Three-dimensional molecular structure editing method and device

Abstract

The invention relates to the technical field of molecular editing, and provides a three-dimensional molecular structure editing method and device, and the method comprises the steps: extracting a three-dimensional structure of a to-be-edited molecule from an original structure; mapping the three-dimensional structure of the to-be-edited molecule to a two-dimensional structure based on manifold embedding to obtain a two-dimensional structure of the to-be-edited molecule; in response to an editing result obtained by editing the two-dimensional structure of the to-be-edited molecule by a user, mapping the editing result back to three dimensions based on an inverse process of manifold embedding to obtain a mapping result; and obtaining an editing result according to the mapping result and the part, except the to-be-edited molecule, in the original structure. According to the method, the three-dimensional molecular structure is smoothly and continuously mapped to the two-dimensional plane in each part through the manifold embedding method, so that the structure mapping relation between the two-dimensional model and the three-dimensional model of the molecule to be edited is determined, and the information loss rate is low.

Description

Technical Field

[0001] The present invention relates to the technical field of molecular editing, and in particular to a method and device for editing a three-dimensional molecular structure. Background Art

[0002] Common three-dimensional molecular structure editing software, such as PyMol, Avogadro and other software, can edit molecules in a three-dimensional user interface; common two-dimensional molecular structure editing software, such as ChemDraw, Marvin and other software, can edit molecules in a two-dimensional user interface.

[0003] However, when using existing molecular structure editing software to convert between three-dimensional and two-dimensional structures, the coordinate information of atoms will be lost: for example, when using ChemDraw to convert a three-dimensional structure to a two-dimensional structure, the corresponding two-dimensional structure is generated only based on the topological structure of the molecule, and the three-dimensional coordinates are lost; for example, when using Chem3D to convert a two-dimensional structure to a three-dimensional structure, the conformation is searched only based on the topological structure of the molecule, and the correlation in the two-dimensional coordinates is lost.

[0004] In summary, the existing three-dimensional molecular structure editing technology has the problem of high information loss rate. Summary of the invention

[0005] The present invention provides a three-dimensional molecular structure editing method and device, which are used to solve the defect of high information loss rate in the prior art and realize three-dimensional molecular structure editing with low information loss rate.

[0006] The present invention provides a three-dimensional molecular structure editing method, comprising the following steps: Extracting the three-dimensional structure of the molecule to be edited from the original structure; Mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; In response to obtaining an editing result obtained by a user after editing the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on an inverse process of the manifold embedding to obtain a mapping result; An editing result is obtained according to the mapping result and the portion of the original structure except the molecule to be edited.

[0007] According to a three-dimensional molecular structure editing method provided by the present invention, the three-dimensional structure of the molecule to be edited is mapped to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited, which specifically includes: Calculating the nearest neighbor atoms of each atom of the molecule to be edited to obtain a nearest neighbor atom set; The two-dimensional structure of the molecule to be edited is obtained according to the nearest neighbor atom set.

[0008] According to a three-dimensional molecular structure editing method provided by the present invention, calculating the nearest neighbor atom of each atom of the molecule to be edited specifically includes: Screening, according to first data in a pre-constructed biochemical information database, atoms that can be connected to the target atom of the molecule to be edited through a chemical bond as nearest neighbor atoms; The first data at least includes structural information of standard amino acids, standard nucleic acids, standard sugars and registered small molecules.

[0009] According to a three-dimensional molecular structure editing method provided by the present invention, calculating the nearest neighbor atom of each atom of the molecule to be edited specifically includes: Calculate the sum of the interatomic distance and the covalent radius between the target atom of the molecule to be edited and the pre-selected candidate atom according to the second data in the pre-constructed biochemical information database; Calculating the difference between the interatomic distance and the sum of the covalent radii, and if the difference is within a preset range, taking the candidate atom as the nearest neighbor atom; The second data at least includes the covalent radius of each element.

[0010] According to a three-dimensional molecular structure editing method provided by the present invention, calculating the nearest neighbor atom of each atom of the molecule to be edited specifically includes: When the target atom is an alkaline earth metal element or a transition metal element, a target matrix is ​​constructed according to the atomic spatial arrangement within a preset range around the target atom; wherein the target matrix is ​​composed of a plurality of three-dimensional matrices describing the atomic spatial arrangement; Inputting the target matrix into a pre-built classification model to obtain a classification result; When the classification result is that the target atom is a complex ion, the opponent atom participating in the complex is taken as the nearest neighbor atom; The classification model is obtained by training a convolutional neural network using a target matrix of atomic samples.

[0011] According to a three-dimensional molecular structure editing method provided by the present invention, obtaining an editing result according to the mapping result and a portion of the original structure except the molecule to be edited specifically includes: Merging the mapping result and the portion of the original structure except the molecule to be edited to obtain a merged result; The merging result is optimized to obtain an editing result.

[0012] According to a three-dimensional molecular structure editing method provided by the present invention, the merging result is optimized to obtain an editing result, which specifically includes: Calculating the total energy of the combined result and the energy gradient of each atom; The three-dimensional coordinates of the atoms are taken as independent variables, and the total energy is taken as the optimization target to perform optimization so as to minimize the total energy and obtain the editing result.

[0013] The present invention also provides a three-dimensional molecular structure editing device, comprising the following modules: A molecule extraction unit, used to extract the three-dimensional structure of the molecule to be edited from the original structure; A first mapping unit, used for mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; A second mapping unit is used for, in response to obtaining an editing result obtained by a user after editing the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; An editing result unit is used to obtain an editing result according to the mapping result and the part of the original structure except the molecule to be edited.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the three-dimensional molecular structure editing methods described above is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the three-dimensional molecular structure editing methods described above.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the three-dimensional molecular structure editing methods described above.

[0017] The three-dimensional molecular structure editing method and device provided by the present invention extract the three-dimensional structure of the molecule to be edited from the original structure; map the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; in response to obtaining the editing result obtained by the user after editing the two-dimensional structure of the molecule to be edited, map the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; and obtain the editing result based on the mapping result and the part of the original structure except the molecule to be edited. The present invention uses the manifold embedding method to smoothly and continuously map the three-dimensional molecular structure to the two-dimensional plane in each part, thereby determining the structural mapping relationship between the two-dimensional model and the three-dimensional model of the molecule to be edited, avoiding the overlapping display of atoms and chemical bonds on the two-dimensional plane caused by the overall molecular structure caused by the mapping method in the prior art, and the information loss rate is low. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 This is one of the flow diagrams of the three-dimensional molecular structure editing method provided by the present invention; Figure 2 This is the second schematic diagram of the process of the three-dimensional molecular structure editing method provided by the present invention; Figure 3 This is the third flow chart of the three-dimensional molecular structure editing method provided by the present invention; Figure 4 This is the fourth flow chart of the three-dimensional molecular structure editing method provided by the present invention; Figure 5 It is a three-view drawing of a molecule to be edited in one embodiment of the three-dimensional molecular structure editing method provided by the present invention; Figure 6 is a schematic diagram of a two-dimensional structure of a molecule to be edited in one embodiment of the three-dimensional molecular structure editing method provided by the present invention; Figure 7 This is a schematic diagram of mapping results of an embodiment of the three-dimensional molecular structure editing method provided by the present invention; Figure 8 is a three-view drawing of a molecule to be edited according to another embodiment of the three-dimensional molecular structure editing method provided by the present invention; Fig. 9 is a schematic diagram of a two-dimensional structure of a molecule to be edited in another embodiment of the three-dimensional molecular structure editing method provided by the present invention; Fig.10 is a schematic diagram of mapping results of another embodiment of the three-dimensional molecular structure editing method provided by the present invention; Fig.11 It is a schematic diagram of the structure of the three-dimensional molecular structure editing device provided by the present invention; Fig.12 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Combine the following Figure 1-Figure 10 Describe the three-dimensional molecular structure editing method of the present invention, Figure 1 is one of the flow charts of the three-dimensional molecular structure editing method provided by the present invention, such as Figure 1 As shown, the method includes: Step 110: extracting the three-dimensional structure of the molecule to be edited from the original structure.

[0022] It should be noted that the original structure is the original three-dimensional structure of the molecule to be edited selected by the user. The molecule to be edited is the molecule that the user needs to edit.

[0023] The user determines the molecule to be edited by selection, and the molecule to be edited is separated from the original structure according to the user's selection, so as to obtain the three-dimensional structure of the molecule to be edited.

[0024] Furthermore, the three-dimensional structure of the molecule to be edited may be a digital three-dimensional model, or a set of three-dimensional coordinates of several atoms of the molecule to be edited, and the present invention does not impose any limitation on this.

[0025] The present invention aims to select a specified molecular structure using a three-dimensional user operation interface, so that it is displayed as a planar structure that is easy to observe and edit in a two-dimensional user operation interface. After editing in the two-dimensional user operation interface, the three-dimensional structure is updated accordingly while maintaining the original three-dimensional coordinates and the relative relationship of the two-dimensional coordinates.

[0026] The following is an example to further illustrate the meaning of "maintaining the original three-dimensional coordinates and maintaining the relative relationship of the two-dimensional coordinates". Figure 2 As shown, Figure 2 The three-dimensional structure of two water molecules (denoted as A and B) is shown. Another water molecule (denoted as C) is added by editing the two-dimensional structure. Figure 3 As shown, we expect to obtain a three-dimensional structure of three water molecules, such as Figure 4 As shown, the three-dimensional coordinates of water molecules A and B remain unchanged, while the newly added water molecule C has its oxygen atom facing water molecule A, and its hydrogen atom facing water molecule B, which is consistent with the two-dimensional structure.

[0027] Step 120: Mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited.

[0028] In step 120, the separated three-dimensional structure is mapped into a two-dimensional structure using manifold embedding. The present invention does not limit the specific manifold embedding method used, and a manifold embedding method such as local linear embedding, distance-preserving feature mapping, and local tangent space alignment can be selected. The function of this method in the present invention is to convert three-dimensional coordinates into two-dimensional coordinates.

[0029] It should be noted that in molecular modeling, usually only the editing of local structures is concerned. For example, for protein structures, usually only the editing of a local amino acid structure is concerned; for small molecule compounds, usually only the editing of a local functional group structure is concerned. If the overall three-dimensional structure is simply mapped to a two-dimensional plane using trivial methods such as parallel projection or perspective projection, the overall configuration of the molecule will cause its overlap on the two-dimensional plane, which is not conducive to the user's editing of its structure. On the other hand, for any local part, atoms can only form chemical bonds with a small number of atoms, so projecting it to a two-dimensional plane in a trivial way can still make it easy for users to edit its structure. This relationship between the whole and the local can correspond to the concept of manifold in mathematics. For example, a smooth manifold in a high-dimensional space can always be mapped to a low-dimensional space smoothly and continuously locally. If the mapping is reversible, it is further guaranteed that there will be no overlap during the mapping process. In the present invention, the manifold embedding method is used to smoothly and continuously map the three-dimensional molecular structure to the two-dimensional plane in each part, and such mapping is reversible, thereby avoiding the overlapping display of atoms and chemical bonds on the two-dimensional plane caused by the overall molecular structure caused by ordinary mapping methods.

[0030] It should be emphasized that after the two-dimensional structure of the molecule to be edited is obtained in step 120, the two-dimensional structure is displayed on the two-dimensional user operation interface, waiting for the user to edit the two-dimensional structure.

[0031] Step 130: in response to obtaining the editing result obtained by the user after editing the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result.

[0032] In step 130 , in response to the user completing the editing of the two-dimensional structure of the molecule to be edited, the two-dimensional structure edited by the user is mapped back to the three-dimensional structure through the inverse process of the manifold embedding in step 120 , that is, a mapping result is obtained.

[0033] The following takes the local linear embedding method as an example to further illustrate the specific steps of step 120 to step 130: 1) Assume that in step 110, a set of three-dimensional coordinates of several atoms is separated. , where n represents the number of atoms.

[0034] 2) Order is the coefficient of the i-th atom and the j-th nearest neighbor atom, and satisfies the normalization condition , where k is a constant; let is the atom closest to the jth atom and the ith atom. , let the summation expression Minimum.

[0035] 3) Let the set of two-dimensional coordinates of all atoms be , where n represents the number of atoms, and the optimization , let the summation expression Minimum, in the formula Indicates the atomic The nearest atom.

[0036] 4) Collect As output in the main editing step 120, the two-dimensional structure of the molecule to be edited.

[0037] 5) Mark the two-dimensional coordinates after the user completes editing as a collection , where m represents the number of atoms.

[0038] 6) Order is the coefficient of the i-th atom and the j-th nearest neighbor atom, and satisfies the normalization condition , where k is a constant; let is the atom closest to the jth atom and the ith atom. , let the summation expression Minimum.

[0039] 7) After editing, the three-dimensional coordinates of the atom are marked as a set , where m represents the number of atoms, and the three-dimensional coordinates are missing due to user editing. , by the expression Given, where is the jth atom that is closest to the ith atom.

[0040] 8) Collect The output of step 130 is the mapping result.

[0041] Step 140: Obtain an editing result according to the mapping result and the portion of the original structure except the molecule to be edited.

[0042] In step 140, the mapped three-dimensional structure is merged with the original molecular structure that does not require editing, that is, the final editing result is obtained, and the final editing result is returned to the user, for example, by displaying the final editing result through a three-dimensional user operation interface.

[0043] Compared with the structure editing software or method based on the three-dimensional user operation interface in the prior art, the three-dimensional molecular structure editing method provided by the present invention has the advantages of easy observation and editing. For example, for the determination of chiral atoms in a molecule, the use of a planar chemical formula in a wedge form or the like in a two-dimensional user operation interface helps users with certain chemical experience to quickly determine the chirality of the atom; while in a three-dimensional user operation interface, the user needs to drag and rotate the molecular structure to determine the chirality of the atom.

[0044] Compared with the structure editing software or method based on the two-dimensional user operation interface in the prior art, the three-dimensional molecular structure editing method provided by the present invention has the advantage that the generated three-dimensional atomic coordinates are close to the original atomic coordinates. For example, for a small molecule with an existing three-dimensional structure, when it is necessary to replace or derivatize the functional group at a specified position while maintaining its three-dimensional structural coordinates, the method according to the present invention can effectively generate the corresponding structure; however, the existing editing software or method needs to use constrained optimization and other methods after generating the three-dimensional structural coordinates to obtain a structure close to that of the present method.

[0045] The two-dimensional model provided to the user in the present invention is determined by the three-dimensional model provided by the user, and the structural mapping relationship between the two-dimensional model and the three-dimensional model is finally determined by the method claimed in the present invention. There is no need to use a pre-built structural mapping relationship. For the specific three-dimensional model provided by the user, a more suitable structural mapping relationship between the two-dimensional model and the three-dimensional model can be generated. There is no need to exhaustively construct possible structural mapping relationships, thereby having better generalization capabilities in the application of molecular structure modeling.

[0046] The following further describes step 120. In some embodiments, mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited specifically includes: Step 121: Calculate the nearest neighbor atoms of each atom of the molecule to be edited to obtain a nearest neighbor atom set; Step 122: Obtain the two-dimensional structure of the molecule to be edited according to the nearest neighbor atom set.

[0047] Specifically, in step 121, the nearest neighbor atoms are a set of atoms that must be obtained by the above-mentioned manifold embedding algorithm.

[0048] In some embodiments, the manifold embedding algorithm mostly uses the ball tree algorithm, a general algorithm for multi-point problems, to calculate the nearest atoms. It can be understood that the ball tree algorithm is used to find a number of points in space with the shortest Euclidean distance from a certain point. For example, the corresponding implementation of the manifold embedding algorithm in the Python software package scikit-learn uses the ball tree algorithm.

[0049] In other embodiments, as a variant of the widely used manifold algorithm, a knowledge-based expert system is used to solve the set of the nearest atoms of a certain atom. This algorithm variant is more suitable for the field of biochemistry. The present invention will further illustrate the knowledge-based expert system in subsequent embodiments.

[0050] In step 122, the two-dimensional structure of the molecule to be edited is obtained according to the nearest neighbor atom set.

[0051] Based on the above embodiment, the knowledge-based expert system is further described. Specifically, the knowledge-based expert system includes the following three independent algorithm units that calculate in parallel to solve the nearest neighbor atom set of a certain atom.

[0052] The algorithm unit 1 is used to ensure that the known molecule gives a set of nearest neighbor atoms according to its covalent bond connection mode. In this embodiment, the nearest neighbor atoms of each atom of the molecule to be edited are calculated, specifically including: Screening, according to first data in a pre-constructed biochemical information database, atoms that can be connected to the target atom of the molecule to be edited through a chemical bond as nearest neighbor atoms; The first data at least includes structural information of standard amino acids, standard nucleic acids, standard sugars and registered small molecules.

[0053] Specifically, the structure information of standard amino acids, standard nucleic acids, standard sugars and registered small molecules in a pre-constructed biochemical information database (collectively referred to as the first data) is used to compare the names of atoms with the structures provided by the user, and the atoms connected to the target atoms by chemical bonds are selected as the nearest neighbor atoms.

[0054] The present invention does not impose any specific limitation on the biochemical information database, which may be a public database.

[0055] Furthermore, the selected nearest neighbor atoms form a set, which is used as a return value and exits this algorithm unit.

[0056] Algorithm unit 2 is used to ensure that the common small molecule compound gives a set of nearest neighbor atoms according to its covalent bond connection mode. In this embodiment, the nearest neighbor atoms of each atom of the molecule to be edited are calculated, specifically including: Calculate the sum of the interatomic distance and the covalent radius between the target atom of the molecule to be edited and the pre-selected candidate atom according to the second data in the pre-constructed biochemical information database; Calculating the difference between the interatomic distance and the sum of the covalent radii, and if the difference is within a preset range, taking the candidate atom as the nearest neighbor atom; The second data at least includes the covalent radius of each element.

[0057] Specifically, using the information about the covalent radii of various elements in a pre-constructed biochemical information database (recorded as the second data), for a given atomic pair consisting of a target atom and a pre-selected candidate atom, the difference between the interatomic distance and the sum of the covalent radii of the two is calculated, and if the difference is within a preset range, the opponent candidate atom is regarded as the nearest neighbor atom.

[0058] The present invention does not specifically limit the biochemical information database, which may be a public database. The present invention does not limit the preset range, which may be set according to actual conditions.

[0059] Furthermore, after examining all atoms within a certain cutoff range, the selected nearest neighbor atoms are grouped into a set, which is used as the return value and the algorithm unit is exited.

[0060] Algorithm unit 3 is used to ensure that the complex gives a set of nearest neighbor atoms in a coordination bond connection mode, and non-complexed ions will not involve this algorithm unit. In this embodiment, the nearest neighbor atoms of each atom of the molecule to be edited are calculated, specifically including: When the target atom is an alkaline earth metal element or a transition metal element, a target matrix is ​​constructed according to the atomic spatial arrangement within a preset range around the target atom; wherein the target matrix is ​​composed of a plurality of three-dimensional matrices describing the atomic spatial arrangement; Inputting the target matrix into a pre-built classification model to obtain a classification result; When the classification result is that the target atom is a complex ion, the opponent atom participating in the complex is taken as the nearest neighbor atom; The classification model is obtained by training a convolutional neural network using a target matrix of atomic samples.

[0061] Specifically, for alkaline earth metals and transition metal elements, the atomic spatial arrangement within a certain range around a given atom is gridded into a four-dimensional matrix to obtain a target matrix.

[0062] It should be noted that the elements in the target matrix represent the number of atoms of a certain element contained in a space of a certain length, width and height. The spatial arrangement of an element generates a three-dimensional matrix, and all three-dimensional matrices are stacked as channels into a four-dimensional matrix, namely the target matrix.

[0063] Afterwards, the pre-trained classification model is run on the atom given by the user to predict whether it is a complex ion. For complex ions, the opponent atoms involved in the complex are taken as the nearest neighbor atoms.

[0064] Furthermore, the selected nearest neighbor atoms form a set, which is used as a return value and exits this algorithm unit.

[0065] In addition, the pre-built classification model is obtained by training the target matrix of a large number of atomic samples based on a convolutional neural network, and the model output is a binary variable, and the classification meaning is whether it is a complex ion. The present invention does not limit its specific training process, and the convolutional neural network includes but is not limited to models such as VGG16 and VGG19.

[0066] The following further describes step 140. In some embodiments, obtaining the editing result according to the mapping result and the portion of the original structure excluding the molecule to be edited specifically includes: Step 141: merging the mapping result and the portion of the original structure except the molecule to be edited to obtain a merged result; Step 142: Optimize the merging result to obtain an editing result.

[0067] Specifically, after the mapped three-dimensional structure is merged with the original molecular structure that does not require editing, the merged result may be optimized before the editing result is output.

[0068] In the specific optimization process, the following methods can be used: 1) Rule-based optimization: Check that the bond lengths, bond angles, and dihedral angles in the molecule are within reasonable ranges, and ensure that all atoms have the correct type and charge state.

[0069] 2) Energy-based optimization: Use molecular mechanics force fields (such as AMBER, CHARMM, etc.) to minimize energy to reduce stress inside the molecule. Or simulate the dynamic behavior of the molecule within a given time to make the molecular structure reach the lowest energy state.

[0070] 3) Knowledge-based optimization: Refer to the known structures of similar molecules to correct the structure of the edited molecule. Or use the trained model to predict the rationality of the molecular structure and optimize it accordingly.

[0071] 4) Expert system-based optimization: Experienced chemists manually adjust the molecular structure based on their chemical knowledge and experience, or use an auxiliary decision support system to optimize the structure in combination with the chemist's experience.

[0072] The following further describes step 142. In some embodiments, optimizing the merging result to obtain the editing result specifically includes: Calculating the total energy of the combined result and the energy gradient of each atom; The three-dimensional coordinates of the atoms are taken as independent variables, and the total energy is taken as the optimization target to perform optimization so as to minimize the total energy and obtain the editing result.

[0073] Specifically, molecular mechanics methods, quantum chemistry methods or machine learning methods are used to obtain the total energy of the entire system and the energy gradient of each atom. The total energy is taken as the optimization target and the three-dimensional coordinates of the atoms are taken as the independent variables to optimize the total energy to minimize it.

[0074] Furthermore, in some embodiments, when optimizing with total energy as the optimization target, a multivariable optimization algorithm may be used, including but not limited to the steepest gradient descent, conjugate gradient descent, Newton's method and other multivariable optimization algorithms.

[0075] In order to provide a more detailed description of the three-dimensional molecular structure editing method provided by the present invention, two specific embodiments of performing three-dimensional molecular structure editing using the three-dimensional molecular structure editing method provided by the present invention are given below.

[0076] In one embodiment, the crystal structure 8SI0 of cyclic guanosine monophosphate-adenosine monophosphate (cGAMP) synthase is taken as an example, in which the small molecule ligand cGAMP is arranged in a U-shape, and a non-overlapping two-dimensional structure cannot be obtained by parallel projection. The three-view drawing is as follows: Figure 5 The editor system implemented based on the present invention is embedded in the two-dimensional user operation interface through the manifold to obtain a two-dimensional structure as shown in FIG. Figure 6 By adding a carbon atom in the two-dimensional user operation interface, the editor system implemented based on the present invention maps its two-dimensional coordinates to the three-dimensional structure, and the following is obtained: Figure 7 The three-dimensional structure shown is marked with a blue frame.

[0077] In another embodiment, taking the crystal structure of fatty acid synthase 1LPA as an example, the small molecule ligand dilauroylphosphatidylcholine is arranged in a nearly linear manner, and it is difficult to obtain a non-overlapping two-dimensional structure by parallel projection. Figure 8 The editor system implemented based on the present invention is embedded in the two-dimensional user operation interface through the manifold to obtain a two-dimensional structure as shown in FIG. Fig. 9 By adding a methyl group in the two-dimensional user operation interface, the editor system implemented based on the present invention maps its two-dimensional coordinates to the three-dimensional structure, and obtains the following Fig.10 The three-dimensional structure shown is marked with a blue frame.

[0078] The three-dimensional molecular structure editing method provided by the present invention extracts the three-dimensional structure of the molecule to be edited from the original structure; maps the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; in response to obtaining the editing result obtained by the user after editing the two-dimensional structure of the molecule to be edited, maps the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; and obtains the editing result based on the mapping result and the part of the original structure except the molecule to be edited. The present invention uses the manifold embedding method to smoothly and continuously map the three-dimensional molecular structure to the two-dimensional plane in each part, thereby determining the structural mapping relationship between the two-dimensional model and the three-dimensional model of the molecule to be edited, avoiding the overlapping display of atoms and chemical bonds on the two-dimensional plane caused by the overall molecular structure caused by the mapping method in the prior art, and the information loss rate is low.

[0079] The three-dimensional molecular structure editing device provided by the present invention is described below. The three-dimensional molecular structure editing device described below and the three-dimensional molecular structure editing method described above can be referenced to each other. Fig.11 Schematic diagram of the structure of the three-dimensional molecular structure editing device provided by the present invention, such as Fig.11 As shown, the device includes the following modules: A molecule extraction unit 1110, used to extract the three-dimensional structure of the molecule to be edited from the original structure; A first mapping unit 1120, configured to map the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; A second mapping unit 1130 is configured to, in response to obtaining an editing result obtained by a user after editing the two-dimensional structure of the molecule to be edited, map the editing result back to three dimensions based on an inverse process of the manifold embedding to obtain a mapping result; The editing result unit 1140 is used to obtain the editing result according to the mapping result and the part of the original structure except the molecule to be edited.

[0080] According to a three-dimensional molecular structure editing device provided by the present invention, the three-dimensional structure of the molecule to be edited is mapped to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited, which specifically includes: Calculating the nearest neighbor atoms of each atom of the molecule to be edited to obtain a nearest neighbor atom set; The two-dimensional structure of the molecule to be edited is obtained according to the nearest neighbor atom set.

[0081] According to a three-dimensional molecular structure editing device provided by the present invention, calculating the nearest neighbor atom of each atom of the molecule to be edited specifically includes: Screening, according to first data in a pre-constructed biochemical information database, atoms that can be connected to the target atom of the molecule to be edited through a chemical bond as nearest neighbor atoms; The first data at least includes structural information of standard amino acids, standard nucleic acids, standard sugars and registered small molecules.

[0082] According to a three-dimensional molecular structure editing device provided by the present invention, calculating the nearest neighbor atom of each atom of the molecule to be edited specifically includes: Calculate the sum of the interatomic distance and the covalent radius between the target atom of the molecule to be edited and the pre-selected candidate atom according to the second data in the pre-constructed biochemical information database; Calculating the difference between the interatomic distance and the sum of the covalent radii, and if the difference is within a preset range, taking the candidate atom as the nearest neighbor atom; The second data at least includes the covalent radius of each element.

[0083] According to a three-dimensional molecular structure editing device provided by the present invention, calculating the nearest neighbor atom of each atom of the molecule to be edited specifically includes: When the target atom is an alkaline earth metal element or a transition metal element, a target matrix is ​​constructed according to the atomic spatial arrangement within a preset range around the target atom; wherein the target matrix is ​​composed of a plurality of three-dimensional matrices describing the atomic spatial arrangement; Inputting the target matrix into a pre-built classification model to obtain a classification result; When the classification result is that the target atom is a complex ion, the opponent atom participating in the complex is taken as the nearest neighbor atom; The classification model is obtained by training a convolutional neural network using a target matrix of atomic samples.

[0084] According to a three-dimensional molecular structure editing device provided by the present invention, obtaining an editing result according to the mapping result and a part of the original structure except the molecule to be edited specifically includes: Merging the mapping result and the portion of the original structure except the molecule to be edited to obtain a merged result; The merging result is optimized to obtain an editing result.

[0085] According to a three-dimensional molecular structure editing device provided by the present invention, optimizing the merging result to obtain the editing result specifically includes: Calculating the total energy of the combined result and the energy gradient of each atom; The three-dimensional coordinates of the atoms are taken as independent variables, and the total energy is taken as the optimization target to perform optimization so as to minimize the total energy and obtain the editing result.

[0086] The three-dimensional molecular structure editing device provided by the present invention extracts the three-dimensional structure of the molecule to be edited from the original structure; maps the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; in response to obtaining the editing result obtained by the user after editing the two-dimensional structure of the molecule to be edited, maps the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; and obtains the editing result based on the mapping result and the part of the original structure except the molecule to be edited. The present invention uses the manifold embedding method to smoothly and continuously map the three-dimensional molecular structure to the two-dimensional plane in each part, thereby determining the structural mapping relationship between the two-dimensional model and the three-dimensional model of the molecule to be edited, avoiding the overlapping display of atoms and chemical bonds on the two-dimensional plane caused by the overall molecular structure caused by the mapping method in the prior art, and the information loss rate is low.

[0087] Fig.12 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.12 As shown, the electronic device may include: a processor 1210, a communication interface 1220, a memory 1230 and a communication bus 1240, wherein the processor 1210, the communication interface 1220 and the memory 1230 communicate with each other through the communication bus 1240. The processor 1210 may call the logic instructions in the memory 1230 to execute the three-dimensional molecular structure editing method, which includes: extracting the three-dimensional structure of the molecule to be edited from the original structure; mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; in response to obtaining the editing result obtained by the user editing the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; and obtaining the editing result according to the mapping result and the part of the original structure except the molecule to be edited.

[0088] In addition, the logic instructions in the above-mentioned memory 1230 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0089] On the other hand, the present invention also provides a computer program product, which includes a computer program, and the computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the three-dimensional molecular structure editing method provided by the above methods, and the method includes: extracting the three-dimensional structure of the molecule to be edited from the original structure; mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; in response to obtaining the editing result obtained by the user after completing the editing of the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; and obtaining the editing result based on the mapping result and the part of the original structure except the molecule to be edited.

[0090] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the three-dimensional molecular structure editing method provided by the above-mentioned methods, the method comprising: extracting the three-dimensional structure of the molecule to be edited from the original structure; mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; in response to obtaining the editing result obtained by the user after completing the editing of the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; and obtaining the editing result based on the mapping result and the part of the original structure except the molecule to be edited.

[0091] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0092] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A three-dimensional molecular structure editing method, characterized in that: include: Extracting the three-dimensional structure of the molecule to be edited from the original structure; Mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; In response to obtaining an editing result obtained by a user after editing the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on an inverse process of the manifold embedding to obtain a mapping result; An editing result is obtained according to the mapping result and the portion of the original structure except the molecule to be edited.

2. The three-dimensional molecular structure editing method according to claim 1, characterized in that: Based on the random forest model, the three-dimensional structure of the molecule to be edited is mapped to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited, specifically including: Calculating the nearest neighbor atoms of each atom of the molecule to be edited to obtain a nearest neighbor atom set; The two-dimensional structure of the molecule to be edited is obtained according to the nearest neighbor atom set.

3. The three-dimensional molecular structure editing method according to claim 2, characterized in that: Calculating the nearest neighbor atoms of each atom of the molecule to be edited, specifically including: Screening, according to first data in a pre-constructed biochemical information database, atoms that can be connected to the target atom of the molecule to be edited through a chemical bond as nearest neighbor atoms; The first data at least includes structural information of standard amino acids, standard nucleic acids, standard sugars and registered small molecules.

4. The three-dimensional molecular structure editing method according to claim 2, characterized in that: Calculating the nearest neighbor atoms of each atom of the molecule to be edited, specifically including: Calculate the sum of the interatomic distance and the covalent radius between the target atom of the molecule to be edited and the pre-selected candidate atom according to the second data in the pre-constructed biochemical information database; Calculating the difference between the interatomic distance and the sum of the covalent radii, and if the difference is within a preset range, taking the candidate atom as the nearest neighbor atom; The second data at least includes the covalent radius of each element.

5. The three-dimensional molecular structure editing method according to claim 2, characterized in that: Calculating the nearest neighbor atoms of each atom of the molecule to be edited, specifically including: When the target atom is an alkaline earth metal element or a transition metal element, a target matrix is ​​constructed according to the atomic spatial arrangement within a preset range around the target atom; wherein the target matrix is ​​composed of a plurality of three-dimensional matrices describing the atomic spatial arrangement; Inputting the target matrix into a pre-built classification model to obtain a classification result; When the classification result is that the target atom is a complex ion, the opponent atom participating in the complex is taken as the nearest neighbor atom; The classification model is obtained by training a convolutional neural network using a target matrix of atomic samples.

6. The three-dimensional molecular structure editing method according to claim 1, characterized in that: Obtaining an editing result according to the mapping result and a portion of the original structure except the molecule to be edited, specifically comprising: Merging the mapping result and the portion of the original structure except the molecule to be edited to obtain a merged result; The merging result is optimized to obtain an editing result.

7. The three-dimensional molecular structure editing method according to claim 6, characterized in that: Optimizing the merging result to obtain an editing result specifically includes: Calculating the total energy of the combined result and the energy gradient of each atom; The three-dimensional coordinates of the atoms are taken as independent variables, and the total energy is taken as the optimization target to perform optimization so as to minimize the total energy and obtain the editing result.

8. A three-dimensional molecular structure editing device, characterized in that: include: A molecule extraction unit, used to extract the three-dimensional structure of the molecule to be edited from the original structure; A first mapping unit, used for mapping the three-dimensional structure of the molecule to be edited to two dimensions based on manifold embedding to obtain the two-dimensional structure of the molecule to be edited; A second mapping unit is used for, in response to obtaining an editing result obtained by a user after editing the two-dimensional structure of the molecule to be edited, mapping the editing result back to three dimensions based on the inverse process of the manifold embedding to obtain a mapping result; An editing result unit is used to obtain an editing result according to the mapping result and the part of the original structure except the molecule to be edited.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the three-dimensional molecular structure editing method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the three-dimensional molecular structure editing method according to any one of claims 1 to 7 is implemented.

Images