Method for detecting size of single molecule in solution

Abstract

The invention specifically discloses a method for detecting the size of a single molecule in a solution, and relates to the technical field of nano material detection. The method comprises the following specific steps: S1, capturing video data sets of known monomolecular particles with different standard sizes by adopting an optical tweezers technology, and dividing the video data sets into a training set and a test set according to a certain proportion; s2, constructing a neural network model based on deep learning, carrying out supervised training on the neural network model by adopting a training set, and enabling the model to establish a mapping relation between motion track characteristics and particle sizes by optimizing network parameters; s3, inputting the test set into the preliminarily trained neural network model, and checking whether the size information given by the neural network model is consistent with the known size information or not; and S4, inputting a motion trail video of a to-be-tested single molecule particle into the trained neural network model, and outputting size information of the particle. The method overcomes the difficulty that the size of a single molecule in a solution cannot be measured in the prior art by virtue of a deep learning technology.

Description

Technical Field

[0001] The present invention relates to the technical field of nano material detection, in particular to a method for detecting the size of a single molecule in a solution. Background Art

[0002] In the frontier areas of modern science and technology where development is booming, upconversion nanoparticles (UCNPs) have shined in the fields of sensing and biological applications with their excellent and unique optical properties, becoming a valuable key probe. When irradiated with near-infrared light, UCNPs can stimulate stable and bright visible luminescence. This excellent property makes it occupy an irreplaceable position in many key links of biological experiments. In terms of biomolecule labeling, UCNPs can accurately label target biomolecules, providing clear instructions for tracking the activity trajectory of biomolecules and exploring their functional mechanisms; in the field of cell imaging, with the help of the luminescence properties of UCNPs, researchers can observe the microstructure and physiological processes of cells more clearly and accurately, helping to deepen the study of cell biology.

[0003] However, UCNPs have a problem that is difficult to ignore: in an aqueous solution environment, UCNPs will dissolve, and this dissolution effect is more prominent when the solution concentration is low. In technical application scenarios such as optical tweezers, which require extremely high precision in nanoparticle size measurement, the accuracy of the nanoparticle size is directly related to the reliability and accuracy of the measurement results. In the experimental scenario of single nanoparticles, the precise measurement of UCNPs size becomes an extremely challenging but crucial task. The measurement results play a decisive role in in-depth research on the performance of UCNPs and optimization of related technical applications.

[0004] Traditional nanoparticle size measurement techniques, represented by transmission electron microscopy (TEM), can provide high-resolution images when examining solid-phase samples, enabling accurate measurement of nanoparticle size. However, they suffer from a significant limitation: they cannot directly measure samples in solution. This limitation makes TEM difficult to meet the demand for real-time, in-situ measurement of nanoparticle size in solution in biological experiments.

[0005] The emergence of optical tweezers technology has brought new detection methods to biological experiments. It uses a highly focused laser beam to form a gradient potential field in the focal area. The gradient force generated by this potential field can achieve three-dimensional capture of tiny particles, causing the captured objects to exhibit confined Brownian motion characteristics within a limited space. With its unique advantages of being contactless and non-destructive, optical tweezers technology has been widely used in biological experiments. However, when using optical tweezers to capture upconversion nanoparticles, the effect of the laser will accelerate the dissolution of UCNPs, causing their size to change, making it impossible to directly use optical tweezers to accurately measure the size of UCNPs. Summary of the Invention

[0006] The purpose of this invention is to propose a method for detecting the size of single molecules in solution. By using deep learning technology, videos of optical tweezers capturing single-molecule particles of different sizes are learned. Given a video of optical tweezers capturing particles of unknown size, the trained neural network is used to determine the size of the single-molecule particle, thereby overcoming the problem that existing technologies cannot measure the size of single molecules in solution.

[0007] To achieve the above object, the present invention proposes a method for detecting the size of a single molecule in a solution, the specific steps of which are as follows:

[0008] Step S1: using optical tweezers technology to capture a video dataset of known single-molecule particles of different standard sizes, and dividing it into a training set and a test set according to a certain ratio;

[0009] Step S2: constructing a neural network model based on deep learning, using a training set to perform supervised training on the neural network model, and optimizing network parameters so that the model establishes a mapping relationship between motion trajectory characteristics and particle size;

[0010] Step S3: Put the test set into the preliminarily trained neural network model to check whether the size information given by the neural network model is consistent with the known size information. If not, retrain the neural network model. If consistent, obtain the final neural network model.

[0011] Step S4: input the motion trajectory video data of the single-molecule particle to be measured into the finally trained neural network model, and output the size information of the particle.

[0012] Preferably, in step S1, the single-molecule particles are one or more of upconversion nanoparticles UCNPs, gold nanoparticles, quantum dots or polymer nanoparticles.

[0013] Preferably, in step S1, the optical tweezers technology uses a near-infrared laser with a power range of 5-50 mW.

[0014] Preferably, in step S2, the neural network model is a convolutional neural network CNN, and its network structure includes at least one 3D convolution layer for extracting spatiotemporal features and a fully connected output layer.

[0015] Preferably, in step S2, supervised training includes the following steps:

[0016] Step S21: After inputting the training data set into the neural network model, the mean square error (MSE) between the model-predicted size and the standard size is calculated;

[0017] Step S22: If the error exceeds a preset threshold, the network parameters are dynamically adjusted through the back propagation algorithm until the error converges to within the threshold range.

[0018] Preferably, the preset threshold is 3 nm, and the judgment criterion for error convergence is that the MSE change rate in three consecutive iterations is less than 1%.

[0019] Preferably, in step S4, the output size information is the equivalent spherical diameter of the particle.

[0020] Therefore, the present invention proposes a method for detecting the size of a single molecule in a solution, which has the following beneficial effects:

[0021] (1) Breaking through the limitations of the detection environment: Traditional detection technologies such as transmission electron microscopy (TEM) can only detect solid-phase samples. This invention breaks through this limitation and achieves accurate detection of the size of single-molecule particles in solution. In biological experiments, many biomolecules and nanomaterials are active and functional only in solution. This method makes it possible to directly measure the size of single molecules in solution, providing a powerful tool for studying the characteristics and behavior of biomolecules in their natural state, greatly broadening the application scenarios of detection technology.

[0022] (2) Achieving nanoscale detection accuracy: Successfully overcoming the optical diffraction limit problem, single-molecule particle size detection at the nanoscale can be achieved. This is of great significance to the research of nanomaterials. The ability to accurately measure the size of nanoparticles helps to gain a deeper understanding of the relationship between the structure and properties of nanomaterials, and promotes the innovative application of nanomaterials in electronics, medicine, energy and other fields. For example, in the development of nanomedicines, accurately understanding the size of nanoparticles can optimize the design of drug carriers and improve the targeting and efficacy of drugs.

[0023] (3) Easy and efficient operation: The operation process is extremely simple. One only needs to obtain a video of optical tweezers capturing a single molecule particle and input it into a trained neural network to quickly obtain size information. Compared with traditional detection methods, there is no need for complex sample preparation and expensive equipment, which greatly saves time and labor costs. In high-throughput detection scenarios, it can quickly process a large number of samples, improve detection efficiency, and meet the needs of modern scientific research and industrial production for fast and accurate detection.

[0024] (4) Strong compatibility and scalability: This method has good compatibility with the types of single-molecule particles. It is not only applicable to upconversion nanoparticles, but can also be applied to the size detection of other single-molecule particles such as gold particles. This makes the method widely applicable to different materials and fields, and can provide a universal detection method for multidisciplinary research. In addition, the architecture based on deep learning has strong scalability. With the continuous development and optimization of neural network algorithms, the accuracy and speed of detection can be further improved. It can also adapt to more complex detection needs and scenarios by increasing training data and adjusting the network structure.

[0025] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The figure is a schematic flow chart of a method for detecting the size of a single molecule in a solution according to the present invention. DETAILED DESCRIPTION

[0027] To make the technical solutions, advantages, and objectives of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below. The described embodiments are part of the embodiments of the present invention, not all of them. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Unless otherwise defined, technical or scientific terms used in the present invention shall have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs.

[0029] Example 1

[0030] like Figure 1 FIG. 1 is a flow chart of a method for detecting the size of a single molecule in a solution according to the present invention, and the specific steps are as follows:

[0031] S1. A video dataset of single-molecule particles of known different standard sizes captured using optical tweezers technology is divided into training and test sets according to a certain ratio.

[0032] The single-molecule particles are one or more of upconversion nanoparticles UCNPs, gold nanoparticles, quantum dots or polymer nanoparticles;

[0033] Optical tweezers technology uses near-infrared lasers with a power range of 5-50mW.

[0034] S2. Construct a neural network model based on deep learning, use the training set to supervise the neural network model, and optimize the network parameters to enable the model to establish a mapping relationship between motion trajectory characteristics and particle size;

[0035] The neural network model is a convolutional neural network (CNN), whose network structure includes at least one 3D convolutional layer for extracting spatiotemporal features and a fully connected output layer.

[0036] Supervised training consists of the following steps:

[0037] S21, after inputting the training data set into the neural network model, calculating the mean square error (MSE) between the model predicted size and the standard size;

[0038] S22. If the error exceeds a preset threshold, the network parameters are dynamically adjusted through the back propagation algorithm until the error converges to within the threshold range.

[0039] The preset threshold is 3 nm, and the error convergence criterion is that the MSE change rate is less than 1% in three consecutive iterations.

[0040] S3. Put the test set into the preliminarily trained neural network model to check whether the size information given by the neural network model is consistent with the known size information. If not, retrain the neural network model. If consistent, obtain the final neural network model.

[0041] S4. Input the motion trajectory video of the single-molecule particle to be measured into the finally trained neural network model, and output the equivalent spherical diameter of the particle.

[0042] Example 2

[0043] 1. Training process.

[0044] 1.1. Prepare a variety of single-molecule particle samples of known standard sizes, such as upconversion nanoparticles and gold particles of different sizes.

[0045] 1.2. Using an optical tweezers setup, capture videos of each standard-sized single-molecule particle under appropriate experimental conditions (e.g., specific laser power, solution environment, etc.). During video capture, ensure that the particle's trajectory under the tweezers is clearly recorded.

[0046] 1.3. The collected video data are sorted and labeled, and most of the data sets are selected as training sets according to a certain ratio (e.g. 80%).

[0047] 1.4. Select the CNN neural network model, input the training set data into the neural network, and set the initial network parameters, such as learning rate, number of iterations, etc.

[0048] 1.5. Train the neural network using training methods such as the backpropagation algorithm. During the training process, continuously adjust the network parameters so that the network can learn the relationship between the motion characteristics and size of single-molecule particles of different sizes captured by optical tweezers, thereby obtaining a preliminarily trained neural network.

[0049] 2. Testing process.

[0050] 2.1. From the video dataset of single-molecule particles of known standard size captured by optical tweezers, select the remaining part not used for training (e.g., 20%) as the test set.

[0051] 2.2. Input the test set data into the preliminarily trained neural network and let the neural network predict the size of the single-molecule particles in the test set.

[0052] 2.3. Compare the size information predicted by the neural network with the known true size information in the test set. If the error between the predicted and true sizes is within an acceptable range (the error threshold can be set according to specific experimental requirements), the initially trained neural network is considered qualified and the final trained neural network is obtained. If the error exceeds the acceptable range, the neural network parameters are readjusted and the training process is repeated until the test passes.

[0053] 3. Application process.

[0054] 3.1. In actual testing, when using optical tweezers to capture a video of a solution sample containing single-molecule particles of unknown size, the video quality must also be guaranteed to clearly show the movement of the particles.

[0055] 3.2. Input the captured video of single-molecule particles of unknown size into the final trained neural network.

[0056] 3.3. After analysis and calculation, the neural network outputs the size information of unknown single-molecule particles and completes the detection of single-molecule size in the solution.

[0057] Therefore, the present invention provides a method for detecting the size of single molecules in solution. By learning videos of optical tweezers capturing single-molecule particles of different sizes, when a video of optical tweezers capturing particles of unknown size is given, the trained neural network is used to give the size of the single-molecule particle, thereby overcoming the problem that existing technologies cannot measure the size of single molecules in solution.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting the size of a single molecule in a solution, characterized in that: Here are the steps: Step S1: using optical tweezers technology to capture a video dataset of known single-molecule particles of different standard sizes, and dividing it into a training set and a test set; Step S2: constructing a neural network model based on deep learning, using a training set to perform supervised training on the neural network model, and optimizing network parameters so that the model establishes a mapping relationship between motion trajectory characteristics and particle size; Step S3: Put the test set into the preliminarily trained neural network model to check whether the size information given by the neural network model is consistent with the known size information. If not, retrain the neural network model. If consistent, obtain the final neural network model. Step S4: input the motion trajectory video data of the single-molecule particle to be measured into the finally trained neural network model, and output the size information of the particle.

2. The method for detecting the size of a single molecule in a solution according to claim 1, wherein: In step S1, the single-molecule particles are one or more of upconversion nanoparticles UCNPs, gold nanoparticles, quantum dots or polymer nanoparticles.

3. The method for detecting the size of a single molecule in a solution according to claim 1, wherein: In step S1, the optical tweezers technology uses near-infrared laser with a power range of 5-50mW.

4. The method for detecting the size of a single molecule in a solution according to claim 1, wherein: In step S2, the neural network model is a convolutional neural network (CNN), whose network structure includes at least one 3D convolutional layer for extracting spatiotemporal features and a fully connected output layer.

5. The method for detecting the size of a single molecule in a solution according to claim 1, wherein: In step S2, supervised training includes the following steps: Step S21: After inputting the training data set into the neural network model, the mean square error (MSE) between the model-predicted size and the standard size is calculated; Step S22: If the error exceeds a preset threshold, the network parameters are dynamically adjusted through the back propagation algorithm until the error converges to within the threshold range.

6. The method for detecting the size of a single molecule in a solution according to claim 5, characterized in that: The preset threshold is 3 nm, and the judgment standard for error convergence is that the MSE change rate in three consecutive iterations is less than 1%.

7. The method for detecting the size of a single molecule in a solution according to claim 1, wherein: In step S4, the output size information is the equivalent spherical diameter of the particle.

Images