. For example, an MMP9 inhibitor, thiamine or biarylsuccinic acid inhibitor, etc. is selected from the drug molecule library waiting for selection.

[0032] In step S104, a receptor is selected from the receptor library, and a schematic diagram of pseudo-atomic points (see FIG. 4) inside the receptor target is constructed based on the selected receptor molecular structure to obtain a receptor diagram. Specifically, the pseudo-atomic points of the receptor are obtained using AutoSite (software), and a receptor diagram is obtained based on the pseudo-atomic points. As shown in FIG. 11B, FIG. 14B, and FIG. 17B, the receptor diagram is circular, and includes a plurality of atoms and a plurality of edges in the receptor molecule, and each edge is an edge between any two atoms. The selected receptor is a protein, a nucleic acid, a polysaccharide, etc. For example, MMP9, mouse thiamine pyrophosphate kinase, or IMP-1 metallo-β-lactamase from Pseudomonas aeruginosa is selected from the receptor library. FIG. 5 shows the pseudo-atomic point diagram and the pseudo-atomic diagram of the protein target structure. For example, as shown in FIG. 4 and FIG. 5, if the receptor is a protein, assuming that the pseudo-atomic points are points in the protein cavity, the drug molecule is matched with the receptor to determine whether the selected drug molecule inhibits the activity of the selected receptor.

[0033] In step S106, a ligand-receptor similarity diagram is constructed based on the ligand diagram and the receptor diagram, the ligand-receptor similarity diagram includes a plurality of vertices and a plurality of edges, and as shown in FIG. 6, the plurality of vertices are traversed through all points in the ligand diagram and the receptor diagram, and any one of the vertices includes one point in the ligand diagram and one point in the receptor diagram, and each edge is an edge between any two vertices, and determining whether each edge exists. Specifically, determining whether each edge exists further includes determining whether an edge exists between any two vertices according to a distance between any two vertices among the plurality of vertices. Determining whether an edge exists between any two vertices according to a distance between any two vertices among the plurality of vertices further includes determining a first distance between a point in the ligand diagram in the first vertex and a point in the ligand diagram in the second vertex, determining a second distance between a point in the receptor diagram in the first vertex and a point in the receptor diagram in the second vertex, and determining that an edge exists between any two vertices when the first distance and the second distance are both smaller than a predetermined distance. The predetermined distance is in the range of 0.1 Angstroms to 2 Angstroms, and in a preferred embodiment, the predetermined distance is in the range of 0.5 Angstroms to 1 Angstrom.

[0034] In step S108, a pharmacophore model is constructed based on the ligand-receptor similarity diagram to determine whether the selected drug molecule inhibits the activity of the selected receptor, and the pharmacophore model calculates the maximum weight group (or maximum group) between the selected drug molecule and the selected receptor to screen the drug molecule compound. Specifically, the maximum weight group between the selected drug molecule and the selected receptor corresponds to the similarity of atoms between the selected drug molecule and the selected receptor. The maximum group is usually to find the complete diagram with the most number of one points in the undirected diagram. For example, as shown in FIG. 7, the maximum group in the ligand-receptor similarity diagram is the group with the most number of vertices in the ligand-receptor similarity diagram.

[0035] The following formula determines the maximum weight group between the selected drug molecule and the selected receptor:

[0036] [ka] In addition, w (i,j) is the vertex x in the ligand-receptor similarity diagram. (i,j) is the weight of i, j, which is determined by the type to which the atom / group corresponding to i and j belongs, that is, by the similarity between the two atoms. For example, the similarity between two atoms such as (C,C) may be 1, and the similarity between two atoms such as (C,O) may be (0.8), and w (i,j),(k,l) is the x in the ligand-receptor similarity diagram. (i,j) and x (k,l) is the edge weight of x (i,j) and x (k,l) If there is an edge connecting x, the edge weight is 1, and (i,j) and x (k,l) If there is no connecting edge, the edge weight is 0, and the K1 term is a coefficient of the maximum weight group, which can be artificially adjusted for that coefficient K1.

[0037] The pharmacophore model is calculated according to the following formula:

[0038] [ka] In addition, w (i,j) is the vertex x in the ligand-receptor similarity diagram. (i,j) is the weight of i, j, which is determined by the type of atom / group that corresponds to i, j, and w (i,j),(k,l) is the x in the ligand-receptor similarity diagram. (i,j) and x (k,l) where K1 is the weight of the edges of the vectors, K2 and K3 are the Lagrange coefficients, and are used to constrain the number of connections of the nodes and the number of solutions.

[0039] Another specific embodiment of the present application provides a molecular docking apparatus based on a coherent Ising machine. As shown in Figure 9, the molecular docking apparatus based on the coherent Ising machine includes a ligand selection module 902, a ligand map construction module 904, a receptor selection module 906, a receptor map construction module 908, a ligand-receptor similarity map construction module 910 and a pharmacophore model 912.

[0040] The ligand selection module 902 is used to select a drug molecule from the drug molecule library waiting for selection. The ligand diagram construction module 904 is used to obtain a ligand diagram by constructing a three-dimensional molecular structure diagram based on the selected drug molecule. The ligand is circular and includes multiple atoms and multiple edges in the drug molecule, and each edge is an edge between any two atoms. The receptor selection module 906 is used to select a receptor from the receptor library. The receptor diagram construction module 908 is used to obtain a receptor diagram by constructing a pseudo-atom point diagram inside the receptor target based on the selected receptor, the receptor diagram is circular and includes multiple atoms and multiple edges in the receptor molecule, and each edge is an edge between any two atoms. The ligand receptor similarity diagram construction module 910 is used to construct a ligand receptor similarity diagram based on the ligand diagram and the receptor diagram, the ligand receptor similarity diagram includes multiple vertices and multiple edges, and any one vertex among the multiple vertices includes a point in the ligand diagram and a point in the receptor diagram, and each edge is an edge between any two vertices, and determines whether each edge exists. Specifically, the ligand-receptor similarity map construction module determines whether an edge exists between any two vertices according to the distance between any two vertices among the multiple vertices. The ligand-receptor similarity map construction module further includes a first distance determination module, a second distance determination module, and an edge determination module, where the first distance determination module is used to determine a first distance between a point of the ligand map in the first vertex and a point of the ligand map in the second vertex, the second distance determination module is used to determine a second distance between a point of the receptor map in the first vertex and a point of the receptor map in the second vertex, and the edge determination module determines that an edge exists between any two vertices when the first distance and the second distance are both smaller than a predetermined distance.The pharmacophore model 912 determines whether the selected drug molecule inhibits the activity of the selected receptor by constructing a pharmacophore model based on the ligand-receptor similarity map, and the pharmacophore model is used to screen the drug molecule compounds by calculating the maximum weight group between the selected drug molecule and the selected receptor.

[0041] In the following, the molecular docking method based on the coherent Ising machine according to the embodiment of the present application will be described in detail with reference to the drawings and specific examples.

[0042] This application uses a coherent Ising machine (CIM) to convert the important molecular docking problem in the drug screening process into the QUBO model (Quadratic Unconstrained Binary Optimization model). Users can convert molecular 3D models into mathematical diagrams, express ligands and receptors in the form of diagrams, and use the QUBO model through mathematical modeling to realize prediction of ligand-receptor binding patterns. Users upload different small molecule and protein structures, and the back-end server converts them, and the optimal calculation results are given by the coherent Ising machine, and the 3D model is presented to the user. For drug screening problems, the coherent Ising machine is faster and more accurate than traditional computers.

[0043] molecule exhibition First, a molecular display is performed and the obtained molecular three-dimensional structure (either a crystal structure or a structure after energy minimization) is invented. Then, the molecular 3D structure is simplified into the form of a mathematical diagram, and the simplified structure is represented as a distance matrix / weighted edge diagram G. L , G P Presented as:

[0044] Constructing a pseudo-atomic map inside the receptor target A series of pseudo atom maps was generated using AutoSite and used to represent the atom maps most likely to occur in the receptor target region.

[0045] Constructing a ligand-receptor similarity map The receptor and ligand are graphed according to the above procedure, and the resulting ligand-receptor similarity graph is G, whose set of points is as follows:

[0046] [ka] In addition, V P is a point on the receptor diagram, and V L is a point on the ligand diagram, and x (i,j) The weight of is determined by the types of atoms represented by i and j.

[0047] The edge between vertices x(i,j) and x(k,l) of the ligand-receptor similarity diagram (i.e., the binding interaction diagram) is determined by the distance between (i,k) and (j,l). If the distance between (i,k) and the distance between (j,l) are less than 0.5 angstroms, it is determined that there is an edge between these two vertices; otherwise, there is no edge between these two vertices.

[0048] Then, we simplify the problem to find the maximum group in the ligand-receptor similarity diagram. Given any diagram G=(V, E), if U ⊆ V and for any two vertices u,v∈U there is also |u,v|∈E, then U is said to be a complete subdiagram of G, a complete subdiagram of G is said to be a group of G, and the maximum group of G refers to the maximum complete subdiagram of G, and the problem of finding the maximum group of any diagram is an NP-hard problem. The atomic similarity between the ligand and the receptor corresponding to the maximum group in the binding interaction diagram is the optimal docking method based on the weights (the weights are calculated by historical data, and the quality of the final optimal docking is determined by the quality of the weight calculation method).

[0049] Mathematical Model The ligand diagram is: [ka] The receptor diagram is: [ka] The ligand-receptor analogy diagram is: [ka] Maximize: [ka] The solution satisfies the constraints (each atom of the ligand / receptor overlaps with at most one atom of the receptor / ligand) and w (i,j) is the vertex x (i,j) is the weight of i,j, which is determined by the type to which the atom / group corresponding to i,j belongs.

[0050] w (i,j),(k,l) is the x in the ligand-receptor similarity diagram. (i,j) and x (k,l) is the edge weight of x (i,j) and x (k,l) If there is an edge connecting (i,j),(k,l) = 1, and x (i,j) and x (k,l) If there is no connecting edge, w (i,j),(k,l) = 0. x (i,j) and x (k,l) Whether there is a connecting edge between (i,j) and (k,l) is determined by the distance between (i,j) and (k,l).

[0051] Converting to the QUBO model, in the following equation, the K1 term is the coefficient of the maximum weight group, and K2 and K3 are Lagrange coefficients used to constrain the number of node connections and the number of solutions.

[0052] [ka] 5. As shown in Figure 8, when the user uses the system, the user only needs to input the selected drug library file and the target protein structure file. The initial structure and pharmacophore structure are displayed on the user side, which is convenient for the user to perform visual three-dimensional operation. After that, once the target range is determined, the data is transmitted to the server. The server builds a QUBO model based on the above parameters and transmits it to a coherent Ising machine (quantum computer) to calculate the optimal solution. The entire process takes much less time when using a quantum computer. In addition, the user can view the selected structural model in the user interface and score it. Then, the next experimental verification is performed using these molecules to obtain lead compounds.

[0053] As shown in Figure 19, the laser is used as the pump light source, and is split into two beams by the first coupler, one of which passes through the second harmonic device crystal to undergo the frequency doubling process, and the other beam is used as the injection light. The frequency-doubled light is injected into the fiber loop, and at the same time, the optical parametric oscillator (OPO) is excited to perform the optical parametric oscillation process, generating vacuum-compressed light pulses, which circulate in the fiber loop. At the third coupler side, a part of the light is split and homodyne frequency detection is performed to obtain the compressed phase information of the light pulse, which is provided to the FPGA as a partial input. The data (J ij ) are simultaneously input into the FPGA as programming data, and the calculation unit outputs the modulation signal to the intensity modulator and the phase modulator. One light of the first coupler is modulated by IM and PM, passes through the second coupler and is input into the optical fiber loop again to realize the pulse injection. When the calculation starts, the output of the pump laser gradually increases, and according to the minimum gain principle, the mutually dependent pulse signals will undergo a phase change, resulting in phase collapse, which is measured by the photodetector (PD), and then the input J ij If the design is satisfied, an optimized combination of computation results can be obtained.

[0054] In a first example, the structure of MMP9 in complex with its inhibitor is as follows:

[0055] (1) Select an MMP9 inhibitor from the drug molecule library awaiting screening, construct a three-dimensional molecular structure diagram (see Figure 10A) based on the MMP9 inhibitor, and then obtain a ligand diagram (see Figure 10B) based on the three-dimensional molecular structure diagram, where the ligand diagram includes 22 atoms in the three-dimensional molecular structure diagram and a number of edges between any two atoms, and each edge is an edge between any two atoms among the 22 atoms.

[0056] (2) Select MMP9 from the receptor library, obtain a pseudo-atom diagram (see FIG. 11A) based on MMP9 using AutoSite, and then construct a receptor diagram (see FIG. 11B) based on the pseudo-atom diagram. The receptor diagram includes 51 atoms and multiple edges between any two atoms, where each edge is between any two atoms among the 51 atoms.

[0057] (3) Based on the ligand and receptor maps, vertices in the ligand-receptor similarity map were constructed, and some of the vertices are shown in Table 1.

[0058] [Table 1] (4) Add weights to the points in the diagram, and the weight determination can involve different parameter combinations, one of which is shown in Table 2.

[0059] [Table 2] (5) Determine whether to add an edge to a point in the diagram, and the criteria for this determination are as follows: When determining whether an edge exists between node (i.e., the vertex described above) A(L1,R2) and nodeB(L2,R2), determine whether the distance between (L1,L2) and the distance between (L1,R2) are between 0.1 and 2 angstroms, and if they are within this range, an edge exists between nodeA(L1,R2) and nodeB(L1,R2), and if they are greater than 2 angstroms, no edge exists between these two points.

[0060] (6) Construct the QUBO matrix and obtain the QUBO model.

[0061] (7) Get the maximum weight group by outputting the pairs that are [('l2','r45'),('l4','r25'),('l8','r31'),('l10','r25'),('l13','r45'),('l14','r18'),('l15','r11'),('l16','r28'),('l17','r39'),('l18','r37'),('l19','r28'),('l20','r16')].

[0062] (8) The three-dimensional structure is obtained using the Kabsch rotation matrix.

[0063] Compared with the original results, as shown in FIG. 12, the gray is the original ligand structure, and the black is the ligand structure after docking, and the difference between the two is very close.

[0064] In a second example, the structure of mouse thiamine pyrophosphate kinase and thiamine complex is as follows:

[0065] (1) Select thiamine from the drug molecule library waiting for screening, construct a three-dimensional molecular structure diagram (see Figure 13A) based on thiamine, and then obtain a ligand diagram (see Figure 13B) based on the three-dimensional molecular structure diagram, where the ligand diagram includes 18 atoms in the three-dimensional molecular structure diagram and a number of edges between any two atoms, where each edge is an edge between any two atoms among the 18 atoms.

[0066] (2) Select mouse thiamine pyrophosphate kinase from the receptor library, obtain a pseudo-atom diagram (see FIG. 14A) based on mouse thiamine pyrophosphate kinase using AutoSite, and then construct a receptor diagram (see FIG. 14B) based on the pseudo-atom diagram. The receptor diagram includes 54 atoms and multiple edges between any two atoms, where each edge is between any two atoms among the 54 atoms.

[0067] (3) Based on the ligand and receptor maps, vertices in the ligand-receptor similarity map were constructed, and some of the vertices are shown in Table 3.

[0068] [Table 3] (4) Add weights to the points in the diagram, and the weighting decision may involve different parameter combinations, one of which is shown in Table 4.

[0069] [Table 4] (5) Determine whether to add an edge to a point in the diagram. The criteria for this determination are as follows: When determining whether an edge exists between nodeA(L1,R2) and nodeB(L2,R2), determine whether the distance between (L1,L2) and the distance between (L1,R2) are between 0.1 and 2 angstroms. If they are within this range, an edge exists between nodeA(L1,R2) and nodeB(L2,R2), and if they are greater than 2 angstroms, no edge exists between these two points.

[0070] (6) Construct the QUBO matrix and obtain the QUBO model.

[0071] (7) Obtain the maximum weight group by outputting the pairs that are ('l3','r4'), ('l5','r37'), ('l6','r18'), ('l6','r51'), ('l8','r35'), ('l10','r20'), ('l11','r42'), ('l13','r20'), ('l15','r21').

[0072] (8) The three-dimensional structure is obtained using the Kabsch rotation matrix.

[0073] Compared with the original results, as shown in FIG. 15, the gray is the original ligand structure, and the black is the ligand structure after docking, and the difference between the two is very close.

[0074] In a third example, the structure of the IMP-1 metallo-β-lactamase from Pseudomonas aeruginosa complexed with a biarylsuccinic acid inhibitor (11) is shown below.

[0075] (1) Select a biaryl succinic acid inhibitor complex (11) from the drug molecule library waiting for screening, construct a three-dimensional molecular structure diagram (see Figure 16A) based on the biaryl succinic acid inhibitor complex (11), and then obtain a ligand diagram (see Figure 16B) based on the three-dimensional molecular structure diagram, where the ligand diagram includes 25 atoms in the three-dimensional molecular structure diagram and a number of edges between any two atoms, and each edge is an edge between any two atoms among the 25 atoms.

[0076] (2) Select IMP-1 metallo-β-lactamase derived from Pseudomonas aeruginosa from the receptor library, obtain a pseudo-atom diagram (see FIG. 17A) based on IMP-1 metallo-β-lactamase derived from Pseudomonas aeruginosa using AutoSite, and then construct a receptor diagram (see FIG. 17B) based on the pseudo-atom diagram. The receptor diagram includes 23 atoms and multiple edges between any two atoms, and each edge is an edge between any two atoms among the 23 atoms.

[0077] (3) Construct vertices in the ligand-receptor similarity map based on the ligand map and the receptor map, and some of the vertices are shown in Table 5.

[0078] [Table 5] (4) Add weights to the points in the diagram, and the weighting decision may involve different parameter combinations, one of which is shown in Table 6.

[0079] [Table 6] (5) Determine whether to add an edge to a point in the diagram. The criteria for this determination are as follows: When determining whether an edge exists between nodeA (L1, R2) and nodeB (L2, R2), determine whether the distance between (L1, L2) and the distance between (R1, R2) are between 0.1 and 2 angstroms. If they are within this range, an edge exists between nodeA (L1, R2) and nodeB (L2, R2). If they are greater than 2 angstroms, no edge exists between these two points.

[0080] (6) Construct the QUBO matrix and obtain the QUBO model.

[0081] (7) Get the maximum weight group by outputting the pairs that are [('l0','r11'),('l2','r20'),('l5','r16'),('l6','r19'),('l9','r21'),('l17','r12'),('l19','r3'),('l21','r3')].

[0082] (8) The three-dimensional structure is obtained using the Kabsch rotation matrix.

[0083] Compared with the original results, as shown in FIG. 18, the gray is the original ligand structure, and the black is the ligand structure after docking, and the difference between the two is very close.

[0084] Key points of this application This application develops a computing device that can be used for drug screening, accelerated by a quantum computer, and can rapidly calculate drug-protein affinity, helping researchers screen potential lead compounds and aiding the drug development process.

[0085] Benefits of this Application Compared with conventional molecular docking models, this model can solve problems such as heuristic algorithms taking a long time but not being able to obtain standard solutions, and the method of solving molecular simulation problems with the Ising model used in this specification is faster and more accurate. This specification proposes a better model for obtaining molecular binding patterns by utilizing the characteristics of entanglement states, overlap states, and full connections of quantum computers, and provides it for users to use by placing it on the Web. For the pharmaceutical field, this application is more convenient and faster, and can obtain better potential pharmaceutical compounds.

[0086] Compared with conventional techniques, the advantage of the present invention is that it is possible to directly obtain the three-dimensional structure of the ligand and receptor in their bound state, which is the POSE.

[0087] Those skilled in the art should understand that the implementation of all or part of the method steps of the above-mentioned embodiments can be accomplished by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, such as a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0088] The above is merely a preferred embodiment of the present invention, and the scope of the claims of the present invention is not limited thereto. Any conversion or replacement that a person skilled in the art can easily conceive within the technical scope disclosed in the present application falls within the scope of the claims of the present application.

Claims

1. obtaining a ligand diagram by selecting a drug molecule from a library of drug molecules waiting to be screened and constructing a three-dimensional molecular structure diagram based on the selected drug molecule; obtaining a receptor map by selecting a receptor from a receptor library and constructing a pseudo-atom map of the interior of the receptor target based on the selected receptor molecule; constructing a ligand-receptor similarity diagram based on the ligand diagram and the receptor diagram, the ligand-receptor similarity diagram including a plurality of vertices and a plurality of edges, any one of the plurality of vertices including one point of the ligand diagram and one point of the receptor diagram, each edge being an edge between any two vertices, and determining whether each edge exists; determining whether the selected drug molecule inhibits the activity of the selected receptor by constructing a pharmacophore model based on the ligand-receptor similarity map, and the pharmacophore model selects a drug molecule compound by calculating a maximum weight group between the selected drug molecule and the selected receptor.

2. 2. The method for molecular docking based on a coherent Ising machine according to claim 1, wherein determining whether each edge exists further comprises determining whether an edge exists between any two vertices among the plurality of vertices based on a distance between the any two vertices.

3. Determining whether an edge exists between any two vertices among the plurality of vertices based on a distance between the any two vertices includes: determining a first distance between a point of the ligand diagram within a first vertex and a point of the ligand diagram within a second vertex; determining a second distance between a point on the receptor map within the first vertex and a point on the receptor map within the second vertex; 3. The molecular docking method based on a coherent Ising machine according to claim 2, further comprising: determining that an edge exists between any two vertices when the first distance and the second distance are both less than a predetermined distance.

4. The method for molecular docking based on a coherent Ising machine according to claim 3, wherein the predetermined distance is in the range of 0.1 angstroms to 2 angstroms.

5. 2. The molecular docking method based on the coherent Ising machine according to claim 1, wherein the maximum weight group between the selected drug molecule and the selected receptor corresponds to the atomic similarity between the selected drug molecule and the selected receptor.

6. determining a maximum weight group between the selected drug molecule and the selected receptor according to the following formula: 【Chemistry 1】 In addition, （ｉ，ｊ） is the vertex x in the ligand-receptor similarity diagram. （ｉ，ｊ） is the weight of i, j, which is determined by the type to which the atom / group corresponding to i and j belongs; （ｉ，ｊ），（ｋ，ｌ） is the x in the ligand-receptor analogy diagram. （ｉ，ｊ） and x （ｋ，ｌ） is the edge weight of K １ The method of claim 3, wherein the term is a coefficient of the largest weight group.

7. The pharmacophore model is calculated according to the following formula: 【Chemistry 2】 In addition, （ｉ，ｊ） is the vertex x in the ligand-receptor similarity diagram. （ｉ，ｊ） is the weight of i, j, which is determined by the type to which the atom / group corresponding to i and j belongs; （ｉ，ｊ），（ｋ，ｌ） is the x in the ligand-receptor analogy diagram. （ｉ，ｊ） and x （ｋ，ｌ） is the edge weight of K １ The term is the coefficient of the largest weight group, K ２ , K ３ The method for molecular docking based on the coherent Ising machine according to claim 6, characterized in that, σ is a Lagrange coefficient, which is used to constrain the number of connections of the nodes and the number of solutions.

8. the selected receptor comprises a protein, a nucleic acid or a polysaccharide; The molecular docking method based on the coherent Ising machine according to claim 1, characterized in that the selected drug molecules include aspirin, oseltamivir or remdesivir.

9. a ligand selection module used to select drug molecules from a library of drug molecules awaiting selection; a ligand diagram construction module, which is used to obtain a ligand diagram by constructing a three-dimensional molecular structure diagram based on the selected drug molecule; a receptor selection module used to select a receptor from a receptor library; a receptor diagram construction module for constructing a pseudo-atomic point diagram inside the receptor target based on the selected receptor molecule to obtain a receptor diagram; a ligand-receptor similarity diagram construction module for constructing a ligand-receptor similarity diagram based on the ligand diagram and the receptor diagram, the ligand-receptor similarity diagram including a plurality of vertices and a plurality of edges, any one of the plurality of vertices including one point of the ligand diagram and one point of the receptor diagram, each edge being an edge between any two vertices, and determining whether each edge exists; determining whether the selected drug molecule inhibits the activity of the selected receptor by constructing a pharmacophore model based on the ligand-receptor similarity map, the pharmacophore model selecting a drug molecule compound by calculating a maximum weight group between the selected drug molecule and the selected receptor.

10. The ligand-receptor similarity map construction module is used to determine whether an edge exists between any two vertices according to a distance between the any two vertices among the plurality of vertices, and the ligand-receptor similarity map construction module further comprises a first distance determination module, a second distance determination module, and an edge determination module; the first distance determination module is used to determine a first distance between a point of the ligand diagram among the first vertices and a point of the ligand diagram among the second vertices; the second distance determination module is used to determine a second distance between a point on the receptor diagram of the first vertex and a point on the receptor diagram of the second vertex; 10. The molecular docking apparatus based on a coherent Ising machine according to claim 9, wherein the edge determination module determines that an edge exists between any two vertices when the first distance and the second distance are both smaller than a predetermined distance.

Images