DNA damaged basic group detection method, system and terminal based on molecular conductance

Abstract

The invention relates to the technical field of DNA adduct detection, in particular to a DNA damaged basic group detection method and system based on molecular conductance and a terminal.The detection method comprises the steps that an electrode gap between two metal electrodes modified with recognition probes is controlled, and the recognition probes and a DNA adduct form a monomolecular junction; applying a constant bias voltage to the metal electrode to obtain a current signal of the unimolecular junction; and inputting the current signal into a trained DNA damaged basic group detection model, and outputting a detection result of the DNA adduct. A DNA adduct and recognition probes on two metal electrodes are used for molecular recognition through hydrogen bond pairing to form a single-molecule junction of a sandwich structure, and electrons penetrate through the single-molecule junction to generate tunneling current signals under driving of constant bias voltage. And obtaining a detection result of the DNA adduct corresponding to the current signal based on the trained DNA damaged basic group detection model. According to the detection method, higher spatial resolution can be realized, and identification of DNA damaged bases with similar structures is facilitated.

Description

Technical Field

[0001] This invention relates to the field of DNA adduct detection technology, and in particular to a method, system and terminal for detecting DNA damaged bases based on molecular conductivity. Background Technology

[0002] Exogenous factors such as ionizing radiation, ultraviolet radiation, and mutagenic chemicals can all cause varying degrees of damage to DNA, disrupting its structural stability, affecting its accurate expression, and thus exerting a certain degree of toxicity on cells, increasing the risk of gene mutations, particularly the risk of diseases such as cancer. Currently, DNA adducts, as sensitive biomarkers, are important indicators of great interest in cancer diagnosis, precision medicine, toxicity assessment, and epidemiological surveys. However, due to the wide variety, similar structures, and extremely low concentrations of DNA adducts, accurate quantitative analysis of adducts is extremely challenging and has become a bottleneck limiting breakthroughs in related research fields.

[0003] Traditional methods for detecting DNA adducts mainly include 32 Post-labeling methods, high-performance liquid chromatography-mass spectrometry (HPLC-MS), immunoassay, and electrochemical sensing methods are widely used primarily due to their detection sensitivity and accuracy. However, these methods also have some drawbacks, such as: although... 32 Post-labeling (P-labeling) methods offer extremely high sensitivity, but the need for radioactive isotope labeling poses potential health risks to researchers, and the results do not provide structural information about the base adducts. Mass spectrometry relies on expensive instruments, resulting in high operating costs, complex operation, and long analysis cycles. Immunoassays lack versatility and are susceptible to false positives due to similar DNA adduct structures. Electrochemical sensors, based on the electrochemical signals of bases, lack sensitivity to changes in adduct structure, leading to poor selectivity in this type of technique.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method, system and terminal for detecting DNA damage bases based on molecular conductivity, which aims to solve the problems of poor selectivity for adducts and inability to provide structural information of base adducts in existing DNA damage base detection methods.

[0006] The technical solution of the present invention is as follows: A method for detecting DNA damaged bases based on molecular conductivity, comprising the following steps: The electrode gap between two metal electrodes modified with recognition probes is controlled, and the recognition probes form monomolecular knots with DNA adducts; A constant bias voltage is applied to the metal electrode to obtain the current signal of the monomolecular junction; The current signal is input into the trained DNA damage base detection model, and the detection results of DNA adducts are output.

[0007] The aforementioned method for detecting DNA damage bases based on molecular conductivity, wherein the recognition probe comprises at least one of thiolated benzamide, imidazole-formamide, and imidazole-formamide derivatives.

[0008] The aforementioned method for detecting DNA damage bases based on molecular conductivity, wherein the electrode gap is 1nm-5nm.

[0009] The method for detecting DNA damage bases based on molecular conductivity, wherein the constant bias voltage is -0.2V to -0.5V.

[0010] The aforementioned method for detecting DNA damage bases based on molecular conductivity, wherein the training of the trained DNA damage base detection model includes the following steps: The electrode gap between two metal electrodes modified with recognition probes is controlled, and the recognition probes form different monomolecular knots with different DNA adducts; A constant bias voltage is applied to the metal electrode to obtain the current signal of the monomolecular junction. After preprocessing, a signal cluster is obtained. The signal clusters are input into a machine learning model for training to obtain a DNA damage base detection model.

[0011] The aforementioned method for detecting DNA damaged bases based on molecular conductivity, wherein the preprocessing includes the following steps: The current signal is converted into a digital signal, and the digital signal is then denoised using the Wiener filter function. The noise-reduced digital signal is convolved using a Gaussian window with a preset data point width, and the consecutive data points higher than the preset data point after convolution are identified as a signal cluster. The time-domain signal of the signal cluster is extracted and converted into a frequency-domain signal and a cepstral-domain signal through Fourier transform and 1 / f noise removal. The frequency-domain signal and the cepstral-domain signal are downsampled to several signal intervals respectively to obtain the frequency-domain signal features and the cepstral-domain signal features. Based on the frequency domain signal features and the cepstral domain signal features, the mutual information algorithm is used to screen out the core features that can distinguish different bases or adducts, and these features are used as single base signal clusters or single adduct signal clusters.

[0012] The aforementioned method for detecting DNA damage bases based on molecular conductivity includes, wherein the time-domain signal comprises, signal cluster time-domain width, number of peaks in the signal cluster, maximum peak height of the signal cluster, minimum peak height of the signal cluster, average peak width, average peak height, maximum current value of the signal cluster, and minimum current value of the signal cluster.

[0013] A DNA damage base detection system based on molecular conductivity, comprising: The reaction module is used to control the electrode gap between two metal electrodes modified with recognition probes and to enable the recognition probes to form monomolecular junctions with DNA adducts; The data acquisition module is used to apply a constant bias voltage to the metal electrode and acquire the current signal of the monomolecular junction. The result output module is used to input the current signal into the trained DNA damage base detection model and output the detection results of DNA adducts.

[0014] A terminal includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the molecular conductivity-based DNA damage base detection method.

[0015] A computer-readable storage medium having a computer program stored thereon, the computer-readable storage medium storing a molecular conductivity-based DNA damage base detection program, the molecular conductivity-based DNA damage base detection program, when executed by a processor, implements the steps of the molecular conductivity-based DNA damage base detection method.

[0016] Beneficial Effects: This invention provides a method, system, and terminal for detecting DNA damage bases based on molecular conductivity. The method includes the following steps: controlling the electrode gap between two metal electrodes modified with recognition probes, and forming a monomolecular junction between the recognition probes and DNA adducts; applying a constant bias voltage to the metal electrodes to acquire the current signal of the monomolecular junction; inputting the current signal into a trained DNA damage base detection model, and outputting the detection result of the DNA adduct. This invention utilizes the molecular recognition of DNA adducts and recognition probes on two metal electrodes through hydrogen bonding to form a monomolecular junction with a "recognition probe-DNA adduct-recognition probe" sandwich structure. Under the drive of a constant bias voltage, electrons pass through the monomolecular junction to generate a tunneling current signal. Finally, based on the trained DNA damage base detection model, the detection result of the DNA adduct corresponding to the current signal can be obtained. This detection method uses recognition tunneling to detect DNA adducts. Recognition tunneling has a more stable structure than traditional tunneling, which can achieve higher spatial resolution and is beneficial for identifying DNA damage bases with similar structures. Even at extremely low concentrations of DNA damage bases, significant changes in conductivity can be detected. Meanwhile, this detection method can simultaneously detect multiple DNA adducts, making it a sensitive, rapid, label-free, and low-cost method for detecting DNA damage. Attached Figure Description

[0017] Figure 1 This is a schematic flowchart of a DNA damage base detection method based on molecular conductivity according to the present invention. Figure 2 This is a schematic diagram illustrating the principle of single-molecule junction recognition and tunneling detection. Figure 3 Characterization of gold needle tips obtained by electrochemical etching; Figure 4 Characterization diagram of gold needle tips coated with polyethylene material; Figure 5 Here is the molecular structure diagram of ICA; Figure 6 XPS spectra of the modified electrode and the blank electrode; Figure 7 Infrared spectrum of the modified electrode; Figure 8 This is a graph of the It signal collected in Example 1; Figure 9 Acquire time-domain / frequency-domain / ceptorical-domain transformed spectra for the "signal cluster"; Figure 10 For machine learning 10-fold cross-validation plot; Figure 11 A summary chart of the fitted curves for peak width and peak height; Figure 12This is a summary diagram of the two-dimensional density distribution of peak width and peak height. Figure 13 The optimistic accuracy plot (left) and the prediction accuracy plot (right) obtained after machine learning for multidimensional features. Detailed Implementation

[0018] This invention provides a method, system, and terminal for detecting DNA damage bases based on molecular conductivity. To make the objectives, technical solutions, and effects of this invention clearer and more explicit, the invention is further described in detail below. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] Existing methods for detecting DNA damage bases have poor selectivity for adducts and cannot provide structural information about the base adducts. Given the wide variety and similar structures of DNA adducts, there is an urgent need to develop an analytical method that is highly sensitive, low-cost, and versatile, so as to achieve the simultaneous detection of multiple DNA adducts.

[0020] Based on this, such as Figure 1 As shown, this invention provides a method for detecting DNA damaged bases based on molecular conductivity, comprising the following steps: Step S10: Control the electrode gap between the two metal electrodes modified with recognition probes, and make the recognition probes form monomolecular knots with the DNA adducts; Step S20: Apply a constant bias voltage to the metal electrode and obtain the current signal of the monomolecular junction; Step S30: Input the current signal into the trained DNA damage base detection model and output the detection results of DNA adducts.

[0021] In this embodiment, DNA adducts and recognition probes on two metal electrodes pair via hydrogen bonds for molecular recognition, forming a monomolecular junction with a "recognition probe-DNA adduct-recognition probe" sandwich structure. Under a constant bias voltage, electrons pass through the monomolecular junction, generating a tunneling current signal. Finally, based on a trained DNA damage base detection model, the detection result of the DNA adduct corresponding to the current signal can be obtained. This detection method uses recognition tunneling to detect DNA adducts. Recognition tunneling has a more stable structure than traditional tunneling, achieving higher spatial resolution and facilitating the identification of structurally similar DNA damage bases. Even at extremely low concentrations of DNA damage bases, significant changes in conductivity can be detected. Furthermore, this detection method can simultaneously detect multiple DNA adducts, making it a sensitive, rapid, label-free, and low-cost method for DNA damage detection.

[0022] Specifically, driven by a detection bias voltage, electrons pass through a single-molecule junction, generating a tunneling current signal, which is a typical telegraph noise (TN) signal. Typically, under a certain bias voltage, the peak current intensity and peak width characteristics of the TN signal are related to the HOMO-LUMO band gap of the analyte molecule, and therefore can be used to identify different modified bases. Simultaneously, the frequency domain characteristics of the TN signal are related to the probability of random molecular identification, thus reflecting the concentration information of the analyte. Furthermore, for DNA adducts with similar molecular structures, there may be cases where identification cannot be achieved solely through peak height and peak width characteristics. Therefore, based on current signal characteristics and employing machine learning algorithms, fine structure identification is achieved, resulting in a detection method with good selectivity and the ability to provide structural information about the base adducts. Moreover, by introducing machine learning into the precise differentiation of DNA damaged bases using electrical instruments, it offers advantages such as ease of device fabrication and high-throughput detection. This advantage makes it possible to simultaneously detect extremely low concentrations of DNA damaged bases, providing a crucial analytical technique for fields such as medical diagnosis, drug efficacy evaluation, and genotoxicity screening.

[0023] In some embodiments, the metal electrode includes a tip electrode and a base electrode; both the tip electrode and the base electrode are modified with identification probes.

[0024] In some embodiments, the material of the needle tip electrode is selected from, but is not limited to, gold, platinum, palladium, and silver; the preparation of the needle tip electrode includes the following steps: using a metal wire as the electrode material, the metal wire is etched in an etching solution using a DC or AC power supply, controlling the voltage and current, to prepare a metal needle tip electrode with considerable sharpness; subsequently, the tip of the needle tip electrode is coated with a coating material to ensure that the leakage current of the coated tip in the constructed nanosystem is less than 1 pA. Preferably, the length of the metal wire is 2-4 cm.

[0025] In some embodiments, the material of the substrate electrode is selected from, but not limited to, gold, platinum, palladium, and silver; the preparation of the substrate electrode includes the steps of: depositing an adhesion layer and a metal layer on the surface of a silicon wafer by deposition technology to prepare a substrate electrode with good conductivity; subsequently, cleaning the electrode surface with a solvent and calcining the substrate with a hydrogen flame to remove surface impurities and contaminants.

[0026] In some embodiments, the recognition probe includes, but is not limited to, at least one of thiolated benzamide, imidazole-formamide, and imidazole-formamide derivatives. One end of the recognition probe can self-assemble with a metal electrode to form a monolayer, and the other end can bind to DNA adduct bases via hydrogen bonding to form stable molecular pairings.

[0027] Preferably, the recognition probe is 4(5)-(2-mercaptoethyl)-1H-imidazolium-2-carboxamide. This molecule consists of an imidazolium ring and a rotatable amide group. The N1-H atom on the imidazolium ring and the two N7-H atoms on the amide group can act as hydrogen bond donors, while the N3 atom on the imidazolium ring and the O8 atom on the amide group can act as hydrogen bond acceptors. These hydrogen bond donors and acceptors can form stable hydrogen bond pairings with DNA bases, thereby achieving the capture and recognition of nucleotides. Furthermore, to facilitate the modification of the recognition probe and electrodes, the hydrogen atom on the C5 of the imidazolium ring is replaced with a mercaptoethyl group. The length of the connecting carbon chain can be adjusted as needed, and the connected mercapto group can self-assemble with the electrodes (gold needle tip electrode and gold silicon substrate electrode) to form a monolayer.

[0028] In this embodiment, the present invention utilizes molecular recognition tunneling to detect DNA adducts, and performs electrical detection by constructing single-molecule junctions. Detection equipment includes, but is not limited to, scanning tunneling microscopy with fractured junctions, mechanically controllable fractured junctions, conductive atomic force microscopy, and nanopore detection. The main construct is a single-molecule junction nanosystem for DNA adduct detection, where the DNA adduct and recognition probe are stably bound by hydrogen bonds, exhibiting both recognition stability and versatility.

[0029] Taking gold needle tip electrode and gold silicon wafer electrode as examples, the schematic diagram of single-molecule recognition tunneling detection is as follows: Figure 2 As shown, to achieve stable and efficient single-molecule recognition tunneling, different types of probes are synthesized and gold-sulfur self-assembly is used to covalently immobilize probe molecules on the surfaces of gold silicon wafer electrodes and gold needle tips. Subsequently, DNA adduct bases will randomly bind to the recognition probes via hydrogen bonds, forming a stable sandwich structure of "probe molecule-analyte-probe molecule". When a certain bias voltage is applied, when electrons pass through the single-molecule junction, a TN signal similar to telegraph noise will be generated on the It curve of the detection device, thus obtaining the current signal of molecular recognition tunneling.

[0030] In some embodiments, the DNA-damaging bases include, but are not limited to, at least one of dA, dT, dC, dG, dU, dI, N6-dMeA, 8-OHdG, O6-CMedG, 5-hmcdC, 5-mcdC, and 5-CldC.

[0031] In some embodiments, the preparation of the metal electrode modified with the recognition probe includes: using the metal electrode as a fixation carrier, fixing the recognition probe onto the surface of the metal electrode using a fixation technique to form a hydrogen bond recognition interface; after fixation, cleaning the surface of the probe with a suitable solvent to remove unbound components, followed by drying, ultimately obtaining a modified electrode that can be used for hydrogen bond binding of DNA adducts. Preferably, the fixation technique includes, but is not limited to, physical adsorption, covalent bonding, ligand coupling, self-assembled monolayer, electrochemical reduction, nanomaterial-assisted fixation, and molecular imprinting techniques.

[0032] In some implementations, the binding mode of the recognition probe to the DNA damage bases includes, but is not limited to, hydrogen bonding; the use of electrodes includes, but is not limited to, gold chip electrodes, palladium chip electrodes; and the cleaning of the electrodes includes, but is not limited to, the use of solvents such as anhydrous ethanol and isopropanol.

[0033] In some embodiments, the electrode gap is 1 nm to 5 nm. A nanometer-distance electrode gap is constructed by controlling a constant tunneling current value at the pA level between two metal electrodes modified with recognition probes; when the DNA adduct binds to the recognition probes on the two metal electrodes to form a monomolecular junction, electrons generate a current through the monomolecular junction due to the applied constant bias voltage.

[0034] In some embodiments, the constant bias voltage is -0.2V to -0.5V. Under the action of the constant bias voltage, the tip electrode is controlled to approach the base electrode until a tunneling current of 4pA is generated, after which the tip electrode stops, forming a gap of 1nm-5nm between the tip electrode and the base electrode. At this time, individual DNA adduct molecules in the solution phase will randomly bind to the recognition probe incubated on the electrode to form a monolayer, thereby generating a tunneling current. Preferably, the constant bias voltage is -0.2V.

[0035] Specifically, the detection method provided by this invention is a high time resolution single-molecule conductivity detection method with a signal at the pA level. When a bias voltage is applied, when the DNA adduct binds to the probe via hydrogen bonds, electrons pass through the single-molecule junction of the sandwich structure to generate a typical TN signal. The signal characteristics are closely related to the structure of the analyte molecule. Therefore, the structure of the base adduct can be identified based on the TN signal characteristics, and quantitative analysis can be achieved accordingly.

[0036] In some embodiments, step S30, training the trained DNA damage base detection model, includes the following steps: Step S31: Control the electrode gap between two metal electrodes modified with recognition probes, and make the recognition probes form different monomolecular knots with different DNA adducts; Step S32: Apply a constant bias voltage to the metal electrode, acquire the current signal of the monomolecular junction, and obtain a signal cluster after preprocessing; Step S33: Input the signal cluster into the machine learning model for training to obtain the DNA damage base detection model.

[0037] In this embodiment, due to the presence of thermal motion, the modified bases within the molecular junction are in a metastable state during a single recognition event. Therefore, the TN signal is usually realized in the form of a "signal cluster." A "signal cluster" is considered to be a collection of multiple tunneling signals that occur during a single recognition event due to the generation of a single molecular structure. The "signal cluster" provides richer feature tags. By combining it with machine learning algorithms, the "signal cluster" features in different DNA adduct signals are matched, classified, and counted to achieve synchronous quantitative analysis, thereby meeting the detection needs of actual samples.

[0038] Specifically, this invention combines machine learning to establish a molecular recognition tunneling mechanism. The detection method based on this mechanism can be used for all DNA adduct molecules. Since it is a single-molecule level binding, this detection method has significant advantages in resolving molecular structures and can achieve simultaneous detection of multiple structurally similar molecules, thus having good detection versatility.

[0039] In some embodiments, step S32, the preprocessing, includes the following steps: Step S321: Convert the current signal into a digital signal and use the Wiener filter function to perform noise reduction processing on the digital signal; Step S322: Convolve the denoised digital signal using a Gaussian window with a preset data point width, and identify the consecutive data points higher than the preset data point after convolution as a signal cluster; Step S323: Extract the time-domain signal of the signal cluster and convert it into a frequency-domain signal and a cepstral domain signal after Fourier transform and 1 / f noise removal. Then, downsample the frequency-domain signal and the cepstral domain signal to several signal intervals to obtain the frequency-domain signal features and the cepstral domain signal features. Step S324: Based on the frequency domain signal features and the cepstral domain signal features, use the mutual information algorithm to screen out the core features that can distinguish different bases or adducts, and use them as single base signal clusters or single adduct signal clusters.

[0040] In some implementations, a LabVIEW-written program controls a Data Acquisition Card (DAQ) to read the electrical signals generated by the formation of monolayers and convert them into digital signals to achieve data acquisition. Specifically, different DNA adduct molecules are added to a system with two metal electrodes modified with recognition probes. The It signal changes after different DNA adduct molecules form monolayers with the probes under a constant bias voltage can be obtained. Then, a Support Vector Machine (SVM) algorithm is used to train the raw current data to achieve classification and recognition of DNA adducts.

[0041] Specifically, firstly, the `tdms` file, acquired and saved by the LabVIEW-controlled data acquisition card, is read using the `nptdms` function package. The It signal data is converted into a NumPy array, and then Wiener filtering is used to remove background noise. Next, a Gaussian window with an appropriate data point width (e.g., 4096 acquisition points) is used to convolve the denoised digital signal. A "signal cluster" is defined as a sequence of data points continuously above a suitable value (e.g., 0.1). The time-domain signal within this cluster is then extracted. Subsequently, the time-domain signal of the cluster is transformed using a Fourier transform and subjected to 1 / f noise removal to convert it into a frequency domain and cepstral domain signal. The signal is downsampled to 51 / 61 signal intervals to obtain the frequency domain and cepstral domain signal features of the cluster. Frequency domain features such as power spectral density and downsampled cepstral domain features are extracted. Feature matching, classification, and counting are performed on different DNA adducts. Based on the extracted features, a feature selection method is used to screen out effective features. Finally, a support vector machine algorithm is used for supervised learning and predictive model training on different DNA adducts. Finally, the trained model is used for detection on the test set and for "signal clusters" in independent experiments to obtain optimistic accuracy and prediction accuracy results for different DNA adducts, thereby providing a basis for accurate identification of DNA adducts. Preferably, the feature selection method includes one of mutual information screening, variance thresholding, univariate selection, and random forest.

[0042] In some embodiments, the time-domain signal includes the time-domain width of the signal cluster, the number of peaks in the signal cluster, the maximum peak height of the signal cluster, the minimum peak height of the signal cluster, the average peak width, the average peak height, the maximum current value of the signal cluster, and the minimum current value of the signal cluster. The peak height, peak width, and frequency characteristics of the TN signal are extracted. Since the peak current intensity and peak width characteristics of the TN signal are related to the HOMO-LUMO band gap of the analyte molecule and the frequency is related to the probability of molecule recognition, it can be used to identify different modified bases and reflect their concentration information.

[0043] In addition, the present invention also provides a DNA damage base detection system based on molecular conductivity, comprising: The reaction module is used to control the electrode gap between two metal electrodes modified with recognition probes and to enable the recognition probes to form monomolecular junctions with DNA adducts; The data acquisition module is used to apply a constant bias voltage to the metal electrode and acquire the current signal of the monomolecular junction. The result output module is used to input the current signal into the trained DNA damage base detection model and output the detection results of DNA adducts.

[0044] In this embodiment, the system has the advantage of miniaturization, and the high-throughput tunneling detection chip fabricated using semiconductor technology significantly reduces detection costs. Furthermore, the method for immobilizing DNA damaged bases in this detection system is unaffected by the type and result of the DNA damaged bases or other environmental factors, achieving stable binding of DNA damaged bases with strong reproducibility and enabling label-free detection of DNA damaged bases.

[0045] In addition, the present invention also provides a terminal, the terminal including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the DNA damage base detection method based on molecular conductivity.

[0046] In some embodiments, the memory may be an internal storage unit of the terminal, such as a hard drive or RAM. In other embodiments, the memory may be an external storage device of the terminal, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Further, the memory may include both internal and external storage units. The memory is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory stores a molecular conductivity-based DNA damage base detection program, which can be executed by a processor to realize the molecular conductivity-based DNA damage base detection method of the present invention.

[0047] In some embodiments, the processor may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory or process data, such as executing the DNA damage base detection method based on molecular conductivity.

[0048] Finally, the present invention also provides a computer-readable storage medium, characterized in that it stores a computer program thereon, the computer-readable storage medium storing a DNA damage base detection program based on molecular conductivity, the DNA damage base detection program based on molecular conductivity implementing the steps of the DNA damage base detection method based on molecular conductivity when executed by a processor.

[0049] The following examples further illustrate the present invention in detail. It should also be understood that the following examples are only for further explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-essential improvements and adjustments made by those skilled in the art based on the above description of the present invention are within the scope of protection of the present invention.

[0050] Example 1 In this embodiment, a 0.25mm diameter gold wire was selected as the material for the needle electrode. The gold wire, with a purity of 99.99%, was obtained directly through purchase. During fabrication, the gold wire was first cut to approximately 3cm. Next, a DC power supply was used to electrochemically etch the gold wire. The output voltage was 3.6V. The gold wire was connected to the positive terminal of the DC power supply and fixed in a suitable fixture. The negative terminal of the electrode was connected to a platinum wire ring, which was in contact with and parallel to the surface of the etching solution. The gold wire was perpendicular to the etching solution and pointed towards the center of the platinum wire ring. The etching solution was a 1:1 volume mixture of anhydrous ethanol and concentrated hydrochloric acid. During etching, the gold wire was moved up and down to stabilize the current at a certain reading. When the current returned to zero, the etching was complete, and the gold wire was removed from the etching solution. Figure 3 As shown, a gold needle tip was obtained through electrochemical etching.

[0051] Because the nanosystem used in the test was conducted in solution, the collected signal was at the pA level. Therefore, in order to obtain a gold tip electrode with a leakage current of less than 1 pA, it was chosen to coat the electrochemically etched gold tip with purchased polyethylene microspheres. First, the polyethylene microspheres were placed on a soldering iron at a temperature of 219°C, and after a certain time, they became molten and transparent. At this point, a self-written tip control program was used to control the tip to pierce the polyethylene microsphere at a set speed of approximately 0.1 mm / s and a coating length of approximately 9 mm. Finally, a gold tip electrode coated with polyethylene material but with the tip exposed was obtained (e.g., ...). Figure 4(As shown). Before use, the gold needle tip electrode will undergo electrical performance testing in a scanning tunneling microscope (STM) solution phase (1 mM phosphate buffer solution, pH=7.4), that is, the leakage current must be less than 1 pA under a bias voltage of -0.5V.

[0052] Taking a gold-silicon wafer substrate electrode as an example, a 10nm thick titanium coating was deposited on the silicon wafer surface using an electron beam evaporator (Lesker PVD 75) as an adhesion layer, followed by a 200nm thick gold layer. Before experimental use, the gold-silicon wafer was first immersed in anhydrous ethanol and sonicated for 3-5 minutes, then dried using nitrogen gas. Finally, the surface of the gold-silicon wafer was ignited with a hydrogen flame to ensure the substrate electrode surface was dry and clean, thus obtaining the gold-silicon wafer substrate electrode.

[0053] Taking the 4(5)-(2-mercaptoethyl)-1H-imidazole-2-carboxamide (ICA) molecular probe as an example, its molecular structure is as follows: Figure 5 As shown, the molecule consists of an imidazole ring and a rotatable amide group. The N1-H atom on the imidazole ring and the two N7-H atoms on the amide group can act as hydrogen bond donors, while the N3 atom on the imidazole ring and the O8 atom on the amide group can act as hydrogen bond acceptors. These hydrogen bond donors and acceptors can form stable hydrogen bond pairings with DNA bases, thereby achieving the capture and recognition of nucleotides. Furthermore, to facilitate the modification of the probe molecule and electrodes, the hydrogen atom at C5 of the imidazole ring is replaced with a mercaptoethyl group, where the length of the connecting carbon chain can be adjusted as needed. The connected mercapto group can then self-assemble with the electrodes (gold needle tip electrode and gold silicon substrate electrode) to form a monolayer.

[0054] Taking the incubation of ICA probe molecules as an example, covalent modification was performed on the gold needle tip electrode and the gold silicon wafer substrate electrode. 0.5 mg of MICA powder was dissolved in 2 mL of anhydrous ethanol, and the gold needle tip and gold silicon wafer were immersed in the MICA solution, respectively. After incubation for a period of time, the gold needle tip and gold silicon wafer were removed, and excess solution was gently rinsed off with anhydrous ethanol. The electrode surfaces were then gently dried with nitrogen gas for subsequent use.

[0055] Qualitative and quantitative analyses of the probe molecules modified on the electrode were performed using various characterization and analysis methods, such as... Figure 6As shown, X-ray photoelectron spectroscopy (XPS) was used to analyze the surface composition of the substrate electrode before and after incubation with ICA molecules. The results showed the presence of Au-S bonds on the surface of the gold chip after incubation, indicating that the probe molecules were successfully covalently modified onto the substrate electrode surface; specifically, the XPS spectrum obtained from the electrode surface showed a peak of 163.32 eV belonging to the S 2p region of the Au-S bond. Furthermore, the peaks at and around 399.34 eV in the N 1s region can be attributed to amide nitrogen, confirming the successful immobilization of the ICA probe molecules. Additionally, Figure 7 The Fourier Transform Infrared Spectroscopy (FTIR) spectrum shows that the characteristic peaks of the infrared curve after electrode modification are basically consistent with those of the probe molecule powder infrared curve, further proving the successful modification of the probe molecule; that is, the characteristic peaks of the infrared curve after electrode modification of the probe molecule are basically consistent with those of the infrared curve of its powder, and both show C=O bonds (1654 cm⁻¹). -1 CH bond (1434 cm) -1 ) and CN bond (1117 cm -1 The characteristic peaks indicated that the probe molecules were successfully modified onto the gold electrode surface. XPS and infrared characterization demonstrated that the probe molecules could be firmly modified onto the electrode through covalent self-assembly, thereby providing sites for the recognition of hydrogen bonds of DNA adduct bases, avoiding the generation of stray signals, and constructing a stable single-molecule junction detection system.

[0056] In this embodiment, data collection is achieved by applying a certain bias voltage to a gold-silicon wafer in a scanning tunneling microscope (STM) within a nanometer-spaced system, and then using a DAQ card to acquire and record the It signal generated by the formation of a single molecular junction.

[0057] Specifically, the STM was preset to a current of 4 pA. Under an applied bias of -0.2 V, the STM controlled the gold needle tip to approach the gold-silicon wafer at a speed of 1 µm / s until a tunneling current of 4 pA was generated. The gold needle tip then stopped, creating a gap of several nanometers between it and the gold-silicon wafer. At this point, individual DNA adduct molecules in the solution phase randomly bind to probe molecules incubated on the electrode to form monojunctions, thereby generating the tunneling current. Next, a program written in LabVIEW was used to control the DAQ card to read the generated electrical signal and convert it into a digital signal. The program collected 90 seconds of It signals at a sampling frequency of 50 kHz. During the experiment, the obtained current signal should be higher than the noise level (4 pA). At least 1000 It signal peaks with significantly higher than the noise level needed to be collected for "signal cluster" extraction and feature extraction in machine learning.

[0058] Furthermore, in this embodiment, ICA molecules are used as modified base recognition probe molecules to capture DNA adducts (2'-deoxyadenosine (dA), 2'-deoxythymidine (dT), 2'-deoxycytidine (dC), 2'-deoxyguanosine (dG), 2'-deoxyuridine (dU), 2'-deoxyinosine (dI), N6-methyl-2'-deoxyadenosine (N6-dMeA), 8-hydroxy-2'-deoxyguanosine (8-OHdG), 2'-deoxy-6-O-methyl-guanosine (O6-CMedG), 2'-deoxy-5-(hydroxymethyl)cytidine (5-hmcdC), 5-methyl-2'-deoxycytidine (5-mcdC), 5-chloro-2'-deoxycytidine (5-CldC)) in a 1 mM phosphate buffered solution (PB) solution phase system. When a bias voltage of -0.2V is applied, the bases of the DNA adduct are linked to the probe molecule via hydrogen bonds, generating a current signal. This current signal is amplified by the STM amplifier and then converted from analog to digital. The digital signal is then acquired (e.g., ...). Figure 8 (As shown) The signal is stored as a binary .tdms file using LabVIEW. A Python data processing program can recognize the .tdms file to obtain the original current signal graph. A Gaussian window with a width of 4096 sampling points is used to convolve the Wiener-filtered and denoised It curve. Data points continuously higher than 0.1 after convolution are identified as a "signal cluster". Time-domain signal features, including "signal cluster time-domain width", "number of peaks in the signal cluster", and "maximum / minimum current value of the signal cluster", are extracted. Subsequently, the time-domain signal of the "signal cluster" is converted into a frequency domain signal and a cepstral domain signal after Fourier transform and 1 / f noise removal. The signal is downsampled to 51 / 61 signal intervals respectively to obtain the frequency domain and cepstral domain signal features of the signal cluster, such as... Figure 9 As shown.

[0059] Subsequently, the "signal clusters" containing each feature are saved in sequence. A mutual information algorithm is used to select a subset of core features capable of distinguishing different base / adduct "signal clusters," and these are then used for support vector machine (SVM) machine learning. First, the "signal clusters" of each base / adduct are randomly divided into training and test sets in an 8:2 ratio. The selected features are used to train the "signal clusters" in the training set. Simultaneously, to ensure the robustness of the machine learning model, ten-fold cross-validation is used to verify the accuracy of the training set data (e.g., ...). Figure 10 (As shown).

[0060] Specifically, the dataset is divided into ten parts and reused cyclically. Each dataset is divided into ten mutually exclusive subsets of similar size. The union of nine subsets is used as the training set each time, and the remaining subset is used as the validation set. This process is repeated ten times, and the results from all ten iterations are combined to evaluate the model. Figure 8 As shown, the results of the ten-fold cross-validation demonstrate the good robustness of the machine learning model in this embodiment, with the accuracy of each subset exceeding 90%.

[0061] The trained prediction model is used to classify "signal clusters" on the test set. The predicted labels are compared with their true labels to obtain the model's optimistic accuracy. Finally, the trained model is used to predict and classify "signal clusters" extracted from independent .tdms files that are not used for machine learning. The prediction accuracy is then compared with the true labels to obtain the model's prediction accuracy.

[0062] After summarizing the signal extraction results (including peak height and peak width) of the 12 nucleotides obtained in this embodiment at the same concentration, the LogNormal function was used to fit them, and the resulting fitted curves were summarized to obtain... Figure 11 Based on one-dimensional differentiation, the peak value and width are considered simultaneously to generate a two-dimensional probability density distribution map, such as... Figure 12 As shown. Furthermore, considering the two-dimensional characteristics of peak width and peak height simultaneously to distinguish DNA bases from DNA adduct bases, a density distribution map with log(peak width) on the horizontal axis and log(peak height) on the vertical axis was plotted. This revealed that, to a certain extent, it is possible to distinguish natural nucleosides from their structurally similar adducts, such as... Figure 12 As shown, dG and 8-OHdG, dA and N6-dMeA are distinguished to some extent.

[0063] Furthermore, due to thermal motion, TN signals will exist in the form of "signal clusters." These signal clusters can provide more molecular feature information. Combined with machine learning algorithms, the SVM algorithm is used to match and classify the signal features of different DNA adducts. After training, predictions can be made, yielding optimistic accuracy (left) and prediction accuracy (right). Figure 13 As shown, the optimistic accuracy rate mostly remained above 85%, and the prediction accuracy also performed well. The accuracy rate can demonstrate that this technology can, to a certain extent, accurately identify and distinguish different DNA adduct molecules.

[0064] In summary, this invention provides a method, system, and terminal for detecting DNA damage bases based on molecular conductivity. The method includes the following steps: controlling the electrode gap between two metal electrodes modified with recognition probes, and forming a monomolecular junction between the recognition probes and DNA adducts; applying a constant bias voltage to the metal electrodes to acquire the current signal of the monomolecular junction; inputting the current signal into a trained DNA damage base detection model, and outputting the detection result of the DNA adduct. This invention utilizes the hydrogen bonding between DNA adducts and recognition probes on two metal electrodes to achieve molecular recognition, forming a monomolecular junction with a "recognition probe-DNA adduct-recognition probe" sandwich structure. Under a constant bias voltage, electrons pass through the monomolecular junction, generating a tunneling current signal. Finally, based on the trained DNA damage base detection model, the detection result of the DNA adduct corresponding to the current signal can be obtained. This detection method uses recognition tunneling to detect DNA adducts. Recognition tunneling has a more stable structure than traditional tunneling, achieving higher spatial resolution and facilitating the identification of structurally similar DNA damage bases. Even at extremely low concentrations of DNA damage bases, significant conductivity changes can be detected. Meanwhile, this detection method can simultaneously detect multiple DNA adducts, making it a sensitive, rapid, label-free, and low-cost method for detecting DNA damage.

[0065] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for detecting DNA damage bases based on molecular conductivity, characterized in that, Including the following steps: The electrode gap between two metal electrodes modified with recognition probes is controlled, and the recognition probes form monomolecular knots with DNA adducts; A constant bias voltage is applied to the metal electrode to obtain the current signal of the monomolecular junction; The current signal is input into the trained DNA damage base detection model, and the detection results of DNA adducts are output.

2. The method for detecting DNA damage bases based on molecular conductivity according to claim 1, characterized in that, The identification probe includes at least one of thiolated benzamide, imidazole-formamide, and imidazole-formamide derivatives.

3. The method for detecting DNA damage bases based on molecular conductivity according to claim 1, characterized in that, The electrode gap is 1nm-5nm.

4. The method for detecting DNA damage bases based on molecular conductivity according to claim 1, characterized in that, The constant bias voltage is -0.2V to -0.5V.

5. The method for detecting DNA damage bases based on molecular conductivity according to claim 1, characterized in that, The training of the trained DNA damage base detection model includes the following steps: The electrode gap between two metal electrodes modified with recognition probes is controlled, and the recognition probes form different monomolecular knots with different DNA adducts; A constant bias voltage is applied to the metal electrode to obtain the current signal of the monomolecular junction. After preprocessing, a signal cluster is obtained. The signal clusters are input into a machine learning model for training to obtain a DNA damage base detection model.

6. The method for detecting DNA damage bases based on molecular conductivity according to claim 5, characterized in that, The preprocessing includes the following steps: The current signal is converted into a digital signal, and the digital signal is then denoised using the Wiener filter function. The noise-reduced digital signal is convolved using a Gaussian window with a preset data point width, and the consecutive data points higher than the preset data point after convolution are identified as a signal cluster. The time-domain signal of the signal cluster is extracted and converted into a frequency-domain signal and a cepstral-domain signal through Fourier transform and 1 / f noise removal. The frequency-domain signal and the cepstral-domain signal are downsampled to several signal intervals respectively to obtain the frequency-domain signal features and the cepstral-domain signal features. Based on the frequency domain signal features and the cepstral domain signal features, the mutual information algorithm is used to screen out the core features that can distinguish different bases or adducts, and these features are used as single base signal clusters or single adduct signal clusters.

7. The method for detecting DNA damage bases based on molecular conductivity according to claim 6, characterized in that, The time-domain signal includes the time-domain width of the signal cluster, the number of peaks in the signal cluster, the maximum peak height of the signal cluster, the minimum peak height of the signal cluster, the average peak width, the average peak height, the maximum current value of the signal cluster, and the minimum current value of the signal cluster.

8. A DNA damage base detection system based on molecular conductivity, characterized in that, include: The reaction module is used to control the electrode gap between two metal electrodes modified with recognition probes and to enable the recognition probes to form monomolecular junctions with DNA adducts; The data acquisition module is used to apply a constant bias voltage to the metal electrode and acquire the current signal of the monomolecular junction. The result output module is used to input the current signal into the trained DNA damage base detection model and output the detection results of DNA adducts.

9. A terminal, characterized in that, The terminal includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the DNA damage base detection method based on molecular conductivity as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program thereon, the computer-readable storage medium storing a DNA damage base detection program based on molecular conductivity, the DNA damage base detection program based on molecular conductivity being executed by a processor to implement the steps of the DNA damage base detection method based on molecular conductivity as described in any one of claims 1-7.

Images