Methods, apparatuses, electronic devices, and media of determining a conformation of a molecule

Abstract

Embodiments of the present disclosure relate to methods, apparatuses, electronic devices and media for determining conformations of a molecule. The method includes, based on a two-dimensional structure of the molecule, segmenting the molecule into a plurality of molecular fragments, at least one of the plurality of molecular fragments including a plurality of atoms. The method further includes determining a plurality of dihedral angles between adjacent ones of the plurality of molecular fragments. The method further includes determining a plurality of combinations of the plurality of molecular fragments and the plurality of dihedral angles as a plurality of conformations of the molecule. Through embodiments of the present disclosure, the degrees of freedom of conformation models that need to be modeled when predicting conformations of a molecule can be reduced, and the prediction diversity and prediction accuracy of conformations of a molecule can be improved.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computers, and more specifically, to methods, apparatus, electronic devices, and media for determining the conformation of molecules. Background Technology

[0002] Determining the three-dimensional structure of molecules based on their two-dimensional structure is a fundamental and crucial step in studying the activity and related physicochemical properties of drug molecules. In recent years, artificial intelligence (AI) technology has made groundbreaking progress in predicting the three-dimensional structure of molecules. Compared to traditional experimental methods such as crystal diffraction and cryo-electron microscopy, and computational simulations based on molecular dynamics and density functional theory, AI technology can resolve molecular three-dimensional structures with high throughput, thereby significantly accelerating the development of new drugs.

[0003] Therefore, conformational generation of molecules is an important task for predicting the three-dimensional structure of molecules. The main process of this task is to predict stable three-dimensional molecular conformations based on an input two-dimensional molecular structure diagram. Conformational generation tasks focus on the translational invariance of molecules in three-dimensional space, as well as the bond lengths and bond angles of the chemical bonds between atoms. Summary of the Invention

[0004] Embodiments of this disclosure provide a method, apparatus, electronic device, and computer-readable storage medium for determining the conformation of a molecule.

[0005] According to a first aspect of this disclosure, a method for determining the conformation of a molecule is provided. The method includes dividing the molecule into multiple molecular segments based on its two-dimensional structure, at least one of the molecular segments comprising multiple atoms. The method further includes determining multiple dihedral angles between adjacent molecular segments. The method also includes determining multiple combinations of the multiple molecular segments and the multiple dihedral angles as multiple conformations of the molecule.

[0006] According to a second aspect of this disclosure, an apparatus for determining the conformation of a molecule is provided. The apparatus includes a molecular segmentation module configured to segment the molecule into a plurality of molecular fragments based on its two-dimensional structure, at least one of the molecular fragments comprising a plurality of atoms. The apparatus also includes a dihedral angle determination module configured to determine a plurality of dihedral angles between adjacent molecular fragments. Furthermore, the apparatus includes a conformation combination module configured to determine multiple combinations of the plurality of molecular fragments and the plurality of dihedral angles as multiple conformations of the molecule.

[0007] According to a third aspect of this disclosure, an electronic device is provided. The electronic device includes a processor and a memory coupled to the processor, the memory having instructions stored therein, which, when executed by the processor, cause the electronic device to perform the method according to the first aspect.

[0008] According to a fourth aspect of this disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores one or more computer instructions, wherein the one or more computer instructions are executed by a processor to implement the method according to the first aspect.

[0009] The summary section is provided to present the chosen concepts in a simplified form, which will be further described in the detailed description below. The summary section is not intended to identify key or principal features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter. Attached Figure Description

[0010] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0011] Figure 1 A schematic diagram of an example environment in which a method for determining the conformation of a molecule according to some embodiments of the present disclosure may be implemented is shown;

[0012] Figure 2 A flowchart illustrating a method for determining the conformation of a molecule according to some embodiments of the present disclosure is shown;

[0013] Figure 3 A block diagram illustrating a process for determining the conformation of a molecule according to some embodiments of the present disclosure is shown;

[0014] Figure 4 A schematic diagram of a molecular conformational field according to some embodiments of the present disclosure is shown;

[0015] Figure 5A A flowchart illustrating a method for constructing a fragment library according to some embodiments of this disclosure is shown;

[0016] Figure 5B A block diagram illustrating the process of constructing a fragment library according to some embodiments of the present disclosure is shown.

[0017] Figure 6A A flowchart of a method for determining energy distribution according to some embodiments of the present disclosure is shown;

[0018] Figure 6B A block diagram illustrating a process for determining energy distribution according to some embodiments of the present disclosure is shown;

[0019] Figure 7 A block diagram illustrating a node sampling process according to some embodiments of the present disclosure is shown;

[0020] Figure 8A block diagram of an apparatus for determining the conformation of a molecule according to some embodiments of the present disclosure is shown; and

[0021] Figure 9 A block diagram of an apparatus for determining the conformation of a molecule according to some embodiments of the present disclosure is shown.

[0022] In all the accompanying drawings, the same or similar reference numerals indicate the same or similar elements. Detailed Implementation

[0023] The data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0024] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0025] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0026] Traditional methods for generating molecular conformations cannot handle the rotational and translational invariance of molecular conformations in three-dimensional space. This means that no matter how the entire conformation is rotated or translated in three-dimensional space, the three-dimensional coordinates of individual atoms change, but the represented three-dimensional shape remains unchanged. Therefore, we cannot simply use the three-dimensional coordinates of atoms to represent the three-dimensional shape of molecular conformations. Instead, we must try to find rotational and translational invariants, that is, we must learn the relationships between atoms.

[0027] Traditional molecular conformation generation methods cannot handle situations where a single molecule exists in multiple low-energy stable conformations. However, to ensure the accuracy of the generated conformations, they need to conform to the Boltzmann energy distribution, thus reflecting the distribution of various molecular conformations in nature, while simultaneously considering the diversity of the generated molecular conformations. It is evident that simultaneously satisfying these constraints is difficult.

[0028] Embodiments of this disclosure provide a scheme for determining the conformation of molecules. This scheme proposes a molecular conformation field using three-dimensional molecular fragments and dihedral angles as units, thereby sampling and generating a series of different low-energy molecular conformations in the conformation field space. Specifically, this disclosure proposes a data-driven three-dimensional molecular fragment segmentation algorithm, thereby reducing the internal degrees of freedom of the segmented fragments and thus reducing modeling complexity. Based on this, a random field of molecular conformations using three-dimensional molecular fragments and dihedral angles as units is constructed, and Gibbs sampling is used to generate various low-energy conformations that conform to natural laws within this random field.

[0029] Molecular conformation fields model three-dimensional molecules into dihedral angles between molecular fragments and adjacent molecular fragments, which can effectively solve the problem of rotation invariance. At the same time, random fields can efficiently sample low-energy conformation molecules that conform to the energy distribution and can also generate diverse and highly accurate conformations, effectively solving the problem of energy distribution.

[0030] In the following description, some embodiments will be discussed using molecules having multiple benzene ring structures. However, it should be understood that this is merely to enable those skilled in the art to better understand the principles and ideas of the embodiments of this disclosure, and is not intended to limit the scope of this disclosure in any way.

[0031] Figure 1 A schematic diagram of an example environment 100 in which a method for determining the conformation of a molecule according to some embodiments of the present disclosure can be implemented is shown. Figure 1 As shown, example environment 100 may include computing device 110, which may be, for example, a computing system or a server. Example environment 100 may also include a two-dimensional molecular structure 120. The two-dimensional molecular structure 120 can represent the atomic connections in a molecule in two dimensions, but cannot represent the atomic connections in three-dimensional space. This is because many different three-dimensional molecular structures belong to the same two-dimensional molecular structure; this property can be called the flexibility of molecules.

[0032] Example environment 100 also includes a fragment library 130. Fragment library 130 can be a database storing data as the smallest unit of division. In a molecule, some structures (e.g., a benzene ring) do not require further division (e.g., the benzene ring does not need to be divided into smaller units). The internal structures of these molecular fragments that do not require further division are stable and invariant; that is, the bond lengths and bond angles are fixed, and they can be considered rigid or nearly rigid. The flexibility of a molecule primarily arises from changes in the dihedral angles between some rotatable bonds.

[0033] The computing device 110 receives a two-dimensional molecular structure 120 as input, segments the molecule according to a fragment library 130, and obtains multiple molecular fragments. The example environment 100 also includes three-dimensional molecular conformations. In this disclosure, a conformation can be understood as a three-dimensional molecular structure.

[0034] Molecular conformation not only affects the physical and chemical properties of compounds but also influences the structure and properties of some biological macromolecules, such as proteins, enzymes, and nucleic acids. The conformational isomers of many drug molecules are closely related to their biological activity. Drug receptors generally bind to only one of the many conformations of a drug; this conformation is called the pharmacodynamic conformation. Non-pharmacodynamic conformational isomers of drugs are difficult to bind to drug receptors and are usually ineffective or have no pharmacodynamic effect. Therefore, it is necessary to discover as many molecular conformations as possible.

[0035] refer to Figure 1 Example environment 100 also includes a three-dimensional molecular conformation 140. Computational device 110 models the molecule based on the segmented molecular fragments to obtain a molecular conformation field. Using the molecular conformation field, a three-dimensional molecular conformation 140 with an energy distribution conforming to natural laws is obtained.

[0036] Figure 2 A flowchart of a method 200 for determining the conformation of a molecule according to some embodiments of the present disclosure is shown. Figure 3 A block diagram illustrating a process 300 for determining the conformation of a molecule according to some embodiments of the present disclosure is shown. The following will be combined with... Figure 2 and Figure 3 This describes how to determine the conformation of a molecule.

[0037] At box 202, based on the two-dimensional structure of the molecule, the molecule is divided into multiple molecular fragments, at least one of which includes multiple atoms. For the two-dimensional structure of the molecule, the fragmentation is performed as close as possible to the granularity of atomic groups (which include multiple atoms). For atoms that cannot form atomic groups, a single atom can be used as a molecular fragment. That is, at least one of the multiple molecular fragments includes multiple atoms, while the other molecular fragments can include either multiple atoms or a single atom. For example, see [reference needed]. Figure 3 In the two-dimensional molecular structure 310, at chemical bond 312, the benzene ring 308 in the two-dimensional molecular structure 310 can be segmented from its adjacent atoms or groups. After segmentation, the benzene ring 308 can be represented as a molecular fragment 328 in the molecular conformation field 320. Since a benzene ring includes 6 carbon atoms and 6 hydrogen atoms, the molecular fragment 308 includes 12 atoms. In some embodiments, introducing relevant concepts from graph theory, the molecular fragment 308 can also be referred to as node 308. Similarly, the two-dimensional molecular structure 310 can also be segmented at chemical bonds 314 and 316.

[0038] In the molecular conformation field 320 obtained after completely dividing the two-dimensional molecular structure 310, and also referring to relevant concepts in graph theory, chemical bond 312 can be represented as dihedral angle 322, chemical bond 314 can be represented as dihedral angle 324, and chemical bond 316 can be represented as dihedral angle 326. It can be understood that the dihedral angles in the molecular conformation field 320 are all variables and need to be configured to obtain specific values. Based on the same concept, dihedral angle 322 can also be called node 322, dihedral angle 324 can also be called node 324, and dihedral angle 326 can also be called node 326.

[0039] In some embodiments, a topological graph identifying the connection relationships of multiple molecular segments and multiple dihedrals can be determined, and the multiple molecular segments and multiple dihedrals can be identified as multiple nodes of the topological graph. Therefore, molecular segments, dihedrals and nodes will sometimes not be distinguished in the following text.

[0040] In this way, the two-dimensional molecular structure can be represented as a combination of molecular fragments and dihedral angles. For each molecular fragment, there can be multiple corresponding fragment conformations. For each dihedral angle, there can be multiple angles. Therefore, multiple combinations of multiple molecular fragments and multiple dihedral angles are used to determine multiple conformations of the desired molecule.

[0041] At box 204, multiple dihedral angles between adjacent molecular fragments are determined. (Reference) Figure 3 Box 330 in the diagram illustrates an example of configuring a dihedral. As an example, dihedral 322 can be configured within a range of -180 degrees to 180 degrees (or 0 to 360 degrees). Therefore, in box 330, dihedral 322 can be configured to obtain dihedral 332 with a specific angle. Similarly, dihedrals 324 and 326 can be configured to obtain dihedrals 334 and 336 with specific angles. Examples of configurations will be referenced below. Figure 4 To describe.

[0042] At box 206, multiple combinations of multiple molecular fragments and multiple dihedral angles are defined as multiple conformations of the molecule. As an example, the configured dihedral angles 332, 334, and 336 can be combined with the fragment conformations corresponding to the molecular fragments to obtain a three-dimensional molecular conformation 340. It is understood that such combinations can have many variations due to different configured dihedral angles or different fragment conformations corresponding to the molecular fragments; for the sake of simplicity, they will not be shown one by one here.

[0043] By implementing method 200, a molecular conformational field can be established with molecular fragments having rigid or nearly rigid internal structures as the smallest segmentation units, thereby reducing the degrees of freedom of the conformational model that needs to be modeled when predicting molecular conformations. Due to this segmentation mode, it is unnecessary to consider complex factors such as bond lengths and bond angles between individual atoms, thus reducing a significant amount of redundant predictions. Since the coupling between redundant predictions can lead to numerical instability and model insufficiency, reducing redundant predictions also improves the predictive diversity and accuracy of molecular conformations.

[0044] Figure 4 A schematic diagram of a molecular conformational field according to some embodiments of the present disclosure is shown. Figure 4 The two-dimensional molecular structure 410 in the middle corresponds to Figure 3 The two-dimensional molecular structure 310 in the image has the same two-dimensional molecular structure. After segmenting the two-dimensional molecular structure 410, a molecular conformation field 420 can be obtained. In the molecular conformation field 420, hollow nodes represent molecular segments, such as molecular segment 425, molecular segment 427, molecular segment 428, and molecular segment 429. Solid nodes represent dihedral angles, such as dihedral angle 422, dihedral angle 426.

[0045] The molecular conformational field 420 can be represented using a Markov random field, which can be understood as the probabilistic relationship of each node described by the potential energy distribution between atoms. That is, it considers the single potential energy function of a single segment conformation node (also known as the first energy distribution), the single potential energy function of a single dihedral node (also known as the second energy distribution), and the binary potential energy function between the segment conformation and the dihedral (also known as the third energy distribution). The following will combine... Figure 6A and Figure 6B To describe these three energy distributions.

[0046] As an example, node 428, representing a molecular fragment, can correspond to fragment conformations 462 and 464 shown in box 460. The probabilities of fragment conformations 462 and 464 can be determined based on a function fitted to the energy distribution of the molecular conformation field 420. Therefore, different fragment conformations can be obtained in multiple samplings based on the obtained probabilities. Similarly, nodes 425, 427, and 429, representing molecular fragments, can also have a similar sampling process.

[0047] As an example, an example representing node 422 of a dihedral angle can be seen in the dihedral angle shown in box 450. In box 450, dihedral angle 452 represents the angle between plane 454 and plane 456. Plane 454 represents the plane formed by the straight line 472 formed by two atoms (atoms 470 and 471) of molecular segment 428 and atom 473 of molecular segment 429. Plane 456 represents the plane formed by the straight line 475 formed by two atoms (atoms 473 and 474) of molecular segment 429 and atom 471 of molecular segment 428. It can be seen that plane 454 and plane 456 share a common edge 476.

[0048] The dihedral angle defined in this way indicates the angle at which two adjacent molecular segments are connected. The dihedral angle between two adjacent molecular segments can also be determined by combinations of other atoms, but they must all satisfy the condition of having a common edge.

[0049] Similarly, the probability of the dihedral angle 422's angle value can be determined based on a function fitted to the energy distribution of the molecular conformation field 420. Therefore, different dihedral angle values ​​can be obtained in multiple samplings based on the obtained probabilities. Similarly, nodes 424 and 426 representing the dihedral angle can also have a similar sampling process.

[0050] Figure 5A A flowchart of a method 500 for constructing a fragment library according to some embodiments of the present disclosure is shown. Figure 5B A block diagram illustrating the process of constructing a fragment library according to some embodiments of this disclosure is shown. The following will be combined with... Figure 5A and Figure 5B To describe how to segment two-dimensional molecular structures.

[0051] In some embodiments, the two-dimensional molecular structure can be segmented according to the fragment library 150. The fragment library can be pre-constructed. It is understood that when faced with a two-dimensional molecular structure, there can be multiple segmentation schemes, each of which may correspond to different molecular fragments. Therefore, in some embodiments, the degrees of freedom of each segmentation scheme can be calculated, and the segmentation scheme with the fewest degrees of freedom can be selected as the target segmentation scheme. The segmentation result of the target segmentation scheme is determined as the molecular fragment.

[0052] In some embodiments, at block 502, multiple sample conformations of multiple sample molecules and multiple sample two-dimensional structures corresponding to the multiple sample conformations are obtained. (Reference) Figure 5B The sample molecules are stored in the molecular database 512. The sample molecules can be understood as a variety of common or specially selected organic compounds.

[0053] At box 504, multiple sample conformations and multiple sample two-dimensional structures are segmented to obtain multiple sample fragment conformations and multiple sample molecular fragments corresponding to the multiple sample fragment conformations. As an example, for each sample molecule in the molecular database 512, its two-dimensional molecular structure 520 can be segmented to obtain multiple molecular fragments 540. Alternatively, its three-dimensional molecular conformation 530 can be segmented in parallel or sequentially to obtain multiple fragment conformations 550.

[0054] At box 506, multiple correlations between the plurality of sample fragment conformations and the plurality of sample molecular fragments are determined. As an example, the root mean square deviation (RMSD) between molecular fragment 542 and its corresponding fragment conformations 552 can be calculated, and the calculated RMSDs can be used as the distances between the three-dimensional fragment conformations and molecular fragment 542 to reflect the correlation.

[0055] At box 508, the plurality of sample fragment conformations are clustered based on the plurality of correlations to obtain a plurality of cluster centers. As an example, the plurality of specific fragment conformations of fragment conformation 552 represented by box 570 can be clustered into two classes, and the two cluster centers shown in box 560 represent all the specific fragment conformations in box 570.

[0056] At box 510, multiple cluster centers are added to the fragment library. As an example, two cluster centers from box 560 are added to fragment library 130.

[0057] It is understandable that a two-dimensional molecular structure can be divided into multiple individual atoms without considering any constraints, but this is not what this disclosure aims to achieve. Similarly, a two-dimensional molecular fragment can have diverse three-dimensional conformations. Therefore, a method is needed to divide the molecule using rigid structures as the smallest units as possible. This disclosure hereby proposes the concept of degrees of freedom.

[0058] In some embodiments, the degrees of freedom of a molecular fragment can be determined by the variance of the three-dimensional fragment conformation, and can be determined with reference to the following formula:

[0059] (1)

[0060] in, Represents variance. This segment indicates that... Indicates the root mean square deviation. This represents the average value of all molecular fragments.

[0061] As an example, degrees of freedom can be calculated based on molecular fragment 542 and its corresponding fragment conformation 552. Specifically, fragment conformation 552 includes six specific fragment conformations as shown in box 570. Therefore, six RMSDs can be calculated for molecular fragment 542 and one of the six specific fragment conformations, and these six RMSDs are used as six distances between the three-dimensional fragment conformation and molecular fragment 542. Clustering is performed based on these distances to obtain cluster centers 560. Cluster centers 560 are used as discrete three-dimensional conformational states of molecular fragment 542. Furthermore, the distances between each three-dimensional fragment conformation and its corresponding cluster center are calculated. The variance of the 3D segment is taken as the variance of the segment. The sum of the variances of the individual 3D segments for each segmentation scheme is taken as the degrees of freedom. Specifically, the following formula can be used to determine this:

[0062] (P)

[0063] = (P) (2)

[0064] in, P Indicates the segmentation scheme. (P) Denotes the degrees of freedom of the partitioning scheme. Represents the optimal partitioning scheme ， Indicates when the smallest (P) time P Which segmentation scheme is it? 。

[0065] This segmentation scheme yields the fewest degrees of freedom, which can be understood as dividing the molecule into the smallest functional groups that conform to natural laws. Therefore, it avoids modeling the internal structure between atoms, allowing the diversity of generated conformations to be primarily reflected by the dihedral angles between molecular fragments and the diversity of fragment conformations, thus improving the interpretability of the predicted molecular conformations.

[0066] Figure 6A A flowchart of a method 600 for determining energy distribution according to some embodiments of the present disclosure is shown. Figure 6B A block diagram illustrating a process for determining energy distribution according to some embodiments of the present disclosure is shown. The following will be combined with... Figure 6A and Figure 6B This describes the process of establishing a molecular conformational field.

[0067] At box 602, multiple feature vectors of the plurality of molecular fragments are determined. For example, a multilayer graph neural network (GNN) can be applied to each node for parameterization and information transfer to transform node 612 into vector 622, node 614 into vector 624, and node 616 into vector 626. Based on this, a Markov random field (MRF), i.e., a molecular conformation field, is constructed.

[0068] In some embodiments, the energy distribution of a molecule can be determined based on multiple eigenvectors and multiple dihedral angles, such that the molecular conformation field conforms to natural laws, such as the Boltzmann energy distribution.

[0069] At box 604, for each of the plurality of molecular fragments, a first energy distribution 640 associated with each molecular fragment is determined. At box 606, for each of the plurality of dihedral angles, a second energy distribution 650 associated with each dihedral angle is determined. As an example, the energy distribution representing the nodes of a molecular fragment can be expressed by the following formula:

[0070] E1= (3)

[0071] Where E1 represents the potential energy distribution associated with molecular segment i; This represents the partition function, used for normalization. Represents the i-th molecular fragment. It represents potential energy.

[0072] Similarly, the energy distribution of nodes representing molecular fragments can be expressed by the following formula:

[0073] E2= (4)

[0074] Where E2 represents the potential energy distribution associated with the dihedral angle i; This represents the partition function, used for normalization. This represents the i-th dihedral angle. It represents potential energy.

[0075] At box 608, a third energy distribution 660 associated with each molecular fragment and each dihedral angle is determined based on each molecular fragment and each dihedral angle.

[0076] As an example, since formula (3) only considers the potential function associated with the molecular fragment, and formula (4) only considers the potential function associated with the dihedral angle, a virtual node (e.g., node 630 represented by the solid cube) is introduced between the molecular fragment node and the dihedral angle node to improve the accuracy of the model. A binary potential function associated with both the molecular fragment and the dihedral angle can be used for the virtual node 630 structure, which can be expressed by formula (5):

[0077] E3= (5)

[0078] Where E3 represents the binary potential energy distribution associated with molecular segment i and dihedral angle j; This represents the partition function, used for normalization.

[0079] As an example, a binary potential function 632 can be established for virtual node 630, and a binary potential function 636 can be established for virtual node 634. Furthermore, for simplicity, a grid block can be used to represent the MRF, and the relative relationship of the virtual node within the entire MRF can be shown by referring to grids 632 and 636.

[0080] At box 610, based on the first energy distribution 640, the second energy distribution 650, and the third energy distribution 660, a joint probability distribution associated with multiple molecular fragments and multiple dihedral angles is determined as the energy distribution. As an example, the joint probability distribution obtained by using formulas (3), (4), and (5) as basic units and summing the corresponding potential functions of each node in the entire molecular conformation field can be used as the energy distribution. The formula for the joint probability distribution can be found in formula (6).

[0081]

[0082] + + (6)

[0083] Where P represents probability. This represents the first molecular fragment node. The NFth molecular fragment node; This represents the first dihedral corner node; This represents the Nd-th dihedral node;

[0084] This indicates traversing the entire molecular fragment node. This indicates that the entire dihedral corner node is traversed. This indicates traversing the entire virtual node.

[0085] The molecular conformation field established in this way can reduce the degrees of freedom required for modeling, making the molecular conformation field regularized. It can efficiently learn relevant laws from data without having to consider geometric constraints between atoms and multivariable coupling, thus reducing numerical instability problems, reducing the prediction burden, and improving prediction accuracy.

[0086] Figure 7A block diagram of a node sampling process 700 according to some embodiments of the present disclosure is shown. After establishing the potential energy function describing the molecular conformational field, the diagram illustrates how various sets of suitable fragment conformations and dihedral angle values ​​(also referred to as a set of configurations) are obtained through Gibbs sampling, thereby assembling them into the final molecular idol.

[0087] At box 702, a set of configurations is first randomly initialized, for example, by randomly assigning values ​​to each node. Then, a node to be sampled is selected, the remaining nodes are fixed, and the configuration of a node is sampled according to the probability distribution of that node in the random field, and the node is updated. For example, at box 710, node 712 is sampled. At this point, the remaining nodes are fixed, and the joint probability distribution function of the entire molecular conformation field will become the marginal probability distribution for node 712. Then, a specific value is randomly sampled according to the updated marginal probability distribution, for example, with conditional probability. Sample a value and assign it to node 712.

[0088] For example, at box 720, other unsampled nodes can be kept unchanged, and sampling can be performed on node 722. For example, using conditional probability... Sample a value and assign it to node 722. Similarly, at box 730, we can keep the other unsampled nodes unchanged and sample node 732. For example, with conditional probability. Sample a value and assign it to node 722. Continue sampling all nodes sequentially, ensuring that each node in the molecular conformation field has a specific value, for example, until frame 740, at which point sampling ends. The process of sampling all nodes in the molecular conformation field once can be called one round.

[0089] When the number of sampling rounds is sufficiently large, the configuration obtained from each sampling round will also conform to the potential energy distribution of the molecule. In some embodiments, it can be assumed that the molecular conformational field after N sampling rounds conforms to the potential energy distribution. Therefore, multiple nodes obtained from multiple sampling rounds after N rounds can be assembled into multiple conformations of the molecule.

[0090] This sampling method allows for the generation of molecular conformations that conform to the potential energy distribution, which is crucial for studying molecular dynamics evolution and the state and flexibility of drug molecules in real-world environments. Because molecules in nature coexist in multiple conformations and interconvert, their distribution conforms to the Boltzmann potential energy distribution. Therefore, the method for generating molecular conformations needs to be rapid and conform to the energy distribution to provide important support for downstream molecular simulations.

[0091] In some embodiments, a predetermined number of fragment conformations and dihedral angles can be determined. For example, after N rounds of sampling to obtain a stable potential energy distribution, M more rounds of sampling can be performed, and the obtained fragment conformations and dihedral angles can be clustered to obtain multiple cluster centers. In this way, multiple combinations of these multiple cluster centers can be determined as multiple conformations of the molecule. This method can make the obtained conformations more focused, but may lose some diversity. Therefore, it can be selected as needed, for example when accuracy is required much more than diversity.

[0092] Figure 8 A block diagram of an apparatus for determining the conformation of a molecule according to some embodiments of the present disclosure is shown. Figure 8 As shown, device 800 includes a molecular segmentation module 802 configured to segment the molecule into multiple molecular fragments based on the two-dimensional structure of the molecule, at least one of the multiple molecular fragments comprising multiple atoms. Device 800 also includes a dihedral angle determination module 802 configured to determine multiple dihedral angles between adjacent molecular fragments among the multiple molecular fragments. Device 800 further includes a conformation combination module 806 configured to determine multiple combinations of the multiple molecular fragments and multiple dihedral angles as multiple conformations of the molecule. Device 800 may also include other modules to achieve the same function as method 200; for simplicity, these will not be elaborated upon here.

[0093] It is understood that the apparatus 800 of this disclosure can achieve at least one of the many advantages that can be achieved by the methods or processes described above. For example, a molecular conformational field of molecular fragments can be established, thereby reducing the degrees of freedom of the conformational model that needs to be modeled when predicting the conformation of molecules. The apparatus 800 also reduces redundant predictions and improves the predictive diversity and accuracy of molecular conformations.

[0094] In summary, the technical solution disclosed herein can accurately predict multiple conformations of three-dimensional molecules. This is a fundamental step and key link in the study of downstream drug molecules, which helps in various tasks such as drug activity prediction, drug affinity prediction, and drug target binding analysis, thereby empowering new drug development, shortening the research and development cycle, and reducing research and development costs.

[0095] Figure 9 A block diagram of an apparatus 900 for determining the conformation of molecules according to some embodiments of the present disclosure is shown. Apparatus 900 may be the apparatus or device described in the embodiments of the present disclosure. Figure 9As shown, device 900 includes a central processing unit (CPU) and / or a graphics processing unit (GPU) 901, which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) 902 or loaded from storage unit 908 into random access memory (RAM) 903. Various programs and data required for the operation of device 900 can also be stored in RAM 903. CPU / GPU 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904. Although not shown in... Figure 9 As shown, device 900 may also include a coprocessor.

[0096] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0097] The various methods or processes described above can be executed by CPU / GPU 901. For example, in some embodiments, the methods can be implemented as computer software programs tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by CPU / GPU 901, one or more steps or actions in the methods or processes described above can be performed.

[0098] In some embodiments, the methods and processes described above can be implemented as a computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of this disclosure.

[0099] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example, but not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0100] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, a local area network (LAN), a wide area network (WAN), and / or a wireless network, to an external computer or external storage device. The network may include copper cables, fiber optic cables, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0101] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​and conventional procedural programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to execute the computer-readable program instructions, thereby implementing various aspects of this disclosure.

[0102] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0103] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0104] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0105] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical applications, or technical improvements to the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

[0106] The following are some example implementations of this disclosure.

[0107] Example 1. A method for determining the conformation of a molecule, comprising:

[0108] Based on the two-dimensional structure of the molecule, the molecule is divided into multiple molecular segments, and at least one of the multiple molecular segments includes multiple atoms;

[0109] Determine multiple dihedral angles between adjacent molecular segments in the plurality of molecular segments; and

[0110] The plurality of molecular fragments and the plurality of combinations of the plurality of dihedral angles are determined as the plurality of conformations of the molecule.

[0111] Example 2. The method according to Example 1, wherein dividing the molecule into multiple molecular fragments includes:

[0112] Based on the fragment library of the molecule, multiple segmentation schemes for the molecule are determined;

[0113] The segmentation scheme with the fewest degrees of freedom among the multiple segmentation schemes is determined as the target segmentation scheme; and

[0114] Based on the target segmentation scheme, the plurality of molecular fragments of the molecule are determined.

[0115] Example 3. The method according to any one of Examples 1-2, wherein the fragment library is constructed in the following manner:

[0116] Obtain multiple sample conformations of multiple sample molecules and multiple sample two-dimensional structures corresponding to the multiple sample conformations;

[0117] The plurality of sample conformations and the plurality of sample two-dimensional structures are segmented to obtain a plurality of sample fragment conformations and a plurality of sample molecular fragments corresponding to the plurality of sample fragment conformations;

[0118] Determine multiple correlations between the conformations of the multiple sample fragments and the multiple molecular fragments of the multiple samples;

[0119] Clustering the multiple sample fragment conformations based on the multiple correlations to obtain multiple cluster centers; and

[0120] Add the multiple cluster centers to the fragment library.

[0121] Example 4. The method according to any one of Examples 1-3, wherein determining the segmentation scheme with the fewest degrees of freedom among the plurality of segmentation schemes as the target segmentation scheme includes:

[0122] Determine the root mean square deviations of the plurality of fragment conformations of the molecule and the corresponding plurality of cluster centers; and

[0123] The sum of the multiple variances is determined as the degrees of freedom of the segmentation scheme.

[0124] Example 5. The method according to any one of Examples 1-4, wherein determining the plurality of dihedral angles between adjacent molecular segments among the plurality of molecular segments includes:

[0125] Based on the energy distribution of the molecules, the plurality of dihedral angles and the plurality of fragment conformations are determined.

[0126] Example 6. The method according to any one of Examples 1-5 further includes:

[0127] Determine multiple feature vectors of the plurality of molecular fragments; and

[0128] The energy distribution of the molecule is determined based on the plurality of eigenvectors and the plurality of dihedral angles.

[0129] Example 7. The method according to any one of Examples 1-6, wherein determining the energy distribution of the molecule comprises:

[0130] For each molecular fragment, a first energy distribution associated with each molecular fragment is determined;

[0131] For each of the plurality of dihedral angles, a second energy distribution associated with each dihedral angle is determined;

[0132] Based on each molecular fragment and each dihedral angle, a third energy distribution associated with each molecular fragment and each dihedral angle is determined; and

[0133] Based on the first energy distribution, the second energy distribution, and the third energy distribution, a joint probability distribution associated with the plurality of molecular fragments and the plurality of dihedral angles is determined as the energy distribution.

[0134] Example 8. The method according to any one of Examples 1-7, wherein determining the plurality of molecular fragments and the plurality of combinations of the plurality of dihedral angles as the plurality of conformations of the molecule comprises:

[0135] Based on the energy distribution of the molecule, multiple candidate fragment conformations are sampled from the plurality of molecular fragments of the molecule;

[0136] Based on the energy distribution of the molecule, multiple candidate dihedral angles are sampled from the multiple dihedral angles of the molecule;

[0137] The plurality of candidate fragment configurations and the plurality of candidate dihedral angles are combined multiple times; and

[0138] The obtained combinations are determined as the multiple conformations of the molecule.

[0139] Example 9. The method according to any one of Examples 1-8 further includes:

[0140] A topological graph is determined that identifies the connection relationships between the plurality of molecular fragments and the plurality of dihedral angles;

[0141] The plurality of molecular fragments and the plurality of dihedral angles are defined as the plurality of nodes in the topological graph; and

[0142] Based on the energy distribution of the molecules, multiple candidate nodes are sampled from the multiple nodes of the topology graph.

[0143] Example 10. The method according to any one of Examples 1-9, wherein sampling a plurality of candidate nodes in the plurality of nodes of the topology graph based on the energy distribution of the molecule comprises:

[0144] Select one of the plurality of nodes;

[0145] Keep the other nodes among the plurality of nodes unchanged;

[0146] Update the energy distribution of the molecule; and

[0147] Based on the updated energy distribution, the plurality of candidate nodes are determined for the given node.

[0148] Example 11. The method according to any one of Examples 1-10, wherein determining the plurality of candidate fragment conformations among the plurality of fragment conformations of the molecule comprises:

[0149] Based on the fragment library, multiple cluster centers corresponding to the multiple conformations are determined;

[0150] Determine the multiple fragment conformations represented by the multiple cluster centers; and

[0151] Based on the multiple fragment conformations represented by the multiple cluster centers, the multiple candidate fragment conformations are determined as the multiple conformations.

[0152] Example 12. The method according to any one of Examples 1-11, wherein determining the plurality of combinations obtained as the plurality of conformations of the molecule comprises:

[0153] Determine a predetermined number of segment configurations and dihedrals;

[0154] Clustering the predetermined number of fragment conformations and dihedrals to obtain multiple cluster centers; and

[0155] Multiple combinations of the multiple cluster centers are determined as the multiple conformations.

[0156] Example 13. An apparatus for determining the conformation of a molecule, comprising:

[0157] A molecular segmentation module is configured to segment the molecule into multiple molecular fragments based on the two-dimensional structure of the molecule, at least one of the multiple molecular fragments comprising multiple atoms;

[0158] The dihedral angle determination module is configured to determine multiple dihedral angles between adjacent molecular segments in the plurality of molecular segments; and

[0159] The conformation combination module is configured to determine multiple conformations of the molecule by multiple combinations of the plurality of molecular fragments and the plurality of dihedral angles.

[0160] Example 14. The apparatus according to Example 13, wherein the molecular segmentation module comprises:

[0161] The segmentation scheme determination module is configured to determine multiple segmentation schemes for the molecule based on a fragment library of the molecule.

[0162] The target segmentation scheme determination module is configured to determine the segmentation scheme with the fewest degrees of freedom among the plurality of segmentation schemes as the target segmentation scheme; and

[0163] The second molecular fragment determination module is configured to determine the plurality of molecular fragments of the molecule according to the target segmentation scheme.

[0164] Example 15. The apparatus according to any one of Examples 13-14 further includes a fragment library construction module, the fragment library construction module comprising:

[0165] The sample molecule acquisition module is configured to acquire multiple sample conformations of multiple sample molecules and multiple sample two-dimensional structures corresponding to the multiple sample conformations.

[0166] The sample molecule segmentation module is configured to segment the plurality of sample conformations and the plurality of sample two-dimensional structures to obtain a plurality of sample fragment conformations and a plurality of sample molecule fragments corresponding to the plurality of sample fragment conformations;

[0167] The correlation determination module is configured to determine multiple correlations between the conformations of the plurality of sample fragments and the molecular fragments of the plurality of samples;

[0168] The first clustering module is configured to cluster the multiple sample fragment conformations based on the multiple correlations to obtain multiple cluster centers; and

[0169] An add module is configured to add the plurality of cluster centers to the fragment library.

[0170] Example 16. The apparatus according to any one of Examples 13-15, wherein the target segmentation scheme determination module comprises:

[0171] A variance determination module is configured to determine multiple variances of the root mean square deviations of the plurality of fragment conformations of the molecule and the corresponding plurality of cluster centers; and

[0172] The degree-of-freedom determination module is configured to determine the sum of the plurality of variances as the degrees of freedom of the segmentation scheme.

[0173] Example 17. The apparatus according to any one of Examples 13-16, wherein the dihedral angle determining module comprises:

[0174] The second dihedral angle determination module is configured to determine the plurality of dihedral angles and the plurality of fragment conformations based on the energy distribution of the molecule.

[0175] Example 18. The apparatus according to any one of Examples 13-17 further includes:

[0176] A feature vector determination module is configured to determine multiple feature vectors of the plurality of molecular fragments; and

[0177] An energy distribution determination module is configured to determine the energy distribution of the molecule based on the plurality of feature vectors and the plurality of dihedral angles.

[0178] Example 19. The apparatus according to any one of Examples 13-18, wherein the energy distribution determination module comprises:

[0179] The first energy distribution determination module is configured to determine a first energy distribution associated with each molecular fragment;

[0180] The second energy distribution determination module is configured to determine a second energy distribution associated with each of the plurality of dihedral angles;

[0181] The third energy distribution determination module is configured to determine a third energy distribution associated with each molecular fragment and each dihedral angle based on each molecular fragment and each dihedral angle; and

[0182] The joint probability distribution determination module is configured to determine, based on the first energy distribution, the second energy distribution, and the third energy distribution, a joint probability distribution associated with the plurality of molecular fragments and the plurality of dihedral angles as the energy distribution.

[0183] Example 20. The apparatus according to any one of Examples 13-19, wherein the conformational assembly module comprises:

[0184] A candidate fragment conformation sampling module is configured to sample multiple candidate fragment conformations from the plurality of molecular fragments of the molecule based on the energy distribution of the molecule.

[0185] A candidate dihedral angle sampling module is configured to sample multiple candidate dihedral angles from among the multiple dihedral angles of the molecule based on the energy distribution of the molecule;

[0186] The candidate combination module is configured to combine the plurality of candidate fragment configurations and the plurality of candidate dihedrals multiple times; and

[0187] A candidate combination determination module is configured to determine multiple combinations as the multiple conformations of the molecule.

[0188] Example 21. The apparatus according to any one of Examples 13-20 further includes:

[0189] The first topology graph determination module is configured to determine a topology graph that identifies the connection relationships between the plurality of molecular fragments and the plurality of dihedral angles;

[0190] The second topology graph determination module is configured to determine the plurality of molecular fragments and the plurality of dihedral angles as plurality of nodes in the topology graph; and

[0191] The topology graph sampling module is configured to sample multiple candidate nodes in the plurality of nodes of the topology graph based on the energy distribution of the molecules.

[0192] Example 22. The apparatus according to any one of Examples 13-21, wherein the topology map sampling module comprises:

[0193] The node selection module is configured to select one of the plurality of nodes;

[0194] The node retention module is configured to keep other nodes among the plurality of nodes unchanged.

[0195] An energy distribution update module is configured to update the energy distribution of the molecule; and

[0196] The candidate node determination module is configured to determine the plurality of candidate nodes for the one node based on the updated energy distribution.

[0197] Example 23. The apparatus according to any one of Examples 13-22, wherein the second topology determination module comprises:

[0198] The second clustering module is configured to determine multiple cluster centers corresponding to the multiple conformations based on the fragment library;

[0199] The cluster center determination module is configured to determine multiple segment conformations represented by the plurality of cluster centers; and

[0200] The candidate fragment conformation determination module is configured to determine the plurality of candidate fragment conformations as the plurality of conformations based on the plurality of fragment conformations represented by the plurality of cluster centers.

[0201] Example 24. The apparatus according to any one of Examples 13-23, wherein the candidate combination determining module comprises:

[0202] The fragment configuration and dihedral angle determination module is configured to determine a predetermined number of fragment configurations and dihedral angles;

[0203] The third clustering module is configured to cluster the predetermined number of fragment conformations and dihedrals to obtain multiple cluster centers; and

[0204] The cluster center combination module is configured to determine multiple combinations of the plurality of cluster centers as the plurality of conformations.

[0205] Example 25. An electronic device comprising:

[0206] Processor; and

[0207] A memory coupled to the processor, the memory having instructions stored therein, the instructions which, when executed by the processor, cause the electronic device to perform actions, the actions including:

[0208] Based on the two-dimensional structure of the molecule, the molecule is divided into multiple molecular segments, and at least one of the multiple molecular segments includes multiple atoms;

[0209] Determine multiple dihedral angles between adjacent molecular segments in the plurality of molecular segments; and

[0210] The plurality of molecular fragments and the plurality of combinations of the plurality of dihedral angles are determined as the plurality of conformations of the molecule.

[0211] Example 26. The electronic device according to Example 25, wherein cutting the molecule into a plurality of molecular fragments includes:

[0212] Based on the fragment library of the molecule, multiple segmentation schemes for the molecule are determined;

[0213] The segmentation scheme with the fewest degrees of freedom among the multiple segmentation schemes is determined as the target segmentation scheme; and

[0214] Based on the target segmentation scheme, the plurality of molecular fragments of the molecule are determined.

[0215] Example 27. An electronic device according to any one of Examples 25-26, wherein the fragment library is constructed by the following actions:

[0216] Obtain multiple sample conformations of multiple sample molecules and multiple sample two-dimensional structures corresponding to the multiple sample conformations;

[0217] The plurality of sample conformations and the plurality of sample two-dimensional structures are segmented to obtain a plurality of sample fragment conformations and a plurality of sample molecular fragments corresponding to the plurality of sample fragment conformations;

[0218] Determine the conformation of the plurality of sample fragments and the root mean square deviations of the plurality of sample molecular fragments;

[0219] Clustering of the multiple sample fragment conformations based on the multiple root mean square biases to obtain multiple cluster centers; and

[0220] Add the multiple cluster centers to the fragment library.

[0221] Example 28. An electronic device according to any one of Examples 25-27, wherein determining the segmentation scheme with the fewest degrees of freedom among the plurality of segmentation schemes as the target segmentation scheme includes:

[0222] Determine the root mean square deviations of the plurality of fragment conformations of the molecule and the corresponding plurality of cluster centers; and

[0223] The sum of the multiple variances is determined as the degrees of freedom of the segmentation scheme.

[0224] Example 29. An electronic device according to any one of Examples 25-28, wherein determining a plurality of dihedral angles between adjacent molecular segments among the plurality of molecular segments includes:

[0225] Based on the energy distribution of the molecules, the plurality of dihedral angles and the plurality of fragment conformations are determined.

[0226] Example 30. The electronic device according to any one of Examples 25-29, wherein the operation further includes:

[0227] Determine multiple feature vectors of the plurality of molecular fragments; and

[0228] The energy distribution of the molecule is determined based on the plurality of eigenvectors and the plurality of dihedral angles.

[0229] Example 31. An electronic device according to any one of Examples 25-30, wherein determining the energy distribution of the molecule comprises:

[0230] For each molecular fragment, a first energy distribution associated with each molecular fragment is determined;

[0231] For each of the plurality of dihedral angles, a second energy distribution associated with each dihedral angle is determined;

[0232] Based on each molecular fragment and each dihedral angle, a third energy distribution associated with each molecular fragment and each dihedral angle is determined; and

[0233] Based on the first energy distribution, the second energy distribution, and the third energy distribution, a joint probability distribution associated with the plurality of molecular fragments and the plurality of dihedral angles is determined as the energy distribution.

[0234] Example 32. An electronic device according to any one of Examples 25-31, wherein defining the plurality of molecular fragments and the plurality of combinations of the plurality of dihedral angles as the plurality of conformations of the molecule comprises:

[0235] Based on the energy distribution of the molecule, multiple candidate fragment conformations are sampled from the plurality of molecular fragments of the molecule;

[0236] Based on the energy distribution of the molecule, multiple candidate dihedral angles are sampled from the multiple dihedral angles of the molecule;

[0237] The plurality of candidate fragment configurations and the plurality of candidate dihedral angles are combined multiple times; and

[0238] The obtained combinations are determined as the multiple conformations of the molecule.

[0239] Example 33. The electronic device according to any one of Examples 25-32, wherein the operation further includes:

[0240] A topological graph is determined that identifies the connection relationships between the plurality of molecular fragments and the plurality of dihedral angles;

[0241] The plurality of molecular fragments and the plurality of dihedral angles are defined as the plurality of nodes in the topological graph; and

[0242] Based on the energy distribution of the molecules, multiple candidate nodes are sampled from the multiple nodes of the topology graph.

[0243] Example 34. An electronic device according to any one of Examples 25-33, wherein sampling a plurality of candidate nodes in the plurality of nodes of the topology graph based on the energy distribution of the molecules comprises:

[0244] Select one of the plurality of nodes;

[0245] Keep the other nodes among the plurality of nodes unchanged;

[0246] Update the energy distribution of the molecule; and

[0247] Based on the updated energy distribution, the plurality of candidate nodes are determined for the given node.

[0248] Example 35. An electronic device according to any one of Examples 25-34, wherein determining the plurality of candidate fragment conformations among the plurality of fragment conformations of the molecule comprises:

[0249] Based on the fragment library, multiple cluster centers corresponding to the multiple conformations are determined;

[0250] Determine the multiple fragment conformations represented by the multiple cluster centers; and

[0251] Based on the multiple fragment conformations represented by the multiple cluster centers, the multiple candidate fragment conformations are determined as the multiple conformations.

[0252] Example 36. An electronic device according to any one of Examples 25-35, wherein determining the plurality of combinations obtained as the plurality of conformations of the molecule comprises:

[0253] Determine a predetermined number of segment configurations and dihedrals;

[0254] Clustering the predetermined number of fragment conformations and dihedrals to obtain multiple cluster centers; and

[0255] Multiple combinations of the multiple cluster centers are determined as the multiple conformations.

[0256] Example 37. A computer-readable storage medium having stored thereon one or more computer instructions, wherein the one or more computer instructions are executed by a processor to implement the method according to any one of Examples 1 to 12.

[0257] Example 38. A computer program product tangibly stored on a computer-readable medium and comprising computer-executable instructions that, when executed by a device, cause the device to perform the method according to any one of Examples 1 to 12.

[0258] Although this disclosure has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for determining the conformation of a molecule, comprising: Based on the two-dimensional structure of the molecule, the molecule is divided into multiple molecular segments, and at least one of the multiple molecular segments includes multiple atoms; Based on the energy distribution of the molecule, multiple fragment conformations corresponding to the multiple molecular fragments and multiple dihedral angles between adjacent molecular fragments are determined. as well as The plurality of fragment conformations and the plurality of combinations of the plurality of dihedral angles are defined as the plurality of conformations of the molecule. The method further includes: Determine multiple feature vectors of the multiple molecular fragments; For each of the plurality of molecular fragments, determine a first energy distribution associated with each molecular fragment; For each of the plurality of dihedral angles, a second energy distribution associated with each dihedral angle is determined; Based on each molecular fragment and each dihedral angle, a third energy distribution associated with each molecular fragment and each dihedral angle is determined; and Based on the first energy distribution, the second energy distribution, and the third energy distribution, a joint probability distribution associated with the plurality of molecular fragments and the plurality of dihedral angles is determined as the energy distribution.

2. The method according to claim 1, wherein dividing the molecule into multiple molecular fragments comprises: Based on the fragment library of the molecule, multiple segmentation schemes for the molecule are determined; The segmentation scheme with the fewest degrees of freedom among the multiple segmentation schemes is determined as the target segmentation scheme; as well as Based on the target segmentation scheme, the plurality of molecular fragments of the molecule are determined.

3. The method according to claim 2, wherein the fragment library is constructed in the following manner: Obtain multiple sample conformations of multiple sample molecules and multiple sample two-dimensional structures corresponding to the multiple sample conformations; The plurality of sample conformations and the plurality of sample two-dimensional structures are segmented to obtain a plurality of sample fragment conformations and a plurality of sample molecular fragments corresponding to the plurality of sample fragment conformations; Determine multiple correlations between the conformations of the multiple sample fragments and the multiple molecular fragments of the multiple samples; Based on the multiple correlations, the multiple sample fragment conformations are clustered to obtain multiple cluster centers; as well as Add the multiple cluster centers to the fragment library.

4. The method according to claim 3, wherein determining the segmentation scheme with the fewest degrees of freedom among the plurality of segmentation schemes as the target segmentation scheme includes: Determine multiple variances of the root mean square deviations of the multiple fragment conformations of the molecule and the corresponding multiple cluster centers; as well as The sum of the multiple variances is determined as the degrees of freedom of the segmentation scheme.

5. The method according to any one of claims 1-4, wherein determining the plurality of fragment conformations and the plurality of combinations of the plurality of dihedral angles as the plurality of conformations of the molecule comprises: Based on the energy distribution of the molecule, multiple candidate fragment conformations are sampled from the plurality of molecular fragments of the molecule; Based on the energy distribution of the molecule, multiple candidate dihedral angles are sampled from the multiple dihedral angles of the molecule; The plurality of candidate fragment configurations and the plurality of candidate dihedral angles are combined multiple times; and The obtained combinations are determined as the multiple conformations of the molecule.

6. The method according to claim 5, further comprising: A topological graph is determined that identifies the connection relationships between the plurality of molecular fragments and the plurality of dihedral angles; The plurality of molecular fragments and the plurality of dihedral angles are defined as the plurality of nodes in the topological graph; as well as Based on the energy distribution of the molecules, multiple candidate nodes are sampled from the multiple nodes of the topology graph.

7. The method of claim 6, wherein sampling a plurality of candidate nodes among the plurality of nodes in the topology graph based on the energy distribution of the molecule comprises: Select one of the plurality of nodes; Keep the other nodes among the plurality of nodes unchanged; Update the energy distribution of the molecule; as well as Based on the updated energy distribution, the plurality of candidate nodes are determined for the given node.

8. The method according to claim 6, further comprising: Based on the fragment library, multiple cluster centers corresponding to the multiple conformations are determined; Determine the multiple fragment conformations represented by the multiple cluster centers; as well as Based on the multiple fragment conformations represented by the multiple cluster centers, the multiple candidate fragment conformations are determined as the multiple conformations.

9. The method of claim 5, wherein determining the plurality of combinations obtained as the plurality of conformations of the molecule comprises: Determine a predetermined number of segment configurations and dihedrals; Cluster the predetermined number of fragment conformations and dihedrals to obtain multiple cluster centers; as well as Multiple combinations of the multiple cluster centers are determined as the multiple conformations.

10. An apparatus for determining the conformation of a molecule, comprising: A molecular segmentation module is configured to segment the molecule into multiple molecular fragments based on the two-dimensional structure of the molecule, at least one of the multiple molecular fragments comprising multiple atoms; The dihedral angle determination module is configured to determine, based on the energy distribution of the molecule, multiple fragment conformations corresponding to the multiple molecular fragments and multiple dihedral angles between adjacent molecular fragments in the multiple molecular fragments; as well as The conformational combination module is configured to determine multiple conformations of the molecule by combining the plurality of fragment conformations and the plurality of dihedral angles. The device further includes: The feature vector determination module is configured to determine multiple feature vectors of the plurality of molecular fragments; The first energy distribution determination module is configured to determine a first energy distribution associated with each of the plurality of molecular fragments; The second energy distribution determination module is configured to determine a second energy distribution associated with each of the plurality of dihedral angles; The third energy distribution determination module is configured to determine a third energy distribution associated with each molecular fragment and each dihedral angle based on each molecular fragment and each dihedral angle; and The joint probability distribution determination module is configured to determine, based on the first energy distribution, the second energy distribution, and the third energy distribution, a joint probability distribution associated with the plurality of molecular fragments and the plurality of dihedral angles as the energy distribution.

11. An electronic device, comprising: processor; as well as A memory coupled to the processor, the memory having instructions stored therein, which, when executed by the processor, cause the electronic device to perform the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having stored thereon computer-executable instructions, wherein the computer-executable instructions are executed by a processor to implement the method according to any one of claims 1 to 9.

Images