Biomacromolecule and ligand affinity prediction method, system, equipment and medium

Abstract

The invention discloses a biomacromolecule and ligand affinity prediction method, system and equipment and a medium. The method comprises the following steps: screening a plurality of candidate ligands for a molecular pocket of a biomacromolecule in a preset ligand library based on a pre-trained affinity prediction model; performing fine tuning on the affinity prediction model to obtain a fine tuning affinity prediction model; predicting a first affinity of the molecular pocket and each candidate ligand based on the fine tuning affinity prediction model, and selecting a preset number of target ligands from the plurality of candidate ligands based on the first affinity. According to the method, virtual screening and lead compound optimization are jointly studied through a pre-trained affinity prediction model, a global molecular pocket-ligand activity relationship is established through virtual screening, and meanwhile, the influence of fine structural differences on affinity is deeply studied by utilizing lead compound optimization data; the binding affinity between the biomacromolecules and the ligands can be predicted more accurately, and the experiment success rate is increased.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence and drug discovery technology, and in particular to a method, system, device and medium for predicting the affinity between a biomacromolecule and a ligand. Background Art

[0002] With major breakthroughs in the field of protein structure prediction, models represented by AlphaFold and RoseTTAFold can accurately predict the three-dimensional structure of almost all proteins. Recently, AlphaFold3 has shown significant progress in predicting the structure of protein-ligand complexes, greatly accelerating the drug discovery process by providing new insights into the interactions between disease-related proteins and ligands. In the drug discovery pipeline, the next challenge after structure prediction is to predict protein-ligand binding affinity. In particular, protein-ligand binding affinity reflects the strength of the interaction between potential ligands and disease-related proteins, and is a key factor in determining drug efficacy and targeting selectivity.

[0003] In early drug discovery, protein-ligand binding affinity is mainly used in two scenarios of early drug development: virtual screening and lead compound optimization. Virtual screening aims to discover active compounds that can bind to target human proteins from large-scale compound libraries; while lead compound optimization is committed to optimizing the structure of these active ligands to improve their binding affinity and drug-like properties, thereby improving their efficacy. In recent years, machine learning methods have made significant progress in both fields, greatly improving computational efficiency while maintaining comparable performance to traditional computational chemistry methods.

[0004] However, existing machine learning methods generally study virtual screening and lead compound optimization separately, while ignoring the interdependence and complementarity between virtual screening and lead compound optimization tasks, which in turn affects the accuracy of affinity prediction between biomacromolecules and ligands.

[0005] Therefore the prior art still needs to be improved and enhanced. Summary of the invention

[0006] The technical problem to be solved by the present application is to provide a method, system, device and medium for predicting the affinity between a biomacromolecule and a ligand in view of the deficiencies of the prior art.

[0007] In order to solve the above technical problems, the first aspect of the present application provides a method for predicting the affinity between a biomacromolecule and a ligand, wherein the method for predicting the affinity between a biomacromolecule and a ligand specifically comprises: Based on the pre-trained affinity prediction model, several candidate ligands are screened for the molecular pockets of biomacromolecules in the preset ligand library; Based on the molecular pocket and several known affinity ligands corresponding to the molecular pocket, fine-tuning the pre-trained affinity prediction model to obtain a fine-tuned affinity prediction model; The first affinity between the molecular pocket and each candidate ligand is predicted based on the fine-tuned affinity prediction model, and a preset number of target ligands are selected from a number of candidate ligands based on the first affinity.

[0008] The method for predicting the affinity between a biomacromolecule and a ligand, wherein the pre-trained affinity prediction model is used to screen a number of candidate ligands for the molecular pocket of the biomacromolecule in a preset ligand library, specifically comprising: Acquire a first pocket embedding vector of a molecular pocket of a biomacromolecule and a first ligand embedding vector of each preset ligand in the preset ligand library based on a pre-trained affinity prediction model; The second affinity between the molecular pocket and each preset ligand is predicted based on the first pocket embedding vector and the first ligand embedding vector, and a number of candidate ligands are selected from the preset ligand library based on the second affinity.

[0009] In the method for predicting the affinity between a biomacromolecule and a ligand, the step of obtaining the first pocket embedding vector of the molecular pocket of the biomacromolecule based on the pre-trained affinity prediction model specifically comprises: Constructing a molecular pocket-ligand knowledge graph for the molecular pocket of the biomacromolecule; The pocket nodes in the molecular pocket-ligand knowledge graph are converted into second pocket embedding vectors through the pre-trained affinity prediction model, and the ligand nodes in the molecular pocket-ligand knowledge graph are converted into second ligand embedding vectors to obtain an embedded knowledge graph; The embedded knowledge graph is input into a preset graph neural network, and the first pocket embedding vector of the molecular pocket of the biological macromolecule is output through the graph neural network.

[0010] In the method for predicting the affinity between a biomacromolecule and a ligand, the step of constructing a molecular pocket-ligand knowledge graph for the molecular pocket of the biomacromolecule specifically comprises: Obtaining similar molecular pockets and known ligands of the molecular pocket corresponding to the biomacromolecule; The molecular pocket and similar molecular pockets are used as pocket nodes, and the known ligand is used as a ligand node; Based on the structural similarity between the pocket nodes and the affinity between the ligand node and the pocket node, edges are constructed between the pocket node and the ligand node to obtain a molecular pocket-ligand knowledge graph.

[0011] The method for predicting the affinity of a biomacromolecule and a ligand, wherein the pre-trained affinity prediction model is fine-tuned based on the molecular pocket and a number of known affinity ligands corresponding to the molecular pocket to obtain a fine-tuned affinity prediction model specifically comprises: encoding the molecular pocket into a third pocket embedding vector by the pre-trained affinity prediction model; encoding each known affinity ligand into a third ligand embedding vector by the pre-trained affinity prediction model; Predicting the third affinity of the molecular pocket to each known affinity ligand based on the third pocket embedding vector and the third ligand embedding vector; constructing a first ranking loss term based on the third affinity of the molecular pocket to each known affinity ligand and the first experimental affinity of the molecular pocket to each known affinity ligand; The pre-trained affinity prediction model is fine-tuned based on the first ranking loss term to obtain a fine-tuned affinity prediction model.

[0012] The method for predicting the affinity between a biomacromolecule and a ligand, wherein the step of predicting the first affinity between the molecular pocket and each candidate ligand based on the fine-tuned affinity prediction model specifically comprises: encoding the molecular pocket into a fourth pocket embedding vector by the fine-tuned affinity prediction model; encoding each candidate ligand as a fourth ligand embedding vector by the fine-tuned affinity prediction model; The first affinity of the molecular pocket to each ligand with known affinity is predicted based on the fourth pocket embedding vector and the fourth ligand embedding vector.

[0013] The affinity prediction method for biomacromolecules and ligands, wherein the affinity prediction model includes a pocket encoder for encoding a molecular pocket into an embedded vector and a ligand encoder for encoding a ligand into an embedded vector.

[0014] In the method for predicting the affinity between a biomacromolecule and a ligand, the pre-training process of the pre-trained affinity prediction model specifically includes: Acquire a training data set, wherein the training data set includes a plurality of experimental measurement groups, each of which includes a training molecular pocket, a plurality of training ligands, and a second experimental affinity of each training ligand to the training molecular pocket; encoding the training molecule pocket in each experimental measurement group into a training pocket embedding vector through the affinity prediction model to be pre-trained, and encoding each training ligand in each experimental measurement group into a training ligand embedding vector; Determining a fourth affinity between the training molecule pocket in each experimental measurement group and each training ligand in each experimental measurement group based on the training pocket embedding vector and the training ligand embedding vector; A loss function term is constructed based on the fourth affinity and the second experimental affinity, and the affinity prediction model to be pre-trained is trained based on the loss function term to obtain a pre-trained affinity prediction model.

[0015] In the method for predicting the affinity between a biomacromolecule and a ligand, the step of obtaining a training data set specifically comprises: Collect experimentally determined known affinity data between biomacromolecules and ligands; Dividing the acquired known affinity data into a number of initial experimental measurement groups based on the same biomacromolecule and experimental measurement method, and matching a molecular pocket for each of the initial experimental measurement groups; Using the molecular pocket matched by each of the initial experimental measurement groups as a training molecular pocket, the ligand in each of the initial experimental measurement groups as a training ligand, and the experimentally determined affinity in the initial experimental measurement group as a second experimental affinity, to obtain an experimental measurement group corresponding to each of the initial experimental measurement groups; The data set formed by the experimental measurement group is used as a training data set.

[0016] The method for predicting the affinity of a biomacromolecule and a ligand, wherein matching molecular pockets for each of the initial experimental measurement groups specifically comprises: Selecting a biomacromolecule-ligand complex structure for each biomacromolecule in the initial experimental measurement group based on biomacromolecule similarity in a known biomacromolecule structure database; respectively calculating the ligand similarity between the crystal ligand bound to the biomacromolecule-ligand complex structure selected in each of the initial experimental measurement groups and the ligand in each of the initial experimental measurement groups; The target crystal ligands are matched for each initial experimental measurement group based on ligand similarity, and the molecular pocket corresponding to the target crystal ligand matched by each initial experimental measurement group is used as the molecular pocket corresponding to each initial experimental measurement group.

[0017] In the method for predicting the affinity between a biomacromolecule and a ligand, the step of constructing a loss function based on the fourth affinity and the second experimental affinity specifically comprises: constructing a second ranking loss term for each experimental measurement value group based on a fourth affinity of the training molecule pocket in each experimental measurement group to the training ligand in each experimental measurement group and a second experimental affinity; constructing a comparative loss term based on the fourth affinity between the training molecule pocket in each experimental measurement group and each training ligand in each experimental measurement group; The loss function term is calculated based on the second ranking loss term of each experimental measurement value group and the contrast loss term.

[0018] The second aspect of the present application provides a system for predicting the affinity between a biomacromolecule and a ligand, wherein the system for predicting the affinity between a biomacromolecule and a ligand specifically comprises: A screening module is used to screen several candidate ligands for the molecular pocket of the biomacromolecule in a preset ligand library based on a pre-trained affinity prediction model; A fine-tuning module, used for fine-tuning the pre-trained affinity prediction model based on the molecular pocket and a plurality of known affinity ligands corresponding to the molecular pocket to obtain a fine-tuned affinity prediction model; The lead compound optimization module is used to predict the first affinity between the molecular pocket and each candidate ligand based on the fine-tuned affinity prediction model, and select a preset number of target ligands from a number of candidate ligands based on the first affinity.

[0019] The affinity prediction system for biomacromolecules and ligands, wherein the affinity prediction model includes a pocket encoder for encoding a molecular pocket into an embedded vector and a ligand encoder for encoding a ligand into an embedded vector.

[0020] The third aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in any of the above-mentioned methods for predicting the affinity between a biomacromolecule and a ligand, and / or, The computer-readable storage medium stores the affinity prediction system between biological macromolecules and ligands as described above.

[0021] A fourth aspect of the present application provides a terminal device, comprising: a processor and a memory; The memory stores a computer-readable program executable by the processor; When the processor executes the computer-readable program, the processor implements the steps in any of the above-mentioned methods for predicting the affinity between a biomacromolecule and a ligand.

[0022] Beneficial effects: Compared with the prior art, the present application provides a method, system, device and medium for predicting the affinity between a biomacromolecule and a ligand, the method comprising: based on a pre-trained affinity prediction model, screening a number of candidate ligands for the molecular pocket of the biomacromolecule in a preset ligand library; based on the molecular pocket and a number of known affinity ligands corresponding to the molecular pocket, fine-tuning the pre-trained affinity prediction model to obtain a fine-tuned affinity prediction model; based on the fine-tuned affinity prediction model, predicting the first affinity between the molecular pocket and each candidate ligand, and selecting a preset number of target ligands from a number of candidate ligands based on the first affinity. The present application combines virtual screening and lead compound optimization through a pre-trained affinity prediction model, establishes a global molecular pocket-ligand activity relationship through virtual screening, and uses lead compound optimization data to deeply study the influence of subtle structural differences on affinity, which can more accurately predict the binding affinity between biomacromolecules and ligands, improve the success rate of experiments, and reduce research and development costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 A flow chart of a method for predicting affinity between a biomacromolecule and a ligand provided in an embodiment of the present application.

[0025] Figure 2 Schematic flow chart of the screening process for candidate ligands.

[0026] Figure 3 Schematic flow chart of the pre-training process for the affinity prediction model.

[0027] Figure 4 Schematic flow chart of the fine-tuning process of the affinity prediction model and the prediction process of affinity by fine-tuning the affinity prediction model.

[0028] Figure 5 This is a principle block diagram of the affinity prediction system for biomacromolecules and ligands provided in the embodiments of the present application.

[0029] Figure 6 A functional block diagram of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0030] The embodiments of the present application provide a method, system, device and medium for predicting the affinity between a biomacromolecule and a ligand. In order to make the purpose, technical solution and effect of the present application clearer and more specific, the present application is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0031] Those skilled in the art will appreciate that, unless expressly stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" to another element, it may be directly connected to the other element, or there may be intermediate elements. In addition, the "connection" used here may include wireless connections. The term "and / or" used here includes all or any unit and all combinations of one or more associated listed items.

[0032] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with those in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as here.

[0033] It should be understood that the sequence numbers and sizes of the steps in this embodiment do not mean the order of execution. The execution order of each process is determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0034] The application content is further explained below through the description of embodiments in conjunction with the accompanying drawings.

[0035] This embodiment provides a method for predicting the affinity between a biomacromolecule and a ligand. Figure 1 As shown, the method includes: S10. Based on the pre-trained affinity prediction model, several candidate ligands are screened for the molecular pockets of the biomacromolecule in the preset ligand library.

[0036] Specifically, biomacromolecules refer to macromolecules such as proteins, nucleic acids, polysaccharides, etc. that exist in biological cells, and ligands are small molecules used to bind to biomacromolecules, such as some small molecules of traditional Chinese medicine ingredients, antibiotics, etc. Molecular pockets are partial structures of biomacromolecules, and molecular pockets can bind to ligands, that is, molecular pockets are binding sites in biomacromolecules that can bind to ligands. Compared with the entire biomacromolecule, molecular pockets can provide more detailed binding site information, thereby more accurately predicting the affinity between biomacromolecules and ligands.

[0037] The affinity prediction model is a pre-trained deep learning network model. The pre-trained affinity prediction model can be used for virtual screening, that is, to screen several candidate ligands for the molecular pocket of the biomacromolecule in the preset ligand library. When performing virtual screening through the pre-trained affinity prediction model, the pre-trained affinity prediction model can be used to extract the embedding vectors of the molecular pocket of the biomacromolecule and the preset ligand in the preset ligand library, respectively, and then affinity prediction can be performed by aligning the embedding vectors of the two, and finally the candidate ligands can be screened based on the predicted affinity.

[0038] Exemplarily, the method of screening several candidate ligands for the molecular pocket of the biomacromolecule in the preset ligand library based on the pre-trained affinity prediction model specifically includes: S11, acquiring a first pocket embedding vector of a molecular pocket of a biomacromolecule and a first ligand embedding vector of each preset ligand in the preset ligand library based on a pre-trained affinity prediction model; S12. Predicting the second affinity between the molecular pocket and each preset ligand based on the first pocket embedding vector and the first ligand embedding vector, and selecting a number of candidate ligands from the preset ligand library based on the second affinity.

[0039] Specifically, in step S11, the first pocket embedding vector and the first ligand embedding vector are vector representations in the same vector space, that is, the vector dimensions of the first pocket embedding vector and the first ligand embedding vector are the same, for example, the vector dimensions of the first pocket embedding vector and the first ligand embedding vector are both 128 dimensions, etc. The first pocket embedding vector may be obtained by directly mapping the three-dimensional structure and polar surface of the molecular pocket to the vector space through the affinity prediction model, or may be obtained by first extracting the embedding vector corresponding to the molecular pocket through the affinity prediction model, and then optimizing the embedding vector corresponding to the molecular pocket extracted through the affinity prediction model by using the prior knowledge that molecular pockets with similar structures may share similar ligands.

[0040] In a specific implementation, the obtaining of the first pocket embedding vector of the molecular pocket of the biomacromolecule based on the pre-trained affinity prediction model specifically includes: Constructing a molecular pocket-ligand knowledge graph for the molecular pocket of the biomacromolecule; The pocket nodes in the molecular pocket-ligand knowledge graph are converted into second pocket embedding vectors through the pre-trained affinity prediction model, and the ligand nodes in the molecular pocket-ligand knowledge graph are converted into second ligand embedding vectors to obtain an embedded knowledge graph; The embedded knowledge graph is input into a preset graph neural network, and the first pocket embedding vector of the molecular pocket of the biological macromolecule is output through the graph neural network.

[0041] Specifically, the molecular pocket-ligand knowledge graph is used as the prior knowledge of the molecular pocket, which includes the molecular pocket, similar molecular pockets with similar structures to the molecular pocket, and several known ligands. The relationship between the molecular pocket and similar molecular pockets and known ligands can be obtained through the molecular pocket-ligand knowledge graph. In other words, the molecular pocket is provided with prior knowledge through the affinity (intensity of interaction) relationship between the same molecular pocket and similar molecular pockets and known ligands in the molecular pocket-ligand knowledge graph, and the pocket node extracted by the pocket encoder in the affinity prediction model is converted into a second pocket embedding vector for optimization based on the prior knowledge, so as to achieve rapid and accurate screening of massive candidate ligands.

[0042] Exemplarily, the step of constructing a molecular pocket-ligand knowledge graph for the molecular pocket of the biomacromolecule specifically includes: Obtaining similar molecular pockets and known ligands of the molecular pocket corresponding to the biomacromolecule, and using the molecular pocket and the similar molecular pocket as pocket nodes, and using the known ligand as a ligand node; Based on the structural similarity between the pocket nodes and the affinity between the ligand node and the pocket node, edges are constructed between the pocket node and the ligand node to obtain a molecular pocket-ligand knowledge graph.

[0043] Specifically, the similar molecular pocket is a molecular pocket with a structure similar to that of the molecular pocket, and the known ligand is a ligand known to be able to bind to the molecular pocket. For example, the molecular pocket is a binding pocket of a human sodium ion channel protein, the similar molecular pocket can be a binding pocket of an ion channel protein with a similar structure, and the known ligand can be a known sodium ion channel blocker (such as lidocaine, mexiletine, etc.).

[0044] After obtaining the similar molecular pockets and known ligands corresponding to the molecular pockets of the biomacromolecule, the molecular pockets and similar molecular pockets are used as pocket nodes, and the known ligands are used as ligand nodes. Then, edges are constructed between the nodes to form a molecular pocket-ligand knowledge graph, wherein the edges may include edges from pocket nodes to pocket nodes for reflecting the structural similarity between molecular pockets, and may also include edges from ligand nodes to pocket nodes for reflecting the affinity between pocket nodes and ligand nodes. To this end, when constructing edges between nodes, the structural similarity between pocket nodes and the affinity between the ligand nodes and the pocket nodes may be used as the basis, wherein the structural similarity between pocket nodes may be determined by calculating the similarity scores (e.g., sequence similarity) between pocket nodes, and the edges from ligand nodes to pocket nodes may be connected based on the affinity data measured experimentally. Specifically, the structural similarity threshold between pocket nodes and the affinity threshold between the ligand node and the pocket node can be set in advance, and then edges are established between pocket nodes whose structural similarity between pocket nodes is greater than the structural similarity threshold, and edges are established between ligand nodes and pocket nodes whose affinity from ligand nodes to pocket nodes is greater than the affinity threshold.

[0045] For example: Assume that the pocket nodes include human sodium channel and rat sodium channel, the ligand node includes lidocaine, the experimentally measured affinity of lidocaine to sodium channel is greater than the affinity threshold, and the binding pocket sequence similarity between human sodium channel and rat sodium channel is higher than the structural similarity threshold. Then, establish edges between lidocaine and human sodium channel, and between human sodium channel and rat sodium channel.

[0046] Further, if Figure 2As shown, after constructing a molecular pocket-ligand knowledge graph for the molecular pocket of the biomacromolecule, in order to obtain the embedding vector of the molecular pocket, the pocket node and the ligand node in the pocket-ligand knowledge graph can be converted into an embedding vector, that is, the nodes in the pocket-ligand knowledge graph are converted into embedding vectors to obtain the embedded knowledge graph. Specifically, when converting the nodes in the pocket-ligand knowledge graph into embedding vectors, a pre-trained affinity prediction model can be used, wherein the affinity prediction model includes a pocket encoder (e.g., Uni-Mol) for encoding the molecular pocket into an embedding vector and a ligand encoder (e.g., Uni-Mol, SmilesBERT) for encoding the ligand into an embedding vector, the pocket node is encoded by the pocket encoder to convert the pocket node into a second pocket embedding vector, and the ligand node is encoded by the ligand encoder to convert the ligand node into a second ligand embedding vector. In addition, after obtaining the embedded knowledge graph, the embedded knowledge graph can be input into a preset graph neural network, which is a heterogeneous graph neural network (such as GraphRec). The information in the embedded knowledge graph is integrated through the graph neural network to output the first pocket embedding vector corresponding to the molecular pocket. In this way, the embedding vector corresponding to the molecular pocket can be optimized in combination with prior knowledge, so that the first pocket embedding vector corresponding to the molecular pocket carries more accurate knowledge information.

[0047] Further, in step S22, after obtaining the first pocket embedding vector of the molecular pocket and the first ligand embedding vector of each preset ligand in the preset ligand library, the molecular pocket and each preset ligand can be predicted based on the first pocket embedding vector and the first ligand embedding vector, and the second affinity between the molecular pocket and the preset ligand can be determined by vector alignment. For example, vector alignment can be performed using cosine similarity, that is, the cosine similarity between the first pocket embedding vector and each first ligand embedding vector can be calculated, and the calculated cosine similarity can be used as the second affinity between the molecular pocket and the preset ligand. Moreover, if the first ligand embedding vector is closer to the first pocket embedding vector, the cosine similarity between the first pocket embedding vector and the first ligand embedding vector is greater, which means that the non-covalent interaction between the molecular pocket and the preset ligand is also higher, and thus the second affinity between the molecular pocket and the preset ligand is higher. Conversely, if the first ligand embedding vector is farther from the first pocket embedding vector, the cosine similarity between the first pocket embedding vector and the first ligand embedding vector is smaller, which means that the non-covalent interaction between the molecular pocket and the preset ligand is also lower, and thus the second affinity between the molecular pocket and the preset ligand is lower. In addition, after obtaining the second affinity between the molecular pocket and each preset ligand, several candidate ligands can be selected from the preset ligand library for the molecular pocket in order of the second affinity from high to low, which can quickly search for high-affinity active ligands for the molecular pocket, and use N active ligands with high similarity as candidate ligands. Its search speed exceeds the traditional method (such as Vina Docking) by six orders of magnitude.

[0048] The embodiment of the present application constructs a molecular pocket-ligand knowledge graph for the molecular pocket, and converts the pocket nodes and ligand nodes in the molecular pocket-ligand knowledge graph into embedded vectors in the same vector space through a pre-trained affinity prediction model, and then uses a graph neural network to learn the information of the embedded knowledge graph to obtain the first pocket embedding vector of the molecular pocket, which can make full use of the prior knowledge that similar ligands may be shared between protein pockets with similar structures, thereby improving the accuracy of screening. At the same time, after obtaining the first pocket embedding vector and the first ligand embedding vector of the preset ligand in the preset ligand library, the affinity is directly predicted by vector alignment, and a number of active ligands are selected as candidate ligands for the molecular pocket from the preset ligand library in order of affinity from high to low, thereby greatly improving the screening efficiency of candidate ligands.

[0049] The above is a description of the virtual screening process in the embodiment of the present application. A pre-trained affinity prediction model is used in the virtual screening process. The following is a specific description of the pre-training process of the affinity prediction model to be pre-trained using a training data set to train the affinity prediction model to obtain the affinity prediction model, wherein the affinity prediction model to be pre-trained has the same model structure as the affinity prediction model, such as Figure 3As shown, the pre-trained affinity prediction models include pocket encoders (such as Uni-Mol, etc.) and ligand encoders (such as Uni-Mol, SmilesBERT, etc.), and the difference between the two lies in the model parameters. The affinity prediction model to be pre-trained uses the initial model parameters, and the pre-trained affinity prediction model uses the model parameters trained with the training data set.

[0050] Exemplarily, the pre-training process of the pre-trained affinity prediction model specifically includes: H10, obtain training data set; H20, encoding the training molecule pocket in each experimental measurement group into a training pocket embedding vector through the affinity prediction model to be pre-trained, and encoding each training ligand in each experimental measurement group into a training ligand embedding vector; H30, determining the fourth affinity between the training molecule pocket in each experimental measurement group and each training ligand in each experimental measurement group based on the training pocket embedding vector and the training ligand embedding vector; H40. Construct a loss function term based on the fourth affinity and the second experimental affinity, and train the affinity prediction model to be pre-trained based on the loss function term to obtain a pre-trained affinity prediction model.

[0051] Specifically, in step H10, the training data set is used to train the affinity prediction model to be pre-trained to obtain the pre-trained affinity prediction model, wherein the training data set includes several experimental measurement groups, each of which includes a training molecular pocket, several training ligands, and a second experimental affinity of each training ligand with the training molecular pocket, the second experimental affinity is obtained by measuring by experimental means, and the experimental means corresponding to the training ligands in each experimental measurement group are the same. In addition, for any two experimental measurement groups in the several experimental measurement groups, the training molecular pockets in the two experimental measurement groups may be the same or different. For example, the several experimental measurement groups include experimental measurement group A, experimental measurement group B, and experimental measurement group C, and the training molecular pockets in the three experimental groups may be the same or different, but the experimental means corresponding to each training ligand in each experimental measurement group are the same.

[0052] Based on this, in one implementation, obtaining the training data set specifically includes: Obtain experimentally measured known affinity data between biomacromolecules and ligands; A biomacromolecule, a ligand whose affinity between the biomacromolecule and the ligand is measured by the same experimental measurement method, and known affinity data obtained by measurement are used as an initial experimental measurement group to obtain a plurality of initial experimental measurement groups, and a molecular pocket of the biomacromolecule is matched for each of the initial experimental measurement groups; Using the molecular pocket matched by each of the initial experimental measurement groups as a training molecular pocket, the ligand in each of the initial experimental measurement groups as a training ligand, and the experimentally determined affinity in the initial experimental measurement group as a second experimental affinity, to obtain an experimental measurement group corresponding to each of the initial experimental measurement groups; The data set formed by the experimental measurement group is used as a training data set.

[0053] Specifically, the known affinity data can be collected through relevant databases (such as ChEMBL, BindingDB), relevant literature libraries and wet experiments. The known affinity data include biological macromolecules, ligands, experimental measurement methods for the affinity of biological macromolecules and ligands, and experimentally measured affinity. The experimentally measured affinity is obtained by experimentally measuring the affinity of biological macromolecules and ligands using the experimental measurement method.

[0054] After the known affinity data are collected, the known affinity data are divided based on the same biological macromolecule and experimental measurement method, so that the known affinity data with the same biological macromolecule and the same experimental measurement method are divided into an initial experimental measurement group. For example, the collected known affinity data include the known affinity data of 10 different ligands for human sodium ion channel protein measured by the same experimental measurement method, then the known affinity data corresponding to these 10 ligands are divided into an initial experimental measurement group, and the initial experimental measurement group is about human sodium ion channel protein.

[0055] Further, after the division to obtain a number of initial experimental measurement groups, a molecular pocket is matched for each initial experimental measurement group, wherein the molecular pocket is a binding position where the ligand can bind to the biological macromolecule, and by matching the molecular pocket for the biological macromolecule, more detailed binding point information can be provided. Specifically, when matching the molecular pocket for each initial experimental measurement group, the molecular pocket can be matched for each initial experimental measurement group in a preset biological macromolecule database, or the molecular pocket can be matched for each initial experimental measurement group through human-computer interaction, or the molecular pocket can be matched for the initial experimental measurement group through a large language model, etc.

[0056] In a specific implementation, matching molecular pockets for each of the initial experimental measurement groups specifically includes: Selecting a biomacromolecule-ligand complex structure for each biomacromolecule in the initial experimental measurement group based on biomacromolecule similarity in a known biomacromolecule structure database; respectively calculating the ligand similarity between the crystal ligand bound to the biomacromolecule-ligand complex structure selected in each of the initial experimental measurement groups and the ligand in each of the initial experimental measurement groups; The target crystal ligands are matched for each initial experimental measurement group based on ligand similarity, and the molecular pocket corresponding to the target crystal ligand matched by each initial experimental measurement group is used as the molecular pocket corresponding to each initial experimental measurement group.

[0057] Specifically, the known biomacromolecule structure database is an existing biomacromolecule structure database (such as PDB, etc.), and the known biomacromolecule structure database includes several biomacromolecule-ligand complex structures, that is, the known biomacromolecule structures in the known biomacromolecule structure database include biomacromolecules and crystal ligands. To this end, when matching molecular pockets for the initial experimental measurement group, several biomacromolecule-ligand complex structures can be selected from the known biomacromolecule structure database based on biomacromolecule similarity (such as sequence similarity exceeding 40%, etc.), and then based on ligand similarity (such as ligand similarity greater than a preset ligand similarity threshold, etc.), target crystal ligands are selected for biomacromolecules in the candidate biomacromolecule-ligand complex structures, and finally the molecular pocket corresponding to the target crystal ligand is used as the molecular pocket corresponding to the initial experimental measurement group. Among them, biomacromolecule similarity refers to the sequence similarity between biomacromolecules in the initial experimental measurement group and biomacromolecules in the known biomacromolecule structure, and ligand similarity refers to the similarity between ligands in the initial experimental measurement group and crystal ligands in the known biomacromolecule structure (such as Tanimoto similarity, etc.).

[0058] For example: Assuming that the initial experimental measurement group includes known affinity data of human sodium channel protein and 10 different ligands, and the known biological macromolecule structure database is PDB, then the sequence similarity between human sodium channel protein and the biological macromolecules in each PDB structure is calculated first, and the PDB structure with a sequence similarity of more than 40% is selected as the candidate PDB structure; then, the ligand similarity (such as Tanimoto similarity) between the crystal ligand in the candidate PDB structure and the 10 ligands in the initial experimental measurement group is calculated respectively. When the ligand similarity with any of the 10 ligands exceeds a certain threshold, the molecular pocket of the crystal ligand is used as the molecular pocket of the initial experimental measurement group. In addition, it should be noted that when multiple molecular pockets are matched, the molecular pocket of the crystal ligand with the highest ligand similarity is used as the molecular pocket of the initial experimental measurement group.

[0059] Further, when a molecular pocket is matched for each initial experimental measurement group, the molecular pocket matched by each initial experimental measurement group is used as the training molecular pocket corresponding to each initial experimental measurement group, and the ligand in each initial experimental measurement group is used as the training ligand corresponding to each initial experimental measurement group, and the experimentally determined affinity between the ligand in each initial experimental measurement group and the biological macromolecule is used as the second experimental affinity corresponding to each initial experimental measurement group; then, the set consisting of the training molecular pocket, the training ligand and the second experimental affinity corresponding to each initial experimental measurement group is used as the experimental measurement group corresponding to each initial experimental measurement group to obtain a plurality of experimental measurement groups; finally, the data set consisting of all the experimental measurement groups is used as the training data set.

[0060] The embodiment of the present application constructs an experimental measurement group based on biomacromolecules and experimental measurement methods, thereby ensuring the comparability of affinity data within the experimental measurement group and avoiding batch effects caused by different experimental methods (for example, pH value has an impact on the affinity results of ligands and human sodium ion channel proteins, and it is unreasonable to directly compare data at different pH values). At the same time, the present application constructs a large-scale training data set for the experimental measurement group by collecting known affinity data from relevant databases (such as ChEMBL, BindingDB), relevant literature libraries, and wet experiments, covering more than 400,000 ligands and 40,000 protein pockets, ensuring the richness and diversity of the training data set.

[0061] Further, in step H20, after obtaining the training data set, the affinity prediction model to be pre-trained is trained based on the training data set. Specifically, each experimental measurement group in the training data set is input into the affinity prediction model to be pre-trained, and each training molecular pocket in the experimental measurement group is encoded into a training pocket embedding vector through the pocket encoder in the affinity prediction model to be pre-trained, and each training ligand in the experimental measurement group is encoded into a training ligand embedding vector through the ligand encoder in the affinity prediction model to be pre-trained.

[0062] In addition, it should be noted that when training the pre-trained affinity prediction model based on the training data set, the training data set can be divided into several training batches, each training batch includes a preset number of experimental measurement value groups, and then each training batch is used as training data for a training iteration process, and the training process of each training batch is the same, wherein the number of experimental measurement groups included in each training batch can be determined according to actual needs, and no specific limitation is made here.

[0063] Further, in step H30, after obtaining the training pocket embedding vector and the training ligand embedding vector, the fourth affinity between the training molecular pocket and the training ligand can be predicted by vector alignment. For example, the fourth affinity between the training molecular pocket and the training ligand can be predicted by calculating the cosine similarity between the training pocket embedding vector and the training ligand embedding vector.

[0064] Further, in step H40, after obtaining the fourth affinity between the training molecular pocket and the training ligand, a loss function term can be constructed based on the fourth affinity and the second experimental affinity, wherein the loss function term can include a permutation loss term between the training molecular pocket and the training ligand in the experimental measurement group, can also include a comparison loss term between the experimental measurement groups, can also include both a permutation loss term between the training molecular pocket and the training ligand in the experimental measurement group and a comparison loss term between the experimental measurement groups, and so on.

[0065] In a specific implementation, constructing a loss function term based on the fourth affinity and the second experimental affinity specifically includes: constructing a second ranking loss term for each experimental measurement value group based on a fourth affinity of the training molecule pocket in each experimental measurement group to the training ligand in each experimental measurement group and a second experimental affinity; constructing a comparative loss term based on the fourth affinity between the training molecule pocket in each experimental measurement group and each training ligand in each experimental measurement group; The loss function term is calculated based on the second ranking loss term of each experimental measurement value group and the contrast loss term.

[0066] Specifically, the second ranking loss term is used for list ranking learning, that is, ranking learning is performed for the training molecular pockets and training ligands in each experimental measurement group, so that the affinity ranking predicted by the model is as consistent as possible with the affinity ranking measured experimentally. Wherein, the determination process of the second ranking loss term can be to first adopt a ranking method, perform affinity ranking based on the fourth affinity between the training molecular pockets and the training ligands in the experimental measurement group to form a predicted affinity sequence, and perform ranking based on the second experimental affinity between the training molecular pockets and the training ligands in the experimental measurement group to form an experimental affinity sequence, and then calculate the second ranking loss term based on the predicted affinity sequence and the experimental affinity sequence, wherein the ranking method can adopt an existing ranking method, for example, the Plackett-Luce model, etc.

[0067] The contrast loss term is used for contrast learning in the entire training data set, so that the pre-trained affinity prediction model can quickly distinguish between active ligands and inactive ligands. When determining the contrast loss term, the training molecule pocket-training ligand pair with the second experimental affinity is used as a positive sample (for example, the binding pocket of a sodium channel blocker and the human sodium channel protein), and the training molecule pocket-training ligand pair without the second experimental affinity is used as a negative sample. This can maximize the similarity of the positive sample pair and minimize the similarity of the negative sample pair, so that the pre-trained affinity prediction model can distinguish between active ligands and inactive ligands, wherein the active ligand is the ligand ranked in descending order of affinity as the top N as the active ligand.

[0068] Furthermore, after obtaining the second ranking loss term and the contrast loss term, the second ranking loss term and the contrast loss term can be weighted to determine the loss function term, so that the loss function term combines list ranking learning and contrast learning. List ranking learning enables the pre-trained affinity prediction model to accurately predict the affinity ranking of ligands in the same experimental measurement group (for example, the activity ranking of 10 sodium channel blockers), while contrast learning enables the pre-trained affinity prediction model to distinguish active and inactive ligands (for example, distinguishing sodium channel blockers from non-blockers), which makes the pre-trained affinity prediction model suitable not only for virtual screening scenarios, but also for lead compound optimization scenarios, and can utilize the intrinsic association between virtual screening scenarios and lead compound optimization scenarios to improve the effect in virtual screening scenarios and lead compound optimization scenarios. Of course, the pre-trained affinity prediction model is also applicable to affinity prediction scenarios, that is, the embedding vector of the molecular pocket and the embedding vector of the ligand are encoded by the pre-trained affinity prediction model, and then the affinity between the molecular pocket and the ligand is predicted directly by vector alignment.

[0069] S20. Based on the molecular pocket and several known affinity ligands corresponding to the molecular pocket, fine-tune the pre-trained affinity prediction model to obtain a fine-tuned affinity prediction model.

[0070] Specifically, after applying the pre-trained affinity prediction model to a virtual screening scenario to obtain several candidate ligands, several active ligands are obtained. The pre-trained affinity prediction model and several active ligands can be further applied to optimize lead compounds. In order to ensure the prediction performance of the pre-trained affinity prediction model in the new scenario, Figure 4As shown, the pre-trained affinity prediction model can be fine-tuned based on known experimental data, and then the fine-tuned affinity prediction model can be used to optimize the lead compound. When the pre-trained affinity prediction model is fine-tuned, the pre-trained affinity prediction model is controlled to perform ranking learning, and by learning the relative affinity ranking relationship between ligands, the fine-tuned affinity prediction model can have better prediction performance on the molecular pocket.

[0071] Exemplarily, the fine-tuning of the pre-trained affinity prediction model based on the molecular pocket and several known affinity ligands corresponding to the molecular pocket to obtain the fine-tuned affinity prediction model specifically includes: encoding the molecular pocket into a third pocket embedding vector by a pocket encoder in a pre-trained affinity prediction model; Each known affinity ligand is encoded into a third ligand embedding vector by a ligand encoder in a pre-trained affinity prediction model; Predicting the third affinity of the molecular pocket to each known affinity ligand based on the third pocket embedding vector and the third ligand embedding vector; constructing a first ranking loss term based on the third affinity of the molecular pocket to each known affinity ligand and the first experimental affinity of the molecular pocket to each known affinity ligand; The pre-trained affinity prediction model is fine-tuned based on the first ranking loss term to obtain a fine-tuned affinity prediction model.

[0072] Specifically, the known affinity ligand is a ligand that is experimentally tested with the molecular pocket, that is, several ligands and the molecular pocket are experimentally measured in advance to obtain the affinity between several ligands and the molecular pocket, and the several ligands are used as known affinity ligands, wherein the several ligands for lead compound optimization include several candidate ligands obtained from virtual screening. That is to say, in addition, the known affinity ligand is a candidate ligand among the several candidate ligands screened in step S10, that is, after screening several candidate ligands, the several candidate ligands can be divided into two candidate ligand groups, one candidate ligand group is used as an experimental ligand group, and the other is used as an optimized ligand group. Then, the known affinity between the molecular pocket and each candidate ligand in the experimental ligand group is determined experimentally, and the candidate ligands in the experimental ligand group are used as known affinity ligands for fine-tuning the affinity prediction model. The embodiments of the present application obtain affinity data of a small number of candidate ligands among several candidate ligands through experiments, and use the affinity data to train a pre-trained affinity model to provide knowledge information of the relative affinity ranking relationship between the ligands of the structural class of compounds, so that the fine-tuned affinity prediction model can learn the relative affinity ranking relationship between the ligands of the structural class of compounds instead of directly predicting the absolute affinity value, so that the fine-tuned affinity prediction model can be used to accurately predict the affinity of the structural analog.

[0073] In addition, the third pocket embedding vector is obtained by encoding through the pocket encoder and the third ligand embedding vector is obtained by encoding through the ligand encoder. The third affinity is obtained by adopting a vector alignment method (such as calculating the cosine similarity of the third pocket embedding vector and the third ligand embedding vector). These processes are the same as the prediction process of the above-mentioned embedding vectors and affinities, and will not be repeated here.

[0074] After obtaining the third affinity between the molecular pocket and each known affinity ligand, a ranking method can be used to determine the predicted affinity sequence based on the third affinity, and to determine the experimental affinity sequence based on the first experimental affinity. Finally, the first ranking loss item is determined based on the predicted affinity sequence and the experimental affinity sequence, and the affinity prediction model is fine-tuned based on the first ranking loss item, so that the fine-tuned affinity prediction model can have better prediction performance on the molecular pocket. The process of determining the predicted affinity sequence and the experimental affinity sequence by the sampling ranking method is the same as the ranking method used in determining the second ranking loss item mentioned above. For details, please refer to the description of the second ranking loss item, which will not be repeated here.

[0075] For example: assuming that the binding affinity data of lidocaine and its 10 structural analogs have been experimentally measured, then in the fine-tuning process, the embedding vector similarity between these 10 ligands and the sodium ion channel protein pocket will be calculated as the third affinity, and then the third affinity will be compared with the first experimental affinity in the experimental data and the first ranking loss term (such as the Plackett-Luce model) will be calculated. Finally, the parameters of the pocket encoder and the ligand encoder in the affinity prediction model are adjusted according to the first ranking loss term to fine-tune the pre-trained affinity prediction model to obtain a fine-tuned affinity prediction model.

[0076] S30, predicting the first affinity between the molecular pocket and each candidate ligand based on the fine-tuned affinity prediction model, and selecting a preset number of target ligands from a number of candidate ligands based on the first affinity.

[0077] Specifically, the first affinity is determined based on the fine-tuned affinity prediction model, and its determination process is the same as the above-mentioned affinity determination process, that is, the molecular pocket is first encoded as a fourth pocket embedding vector through the fine-tuned affinity prediction model; each candidate ligand is encoded as a fourth ligand embedding vector through the fine-tuned affinity prediction model; finally, based on the fourth pocket embedding vector and the fourth ligand embedding vector, the first affinity of the molecular pocket and each known affinity ligand is predicted, and a preset number of target ligands are selected from a number of candidate ligands based on the first affinity, wherein the preset number of target ligands are the first preset number of candidate ligands in descending order according to the first affinity.

[0078] The embodiment of the present application uses a small amount of experimental data to fine-tune the pre-trained affinity prediction model, and then selects the target ligand for the molecular pocket by fine-tuning the affinity prediction model, so that the structure-affinity relationship information of the known compound can be fully utilized to improve the prediction accuracy of the affinity between the candidate ligand and the molecular pocket. In addition, as new experimental data is generated, the fine-tuned affinity prediction model can be fine-tuned again through the new experimental data, so that the affinity prediction model can be continuously updated and optimized, so that the affinity of the newly designed compound can be predicted by the fine-tuned affinity prediction model, so as to adjust the optimization method in time, which can not only significantly shorten the optimization cycle of the lead compound, but also reduce the number of compounds that need to be synthesized and tested, improve the success rate of the experiment, and thus reduce the cost of research and development.

[0079] In summary, this embodiment provides a method for predicting the affinity of a biomacromolecule and a ligand, the method comprising: based on a pre-trained affinity prediction model, screening a number of candidate ligands for the molecular pocket of the biomacromolecule in a preset ligand library; based on the molecular pocket and a number of known affinity ligands corresponding to the molecular pocket, fine-tuning the pre-trained affinity prediction model to obtain a fine-tuned affinity prediction model; based on the fine-tuned affinity prediction model, predicting the first affinity of the molecular pocket and each candidate ligand, and selecting a preset number of target ligands from a number of candidate ligands based on the first affinity. This application uses a pre-trained affinity prediction model that combines permutation learning and contrast learning to jointly study virtual screening and lead compound optimization, establishes a global molecular pocket-ligand activity relationship through virtual screening, and uses lead compound optimization data to deeply study the influence of subtle structural differences on affinity, which can more accurately predict the binding affinity between biomacromolecules and ligands, improve the success rate of experiments, and reduce R&D costs.

[0080] Based on the above-mentioned method for predicting the affinity between a biomacromolecule and a ligand, this embodiment provides a system for predicting the affinity between a biomacromolecule and a ligand, such as Figure 5 As shown, the affinity prediction system for biomacromolecules and ligands specifically includes: A screening module 100 is used to screen a number of candidate ligands for the molecular pocket of the biomacromolecule in a preset ligand library based on a pre-trained affinity prediction model; A fine-tuning module 200, configured to fine-tune the pre-trained affinity prediction model based on the molecular pocket and a plurality of known affinity ligands corresponding to the molecular pocket to obtain a fine-tuned affinity prediction model; The lead compound optimization module 300 is used to predict the first affinity between the molecular pocket and each candidate ligand based on the fine-tuned affinity prediction model, and select a preset number of target ligands from a number of candidate ligands based on the first affinity.

[0081] Based on the above-mentioned method for predicting the affinity between a biological macromolecule and a ligand, this embodiment provides a computer-readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the steps in the method for predicting the affinity between a biological macromolecule and a ligand as described in the above-mentioned embodiment.

[0082] Based on the above-mentioned method for predicting the affinity between biomacromolecules and ligands, the present application also provides a terminal device, such as Figure 6As shown, it includes at least one processor 20; a display screen 21; and a memory 22, and may also include a communications interface 23 and a bus 24. The processor 20, the display screen 21, the memory 22, and the communications interface 23 may communicate with each other through the bus 24. The display screen 21 is configured to display a preset user guide interface in the initial setting mode. The communications interface 23 may transmit information. The processor 20 may call the logic instructions in the memory 22 to execute the method in the above embodiment.

[0083] In addition, the logic instructions in the memory 22 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.

[0084] The memory 22 is a computer-readable storage medium that can be configured to store software programs, computer executable programs, such as program instructions or modules corresponding to the methods in the embodiments of the present disclosure. The processor 20 executes functional applications and data processing by running the software programs, instructions or modules stored in the memory 22, that is, implementing the methods in the above embodiments.

[0085] The memory 22 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 22 may include a high-speed random access memory and may also include a non-volatile memory. For example, a variety of 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 disk or an optical disk, may also be a transient storage medium.

[0086] In addition, the specific process of loading and executing the multiple instruction processors in the above-mentioned storage medium and the terminal device has been described in detail in the above-mentioned method, and will not be described one by one here.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application 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 application.

Claims

1. A method for predicting the affinity between a biomacromolecule and a ligand, characterized in that: The method for predicting the affinity between a biomacromolecule and a ligand specifically includes: Based on the pre-trained affinity prediction model, several candidate ligands are screened for the molecular pockets of biomacromolecules in the preset ligand library; Based on the molecular pocket and several known affinity ligands corresponding to the molecular pocket, fine-tuning the pre-trained affinity prediction model to obtain a fine-tuned affinity prediction model; The first affinity between the molecular pocket and each candidate ligand is predicted based on the fine-tuned affinity prediction model, and a preset number of target ligands are selected from a number of candidate ligands based on the first affinity.

2. The method for predicting affinity between a biomacromolecule and a ligand according to claim 1, characterized in that: The method of screening a number of candidate ligands for the molecular pocket of the biomacromolecule in the preset ligand library based on the pre-trained affinity prediction model specifically includes: Acquire a first pocket embedding vector of a molecular pocket of a biomacromolecule and a first ligand embedding vector of each preset ligand in the preset ligand library based on a pre-trained affinity prediction model; The second affinity between the molecular pocket and each preset ligand is predicted based on the first pocket embedding vector and the first ligand embedding vector, and a number of candidate ligands are selected from the preset ligand library based on the second affinity.

3. The method for predicting affinity between a biomacromolecule and a ligand according to claim 2, characterized in that: The method of obtaining the first pocket embedding vector of the molecular pocket of the biomacromolecule based on the pre-trained affinity prediction model specifically includes: Constructing a molecular pocket-ligand knowledge graph for the molecular pocket of the biomacromolecule; The pocket nodes in the molecular pocket-ligand knowledge graph are converted into second pocket embedding vectors through the pre-trained affinity prediction model, and the ligand nodes in the molecular pocket-ligand knowledge graph are converted into second ligand embedding vectors to obtain an embedded knowledge graph; The embedded knowledge graph is input into a preset graph neural network, and the first pocket embedding vector of the molecular pocket of the biological macromolecule is output through the graph neural network.

4. The method for predicting affinity between a biomacromolecule and a ligand according to claim 3, characterized in that: Constructing a molecular pocket-ligand knowledge graph for the molecular pocket of the biomacromolecule specifically includes: Obtaining similar molecular pockets and known ligands of the molecular pocket corresponding to the biomacromolecule; The molecular pocket and similar molecular pockets are used as pocket nodes, and the known ligand is used as a ligand node; Based on the structural similarity between the pocket nodes and the affinity between the ligand node and the pocket node, edges are constructed between the pocket node and the ligand node to obtain a molecular pocket-ligand knowledge graph.

5. The method for predicting affinity between a biomacromolecule and a ligand according to claim 1, characterized in that: The step of fine-tuning the pre-trained affinity prediction model based on the molecular pocket and a plurality of known affinity ligands corresponding to the molecular pocket to obtain a fine-tuned affinity prediction model specifically includes: encoding the molecular pocket into a third pocket embedding vector by the pre-trained affinity prediction model; encoding each known affinity ligand into a third ligand embedding vector by the pre-trained affinity prediction model; Predicting the third affinity of the molecular pocket to each known affinity ligand based on the third pocket embedding vector and the third ligand embedding vector; constructing a first ranking loss term based on the third affinity of the molecular pocket to each known affinity ligand and the first experimental affinity of the molecular pocket to each known affinity ligand; The pre-trained affinity prediction model is fine-tuned based on the first ranking loss term to obtain a fine-tuned affinity prediction model.

6. The method for predicting affinity between a biomacromolecule and a ligand according to claim 1, characterized in that: The predicting the first affinity between the molecular pocket and each candidate ligand based on the fine-tuned affinity prediction model specifically includes: encoding the molecular pocket into a fourth pocket embedding vector by the fine-tuned affinity prediction model; encoding each candidate ligand as a fourth ligand embedding vector by the fine-tuned affinity prediction model; The first affinity of the molecular pocket to each ligand with known affinity is predicted based on the fourth pocket embedding vector and the fourth ligand embedding vector.

7. The method for predicting affinity between a biomacromolecule and a ligand according to claim 1, characterized in that: The affinity prediction model includes a pocket encoder for encoding a molecular pocket into an embedding vector and a ligand encoder for encoding a ligand into an embedding vector.

8. The method for predicting the affinity between a biomacromolecule and a ligand according to any one of claims 1 to 7, characterized in that: The pre-training process of the pre-trained affinity prediction model specifically includes: Acquire a training data set, wherein the training data set includes a plurality of experimental measurement groups, each of which includes a training molecular pocket, a plurality of training ligands, and a second experimental affinity of each training ligand to the training molecular pocket; encoding the training molecule pocket in each experimental measurement group into a training pocket embedding vector through the affinity prediction model to be pre-trained, and encoding each training ligand in each experimental measurement group into a training ligand embedding vector; Determining a fourth affinity between the training molecule pocket in each experimental measurement group and each training ligand in each experimental measurement group based on the training pocket embedding vector and the training ligand embedding vector; A loss function term is constructed based on the fourth affinity and the second experimental affinity, and the affinity prediction model to be pre-trained is trained based on the loss function term to obtain a pre-trained affinity prediction model.

9. The method for predicting affinity between a biomacromolecule and a ligand according to claim 8, characterized in that: The obtaining of the training data set specifically includes: Collect experimentally determined known affinity data between biomacromolecules and ligands; Dividing the acquired known affinity data into a number of initial experimental measurement groups based on the same biomacromolecule and experimental measurement method, and matching a molecular pocket for each of the initial experimental measurement groups; Using the molecular pocket matched by each of the initial experimental measurement groups as a training molecular pocket, the ligand in each of the initial experimental measurement groups as a training ligand, and the experimentally determined affinity in the initial experimental measurement group as a second experimental affinity, to obtain an experimental measurement group corresponding to each of the initial experimental measurement groups; The data set formed by the experimental measurement group is used as a training data set.

10. The method for predicting affinity between a biomacromolecule and a ligand according to claim 9, characterized in that: The matching of molecular pockets for each of the initial experimental measurement groups specifically includes: Selecting a biomacromolecule-ligand complex structure for each biomacromolecule in the initial experimental measurement group based on biomacromolecule similarity in a known biomacromolecule structure database; respectively calculating the ligand similarity between the crystal ligand bound to the biomacromolecule-ligand complex structure selected in each of the initial experimental measurement groups and the ligand in each of the initial experimental measurement groups; The target crystal ligands are matched for each initial experimental measurement group based on ligand similarity, and the molecular pocket corresponding to the target crystal ligand matched by each initial experimental measurement group is used as the molecular pocket corresponding to each initial experimental measurement group.

11. The method for predicting affinity between a biomacromolecule and a ligand according to claim 8, characterized in that: The step of constructing a loss function term based on the fourth affinity and the second experimental affinity specifically includes: constructing a second ranking loss term for each experimental measurement value group based on a fourth affinity of the training molecule pocket in each experimental measurement group to the training ligand in each experimental measurement group and a second experimental affinity; constructing a comparative loss term based on the fourth affinity between the training molecule pocket in each experimental measurement group and each training ligand in each experimental measurement group; The loss function term is calculated based on the second ranking loss term of each experimental measurement value group and the contrast loss term.

12. A system for predicting affinity between a biomacromolecule and a ligand, characterized in that: The affinity prediction system for biomacromolecules and ligands specifically includes: A screening module is used to screen several candidate ligands for the molecular pocket of the biomacromolecule in a preset ligand library based on a pre-trained affinity prediction model; A fine-tuning module, used for fine-tuning the pre-trained affinity prediction model based on the molecular pocket and a plurality of known affinity ligands corresponding to the molecular pocket to obtain a fine-tuned affinity prediction model; The lead compound optimization module is used to predict the first affinity between the molecular pocket and each candidate ligand based on the fine-tuned affinity prediction model, and select a preset number of target ligands from a number of candidate ligands based on the first affinity.

13. The system for predicting affinity between biomacromolecules and ligands according to claim 12, characterized in that: The affinity prediction model includes a pocket encoder for encoding a molecular pocket into an embedding vector and a ligand encoder for encoding a ligand into an embedding vector.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the method for predicting the affinity between a biomacromolecule and a ligand as described in any one of claims 1 to 11; and / or The computer-readable storage medium stores the affinity prediction system between a biomacromolecule and a ligand as described in any one of claims 12 to 13.

15. A terminal device, characterized in that: include: Processor and memory; The memory stores a computer-readable program executable by the processor; When the processor executes the computer-readable program, the steps of the method for predicting the affinity between a biomacromolecule and a ligand as described in any one of claims 1 to 10 are implemented.

Images