Method and device for intelligently and automatically processing data of cryoelectron microscope

Abstract

The invention relates to an intelligent automatic processing method and device for cryoelectron microscope data, target task information of a cryoelectron microscope data processing task is input into a large language model, and operation information used for realizing a task target is generated through the large language model based on basic parameter information and a local knowledge base in the target task information. The operation information is sent to the main agent, the working agents are sequentially controlled to execute corresponding operation steps based on step parameters in an iteration mode, corresponding feedback information is generated after iteration is finished, operation of the current working agent is finished in positive feedback, and the next working agent is controlled to execute the next operation step; if negative feedback exists, step parameters corresponding to the current operation step are updated, and the current working agent is controlled to execute the current operation step again based on the updated step parameters. According to the method, the intelligent agent is automatically indicated to execute the operation through the large language model, the cryoelectron microscope data is automatically processed without depending on user experience, and the processing process efficiency and the accuracy of the processing result are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a method and device for intelligent automatic processing of cryo-electron microscopy data. Background Art

[0002] Generally speaking, the processing of cryo-electron microscopy data usually includes the following operations: image drift correction, estimation of the defocus parameter CTF, selection and extraction of protein particles, 2D classification and 3D classification of protein particles, 3D reconstruction of protein particles, etc. Each of these operations involves the setting of numerous parameters, and the accuracy and rationality of the parameter settings directly determine the quality of each step of data processing, and thus determine the accuracy and precision of the final generated protein 3D structure. In this data processing process, parameter setting is highly dependent on expert experience and knowledge. For the same set of cryo-electron microscopy data, the user's experience and knowledge structure directly determine the quality of data processing. Summary of the Invention

[0003] In view of this, the present disclosure proposes a method and device for intelligent automatic processing of cryo-electron microscopy data, aiming to break free from the limitations of user experience and automatically and efficiently realize the processing of cryo-electron microscopy data.

[0004] According to a first aspect of the present disclosure, a method for intelligent automatic processing of cryo-electron microscopy data is provided, the method comprising:

[0005] Determining target task information of a cryo-electron microscopy data processing task, wherein the target task information includes task objectives and basic parameter information;

[0006] Inputting target task information into a large language model, and generating operation information for achieving the task goal through the large language model based on the basic parameter information and a local knowledge base, wherein the operation information includes at least one operation step with an execution order and a step parameter corresponding to each operation step;

[0007] The operation information is sent to the master agent, which controls the working agents to perform the corresponding operation steps based on the step parameters in an iterative manner according to the operation sequence, and generates corresponding feedback information after each iteration;

[0008] In response to the feedback information being positive feedback, ending the operation of the current working agent and controlling the working agent corresponding to the next operation step to perform the operation step;

[0009] In response to the feedback information being negative feedback, the step parameters corresponding to the current operation step are updated, and the current working agent is controlled to execute the current operation step again based on the updated step parameters.

[0010] In a possible implementation, generating corresponding feedback information after each iteration includes:

[0011] Determine the operation results obtained by the working agent performing the corresponding operation steps according to the corresponding step parameters during this iteration;

[0012] Corresponding feedback information is generated based on the corresponding operation results of the working agent through the large language model.

[0013] In a possible implementation, generating corresponding feedback information after each iteration further includes:

[0014] Sending the operation result to the master agent through the working agent;

[0015] The operation result is forwarded to the large language model through the main agent.

[0016] In a possible implementation, generating corresponding feedback information based on the operation result corresponding to the working agent by using the large language model includes:

[0017] Determine whether the corresponding operation result meets the expected result based on the local knowledge base using a large language model;

[0018] In response to the operation result meeting the expected result, determining the feedback information as positive feedback;

[0019] In response to the operation result not meeting the expected result, the feedback information is determined to be negative feedback.

[0020] In a possible implementation, the step parameters include calculation steps and a parameter list.

[0021] In a possible implementation, the operation steps include at least one of drift correction, CTF parameter estimation, protein particle selection and extraction, protein particle 2D classification, and protein three-dimensional structure reconstruction.

[0022] According to a second aspect of the present disclosure, a device for intelligent automatic processing of cryo-electron microscopy data is provided, the device comprising:

[0023] A task information determination module is used to determine target task information of a cryo-electron microscopy data processing task, wherein the target task information includes task objectives and basic parameter information;

[0024] an operation information determination module, configured to input target task information into a large language model, and generate operation information for achieving the task goal through the large language model based on the basic parameter information and a local knowledge base, wherein the operation information includes at least one operation step having an execution order and a step parameter corresponding to each operation step;

[0025] The task execution module is used to send the operation information to the master agent, control the working agents to perform the corresponding operation steps based on the step parameters in an iterative manner according to the operation sequence, and generate corresponding feedback information after each iteration;

[0026] A first feedback module is configured to terminate the operation of the current working agent in response to the feedback information being positive feedback, and control the working agent corresponding to the next operation step to perform the operation step;

[0027] The second feedback module is used to update the step parameters corresponding to the current operation step in response to the feedback information being negative feedback, and control the current working agent to execute the current operation step again based on the updated step parameters.

[0028] In a possible implementation, the task execution module is further configured to:

[0029] Determine the operation results obtained by the working agent performing the corresponding operation steps according to the corresponding step parameters during this iteration;

[0030] Corresponding feedback information is generated based on the corresponding operation results of the working agent through the large language model.

[0031] In a possible implementation, the task execution module is further configured to:

[0032] Sending the operation result to the master agent through the working agent;

[0033] The operation result is forwarded to the large language model through the main agent.

[0034] In a possible implementation, the task execution module is further configured to:

[0035] Determine whether the corresponding operation result meets the expected result based on the local knowledge base using a large language model;

[0036] In response to the operation result meeting the expected result, determining the feedback information as positive feedback;

[0037] In response to the operation result not meeting the expected result, the feedback information is determined to be negative feedback.

[0038] In a possible implementation, the step parameters include calculation steps and a parameter list.

[0039] In a possible implementation, the operation steps include at least one of drift correction, CTF parameter estimation, protein particle selection and extraction, protein particle 2D classification, and protein three-dimensional structure reconstruction.

[0040] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0041] According to a fourth aspect of the present disclosure, a non-volatile computer-readable storage medium is provided, on which computer program instructions are stored, wherein the computer program instructions implement the above method when executed by a processor.

[0042] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0043] In an embodiment of the present disclosure, the target task information of the cryo-electron microscopy data processing task is input into a large language model, and the large language model generates operation information for achieving the task goal based on the basic parameter information and the local knowledge base therein. The operation information is sent to the main agent, and the working agents are sequentially controlled to perform corresponding operation steps based on the step parameters in an iterative manner. After the iteration is completed, corresponding feedback information is generated. Positive feedback ends the operation of the current working agent and controls the next working agent to perform the next operation step; negative feedback updates the step parameters corresponding to the current operation step and controls the current working agent to perform the current operation step again based on the updated step parameters. The present disclosure automatically instructs the agent to perform operations through a large language model, and automatically processes cryo-electron microscopy data without relying on user experience, thereby improving the efficiency of the processing process and the accuracy of the processing results.

[0044] Further features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure and, together with the description, serve to explain the principles of the disclosure.

[0046] Figure 1 A flowchart showing a method for intelligent automatic processing of cryo-electron microscopy data according to an embodiment of the present disclosure is shown;

[0047] Figure 2 A schematic diagram illustrating an intelligent automatic processing process of cryo-electron microscopy data according to an embodiment of the present disclosure;

[0048] Figure 3 A schematic diagram showing an intelligent automatic processing device for cryo-electron microscopy data according to an embodiment of the present disclosure is shown;

[0049] Figure 4 A schematic diagram illustrating an electronic device according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0050] Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise indicated.

[0051] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0052] In addition, numerous specific details are provided in the following detailed description to better illustrate the present disclosure. Those skilled in the art will appreciate that the present disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art are not described in detail in order to highlight the main points of the present disclosure.

[0053] The method for intelligent automatic processing of cryo-electron microscopy data of the embodiment of the present disclosure can be executed by an electronic device such as a terminal device or a server. Among them, the terminal device can be any fixed or mobile terminal such as user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, vehicle-mounted device, wearable device, etc. The server can be a single server or a server cluster composed of multiple servers. Any electronic device can implement the method for intelligent automatic processing of cryo-electron microscopy data of the embodiment of the present disclosure by calling computer-readable instructions stored in a memory through a processor.

[0054] Figure 1 FIG. 1 is a flow chart showing a method for intelligent automatic processing of cryo-electron microscopy data according to an embodiment of the present disclosure. Figure 1 As shown, the intelligent automatic processing method of cryo-electron microscopy data in the embodiment of the present disclosure may include the following steps S10-S50.

[0055] Step S10: Determine the target task information of the cryo-electron microscopy data processing task.

[0056] In one possible implementation, the target task information of a cryo-EM data processing task can be determined by an electronic device. The cryo-EM data processing task is used to generate a cryo-EM image that meets the expected requirements, and the target task information includes the task objective and basic parameter information of the cryo-EM data processing task. For example, when the cryo-EM data processing task is a 3D Refinement task for reconstructing the three-dimensional structure of a protein, the task objective can be to generate a cryo-EM image with a reconstructed resolution of no less than 5 angstroms. The basic parameter information can include the path of the data to be processed, the pixel size of the protein particles, the camera voltage, spherical aberration, and other data parameters required to generate the cryo-EM image.

[0057] Optionally, the electronic device may determine the target task information by receiving the target task information input by a user through human-computer interaction, receiving the target task information sent by another device, or generating the target task information according to a preset rule. The target task information may be described in natural language.

[0058] Step S20: Input the target task information into the large language model, and generate operation information for achieving the task goal based on the basic parameter information and the local knowledge base through the large language model.

[0059] In one possible implementation, after determining the target task information, the electronic device inputs the target task information into a large language model. The large language model can then automatically analyze the user's needs and goals based on the target task information. Furthermore, the large language model combines a preset local knowledge base with basic parameter information in the target task information to formulate a plan for achieving the task objectives and output corresponding operational information. The local knowledge base can be a knowledge base built based on the experience of experts in the field, including multiple parameters previously set by experts in the field when processing cryo-electron microscopy data.

[0060] Optionally, the operation information output by the large language model in the embodiment of the present disclosure may include at least one operation step with an execution order and step parameters corresponding to each operation step. Among them, the operation steps may include at least one of drift correction, CTF parameter estimation, protein particle selection and extraction, protein particle 2D classification, and protein three-dimensional structure reconstruction, and each operation step has a corresponding execution order. The step parameters may include a calculation step of at least one calculation included in the operation step and a parameter list consisting of multiple parameters for calculation. Exemplarily, for the 3D Refinement task of reconstructing the three-dimensional structure of a protein, the parameter list in the step parameters may include the data block Patch size for drift correction, the image scaling factor binFactor, the resolution range in the CTF parameter estimation, the search step size, the window size, etc.

[0061] Step S30: Send the operation information to the main agent, control the working agents in an iterative manner in the operation sequence to perform the corresponding operation steps based on the step parameters, and generate corresponding feedback information after each iteration.

[0062] In one possible implementation, the electronic device can send the operation information to the main agent after determining the operation information including multiple operation steps and step parameters corresponding to each operation step through a large language model. After receiving the operation information, the main agent controls the working agent in sequence in an iterative manner to perform the corresponding operation steps based on the step parameters, and generates corresponding feedback information after each iteration. Each working agent is used to perform different operation steps. In one iteration, the main agent controls a working agent to perform its corresponding operation steps, that is, in each iteration, the current working agent performs at least one calculation step according to the step parameters of the corresponding operation step to obtain the corresponding operation result. The order in which the main agent controls the working agents is determined according to the operation order of the operation steps that the working agents can perform.

[0063] Alternatively, an agent is a system that can perceive its environment and make decisions based on that information to achieve a specific goal. It can be a simple software program or a complex robot, and is widely used in fields such as autonomous driving, automated control, and game AI. The core of an agent lies in its decision-making ability to select the optimal action path in a complex and dynamic environment. Therefore, an agent possesses the capabilities of perception, decision-making, learning, and action, and can use these capabilities to achieve a specific task.

[0064] Furthermore, after each iteration, the electronic device can generate feedback information based on the operation results generated by the operation steps executed during the iteration, used to provide feedback on the quality of the operation process. Optionally, the content of the operation results generated by different operation steps varies. For example, if the operation step is drift correction, the operation result is the amount of image drift in the X and Y directions. If the operation step is CTF parameter estimation, the operation result can be the amount of defocus obtained by estimating the correlation coefficient of the fitting results. The feedback information can be generated based on the large language model's judgment of the quality of the operation results. For example, after each iteration, the electronic device can determine the operation results obtained by the working agent performing the corresponding operation steps according to the corresponding step parameters during the iteration. The large language model then generates corresponding feedback information based on the corresponding operation results of the working agent. The large language model can obtain the operation results by forwarding them through the main agent, that is, the working agent sends the operation results to the main agent after completing the operation, and the main agent then forwards the operation results to the large language model.

[0065] Furthermore, after receiving the operation result obtained in the current iteration process forwarded by the main agent, the large language model can determine whether the corresponding operation result meets the expected result based on the local knowledge base. If the operation result meets the expected result, the feedback information is determined to be positive feedback. If the operation result does not meet the expected result, the feedback information is determined to be negative feedback. The expected result is a threshold summarized by the large language model based on the local knowledge base. If the operation result meets the threshold, it is considered that the feedback result is consistent with the expectation or exceeds the expectation, and the feedback information is determined to be positive feedback; if the operation result does not meet the threshold, it is considered that the feedback result is inferior to the result summarized by the large model based on the expert knowledge base, and the feedback information is determined to be negative feedback.

[0066] Step S40, in response to the feedback information being positive feedback, the operation of the current working agent is terminated, and the working agent corresponding to the next operation step is controlled to perform the operation step. In a possible implementation, when the electronic device determines that the feedback information is positive feedback, the operation of the current agent is terminated through the main agent, and in the next iteration process, the working agent corresponding to the next operation step is controlled to perform the next operation step. Take working agent 1, working agent 2, and working agent 3 as an example, in which working agent 1, working agent 2, and working agent 3 are respectively used to perform operation step 1, operation step 2, and operation step 3, and the execution order of the three operation steps is operation step 1, operation step 2, and operation step 3. In the case where the main agent controls working agent 1 to perform operation step 1 in the current iteration process, if the feedback information is positive feedback after the execution is completed, the main agent determines that the execution result meets expectations, terminates the execution of operation step 1, and controls working agent 2 to perform the corresponding operation step 2 in the next iteration process.

[0067] Step S50: In response to the feedback information being negative feedback, update the step parameters corresponding to the current operation step, and control the current working agent to execute the current operation step again based on the updated step parameters.

[0068] In one possible implementation, if the electronic device determines that the feedback information is negative, it can use the large semantic model to analyze the local knowledge base to readjust the step parameters of the current operation step and then send them to the master agent. After receiving the updated step parameters, the master agent controls the current working agent to execute the current operation step again based on the updated step parameters during the next iteration.

[0069] Furthermore, the embodiment of the present disclosure ends the iteration process after achieving the task goal in the target task information.

[0070] Figure 2 A schematic diagram showing an intelligent automatic processing process of cryo-electron microscopy data according to an embodiment of the present disclosure is shown. Figure 2 As shown, the target task information of the cryo-electron microscopy data processing task described in natural language is input into the large language model by an electronic device. The large language model generates operation information for achieving the task goal based on the basic parameter information and the local knowledge base. The operation information includes at least one operation step with an execution order and the step parameters corresponding to each operation step (including calculation steps and parameter lists). Further, the master agent (Master) controls the worker agent (Worker) in an iterative manner according to the operation order to perform the corresponding operation steps based on the step parameters. After each iteration, the worker agent feeds back the calculation results to the master agent, which are then forwarded to the large language model to generate corresponding feedback information. If the feedback information is positive feedback that meets the expectations, the operation of the current worker agent is terminated and the worker agent corresponding to the next operation step is controlled to perform the operation step. If the feedback information is negative feedback that does not meet the expectations, the step parameters corresponding to the current operation step are updated, and the current worker agent is controlled to perform the current operation step again based on the updated step parameters. The iterative process ends when the preset task purpose is reached, completing the cryo-electron microscopy data processing task.

[0071] Taking the 3D Refinement process of protein 3D structure reconstruction in cryo-electron microscopy data processing as an example, the entire data processing process starts with raw data preparation and requires sequential drift correction, CTF parameter estimation, protein particle selection and extraction, protein particle 2D classification, and protein 3D structure reconstruction. In the MAP method, the execution process steps of 3D Refinement are as follows:

[0072] 1) The user informs the large model through natural language that it needs to perform a 3DRefinement operation for cryo-electron microscopy, with the goal of reconstructing a resolution of no less than 5 angstroms, and provides the path of the data to be processed, the pixel size of the protein particles, the camera voltage, spherical aberration and other data parameters.

[0073] 2) The large model background automatically analyzes the user's needs and goals, combines the local expert knowledge base, and formulates the operational steps to achieve the task goals (such as performing drift correction, CTF parameter estimation, selecting protein particles, etc.) and a parameter list for each operation step (such as the data block patch size for drift correction, image scaling factor binFactor, resolution range in CTF parameter estimation, search step size, window size, etc.), and sends them to the main intelligent agent Master.

[0074] 3) After receiving the large model operation step sequence and the parameter list corresponding to each step, the Master starts the worker agent corresponding to the corresponding step to perform the corresponding calculation operation.

[0075] 4) After the Worker completes execution, it feeds its output results back to the Master. For example, the drift correction Worker sends the drift of each photo in the X and Y directions to the Worker, and the CTF parameter estimation module sends the correlation coefficient of the fitting result and the estimated defocus amount to the Master.

[0076] 5) After receiving the feedback from the Worker, the Master forwards it to the big model.

[0077] 6) The big model combines its own expert knowledge base to analyze and compare the feedback results of each Worker. If the result fed back by a Worker is consistent with or exceeds expectations, the Master is informed that the Worker can stop its own calculation task; if the result fed back by a Worker is inferior to the result summarized by the big model based on the expert knowledge base, the parameter value of the operation step is adjusted based on the knowledge obtained from its own analysis of the expert knowledge base and sent to the Master, and steps 3), 4), 5), and 6) are repeated until the resolution of the final reconstructed protein three-dimensional structure is higher than 5 angstroms.

[0078] Based on the above technical features, the embodiment of the present disclosure automatically instructs the intelligent agent to perform operations through a large language model, and provides result feedback after each step of the operation is performed, and adjusts the operation results according to the feedback results. In the above-mentioned processing process, the user only needs to interact with the large language model in a natural language manner. Once the large language model understands the user's needs and goals, it can automatically execute the subsequent data processing process. The user does not need to intervene later and only needs to wait for the processing results. Therefore, the embodiment of the present disclosure does not rely on user experience to automatically process cryo-electron microscopy data, thereby improving the efficiency of the processing process and the accuracy of the processing results. Furthermore, in the process of processing data, the large language model and the intelligent agent can also display intermediate calculation results to the user, so that the user can interact with the large language model as needed to determine the subsequent calculation process, thereby further improving the accuracy of the final result.

[0079] Figure 3 A schematic diagram of an intelligent automatic processing device for cryo-electron microscopy data according to an embodiment of the present disclosure is shown. Figure 3 As shown, the intelligent automatic processing device for cryo-electron microscopy data of the embodiment of the present disclosure may include:

[0080] A task information determination module 30 is used to determine target task information of a cryo-electron microscopy data processing task, wherein the target task information includes task objectives and basic parameter information;

[0081] An operation information determination module 31 is configured to input target task information into a large language model, and generate operation information for achieving the task goal based on the basic parameter information and a local knowledge base through the large language model, wherein the operation information includes at least one operation step having an execution order and step parameters corresponding to each operation step;

[0082] The task execution module 32 is used to send the operation information to the master agent, control the working agents to perform the corresponding operation steps based on the step parameters in an iterative manner according to the operation sequence, and generate corresponding feedback information after each iteration;

[0083] The first feedback module 33 is configured to terminate the operation of the current working agent in response to the feedback information being positive feedback, and control the working agent corresponding to the next operation step to perform the operation step;

[0084] The second feedback module 34 is used to update the step parameters corresponding to the current operation step in response to the feedback information being negative feedback, and control the current working agent to execute the current operation step again based on the updated step parameters.

[0085] In a possible implementation, the task execution module 32 is further configured to:

[0086] Determine the operation results obtained by the working agent performing the corresponding operation steps according to the corresponding step parameters during this iteration;

[0087] Corresponding feedback information is generated based on the corresponding operation results of the working agent through the large language model.

[0088] In a possible implementation, the task execution module 32 is further configured to:

[0089] Sending the operation result to the master agent through the working agent;

[0090] The operation result is forwarded to the large language model through the main agent.

[0091] In a possible implementation, the task execution module 32 is further configured to:

[0092] Determine whether the corresponding operation result meets the expected result based on the local knowledge base using a large language model;

[0093] In response to the operation result meeting the expected result, determining the feedback information as positive feedback;

[0094] In response to the operation result not meeting the expected result, the feedback information is determined to be negative feedback.

[0095] In a possible implementation, the step parameters include calculation steps and a parameter list.

[0096] In a possible implementation, the operation steps include at least one of drift correction, CTF parameter estimation, protein particle selection and extraction, protein particle 2D classification, and protein three-dimensional structure reconstruction.

[0097] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments. For the sake of brevity, it will not be repeated here.

[0098] The present disclosure also provides a computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions implement the above method when executed by a processor. The computer-readable storage medium may be a volatile or non-volatile computer-readable storage medium.

[0099] An embodiment of the present disclosure further proposes an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0100] An embodiment of the present disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0101] Figure 4 1 shows a schematic diagram of an electronic device 1900 according to an embodiment of the present disclosure. For example, the electronic device 1900 can be provided as a server or a terminal device. Figure 4 The electronic device 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932 for storing instructions executable by the processing component 1922, such as an application. The application stored in the memory 1932 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute the instructions to perform the above-described method.

[0102] The electronic device 1900 may further include a power supply component 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to a network, and an input / output interface 1958 (I / O interface). The electronic device 1900 may operate based on an operating system stored in the memory 1932, such as Windows Server 2003. TM , Mac OS X TM , Unix TM ,Linux TM , FreeBSD TM or similar.

[0103] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by the processing component 1922 of the electronic device 1900 to perform the above method.

[0104] The present disclosure may be a system, method and / or computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0105] A computer-readable storage medium can be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: 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), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination thereof. As used herein, a computer-readable storage medium is not to be construed as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through an electrical wire.

[0106] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0107] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, and conventional procedural programming languages ​​such as "C" language or similar programming languages. Computer-readable program instructions may be executed entirely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., utilizing an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer-readable program instructions. The electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0108] Various aspects of the present disclosure are described herein with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0109] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device is generated that implements the functions / actions specified in one or more blocks in the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium, where these instructions cause the computer, programmable data processing device, and / or other device to operate in a specific manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing various aspects of the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0110] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0111] The flow charts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and the part of the module, program segment or instruction contains one or more executable instructions for realizing the prescribed logical function. In some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two consecutive boxes 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 box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the prescribed function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0112] While various embodiments of the present disclosure have been described above, the foregoing description is intended to be illustrative, non-exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or technological improvements in the marketplace, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for intelligent automatic processing of cryo-electron microscopy data, characterized in that: The method comprises: Determining target task information of a cryo-electron microscopy data processing task, wherein the target task information includes task objectives and basic parameter information; Inputting target task information into a large language model, and generating operation information for achieving the task goal through the large language model based on the basic parameter information and a local knowledge base, wherein the operation information includes at least one operation step with an execution order and a step parameter corresponding to each operation step; The operation information is sent to the master agent, which controls the working agents to perform the corresponding operation steps based on the step parameters in an iterative manner according to the operation sequence, and generates corresponding feedback information after each iteration; In response to the feedback information being positive feedback, ending the operation of the current working agent and controlling the working agent corresponding to the next operation step to perform the operation step; In response to the feedback information being negative feedback, the step parameters corresponding to the current operation step are updated, and the current working agent is controlled to execute the current operation step again based on the updated step parameters.

2. The method according to claim 1, characterized in that The generating of corresponding feedback information after each iteration includes: Determine the operation results obtained by the working agent performing the corresponding operation steps according to the corresponding step parameters during this iteration; Corresponding feedback information is generated based on the corresponding operation results of the working agent through the large language model.

3. The method according to claim 2, characterized in that The generating of corresponding feedback information after each iteration also includes: Sending the operation result to the master agent through the working agent; The operation result is forwarded to the large language model through the main agent.

4. The method according to claim 2 or 3, characterized in that The generating corresponding feedback information based on the operation result corresponding to the working agent by the large language model includes: Determine whether the corresponding operation result meets the expected result based on the local knowledge base using a large language model; In response to the operation result meeting the expected result, determining the feedback information as positive feedback; In response to the operation result not meeting the expected result, the feedback information is determined to be negative feedback.

5. The method according to any one of claims 1 to 4, characterized in that The step parameters include calculation steps and parameter lists.

6. The method according to any one of claims 1 to 5, characterized in that The operation steps include at least one of drift correction, CTF parameter estimation, protein particle selection and extraction, protein particle 2D classification, and protein three-dimensional structure reconstruction.

7. An intelligent automatic processing device for cryo-electron microscopy data, characterized in that: The device comprises: A task information determination module is used to determine target task information of a cryo-electron microscopy data processing task, wherein the target task information includes task objectives and basic parameter information; an operation information determination module, configured to input target task information into a large language model, and generate operation information for achieving the task goal through the large language model based on the basic parameter information and a local knowledge base, wherein the operation information includes at least one operation step having an execution order and a step parameter corresponding to each operation step; The task execution module is used to send the operation information to the master agent, control the working agents to perform the corresponding operation steps based on the step parameters in an iterative manner according to the operation sequence, and generate corresponding feedback information after each iteration; A first feedback module is configured to terminate the operation of the current working agent in response to the feedback information being positive feedback, and control the working agent corresponding to the next operation step to perform the operation step; The second feedback module is used to update the step parameters corresponding to the current operation step in response to the feedback information being negative feedback, and control the current working agent to execute the current operation step again based on the updated step parameters.

8. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to implement the method according to any one of claims 1 to 6 when executing the instructions stored in the memory.

9. A non-volatile computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images