Protein design method based on intelligent agent and related product

Abstract

The invention provides a protein design method based on an intelligent agent and a related product. According to a specific embodiment of the protein design method based on the intelligent agent, through protein design logic summarized by experience, a complex protein design interaction mode is converted into an interaction mode which is easy to understand by a user and has a guiding prompt through the intelligent agent, a developer does not need to learn each protein design tool and algorithm in advance, and the development efficiency is improved. According to the method, only specific requirements need to be put forward, or protein design parameter information (such as a protein sequence region, a design target and the like) can be selectively fed back under the guidance of prompt information of the intelligent body, and then the intelligent body can automatically call a protein design tool to complete protein design work, so that the protein design efficiency is improved; in addition, the protein design process does not depend on the understanding level and actual experience of developers, the reliability of protein design can be improved, and the error probability is reduced.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the technical field of bioinformatics, and in particular to an agent-based protein design method and related products. Background Art

[0002] Protein design is an important technology in the field of bioengineering and drug development. Its research aims to design proteins with specific functions through computational and experimental means. In recent years, with the development of computational biology and artificial intelligence, protein design technology has made significant progress. However, a large number of tools and algorithms need to be called in the AI-based protein design process. Developers need to learn the principles and operation logic behind each tool and algorithm in advance, resulting in high learning and development costs. It also depends on the understanding ability and practical operation experience of specific developers and is prone to errors.

[0003] In addition, protein types are very rich and different protein design strategies vary greatly. Traditional protein design methods need to be designed and developed separately for different types of proteins and different demand scenarios, resulting in high development costs for protein design and lack of scalability.

[0004] Existing technologies mainly rely on structure-based design and empirical methods, which have limitations when dealing with complex protein structures and sequence variations, such as low design efficiency, low success rate, and lack of flexibility. Summary of the invention

[0005] The embodiments of the present disclosure provide an agent-based protein design method, apparatus, electronic device, storage medium, and computer program product.

[0006] In a first aspect, an embodiment of the present disclosure provides an agent-based protein design method, the method comprising:

[0007] Obtain protein design requirement information input by the user interface in this round of dialogue;

[0008] Determining a protein design process based on the protein design requirement information, wherein the protein design process includes at least one protein design task;

[0009] For each protein design task in the protein design process, executing the protein design task by a task execution tool corresponding to the protein design task to obtain a protein design result;

[0010] The protein design results are presented through the user interface.

[0011] In some optional embodiments, the method further comprises:

[0012] In response to detecting a confirmation operation of the user on the protein design result, the protein design requirement information and the protein design result are correspondingly stored as protein design information.

[0013] In some optional embodiments, the method further comprises:

[0014] In response to receiving experimental result data synthesized and verified by wet experiment, the experimental result data is correspondingly stored as experimental result data in the protein design information.

[0015] In some optional embodiments, the method further comprises:

[0016] The experimental result data is presented through the user interface.

[0017] In some optional embodiments, executing the protein design task by a task execution tool corresponding to the protein design task includes:

[0018] The protein design task is executed by calling the protein design parameters in the protein design requirement information through a machine learning model corresponding to the protein design task type; and the method further comprises:

[0019] According to the protein design result and the experimental result data, determining a model to be updated in the machine learning model called in the process of generating the protein design result, and / or determining a protein design parameter to be updated in the protein design parameters called in the process of generating the protein design result;

[0020] Based on the protein design results and the experimental result data, each of the models to be updated and / or each of the design parameters of the proteins to be updated is fine-tuned to update each of the models to be updated and / or each of the design parameters of the proteins to be updated.

[0021] In some optional embodiments, determining the model to be updated and / or the protein design parameters to be updated according to the protein design results and the experimental result data includes:

[0022] Determining whether the experimental result data, relative to the protein design result, has an improvement in at least a preset ratio of the preset evaluation index that meets the corresponding index improvement requirement;

[0023] In response to the determination being yes, determining the protein design parameters involved in the process of generating the protein design result as the protein design parameters to be updated;

[0024] In response to a no determination, determining the protein generation model involved in the process of generating the protein design result as a model to be updated;

[0025] Determine the correlation between the protein design result and the model scoring result in the experimental result data, and in response to determining that the correlation is less than a preset correlation threshold, determine the scoring model involved in the process of generating the protein design result as a model to be updated.

[0026] In some optional embodiments, determining the protein design process based on the protein design requirement information includes: determining the protein design process based on the protein design requirement information by a protein design agent; and

[0027] The protein design task belongs to a preset task type set, which includes: local conditional protein sequence generation, de novo protein sequence generation, protein property scoring, protein structure generation, and protein information retrieval.

[0028] In some optional implementations, each preset task type corresponds to at least one task execution tool; and

[0029] The performing of the protein design task by a task execution tool corresponding to the protein design task comprises:

[0030] The protein design task is executed by a task execution tool corresponding to the task type to which the protein design task belongs.

[0031] In some optional embodiments, the protein design result includes at least one of the following: protein sequence, protein structure information, protein design model score, and protein design evaluation index score.

[0032] In some optional embodiments, the protein design process includes at least one of the following protein design tasks:

[0033] Generate and present, through the user interface, operation prompt information for instructing the user to provide protein design parameter information;

[0034] Acquiring protein design parameter information input by the user through the user interface for the protein design task;

[0035] Updating the protein design requirement information based on the acquired protein design parameter information;

[0036] According to the protein design requirement information, a corresponding protein design tool is determined and called to perform protein design to obtain a protein design result.

[0037] In some optional embodiments, the protein design parameter information includes at least one of the following:

[0038] Protein structure or sequence, protein sequence designable region, protein sequence generation pattern, protein design target requirements, protein design evaluation model, protein design evaluation indicators, and protein properties to be optimized.

[0039] In some optional embodiments, the protein design tool includes at least one of the following: a protein sequence generation tool, a protein expression prediction tool, a protein polymerization prediction tool, and a protein stability prediction tool.

[0040] In a second aspect, an embodiment of the present disclosure provides an agent-based protein design device, the device comprising:

[0041] A requirement acquisition module is configured to acquire protein design requirement information input by the user through the user interface in this round of dialogue;

[0042] A process determination module is configured to determine a protein design process based on the protein design requirement information, wherein the protein design process includes at least one protein design task;

[0043] A process execution module is configured to execute each protein design task in the protein design process through a task execution tool corresponding to the protein design task to obtain a protein design result;

[0044] The result presentation module is configured to present the protein design result through the user interface.

[0045] In some optional embodiments, the agent-based protein design apparatus further comprises:

[0046] The design information storage module is configured to store the protein design requirement information and the protein design result in correspondence as protein design information in response to detecting a confirmation operation of the user on the protein design result.

[0047] In some optional embodiments, the agent-based protein design apparatus further comprises:

[0048] The experimental result storage module is configured to, in response to receiving experimental result data synthesized and verified by a wet experiment, store the experimental result data as experimental result data in the protein design information.

[0049] In some optional embodiments, the agent-based protein design apparatus further comprises:

[0050] The experimental result presenting module is configured to present the experimental result data through the user interface.

[0051] In some optional embodiments, executing the protein design task by a task execution tool corresponding to the protein design task includes:

[0052] The protein design task is executed by calling the protein design parameters in the protein design requirement information through a machine learning model corresponding to the type of the protein design task; and the agent-based protein design device further comprises:

[0053] A model and / or parameter determination module is configured to determine the model to be updated from the machine learning model called in the process of generating the protein design result, and / or determine the protein design parameters to be updated from the protein design parameters called in the process of generating the protein design result according to the protein design result and the experimental result data;

[0054] The updating module is configured to fine-tune each of the models to be updated and / or each of the protein design parameters to be updated based on the protein design results and the experimental result data, so as to update each of the models to be updated and / or each of the protein design parameters to be updated.

[0055] In some optional embodiments, the model and / or parameter determination module is further configured to:

[0056] Determining whether the experimental result data, relative to the protein design result, has an improvement in at least a preset ratio of the preset evaluation index that meets the corresponding index improvement requirement;

[0057] In response to the determination being yes, determining the protein design parameters involved in the process of generating the protein design result as the protein design parameters to be updated;

[0058] In response to a no determination, determining the protein generation model involved in the process of generating the protein design result as a model to be updated;

[0059] Determine the correlation between the protein design result and the model scoring result in the experimental result data, and in response to determining that the correlation is less than a preset correlation threshold, determine the scoring model involved in the process of generating the protein design result as a model to be updated.

[0060] In some optional embodiments, the process determination module is further configured to: determine the protein design process based on the protein design requirement information by a protein design agent; and

[0061] The protein design task belongs to a preset task type set, which includes: local conditional protein sequence generation, de novo protein sequence generation, protein property scoring, protein structure generation, and protein information retrieval.

[0062] In some optional implementations, each preset task type corresponds to at least one task execution tool; and

[0063] The performing of the protein design task by a task execution tool corresponding to the protein design task comprises:

[0064] The protein design task is executed by a task execution tool corresponding to the task type to which the protein design task belongs.

[0065] In some optional embodiments, the protein design result includes at least one of the following: protein sequence, protein structure information, protein design model score, and protein design evaluation index score.

[0066] In some optional embodiments, the protein design process includes at least one of the following protein design tasks:

[0067] Generate and present, through the user interface, operation prompt information for instructing the user to provide protein design parameter information;

[0068] Acquiring protein design parameter information input by the user through the user interface for the protein design task;

[0069] Updating the protein design requirement information based on the acquired protein design parameter information;

[0070] According to the protein design requirement information, a corresponding protein design tool is determined and called to perform protein design to obtain a protein design result.

[0071] In some optional embodiments, the protein design parameter information includes at least one of the following:

[0072] Protein structure or sequence, protein sequence designable region, protein sequence generation pattern, protein design target requirements, protein design evaluation model, protein design evaluation indicators, and protein properties to be optimized.

[0073] In some optional embodiments, the protein design tool includes at least one of the following: a protein sequence generation tool, a protein expression prediction tool, a protein polymerization prediction tool, and a protein stability prediction tool.

[0074] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation manner in the first aspect.

[0075] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by one or more processors, implements the method described in any implementation manner in the first aspect.

[0076] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the method described in any implementation manner in the first aspect.

[0077] In order to reduce the development cost of protein design, the agent-based protein design method, device, electronic device, storage medium and computer program product provided by the embodiments of the present disclosure obtain the protein design requirement information input by the user through the user interface in this round of dialogue; then determine the protein design process based on the protein design requirement information, and the protein design process includes at least one protein design task; then, for each protein design task in the protein design process, execute the protein design task through the task execution tool corresponding to the protein design task to obtain the protein design result; finally, present the protein design result through the user interface. That is, the user only needs to input the protein design requirement in the user interface, and the agent-based protein design method can automatically determine and execute the protein design process until the protein design result is finally obtained and presented. Here, the protein design logic summarized from experience is used to convert the complex protein design interaction mode into an interaction mode with easy-to-understand guidance prompts through the intelligent agent. Developers do not need to learn each protein design tool and algorithm in advance. They only need to put forward specific requirements, or optionally feedback protein design parameter information (such as the region of the protein sequence, design goals, etc.) under the guidance of the prompt information of the intelligent agent. The intelligent agent will automatically call the protein design tool to complete the protein design work, which improves the efficiency of protein design. In addition, the protein design process does not depend on the understanding level and actual experience of the developer, which can improve the reliability of protein design and reduce the probability of error. In addition, the protein design intelligent agent also has self-evolution ability and high scalability by calling different protein design tools, which can meet a wide range of protein design and analysis needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Other features, objects and advantages of the present disclosure will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the drawings:

[0079] Figure 1 is an exemplary system architecture diagram in which an embodiment of the present disclosure may be applied;

[0080] Figure 2Ais a flow chart of one embodiment of an agent-based protein design method according to the present disclosure;

[0081] Figure 2B is a decomposed flow chart according to one embodiment of step 208 of the present disclosure;

[0082] Figure 3 is an architectural diagram of an embodiment of a protein design agent according to the present disclosure;

[0083] Figure 4 is a schematic structural diagram of an embodiment of an agent-based protein design device according to the present disclosure;

[0084] Figure 5 It is a structural diagram of a computer system of an electronic device suitable for implementing an embodiment of the present disclosure. DETAILED DESCRIPTION

[0085] The present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0086] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0087] Figure 1 An exemplary system architecture 100 is shown to which embodiments of the disclosed agent-based protein design methods and related products may be applied.

[0088] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0089] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, protein design applications, voice recognition applications, short video social applications, audio and video conferencing applications, video live broadcast applications, document editing applications, input method applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0090] Terminal devices 101, 102, 103 can be hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices with display screens, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, etc. When terminal devices 101, 102, 103 are software, they can be installed in the terminal devices listed above. It can be implemented as multiple software or software modules (for example, for providing protein design services), or it can be implemented as a single software or software module. It is not specifically limited here.

[0091] In some cases, the agent-based protein design method provided by the present disclosure may be executed by the terminal devices 101, 102, 103, and accordingly, the agent-based protein design apparatus may be disposed in the terminal devices 101, 102, 103. In this case, the system architecture 100 may not include the server 105.

[0092] In some cases, the agent-based protein design method provided in the present disclosure can be jointly performed by the terminal devices 101, 102, 103 and the server 105. For example, the step of "obtaining the protein design requirement information input by the user through the user interface in this round of dialogue" can be performed by the terminal devices 101, 102, 103, and the steps of "determining the protein design process based on the protein design requirement information" can be performed by the server 105. The present disclosure does not limit this. Accordingly, the agent-based protein design device can also be respectively set in the terminal devices 101, 102, 103 and the server 105.

[0093] In some cases, the agent-based protein design method provided in the present disclosure may be executed by the server 105 , and accordingly, the agent-based protein design device may also be arranged in the server 105 . In this case, the system architecture 100 may not include the terminal devices 101 , 102 , and 103 .

[0094] It should be noted that the server 105 can be hardware or software. When the server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the server 105 is software, it can be implemented as multiple software or software modules (for example, for providing protein design services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0095] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0096] Continue to refer Figure 2A, which shows a process 200 of an embodiment of an agent-based protein design method according to the present disclosure, the agent-based protein design method comprises the following steps:

[0097] Step 201, obtaining protein design requirement information input by the user through the user interface in this round of dialogue.

[0098] In this embodiment, the execution subject of the agent-based protein design method (e.g., Figure 1 The terminal device shown in ) can locally obtain the protein design requirement information input by the user through the user interface in this round of dialogue.

[0099] Alternatively, the above-mentioned execution entity (for example, Figure 1 The server shown in ) can also be remotely connected to the execution subject network from other electronic devices (for example, Figure 1 The terminal device shown in ) obtains the protein design requirement information input by the user through the user interface in this round of dialogue.

[0100] It should be noted that, here, proteins include but are not limited to the following protein types: enzymes (used to catalyze biochemical reactions), antibodies (used to specifically recognize and bind to specific antigens), hormones (used to regulate physiological processes), capsid proteins (used to protect viral genetic material), polypeptides (or short-chain biologically active proteins), small proteins (i.e. proteins with small molecular weight and diverse functions), transport proteins (responsible for transporting small molecules and ions), membrane proteins (embedded in cell membranes and involved in signal transduction), structural proteins (used to provide support for cells and tissues), etc.

[0101] In some optional embodiments, the user interface may be a user interface provided by a protein design agent. Here, the protein design agent may be an agent based on a model or algorithm constructed by techniques such as deep learning and natural language processing. The protein design agent may be an intelligent agent.

[0102] The protein design agent can provide a user interface in various implementation methods, for example, through a web page, an application program, a hosted application program, etc. In the above user interface, a user can be prompted to input various forms of information through a dialogue. For example, the user can enter text, take photos and upload images in real time, collect and upload user voice in real time, and upload existing files. The existing files can be, for example: text files, voice files, image files, video files, and protein design related files. As examples, protein design related files can include but are not limited to:

[0103] Protein structure files: for example, PDB files (Protein Data Bank format), CIF files (mmCIF, macromolecular Crystallographic Information File format).

[0104] Dataset files: For example, Excel files (.xls, .xlsx), CSV files (.csv), XML files (.xml), JSON files (.json), text files (.txt), etc., containing data related to protein design.

[0105] Protein sequence files: For example, FASTQ files (.fastq), FASTA files (.fasta), and other sequence and annotation information files.

[0106] Description information: Users can provide additional protein description information, such as protein classification information, epitope information, functional domain information, source species information, etc., to help the system better understand design requirements.

[0107] Configuration file: used to specify the design parameter values ​​required to be set in the protein design tool, such as temperature, pH value, salt concentration, etc.

[0108] The user can input protein design requirement information through the above user interface. Here, the protein design requirement information may include but is not limited to the following information:

[0109] (1) Protein structural characteristics

[0110] Protein shape: specific three-dimensional shapes such as sphere, rod, etc.

[0111] Protein secondary structure: including the types and numbers of secondary structures such as α-helices and β-folds. For example, design a protein containing five α-helices.

[0112] Specific domains: Designing proteins containing specific functional domains or domains, which can be known or newly designed.

[0113] Symmetry: In some applications, it may be necessary to design proteins with specific symmetries, such as dimers, trimers, etc.

[0114] (2) Functional characteristics

[0115] Binding site: Designing a protein so that it can bind to a specific ligand, substrate, or another protein, such as designing a protein that can bind to a specific drug molecule.

[0116] Catalytic activity: Designing enzymes with specific catalytic activity that can catalyze specific chemical reactions.

[0117] Stability: Ensure that the designed protein has sufficient stability under specific environmental conditions (such as temperature, pH value, etc.).

[0118] Kinetic properties: Designing proteins to have specific kinetic properties, such as fast folding or slow unfolding.

[0119] (3) Sequence characteristics

[0120] Amino acid sequence: specifies the amino acid residues at certain positions to meet specific functional or structural requirements.

[0121] Sequence diversity: Sequence diversity is considered during the design process to generate multiple possible protein variants and increase the success rate of the design.

[0122] (4) Design goals

[0123] Specific applications: Design proteins with specific functions based on specific application requirements, such as drug development, biosynthesis, environmental remediation, etc.

[0124] Optimization target: The protein property target that needs to be optimized during the design process, such as the affinity of the designed protein for binding to a specific antigen should be higher than 1 nM, the activity of the enzyme involved should reach xx μmol / min at a specific substrate concentration, the designed protein can maintain activity at a temperature greater than 70°C, the expression level of the protein should be optimized to xxx mg / L, the designed protein should have high selectivity for a specific substrate or target, etc.

[0125] (5) Constraints

[0126] Avoid undesirable conformations: Make sure the designed protein does not contain conformations that cannot fold properly.

[0127] Kinetic accessibility: ensuring that the designed protein can reach the target conformation through dynamic processes from different regions of the conformational space.

[0128] (6) Other information

[0129] Experimental verification: The designed protein needs to be able to be verified by experimental methods, such as liquid chromatography, mass spectrometry, fluorescent labeling, X-ray crystallography, nuclear magnetic resonance, etc.

[0130] Computing resources: During the design process, you may need to consider the limitations of computing resources, such as computing time, memory, etc.

[0131] It should be noted that the above protein design requirement information can be adjusted and supplemented according to specific design goals and application scenarios to ensure that the designed protein can meet the expected functional and performance requirements.

[0132] Step 202: determine the protein design process based on the protein design requirement information.

[0133] Here, the protein design process may include at least one protein design task. Optionally, at least one protein design task may be arranged in a sequence. By executing the protein design tasks in the sequence of the protein design process, the protein design task corresponding to the protein design requirement can be completed.

[0134] In some optional embodiments, here, the protein design process can be determined based on the protein design requirement information by a protein design agent. The protein design agent can be an agent obtained after fine-tuning the general large language model. Specifically, a professional protein design technician can summarize the specific protein design process according to the professional knowledge of protein design for various types of protein design requirements, and accordingly construct a large amount of protein design requirement description text to protein design process training corpus for each protein design requirement, and then use the above training corpus to fine-tune the general large language model so that the general large language model learns to map from protein design requirements to protein design process. Then, a protein design agent is obtained based on the above-mentioned fine-tuned general large language model.

[0135] Furthermore, the protein design process can be determined based on the protein design requirement information through the protein design agent obtained by fine-tuning the general large language model.

[0136] As an example, the protein design requirement information input by the user is "Help me optimize the activity of XX enzyme". The corresponding protein design process may include: Task 1, Task 2 and Task 3, wherein the specific information of Task 1, Task 2 and Task 3 is as follows:

[0137] Task 1: Interact with users through a dialogue interface, guide users to select design regions and design parameters, and update protein design requirement information with the design regions and design parameters selected by users.

[0138] Task 2: Call the protein sequence generation algorithm and enzyme activity, enzyme stability, expression level and other models to generate candidate protein sequences as protein design results based on the updated protein design requirement information.

[0139] Task 3: Present the protein design results through a dialog interface visualization plug-in and prompt the user to confirm whether the presented protein sequence is the final designed protein sequence.

[0140] Step 203: for each protein design task in the protein design process, execute the protein design task by using a task execution tool corresponding to the protein design task to obtain a protein design result.

[0141] The protein design process has been obtained through step 202. Here, for each protein design task in the protein design process, the protein design task can be executed by a task execution tool corresponding to the protein design task to obtain a protein design result.

[0142] Optionally, the protein design agent can perform each protein design task in sequence according to the sequence of protein design tasks in the protein design process to obtain a protein design result. When performing each protein design task, a task execution tool corresponding to the protein design task is first determined, and then the protein design task is executed by the determined task execution tool.

[0143] As an example, the protein design results may include, but are not limited to: protein sequence, protein structure information, protein design model score, and protein design evaluation index score.

[0144] Step 204: presenting the protein design results through a user interface.

[0145] Here, the user interface of the protein design agent can select corresponding visualization plug-ins for presentation according to the various data types included in the protein design results.

[0146] For example, for protein sequences, they can be presented using visualization plug-ins such as Jalview, AlignmentViewer, BioEdit, and openVectorEditor.

[0147] For another example, protein structure information can be presented through currently known tools such as Pymol, LiteMol, Mol*, etc., or various protein structure information visualization plug-ins to be developed in the future.

[0148] As for the scores of protein design model and protein design evaluation index, they can be presented in various pre-designed tables, graphs and other forms.

[0149] In some optional embodiments, the above agent-based protein design method process 200 may further include the following steps 205:

[0150] Step 205 , in response to detecting a user's confirmation operation on the protein design result, the protein design requirement information and the protein design result are correspondingly stored as protein design information.

[0151] In step 204, the protein design result has been presented to the user. If the user feels that the protein design result does not need to be modified, the user can perform a confirmation operation on the protein design result (e.g., a confirmation operation button, or input voice or text to indicate that the currently provided protein design result does not need to be modified), and then the above-mentioned execution body can detect the above-mentioned confirmation operation, and store the protein design requirement information and the protein design result correspondingly as protein design information. Specifically, the protein design information can be stored locally in the above-mentioned execution body or remotely in other electronic devices (e.g., cloud servers) connected to the network of the above-mentioned execution body.

[0152] Optionally, the protein design requirement information and the protein design results can be stored correspondingly as protein design information through a protein design agent.

[0153] By executing step 205, the protein design information confirmed by the user can be stored for subsequent viewing or other processing.

[0154] In some optional embodiments, the above agent-based protein design method may further include the following steps 206:

[0155] Step 206 , in response to receiving the experimental result data synthesized and verified by the wet experiment for the protein design result, the experimental result data is correspondingly stored as the experimental result data in the protein design information.

[0156] That is to say, in practice, after obtaining the protein design results, the user can synthesize and verify the protein design results through wet experiments, and then transmit the obtained experimental result data to the above-mentioned execution entity, and then the above-mentioned execution entity can store the experimental result data as the experimental result data in the protein design information stored in step 205 through the protein design intelligent agent.

[0157] Optionally, the protein design agent can be used to store the experimental result data as the experimental result data in the protein design information.

[0158] Through step 206, the protein design results obtained through steps 201 to 204 can be verified through wet experiments in practice, and the experimental result data can be saved accordingly.

[0159] In some optional implementations, the execution subject may further execute the following step 207 after executing step 206:

[0160] Step 207: presenting the experimental result data through the user interface.

[0161] Here, the experimental result data can be presented through a visualization plug-in, and then the user can obtain the experimental result data of the wet experiment corresponding to the protein design result obtained in step 204, so as to compare and verify the effect of the protein design result.

[0162] In some optional embodiments, in step 203, executing the protein design task by a task execution tool corresponding to the protein design task can be performed as follows: calling the protein design parameters in the protein design requirement information by a machine learning model corresponding to the protein design task type to execute the protein design task. Accordingly, the above-mentioned execution subject can also perform the following steps 208 and 209 after executing step 206:

[0163] Step 208, based on the protein design results and the experimental result data, determine the model to be updated in the machine learning model called in the process of generating the protein design results, and / or determine the protein design parameters to be updated in the protein design parameters called in the process of generating the protein design results.

[0164] Here, the above-mentioned execution entity can determine the model to be updated in the machine learning model called in the process of generating the protein design results according to the protein design results and experimental results data according to the preset model determination rules, and / or determine the protein design parameters to be updated in the protein design parameters called in the process of generating the protein design results according to the preset parameter determination rules.

[0165] Specifically, optionally, step 208 may include: Figure 2B The following steps 2081 to 2085 are shown:

[0166] Step 2081, determining whether the experimental result data, relative to the protein design result, has an improvement in at least a preset ratio of the preset evaluation index that meets the corresponding index improvement requirement.

[0167] If the improvement of the experimental result data relative to a preset evaluation indicator of the protein design result reaches the corresponding indicator improvement requirement, it indicates that according to the experimental result data, the actual performance of the protein design result is significantly higher than the corresponding evaluation indicator in the original protein design result.

[0168] If the experimental result data is relative to the protein design result, and the improvement of at least a preset proportion of the preset evaluation indicators reaches the corresponding indicator improvement requirements, indicating that most of the evaluation indicators are significantly improved or far exceed the design goals, it can be considered that the protein generation model involved in the process of generating the protein design result performs satisfactorily, and the model does not need to be updated, but the protein design parameters involved in the process of generating the protein design result need to be updated, and the process can go to step 2082 for execution.

[0169] As an example, the preset ratio may be 60%.

[0170] As examples, protein design parameters may include, but are not limited to, the following:

[0171] 1. Structural related parameters: for example, backbone structure, secondary structure, tertiary structure, quaternary structure, steric hindrance, etc.

[0172] 2. Sequence-related parameters: such as amino acid composition, sequence diversity, conserved sequences, sequence constraints, etc.

[0173] 3. Function-related parameters: such as binding affinity, catalytic activity, functional groups, functional domains, etc.

[0174] 4. Stability related parameters: such as thermal stability, chemical stability, folding stability, etc.

[0175] 5. Dynamics-related parameters: such as flexibility, kinetic stability, etc.

[0176] 6. Calculate and optimize relevant parameters

[0177] Energy Minimization, Sampling Algorithms, Design Objectives, Constraints, etc.

[0178] 7. Biophysical related parameters: such as solubility, isoelectric point, size and shape, etc.

[0179] 8. Experimental verification of relevant parameters: such as expression level, purification difficulty, functional validation, etc.

[0180] If the experimental result data is relative to the protein design result, and the improvement of the preset evaluation indicators of less than the preset proportion reaches the corresponding indicator improvement requirement, it indicates that there is no obvious improvement in most of the evaluation indicators or they are far from the design goals. It can be considered that the protein generation model involved in the process of generating the protein design result performs poorly and the protein generation model needs to be updated, and the process can go to step 2083 for execution.

[0181] Step 2082: determine the protein design parameters involved in the process of generating the protein design result as the protein design parameters to be updated.

[0182] Step 2083, determining the protein generation model involved in the process of generating the protein design result as the model to be updated.

[0183] Here, the protein generation model involved in the process of generating protein design results may include but is not limited to:

[0184] Conditional generative models: such as conditional generative adversarial networks (cGANs), which can generate protein sequences with specific functions through conditions such as Gene Ontology terms or Enzyme Condition Numbers.

[0185] Diffusion models: Ability to consider non-protein atoms such as DNA or small molecules, thereby generating protein sequences with the necessary conformational flexibility to achieve complex functions.

[0186] Variational Auto Encoder (VAE): It can be used to learn the potential representation of protein sequences and generate new sequences through sampling, thereby designing proteins with specific functions.

[0187] Generative Adversarial Network (GAN): It consists of a generator and a discriminator. The generator generates new protein sequences, and the discriminator determines the authenticity of the sequence. Through adversarial training, the generator is continuously optimized to generate sequences that are closer to real proteins.

[0188] Step 2084, determining the correlation between the protein design results and the model scoring results in the experimental result data.

[0189] Here, the protein design result may include a model scoring result corresponding to at least one scoring model, and correspondingly, the experimental result data may also include model scoring results corresponding to the above-mentioned various scoring models. Furthermore, various similarity calculation methods may be used to calculate and determine the correlation between the protein design result and the model scoring results in the experimental result data.

[0190] As an example, assuming that the protein design results include model scoring results corresponding to N scoring models, the model scoring results corresponding to the N scoring models included in the protein design results can be used to generate an N-dimensional first vector; then, the model scoring results corresponding to the N scoring models included in the experimental result data can be used to generate an N-dimensional second vector; finally, various vector similarity calculation methods are used to calculate the similarity between the first vector and the second vector, and the similarity calculated above is used as the correlation between the protein design results and the model scoring results in the experimental result data.

[0191] As an example, the model scoring results may include, but are not limited to:

[0192] In protein design, model scoring is a key step in evaluating the quality and function of generated protein sequences. The following are some commonly used model scoring methods and indicators:

[0193] 1. Sequence-based scores: such as sequence similarity, language model scores, evolutionary conservation, etc.

[0194] 2. Structure-based scores: such as Structure Prediction Scores, Rosetta Energy Scores, Structural Similarity, etc.

[0195] 3. Function-based scoring: such as binding affinity prediction, functional validation, etc.

[0196] 4. Comprehensive scoring: such as Composite Scoring Frameworks, Multi-Dimensional Evaluation, etc.

[0197] 5. Other scoring indicators: such as solubility, sequence diversity, experimental validation, etc.

[0198] Step 2085: In response to determining that the correlation is less than a preset correlation threshold, the scoring model involved in the process of generating the protein design result is determined as a model to be updated.

[0199] If the correlation calculated in step 2084 is less than the preset correlation threshold, it indicates that the scoring model involved in the process of generating the protein design result performs poorly, so the scoring model involved in the process of generating the protein design result can be determined as the model to be updated.

[0200] Here, various parameter optimization methods can be used to fine-tune each model to be updated based on the protein design results and experimental results data to update each model to be updated.

[0201] Step 209: fine-tune each model to be updated and / or each protein design parameter to be updated based on the protein design result and the experimental result data, so as to update each model to be updated and / or each protein design parameter to be updated.

[0202] Here, various parameter optimization methods can be used to fine-tune each model to be updated and / or each protein design parameter to be updated based on the protein design results and experimental results data, so as to update each model to be updated and / or each protein design parameter to be updated.

[0203] By executing steps 208 and 209, the protein generation model, protein design parameters or protein scoring model called in the protein design process can be automatically optimized and iterated through the experimental result data of the wet experiment when the performance is poor, thereby achieving closed-loop feedback, so that the protein design intelligent agent can continuously evolve and adapt to new design requirements and challenges.

[0204] In addition, most pharmaceutical companies currently analyze and use experimental data results within the project, and the value of the data is not fully reflected. Here, the above optional methods can be used to transform the experimental result data into new knowledge for protein agents by updating the model or updating the parameters, so that protein agents can accumulate knowledge from historical experimental result data and continuously improve.

[0205] In some optional embodiments, step 202, determining the protein design process based on the protein design requirement information, can be performed as follows: determining the protein design process based on the protein design requirement information by a protein design agent.

[0206] Here, the protein design task may belong to a set of preset task types, which may include: local conditional protein sequence generation, de novo protein sequence generation, protein property scoring, protein structure generation, and protein information retrieval. Among them:

[0207] In the local conditional sequence generation type of protein design tasks, users are required to select the protein design region, constraints, and optimization goals, and then generate the protein sequence based on the above user selections.

[0208] In the protein design task of generating protein sequence from scratch, the user is required to select protein design constraints and optimization objectives, and then generate the protein sequence based on the above user selections.

[0209] In protein design tasks of the protein property scoring type, protein properties are scored based on protein sequence or structure.

[0210] In protein design tasks of the protein structure generation type, protein structure is generated based on protein sequence. A protein structure prediction model based on deep learning can be used, or a protein structure prediction model based on force field calculation methods such as molecular dynamics simulation can be used.

[0211] In protein design tasks of the protein information retrieval type, users need to input a sequence or structure, and then perform related protein sequence retrieval based on sequence similarity retrieval, structural similarity retrieval, protein function retrieval, etc.

[0212] Optionally, each of the above-mentioned preset task types may correspond to at least one task execution tool. Accordingly, executing the corresponding protein design task may be: executing the protein design task by the task execution tool corresponding to the task type to which the protein design task belongs. Furthermore, in the future, new task execution tools corresponding to new protein design tasks (e.g., new protein sequence generation algorithms, new protein scoring models, etc.) will be developed to achieve high scalability of the task execution tools corresponding to the protein design tasks.

[0213] The above optional method classifies protein design tasks by summarizing the experience based on the protein design process. Under each task category, the corresponding design task can be implemented by calling multiple execution tools, thereby achieving high scalability of the execution tools.

[0214] In some optional embodiments, the protein design process determined in step 202 may include at least one of the following protein design tasks:

[0215] Generation and presentation tasks: Generate and present, through the user interface, operational prompt information for instructing the user to provide protein design parameter information.

[0216] Parameter acquisition task: obtain the protein design parameter information input by the user through the user interface for the protein design task.

[0217] Requirements update task: Update protein design requirements information based on the obtained protein design parameter information.

[0218] Protein design task: Determine and call the corresponding protein design tool to perform protein design based on the protein design requirement information to obtain the protein design results.

[0219] It should be noted that the above-mentioned protein design tasks can be combined in various ways according to the protein design requirement information. Accordingly, in step 203, the protein design tasks in the above-mentioned combined protein design process can be executed by the protein design agent.

[0220] For example, the protein design requirement information is "Help me optimize the activity of XX enzyme". The corresponding protein design process can include the following tasks arranged in order:

[0221] Generation and presentation tasks: Generate and present, through a user interface, operational prompt information for instructing the user to provide a selection of a protein design region.

[0222] Parameter acquisition task: Acquire the protein design region selected by the user through the user interface for the above generation and presentation tasks.

[0223] Generation and presentation tasks: Generate and present, through the user interface, operational prompt information to instruct users to specify protein design parameters.

[0224] Parameter acquisition task: obtain the parameter values ​​of the protein design parameters set by the user through the user interface for the above generation and presentation tasks.

[0225] Requirements update task: Update protein design requirements information based on the obtained protein design region and protein design parameter information.

[0226] Protein design task: Determine and call the protein sequence generation algorithm and enzyme activity, enzyme stability, expression level and other models based on the updated protein design requirement information, and generate candidate protein sequences as protein design results based on the updated protein design requirement information.

[0227] As an important component of the agent-based protein design method provided by the above embodiments of the present disclosure, the protein design agent is described below for convenience of understanding the protein design agent. Figure 3 , Figure 3 An exemplary architecture diagram of a protein design agent is shown.

[0228] like Figure 3 As shown, the protein design agent includes: a user interface, a design process determination module, a protein design tool module and a data storage module.

[0229] The user interacts with the protein design agent through the user interface. The user interface provides a chat window, through which the user can provide protein design requirement information to the protein design agent. After receiving the protein design requirement information, the protein design agent can use the design process determination module to generate a protein design process, such as Figure 3 The following protein design process is shown: Task A, Task B, Task C and Task D. Among them:

[0230] Task A: The user is prompted to upload the designed antibody structure or sequence through the user interface. If the user enters the designed antibody structure or sequence in the user interface, the above antibody structure or sequence can be obtained and updated to the protein design requirements.

[0231] Task B: Prompt the user to select an antibody designable region through the user interface. If the user selects an antibody designable region in the user interface, the above antibody designable region can be obtained and updated to the protein design requirements.

[0232] Task C: Prompt the user to select a new sequence generation mode through the user interface, such as structure-based generation or sequence-based generation. If the user selects a new sequence generation mode in the user interface, the new sequence generation mode can be obtained and updated to the protein design requirements.

[0233] Task D: Prompt the user to select design target requirements through the user interface, such as expression level, stability, affinity, multimerization, etc. If the user selects a specific design target requirement in the user interface, the above design target requirements can be obtained and updated to the protein design requirements.

[0234] Figure 3 Only Task A, Task B, Task C and Task D are shown in the example. In practice, for example, Task E ( Figure 3 Task E is to call specific tools, algorithms or models in the corresponding protein design tool module based on the updated protein design requirements (for example, Figure 3 The protein design tool can be used to design proteins using the protein sequence generation algorithm, protein expression prediction model, protein multimerization prediction model, and protein stability prediction model shown in the figure to obtain protein design results. Of course, the user can also be guided to provide other protein design parameters and then call the protein design tool to perform protein design.

[0235] After the protein design agent obtains the protein design results, it can present the above protein design results through the user interface. Figure 3As shown, the protein structure information can be presented through the protein structure visualization plug-in, the protein sequence can be presented through the protein sequence visualization plug-in, and the model score in the protein design result can be presented through the model score plug-in. The user can evaluate the above-presented protein design results. If it is confirmed that the currently presented protein design result has met the requirements, the above-mentioned protein design result can be confirmed through the user interface, and then the above-mentioned protein design requirement information and the protein design result can be correspondingly stored as protein design information through the data storage module.

[0236] In addition, the protein design result can also be verified by a wet experiment verification device, such as verifying the protein sequence in the protein design result. After obtaining the experimental result data, the wet experiment verification device can send the experimental result data to the protein design intelligent agent, and through the data storage module, the experimental result data in the corresponding protein design information is saved, and presented by the experimental result data plug-in in the user interface. At the same time, the protein design intelligent agent can also determine whether the protein design result reaches the design goal based on the protein design requirement information and the experimental result data. If it is determined that the design goal is not reached, the machine learning model (such as protein generation model or scoring model, etc.) and / or the protein design parameter called in the process of generating the protein design result can also be fine-tuned to achieve closed-loop feedback.

[0237] The above-mentioned embodiment of the present disclosure provides an agent-based protein design method, which obtains the protein design requirement information input by the user through the user interface in this round of dialogue; then, through the protein design agent, the protein design process is determined based on the protein design requirement information, and the protein design process includes at least one protein design task; then, for each protein design task in the protein design process, the protein design task is executed by the task execution tool corresponding to the protein design task to obtain the protein design result; finally, the protein design result is presented through the user interface. That is, the user only needs to input the protein design requirement in the user interface provided by the protein design agent, and the protein design agent can automatically determine and execute the protein design process until the protein design result is finally obtained and presented. Here, the protein design logic summarized from experience is used to convert the complex protein design interaction mode into an interaction mode that is easy for users to understand and has guided prompts through the intelligent agent. Developers do not need to learn each protein design tool and algorithm in advance. They only need to put forward specific requirements, or optionally provide feedback on protein design parameter information (for example, the region of the protein sequence, design goals, etc.) under the guidance of the intelligent agent's prompt information. The intelligent agent will then automatically call the protein design tool to complete the protein design work, thereby improving the efficiency of protein design. In addition, the protein design process does not depend on the developer's level of understanding and actual experience, which can improve the reliability of protein design and reduce the probability of errors.

[0238] Further references Figure 4As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a protein design device based on an intelligent agent, and the device embodiment is Figure 2A Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0239] like Figure 4 As shown, the agent-based protein design device 400 of this embodiment includes: a requirement acquisition module 401, a process determination module 402, a process execution module 403 and a result presentation module 404. Among them, the requirement acquisition module 401 is configured to obtain the protein design requirement information input by the user through the user interface in this round of dialogue; the process determination module 402 is configured to determine the protein design process based on the protein design requirement information, and the protein design process includes at least one protein design task; the process execution module 403 is configured to execute the protein design task for each protein design task in the protein design process through the task execution tool corresponding to the protein design task to obtain the protein design result; the result presentation module 404 is configured to present the protein design result through the user interface.

[0240] In this embodiment, the specific processing of the demand acquisition module 401, the process determination module 402, the process execution module 403 and the result presentation module 404 of the agent-based protein design device 400 and the technical effects thereof can be referred to in Figure 2A The relevant descriptions of step 201, step 202, step 203 and step 204 in the corresponding embodiment are not repeated here.

[0241] In some optional embodiments, the agent-based protein design apparatus 400 may further include:

[0242] The design information storage module 405 is configured to store the protein design requirement information and the protein design result in correspondence as protein design information in response to detecting a confirmation operation of the user on the protein design result.

[0243] In some optional embodiments, the agent-based protein design apparatus 400 may further include:

[0244] The experimental result storage module 406 is configured to, in response to receiving the experimental result data synthesized and verified by the wet experiment, store the experimental result data as the experimental result data in the protein design information.

[0245] In some optional embodiments, the agent-based protein design apparatus 400 may further include:

[0246] The experiment result presenting module 407 is configured to present the experiment result data through the user interface.

[0247] In some optional embodiments, executing the protein design task by a task execution tool corresponding to the protein design task includes:

[0248] The protein design task is executed by calling the protein design parameters in the protein design requirement information through a machine learning model corresponding to the protein design task type; and the agent-based protein design device 400 may further include:

[0249] A model and / or parameter determination module 408 is configured to determine the model to be updated from the machine learning model called in the process of generating the protein design result, and / or determine the protein design parameters to be updated from the protein design parameters called in the process of generating the protein design result according to the protein design result and the experimental result data;

[0250] The updating module is configured 409 to fine-tune each of the models to be updated and / or each of the protein design parameters to be updated based on the protein design results and the experimental result data, so as to update each of the models to be updated and / or each of the protein design parameters to be updated.

[0251] In some optional implementations, the model and / or parameter determination module 408 may be further configured to:

[0252] Determining whether the experimental result data, relative to the protein design result, has an improvement in at least a preset ratio of the preset evaluation index that meets the corresponding index improvement requirement;

[0253] In response to the determination being yes, determining the protein design parameters involved in the process of generating the protein design result as the protein design parameters to be updated;

[0254] In response to a no determination, determining the protein generation model involved in the process of generating the protein design result as a model to be updated;

[0255] Determine the correlation between the protein design result and the model scoring result in the experimental result data, and in response to determining that the correlation is less than a preset correlation threshold, determine the scoring model involved in the process of generating the protein design result as a model to be updated.

[0256] In some optional embodiments, the process determination module 402 may be further configured to: determine the protein design process based on the protein design requirement information by a protein design agent; and

[0257] The protein design task may belong to a preset task type set, which may include: local conditional protein sequence generation, de novo protein sequence generation, protein property scoring, protein structure generation, and protein information retrieval.

[0258] In some optional implementations, each preset task type corresponds to at least one task execution tool; and

[0259] The sequentially executing each of the protein design tasks may include:

[0260] The protein design tasks are executed in sequence, wherein the protein design tasks are executed by a task execution tool corresponding to the task type to which the protein design tasks belong.

[0261] In some optional embodiments, the protein design result may include at least one of the following: protein sequence, protein structure information, protein design model score, and protein design evaluation index score.

[0262] In some optional embodiments, the protein design process may include at least one of the following protein design tasks:

[0263] Generate and present, through the user interface, operation prompt information for instructing the user to provide protein design parameter information;

[0264] Acquiring protein design parameter information input by the user through the user interface for the protein design task;

[0265] Updating the protein design requirement information based on the acquired protein design parameter information;

[0266] According to the protein design requirement information, a corresponding protein design tool is determined and called to perform protein design to obtain a protein design result.

[0267] In some optional embodiments, the protein design parameter information may include at least one of the following:

[0268] Protein structure or sequence, protein sequence designable region, protein sequence generation pattern, protein design target requirements, protein design evaluation model, protein design evaluation indicators, and protein properties to be optimized.

[0269] In some optional embodiments, the protein design tool may include at least one of the following: a protein sequence generation tool, a protein expression prediction tool, a protein polymerization prediction tool, and a protein stability prediction tool.

[0270] It should be noted that the implementation details and technical effects of each module and unit in the agent-based protein design device provided in the embodiments of the present disclosure can be referred to the description of other embodiments in the present disclosure, and will not be repeated here.

[0271] Reference below Figure 5 , which shows a schematic diagram of the structure of a computer system 500 suitable for implementing the electronic device of the present disclosure. Figure 5 The computer system 500 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0272] like Figure 5 As shown, the computer system 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 to a random access memory (RAM) 503. Various programs and data required for the operation of the computer system 500 are also stored in the RAM 503. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0273] Typically, the following devices may be connected to the I / O interface 505: input devices 506 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, etc.; output devices 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 508 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 509. The communication devices 509 may allow the computer system 500 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 5 The computer system 500 of an electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0274] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0275] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0276] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0277] The computer readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device implements the following Figure 2A The illustrated embodiment and its alternative implementations illustrate an agent-based protein design method.

[0278] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate 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 through 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., through the Internet using an Internet service provider).

[0279] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0280] The units or modules involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a unit or module does not limit the unit itself in some cases. For example, an acquisition module may also be described as a "module for acquiring protein design requirement information input by a user through a user interface in this round of dialogue."

[0281] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

Claims

1. A protein design method based on an intelligent agent, comprising: Acquire protein design requirement information input by the user through the user interface; Determining a protein design process based on the protein design requirement information, wherein the protein design process includes at least one protein design task; For each protein design task in the protein design process, executing the protein design task by a task execution tool corresponding to the protein design task to obtain a protein design result; The protein design results are presented through the user interface.

2. The method according to claim 1, wherein: The method further comprises: In response to detecting a confirmation operation of the user on the protein design result, storing the protein design requirement information and the protein design result in correspondence with each other as protein design information; In response to receiving experimental result data synthesized and verified by wet experiment, the experimental result data is correspondingly stored as experimental result data in the protein design information.

3. The method according to claim 2, wherein: The performing of the protein design task by a task execution tool corresponding to the protein design task comprises: Calling the protein design parameters in the protein design requirement information through a machine learning model corresponding to the protein design task type to execute the protein design task; and The method further comprises: According to the protein design result and the experimental result data, determining a model to be updated in the machine learning model called in the process of generating the protein design result, and / or determining a protein design parameter to be updated in the protein design parameters called in the process of generating the protein design result; Based on the protein design results and the experimental result data, each of the models to be updated and / or each of the design parameters of the proteins to be updated is fine-tuned to update each of the models to be updated and / or each of the design parameters of the proteins to be updated.

4. The method according to claim 3, wherein: Determining the model to be updated and / or the protein design parameters to be updated according to the protein design results and the experimental result data includes: Determining whether the experimental result data, relative to the protein design result, has an improvement in at least a preset ratio of the preset evaluation index that meets the corresponding index improvement requirement; In response to the determination being yes, determining the protein design parameters involved in the process of generating the protein design result as the protein design parameters to be updated; In response to a no determination, determining the protein generation model involved in the process of generating the protein design result as a model to be updated; Determine the correlation between the protein design result and the model scoring result in the experimental result data, and in response to determining that the correlation is less than a preset correlation threshold, determine the scoring model involved in the process of generating the protein design result as a model to be updated.

5. The method according to claim 1, wherein: The determining of the protein design process based on the protein design requirement information comprises: determining the protein design process based on the protein design requirement information by a protein design agent; and The protein design task belongs to a preset task type set, which includes: local conditional protein sequence generation, de novo protein sequence generation, protein property scoring, protein structure generation, and protein information retrieval.

6. The method according to claim 5, wherein: Each preset task type corresponds to at least one task execution tool; as well as The performing of the protein design task by a task execution tool corresponding to the protein design task comprises: The protein design task is executed by a task execution tool corresponding to the task type to which the protein design task belongs.

7. A protein design device based on an intelligent agent, wherein: include: A requirement acquisition module is configured to acquire protein design requirement information input by the user through the user interface in this round of dialogue; A process determination module is configured to determine a protein design process based on the protein design requirement information, wherein the protein design process includes at least one protein design task; A process execution module is configured to execute each protein design task in the protein design process through a task execution tool corresponding to the protein design task to obtain a protein design result; The result presentation module is configured to present the protein design result through the user interface.

8. An electronic device, wherein: include: one or more processors; a storage device having one or more programs stored thereon, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, wherein: A computer program is stored thereon, wherein when the computer program is executed by one or more processors, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, wherein: The method comprises a computer program / instruction, which implements the method according to any one of claims 1 to 6 when executed by a processor.

Images