An automated dual electrode membrane clamp manipulation method

Abstract

The application discloses an automatic two-electrode patch clamp operation method, and relates to the technical field of cell-level micromanipulation, and specifically comprises the following steps: S1, calculating the working space of a two-electrode tip through patch clamp double-mechanical arm motion modeling; S2, acquiring the pixel point coordinates of the two needles in the field of view through a deep learning algorithm for two-dimensional positioning of the needle tip; S3, according to the double-electrode working space obtained in step S1, guiding the placement position of the brain slice and the method for positioning the needle tip in step S2, performing motion control on the double electrode in three stages of coarse-fine-precision by using a Qt Creator tool; and S4, moving and contacting the target cell surface of the double-electrode patch clamp through an automatic program control. The application automatically controls the double electrode to contact the target cell under the working space obtained based on double-arm kinematic modeling by using a deep learning target detection method, and realizes a contact success rate of 90% under an average operation time of 6 minutes.

Description

Technical Field

[0001] This invention relates to the field of cell-level micromanipulation technology, specifically to an automated dual-electrode patch-clamp manipulation method. Background Technology

[0002] Patch-clamp technique, which uses microtubule electrodes in close contact with the cell membrane to record weak electrophysiological signals from single ion channels, is widely recognized as the gold standard for ion channel research. Two-electrode patch-clamp techniques are particularly valued for their ability to use one electrode to stimulate the cell while the other records the signal.

[0003] The key to the two-electrode patch-clamp technique is controlling the coordinated movement of two microtubule electrodes within the confined space of brain tissue to reach the target nerve cell. During this process, the operator is highly susceptible to misjudging the workspace of the electrodes, leading to electrode collisions and damage, ultimately resulting in the failure of the two-electrode patch-clamp operation. Furthermore, the significant difference between the micrometer-scale microneedle tip and the cell diameter and the centimeter-scale operating range means that guiding the two electrodes to the target cell often requires considerable time, reducing experimental efficiency. Therefore, it is essential to design a patch-clamp operation method that, based on the system environment, defines the operating workspace and automates the movement of the two electrodes to the target cell surface while achieving needle tip positioning. Summary of the Invention

[0004] This invention addresses the problems existing in existing dual-electrode patch-clamp cell manipulation methods by proposing an automated dual-electrode patch-clamp manipulation method. It utilizes kinematic methods to establish a dual-robotic arm model of the patch-clamp system environment and determine the workspace of the dual electrodes. It uses Qt Creator tools and deep learning methods to locate the electrodes and performs coarse-fine-precision automated motion control to ultimately move the dual electrodes to the surface of the target cell and make contact with the cell.

[0005] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0006] This invention provides an automated dual-electrode patch-clamp operation method, specifically a method for contacting a cell membrane with a dual-electrode patch-clamp based on automation technology, comprising the following steps:

[0007] S1: Based on the kinematic model and the environmental parameters of the patch-clamp system, a dual-arm model is established. Then, the Monte Carlo random probability model is used to solve the reachable workspace of the end of the robotic arm, i.e. the tip of the microtube electrode. The existence of the overlapping space between the workspace of the dual electrode end and the experimental brain slice placement space can not only prove the feasibility of dual-electrode contact with target cells, but also indicate the placement position of the brain slice in the experiment.

[0008] The robotic arm consists of a series of links connected by joints, and the DH (Denavit-Hartenberg) parameters are used to describe the motion relationships between the links. Each arm has one extension joint and one rotational joint on the XYZ axes. To prevent the two needles from colliding with the objective lens, the left arm's rotational joint is typically fixed at -pi / 6 degrees, and the right arm's rotational joint is fixed at pi / 6 degrees.

[0009] To establish the forward kinematic model, the relative positions of the double-arm base and the worktable were measured. The origin of the left arm base was 6 cm from the worktable plane, and the origin of the right arm base was 13.5 cm from the worktable plane. The projections of the origins of the two arm bases onto the XY plane were 32 cm apart. The diameter of the circular hole in the worktable where the petri dish was placed was 5 cm, and the origin of the left arm base was 12.5 cm from the center of the circular hole on the Y-axis. In the world coordinate system, taking the point 2.5 cm translated from the center of the circular hole towards the positive X-axis as the origin, the coordinates of the origin of the left arm base were (-2.5, -12.5, 6), and the coordinates of the origin of the right arm base were (-2.5, 19.5, 13.5).

[0010] The workspace of a dual-arm robot refers to the set of coordinate points reachable by the end effector of the robotic arm during actual operation. Based on forward kinematics, the position and orientation of the end effector are determined using the angle information of each joint. A Monte Carlo random probability model is then used to solve the workspace of the robotic arm, treating the positions of the electrode tips as random numbers, and approximating the collaborative workspace using a region composed of a sufficient number of random numbers. The existence of the overlap between the workspace of the dual electrode tips and the space for placing the experimental brain slices not only demonstrates the feasibility of dual-electrode contact with target cells but also provides guidance for the placement of the brain slices.

[0011] S2: The electrode tips in the camera image are detected by a deep learning target recognition algorithm, and the coordinates of the center pixel of the rectangular bounding box where the tip is located are taken as the position of the tip. Based on this, the coordinates of the tip can be obtained in real time to accurately control the movement of the dual electrodes.

[0012] Before controlling electrode movement, electrode localization is performed using the FRCNN algorithm. The acquired microscopic images are transmitted to the algorithm. First, a pre-trained convolutional network extracts features and outputs a feature map. Then, the feature extraction network searches for rectangular candidate regions that may contain the tip. Pooling layers and fully connected layers of the convolutional network are used to filter and adjust the bounding boxes with the highest probabilities. Finally, the algorithm returns the two-dimensional coordinates of the center of the rectangular region as the position of the electrode tip (two-dimensional localization). In tip localization, this return value not only provides the relative distance between the two electrodes to avoid collisions, but also allows for precise control of the tip movement step size by calculating the difference between this return value and the target's arrival position.

[0013] S3: Based on the dual-electrode workspace obtained in step S1, guide the placement of brain slices and use the FRCNN algorithm for pinpoint localization in step S2. Use tools such as Qt Creator to perform coarse-fine-precision motion control on the dual electrodes.

[0014] The brain slices are positioned on the worktable based on the workspace obtained through kinematic modeling. The coarse-stage aims to manually adjust the electrode tips into the microscope's field of view before needle tip detection. Using a 4x objective lens with a large field of view provides ample space for adjusting the dual-electrode movement, thus reducing the probability of collisions. In the fine-stage, building upon the coarse-stage control that moves the dual needles to the target coordinate point while still remaining within the microscope's field of view after switching to a 40x objective lens, the FRCNN algorithm is used to identify the needle tip and obtain its coordinates. PID control is then used to gradually move the needle tip towards the target coordinate point. In the fine-stage, the FRCNN algorithm is used to identify the target cell and obtain its XY-plane coordinates. The FRCNN algorithm is then used for dual-needle positioning, and PID control is used to gradually move the needle tip to the target cell's coordinates in the XY-plane. During the automatic, slow descent of the dual electrodes along the Z-axis, a timer is used to detect changes in resistance. When the resistance change matches an empirical value, it is considered that the needle tip has contacted the target cell's cell membrane, and the dual needles are automatically stopped.

[0015] S4: The control process of coarse-fine-precision three stages of the double needles in step S3 is integrated into a program that interacts with the operator. The program is run to automatically control the movement of the double electrodes to the cell membrane.

[0016] The above steps were programmed using Qt Creator to automate the step-by-step movement of the dual electrodes to the target cell surface. The workspace obtained through kinematic modeling guided the placement of the brain slices. The program automatically controlled the movement of the dual electrodes and prompted the operator to switch objectives, ultimately moving the electrode tips directly above the target cell. During the automatic descent of the dual electrodes along the Z-axis, a timer was used to detect changes in resistance at high frequency. When the resistance change matched an empirical value, it was considered that the needle tips had contacted the cell membrane of the target cell, and the dual needles were automatically stopped.

[0017] Compared with existing technologies, one or more of the above technical solutions have the following beneficial effects:

[0018] 1. This invention uses a robotic arm to model and determine the workspace accessible to both electrodes, avoiding the tedious steps of repeated experiments in traditional operations. It utilizes deep learning to locate the needle tip, avoiding significant errors and inaccuracies inherent in human visual observation, and improving the ease of acquiring and controlling the pixel coordinates of the needle tip. Furthermore, this invention uses Qt Creator to program the coarse-fine-fine control of the electrodes, automating the step-by-step movement of the two electrodes to the target cell surface. This avoids the repetitiveness and roughness of manual operation and reduces the possibility of collisions between the two needles in confined spaces.

[0019] 2. This invention achieves a success rate of over 90% with an average operation time of about 6 minutes, which shortens the operation time by about 70% and increases the success rate by about 50% compared to manual dual-electrode patch clamp operation. Attached Figure Description

[0020] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0021] Figure 1 This is a flowchart of the present invention;

[0022] Figure 2 This is a schematic diagram of the dual robotic arm modeling and dual electrode workspace of the present invention;

[0023] Figure 3 This is a flowchart illustrating the three-stage control process of moving the dual electrodes in the coarse-fine-fine stage according to the present invention.

[0024] Figure 4 This is a schematic diagram of the automated control process of the two electrodes contacting the target cells according to the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0026] Example 1:

[0027] An automated dual-electrode patch clamp operation method, the process of which is as follows: Figure 1 As shown, the specific operation is as follows:

[0028] (1) Construction of a dual-electrode patch clamp system.

[0029] The mouse brain slices used in this embodiment were obtained from freshly collected mouse brains in the laboratory. The mouse brains were precisely cut using a cutting instrument to obtain several brain slices of appropriate thickness. The brain slices were placed in a solution containing 140 mmol / L sodium chloride, 4 mmol / L potassium chloride, 2 mmol / L magnesium chloride, 2 mmol / L calcium chloride, 15 mmol / L glucose, and 10 mmol / L HEPES to maintain their viability. Before the experiment, 20 μL of electrode internal solution containing 140 mmol / L potassium chloride, 10 mmol / L sodium chloride, 10 mmol / L HEPES, and 5 mmol / L was injected into the microtube electrodes. In the patch-clamp system, a CCD camera acquired cell images at 60 fps, providing visual feedback throughout the operation. Two robotic arms controlled the movement of dual microtube electrodes with an inner diameter of 0.86 mm and an outer diameter of 1.5 mm, respectively. The electrical signals detected by the microtube electrodes were amplified, converted into digital signals, and finally transmitted to the host computer. The host computer is used for microscopic image processing, electrical signal acquisition, suction pressure control, and motion control of the microscope and manipulators.

[0030] (2) The robotic arm modeling based on patch clamp environmental parameters determines the dual-electrode workspace, such as... Figure 2 As shown.

[0031] In this example, the left arm has a travel range of 2cm × 2cm × 2cm in the XYZ directions, with an execution end of 12.5cm; the right arm has a travel range of 2.5cm × 2.5cm × 2.5cm in the XYZ directions, with an execution end of 17cm. To avoid collisions between the two needles and the objective lens, the left arm rotation joint is generally fixed at -pi / 6 degrees, and the right arm rotation joint is fixed at pi / 6 degrees. The origin of the left arm base is 6cm from the worktable plane, and the origin of the right arm base is 13.5cm from the worktable plane. The projections of the origins of the two arm bases onto the XY plane are 32cm apart. The diameter of the circular hole in the worktable where the culture dish is placed is 5cm, and the origin of the left arm base is 12.5cm from the center of the circular hole on the Y-axis. A world coordinate system is established with the point translating 2.5cm from the center of the circular hole in the positive X-axis direction as the origin. Based on these data, a model is created in MATLAB software as follows: Figure 2 As shown in (a), the reachable workspace of the robotic arm's end effector, the tip of the microtube electrode, is solved using a Monte Carlo random probability model. The position of the needle tip is treated as a random number, and the collaborative workspace is approximated by a region composed of a sufficient number of random numbers. In this process, the forward kinematics solution set of the robotic arm is solved using the fkine function based on the range of motion of each joint. The obtained reachable space of the two arms and the space where the experimental brain slices can be placed are shown in (a). Figure 2 (b) and Figure 2As shown in (c), the overlapping space not only demonstrates the feasibility of dual-electrode contact with target cells, but also suggests that brain slices should be placed on the left side of the culture dish during the experiment.

[0032] (3) Needle tip localization is performed using the FRCNN algorithm.

[0033] In this embodiment, the FRCNN algorithm is used to detect the electrode tip. Tip localization is achieved by calculating the center pixel coordinates of the rectangular bounding box containing the tip. Tip images are acquired and calibrated as a training set to train the FRCNN algorithm. Once the model is built, it can be applied to tip localization in a dual-electrode patch clamp. During the experiment, images are first acquired on the Qt Creator and sent to the algorithm. The images are first processed by a pre-trained convolutional network to extract features and output a feature map. Then, the feature extraction network searches for rectangular candidate regions that may contain the tip. Pooling layers and fully connected layers of the convolutional network are used to filter and adjust the bounding boxes with the highest probability. Finally, the center coordinates of the region containing the object are sent to the Qt Creator. This example demonstrates that this method can accurately obtain the pixel position of the tip in an image. With the microscope objective lens remaining stationary in the XY plane, this tip localization method can obtain the position of both needles in real time and precisely control their movement to any target position.

[0034] (4) Three-stage control of dual-electrode movement, such as: coarse-fine-fine three-stage control, for example Figure 3 As shown.

[0035] In this embodiment, the dual electrodes are controlled to approach the target cells in a three-stage process: coarse, fine, and fine. Before performing needle tip detection, the electrode tips need to be manually adjusted into the microscope's field of view, such as... Figure 3 As shown in Step 1. To allow for a relatively large space for the dual electrodes to move and reduce the probability of collisions, a 4x objective lens is used in the coarse adjustment stage. The fine adjustment stage completes the focusing of the dual needles under a 40x objective lens. The FRCNN algorithm is used to identify the needle tip and obtain its coordinates. PID control is used to gradually move the needle tip to the vicinity of the target coordinate point (1110, 500). This target point is an empirical value, ensuring that the needle tip remains in the field of view after switching to the 40x objective lens. Since the 40x objective lens requires imaging in solution, the objective lens and dual needles are lowered by 2800µm and 3700µm, respectively, before switching to the 40x objective lens and adding solution to connect the objective lens and solution. Focusing of the dual needles is achieved by automatically adjusting the movement of the objective lens in the Z-axis direction, as shown in the results. Figure 3 As shown in step 3. During the refining stage, the objective lens is slowly lowered at a rate of 7µm per step until it focuses on the brain slice. The FRCNN algorithm is used to identify the target cells and obtain their coordinates on the XY plane. The objective lens is then slowly raised to focus on the needle tip. Dual-needle positioning is performed using the FRCNN algorithm, and PID control is used to gradually move the needle tip to the coordinates of the target cell on the XY plane, as shown in step 3. Figure 3As shown in step 7, during the slow descent of the dual electrodes along the Z-axis under automatic control, the change in resistance is detected at high frequency using a timer. When the change in resistance matches the empirical value, it is considered that the needle tip has contacted the cell membrane of the target cell, and the dual needles are automatically stopped.

[0036] (5) Experiments based on automated control of dual-electrode contact with target cells, such as Figure 4 As shown.

[0037] In this embodiment, after preparing brain slices and placing them on the left side of the workbench's operable space, the pre-written automated program is run. The program execution interface is as follows: Figure 4 As shown in (a). By clicking the "Initialize" button, all serial COM ports are connected, ensuring the data acquisition card is connected and performing a self-test. Clicking the second button will pop up a message box prompting the operator to manually focus the dual needles under a 4x scope, as shown. Figure 4 As shown in (b). After clicking buttons three and four to achieve needle tip positioning, PID control moves it to the target coordinates (1110, 500), as follows. Figure 4 (c) shows the process of instructing the operator to switch to a 40x objective lens and add solution, connecting the objective lens to the liquid surface. Then, the dual-needle focusing is achieved by automatically adjusting the up-and-down movement of the objective lens. Figure 4 As shown in (d). Clicking buttons five through nine achieves target cell localization and dual-needle localization. First, control the dual needles to move them directly above the target cell, as shown in... Figure 4 As shown in (e), the dual electrodes then slowly descend along the Z-axis, using a timer to detect changes in resistance at high frequency. When the resistance change matches an empirical value, it is considered that the needle tip has contacted the cell membrane of the target cell, and the movement is stopped. The results are as follows: Figure 4 As shown in (f), the automated two-electrode patch-clamp operation had an average operation time of 6 minutes and a success rate of 90% (9 / 10) in contacting the target cell. The traditional two-electrode patch-clamp operation had an average operation time of 20 minutes, and the success rate was approximately 60% due to the tendency for the two electrodes to collide. The experimental results demonstrate that the automated control method for contacting the cell membrane with the two-electrode patch-clamp based on kinematic modeling and deep learning can avoid electrode collisions, thereby improving the success rate and shortening the operation time.

[0038] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An automated two-electrode membrane patch clamp operation method, characterized by, The method comprises the following steps: S1: calculating the working space of the double electrode tip through membrane patch clamp double mechanical arm motion modeling; S2: obtaining the pixel point coordinates of the double needle in the field of view through the two-dimensional positioning of the needle tip by a deep learning algorithm; S3: according to the double electrode working space obtained in step S1 and the needle tip positioning method in step S2, using Qt Creator tool to control the movement of the double electrode in three stages of coarse-fine-precision; S4: moving the double electrode patch clamp and contacting the target cell surface through an automatic program.

2. The automated dual electrode patch clamp procedure of claim 1, wherein: In the step S1, a world coordinate system is established on the workbench by measuring the membrane patch clamp system environment parameters, the micro mechanical arm for controlling the double electrode is modeled, the reachable working space of the double-arm operating end is calculated by using the forward kinematics method, the intersection of the double-arm common working space and the selectable placement space of the brain slice on the workbench is calculated, and finally the placement space of the brain slice and the working space of the double electrode tip are determined.

3. The automated dual electrode patch clamp procedure of claim 2, wherein: In the step S1, the modeling of the mechanical arm is determined by the DH parameters describing the motion relationship between the connecting rods, the reachable working space of the mechanical arm is solved by the Monte Carlo random probability model, the position of the electrode tip is regarded as a random number, and the region composed of a sufficient number of random numbers is used to approximate the working space. The random number is determined by using the solution set of the random combination of the connecting rod offset distance in the DH parameters of each joint solved by the forward kinematics.

4. The automated dual electrode patch clamp procedure of claim 1, wherein: In the step S2, the two-dimensional positioning of the needle tip is performed by collecting microscopic images when the needle tip is located on the objective focusing plane, and the horizontal position of the needle tip is described by the center point coordinates of the rectangular region frame obtained by the needle tip contour detection.

5. The automated dual electrode patch clamp procedure of claim 1, wherein: In the step S3, during the coarse-fine-precision three-stage movement control of the double electrode, the double needle movement is realized in steps, including manually controlling the double needle to move to the 4x mirror focal plane, automatically controlling the double needle to move to the 40x mirror focal plane, and automatically moving to the top of the target cell and descending to contact the cell.

6. The automated dual electrode membrane chip clamp operation method of claim 5, wherein: In the fine stage, the real-time needle tip positioning and PID controller are used to control the electrode to move to the target position; in the fine stage, the target cell is positioned by the needle tip positioning method, and the double needle is controlled to move to the top of the target cell and contact the cell by using the method used in the fine stage.

7. The automated dual electrode patch clamp procedure of claim 1, wherein: In the step S4, the Qt Creator tool is used to write the coarse-fine-precision three-stage movement control process into a program, and the visual interface generated by the program is clicked to automatically control the movement of the double electrode to the target cell surface and contact the cell in steps.

Images