Early tumor detection method and system based on DNA (Deoxyribose Nucleic Acid) characteristics

Abstract

The invention discloses an early tumor detection method and system based on DNA characteristics, and relates to the technical field of molecular diagnos.By setting standardized time sequence sample collection nodes and synchronously collecting plasma samples at multiple time points, it is ensured that the dynamic process of DNA steady-state change can be captured, and in the aspect of sample treatment, the detection accuracy is improved. The method comprises the following steps: extracting cfDNA by adopting a paramagnetic particle method automatic extraction technology, setting a strict quality control standard, ensuring the quality of a detection sample, constructing an integrated cross-interference-free mixed reaction system, and synchronously capturing oxidative damage markers and methylation modification in the cfDNA by using a specific probe, so that synchronous detection of a two-dimensional molecular marker is realized; accurate two-dimensional time sequence signal data are obtained through a two-channel fluorescence synchronous detection and internal reference correction technology, a normal physiological dynamic reference model is constructed, and based on the model, accurate judgment of the early risk of the tumor is achieved by calculating a time sequence characteristic deviation value.

Description

Technical Field

[0001] This invention relates to the field of molecular diagnostics technology, specifically to a method and system for early detection of tumors based on DNA characteristics. Background Technology

[0002] As a major disease that seriously threatens human health, early diagnosis and treatment of cancer are of great significance for improving patient survival rates and prognosis. In recent years, with the continuous advancement of molecular biology technology, molecular diagnostics has shown great potential in the field of early cancer screening. Among them, cell-free DNA in plasma, as circulating nucleic acid released into the blood by tumor cells, contains a wealth of tumor-related information. In the early stages of tumor development, genomic DNA suffers oxidative damage, leading to the production of oxidative damage markers such as 8-hydroxydeoxyguanosine. These markers can directly reflect the imbalance of genomic oxidative stress. At the same time, methylation modification of CpG islands in tumor-related genes is also a key epigenetic event in the process of tumorigenesis, which can regulate the expression of oncogenes and tumor suppressor genes and play an important role in the initiation and progression of tumors.

[0003] However, existing liquid biopsy technologies for tumors have significant limitations. Most detection protocols focus only on single-dimensional molecular markers, either detecting only oxidative damage or only methylation modification, failing to achieve simultaneous capture and joint analysis of dual-dimensional molecular markers. This single-dimensional detection approach results in an incomplete characterization of DNA homeostasis abnormalities, easily missing important early tumor information. Furthermore, current technologies generally employ a single-time-point static detection mode, without establishing a normal dynamic reference benchmark within the physiological cycle. Since molecular levels in the human body fluctuate due to various physiological factors, this single-time-point detection cannot distinguish between physiological molecular fluctuations and pathological evolution characteristics, leading to insufficient sensitivity in identifying early-stage borderline tumor lesions and a high likelihood of false negatives and false positives. In addition, traditional methods require separate, step-by-step detection for dual-dimensional markers, resulting in high sample consumption, cumbersome procedures, increased time costs, and operational difficulty. Moreover, the lack of dedicated analysis algorithms adapted to time-series dynamic data further limits the accuracy and reliability of the detection results, making it difficult to meet the clinical needs for precise early tumor screening. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method and system for early tumor detection based on DNA characteristics. This system can simultaneously collect plasma samples from multiple time points by setting standardized time-series sample collection nodes, ensuring the capture of the dynamic process of DNA homeostasis changes. In terms of sample processing, it employs automated magnetic bead extraction technology to extract cfDNA and sets strict quality control standards to guarantee the quality of the test samples. An integrated, cross-interference-free mixed reaction system is constructed, and specific probes are used to simultaneously capture oxidative damage markers and methylation modifications in cfDNA, achieving simultaneous detection of dual-dimensional molecular markers. Through dual-channel fluorescence synchronous detection and internal reference calibration technology, accurate dual-dimensional time-series signal data is obtained, and a normal physiological dynamic benchmark model is constructed. Based on this model, by calculating the time-series characteristic deviation value, accurate determination of early tumor risk is achieved.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On one hand, a method for early tumor detection based on DNA characteristics, comprising the following specific steps: S1: Collect peripheral venous blood from the subject according to the preset physiological cycle time nodes, separate and obtain plasma samples, and extract free cfDNA from the plasma; S2: Simultaneous hybridization capture of cfDNA was performed using oxidative damage-specific probes and methylation-specific probes. The oxidative damage-specific probes targeted and bound to the 8-hydroxydeoxyguanosine modification site in cfDNA, while the methylation-specific probes targeted and bound to the CpG island methylation site of tumor-associated genes. S3: Quantitatively detect the captured cfDNA and obtain the oxidative damage abundance and methylation level signals of cfDNA at each time point to form a two-dimensional time-series feature dataset; S4: Based on the dual-dimensional time-series detection data of healthy people during their physiological cycle, establish a benchmark model of dynamic changes in DNA oxidative damage and methylation under normal physiological conditions, and determine the time-series feature deviation threshold. S5: Compare the subject's two-dimensional time-series feature dataset with the benchmark model and calculate the time-series feature deviation value; if the deviation value exceeds the preset threshold, it is determined that the subject has abnormal DNA homeostasis, indicating the risk of early tumor development.

[0006] Furthermore, in S1, four consecutive time-series detection nodes, T0, T1, T2, and T3, are set based on the subject's physiological rhythm. The interval between each node is uniformly 48 hours. 5 mL of peripheral venous blood is collected from the subject using EDTA-K2 anticoagulation vacuum blood collection tubes per node. Within 1 hour after blood collection, the sample is centrifuged at 4°C and 3000g for 15 minutes to complete plasma separation. The cell-free plasma layer is aspirated and transferred to an enzyme-free centrifuge tube, which is then frozen and stored at -80°C for later use.

[0007] Furthermore, in step S1, the specific steps for extracting cell-free cfDNA from plasma are as follows: Automated extraction of plasma cfDNA is performed using a magnetic bead method. 200 μL of thawed plasma sample is taken, and 400 μL of lysis buffer containing 20 mM Tris-HCl, 8 M guanidine salt, and 1% β-mercaptoethanol is added. The sample is vortexed for 30 seconds, then lysed at room temperature for 10 minutes. Subsequently, 20 μL of carboxyl-modified magnetic beads are added and thoroughly mixed. The sample is incubated at room temperature for 15 minutes to allow cfDNA to bind to the magnetic beads. After magnetic separation, the supernatant is discarded, and a solution containing 70% ethanol and 10 mM... Magnetic separation washing was performed using Tris-HCl washing buffer, and the washing was repeated twice to remove impurities. Finally, 50 μL of 10 mM Tris-HCl enzyme-free elution buffer (pH 8.0) was added, and the mixture was incubated at 65°C for 10 min to complete cfDNA elution. After magnetic separation and collection of the eluent, quality control was performed using a Qubit 4.0 real-time fluorescence spectrometer, agarose gel electrophoresis, and a Nanodrop detector. The requirements were that the cfDNA concentration was ≥1 ng / μL, the main band of the fragment was concentrated in 160-200 bp, and the A260 / A280 ratio was 1.8-2.0.

[0008] Furthermore, in S2, a nucleic acid aptamer probe labeled with a 5' end FAM fluorophore and targeting 8-hydroxydeoxyguanosine is selected as an oxidative damage-specific probe, while a recombinant methylation-binding domain protein fusion probe labeled with an N-terminal HEX fluorophore is selected as a methylation-specific probe. A reaction system with a total volume of 50 μL is prepared, containing 10 μL cfDNA template, 2 μL 10 μM oxidative damage probe, 2 μL 10 μM methylation probe, 25 μL 2× hybridization buffer containing 50 mM PBS, 1 M NaCl, and 0.1% Tween-20, and 11 μL enzyme-free water. After mixing, the system is denatured at 95°C for 5 min, quenched on ice for 2 min, and then hybridized at 42°C for 90 min. After hybridization, 10 μL streptavidin magnetic beads are added and incubated at room temperature for 10 min. The free probe is magnetically separated and washed three times with washing buffer, retaining the cfDNA complex bound to the dual-dimensional probe.

[0009] Furthermore, in S3, a dual-channel fluorescence synchronous detection and internal reference correction technique is adopted. The abundance of oxidative damage and the level of methylation are simultaneously acquired through the FAM channel and HEX channel of the real-time fluorescence quantitative PCR instrument. The β-actin housekeeping gene without oxidative damage and methylation is introduced as an internal reference. The fluorescence signal of the β-actin gene is used as a reference to correct the systematic error of the target detection signal. The corrected FAM fluorescence value and HEX fluorescence value of the four time nodes from T0 to T3 are recorded respectively to construct a two-dimensional time series data matrix containing oxidative damage and methylation level.

[0010] Furthermore, in step S4, healthy volunteers are selected, and a training dataset is constructed by acquiring dual-dimensional time-series data through the entire process from time-series sample collection to quantitative detection of time-series signals. The time-series data of healthy individuals is fitted using a dynamic time warping algorithm to generate the average dynamic change trajectory of DNA oxidative damage-methylation under normal physiological conditions. The time-series feature deviation threshold of 1.25 is determined through ROC curve analysis and used as a criterion for distinguishing between normal physiological fluctuations and pathological abnormalities.

[0011] Furthermore, in S4, the specific steps for fitting the time series data of healthy individuals using the dynamic time warping algorithm are as follows: Select a two-dimensional time series dataset of healthy volunteers, preprocess the oxidative damage abundance and methylation level data corresponding to each time series node for each volunteer to obtain standardized individual two-dimensional time series trajectories, use the dynamic time warping algorithm to select the mean of the time series trajectories of some healthy volunteers as the initial reference trajectory, dynamically align the individual time series trajectories of the remaining healthy volunteers with the initial reference trajectory, adjust the correspondence of each time series node by calculating the cumulative distance between each pair of trajectories, and then calculate the mean of oxidative damage abundance and methylation level at each time series node based on all aligned individual time series trajectories, repeat the trajectory alignment and reference trajectory update steps until the reference trajectory tends to stabilize, that is, the trajectory fluctuation after two adjacent updates reaches the preset stability standard, stop the iterative calculation, and finally obtain the reference trajectory as the average dynamic change trajectory of DNA oxidative damage-methylation under normal physiological conditions.

[0012] Furthermore, in step S5, the subject's two-dimensional temporal feature dataset is compared with the benchmark model to calculate the temporal feature deviation value, the calculation formula of which is: ,in, It is the temporal characteristic deviation value, which characterizes the overall degree of deviation between the subject's two-dimensional temporal trajectory and the baseline trajectory of healthy individuals; This is the set total number of time-series detection nodes. It is the sequence number of the time sequence node, used to sequentially refer to each independent sampling time sequence node; For the examinee at the Quantitative values ​​of DNA oxidative damage abundance after internal reference correction at each time point. It is the first in the average dynamic baseline trajectory of healthy people. The baseline value of DNA oxidative damage abundance corresponding to each time node. The examinee was on the Quantitative values ​​of DNA methylation levels after internal control correction at each time point. It is the first in the average dynamic baseline trajectory of healthy people. The baseline values ​​of DNA methylation levels corresponding to each time node are obtained by fitting a dataset of healthy individuals using a dynamic time warping algorithm.

[0013] Furthermore, in S5, It is the time series characteristic deviation value. A value ≤1.25 indicates normal DNA homeostasis and no risk of early-stage tumors; a value <1.25 indicates normal DNA homeostasis and no risk of early-stage tumors. A level ≤1.8 is considered a mild abnormality in DNA homeostasis, suggesting a risk of precancerous lesions, and a follow-up examination is recommended in 3 months. A score >1.8 indicates severe DNA homeostasis abnormalities, suggesting a high risk of early-stage tumors, and further imaging and pathological examinations are recommended.

[0014] On the other hand, a DNA-based early tumor detection system includes: Sample preprocessing module: used for peripheral blood plasma separation, cfDNA extraction and purification, and outputting standardized cfDNA samples; Dual-dimensional probe capture module: integrates reaction systems for oxidative damage-specific probes and methylation-specific probes to achieve simultaneous and specific capture of cfDNA dual-dimensional modification sites; Time-series data acquisition module: Connects to the fluorescence detection unit, acquires oxidative damage abundance and methylation level signals of cfDNA according to preset time nodes, and generates time-series feature dataset; Dynamic model analysis module: It has a built-in two-dimensional time-series benchmark model library for healthy people, and is equipped with a time-series deviation calculation unit to complete the comparison and analysis of the subject's data and the benchmark model; Results output module: Used to output DNA homeostasis abnormality determination results, time-series characteristic deviation curves, and early tumor risk grading reports.

[0015] Compared with existing technologies, this method and system for early tumor detection based on DNA characteristics has the following advantages: This invention employs an innovative technology design that simultaneously captures and dynamically tracks DNA oxidative damage and methylation modification from two dimensions, effectively overcoming the core bottleneck of single-dimensional static detection in existing tumor liquid biopsy techniques. Furthermore, the integrated, cross-interference-free probe reaction system can simultaneously capture 8-hydroxydeoxyguanosine oxidative damage markers and methylation modification characteristics of tumor-related genes in cell-free plasma DNA, avoiding the problem of missing DNA homeostasis abnormalities caused by single-dimensional detection. Based on time-series data from healthy individuals, a dynamic benchmark model is constructed, using a dynamic time warping algorithm to fit the dual-dimensional dynamic trajectory under normal physiological conditions and determining a scientific deviation threshold. This accurately distinguishes between physiological molecular fluctuations and pathological evolution characteristics, effectively reducing false positives and false negatives, and significantly improving the ability to identify early-stage borderline tumor lesions. Simultaneously, this invention uses a non-invasive peripheral blood sampling method, allowing for repeated time-series detection without invasive procedures, resulting in high patient acceptance. It is suitable for large-scale early tumor screening in healthy populations and dynamic monitoring of disease progression in high-risk groups, effectively solving the problem that existing technologies cannot meet the clinical needs of accurate early tumor screening.

[0016] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0018] Figure 1 This is a flowchart of a method for early tumor detection based on DNA characteristics; Figure 2 This is a structural block diagram of an early tumor detection system based on DNA characteristics. Figure 3 This is a flowchart of step S4 in a DNA-based early tumor detection method. Detailed Implementation

[0019] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0020] This invention provides a method and system for early tumor detection based on DNA characteristics. By setting standardized time-series sample collection nodes, plasma samples are collected simultaneously at multiple time points to ensure the capture of the dynamic process of DNA homeostasis changes. In terms of sample processing, automated extraction technology using magnetic beads is employed to extract cfDNA, and strict quality control standards are set to ensure the quality of the test samples. An integrated, cross-interference-free mixed reaction system is constructed, and specific probes are used to simultaneously capture oxidative damage markers and methylation modifications in cfDNA, achieving simultaneous detection of dual-dimensional molecular markers. Through dual-channel fluorescence synchronous detection and internal reference calibration technology, accurate dual-dimensional time-series signal data is obtained, and a normal physiological dynamic benchmark model is constructed. Based on this model, the time-series characteristic deviation value is calculated to achieve accurate determination of early tumor risk.

[0021] like Figure 1 As shown, S1: According to the preset physiological cycle time nodes, peripheral venous blood of the subject is collected, plasma samples are separated and free cfDNA is extracted from the plasma; Based on the subject's natural physiological rhythm, and avoiding interfering factors such as staying up late, strenuous exercise, and acute infection, four consecutive time-series testing nodes, T0, T1, T2, and T3, were set, with a uniform interval of 48 hours between each node to ensure the consistency of the subject's physiological state at each sampling point.

[0022] Peripheral venous blood was collected from subjects using EDTA-K2 anticoagulant vacuum blood collection tubes. The blood volume collected at each time point was strictly controlled to 5 mL. Hemolysis was avoided during the blood collection process. Sample processing was initiated within 1 hour after blood collection. The blood collection tubes were placed stably in the centrifuge, and the temperature was set to 4℃, the speed to 3000g, and the centrifugation time to 15 min. Violent shaking of the equipment was avoided during centrifugation. After centrifugation, the cell-free plasma layer was slowly aspirated using a disposable sterile pipette in a sterile laminar flow hood, avoiding the aspiration of blood cell layer. The plasma was transferred to enzyme-free centrifuge tubes labeled with time point information and immediately placed in an ultra-low temperature freezer at -80℃ for freezing. The tubes should be placed upright during storage to avoid sample leakage, and the freezing time should not exceed 3 months. Repeated freeze-thaw cycles were prohibited during this period.

[0023] Take out the frozen plasma sample from the -80°C refrigerator and place it in the 4°C refrigerator to thaw slowly for about 2 - 3 hours, avoiding rapid thawing at room temperature which may cause cfDNA degradation. After thawing, gently invert the sample 3 times to mix well. Then, pipette 200 μL of the thawed plasma sample into a nuclease-free reaction tube and add 400 μL of pre-cooled lysis buffer. The lysis buffer consists of 20 mM Tris-HCl, 8 M guanidine salt, and 1% β-mercaptoethanol, with the pH adjusted to 8.0. Vortex for 30 s at a frequency of 3000 r / min and let it stand at room temperature for 10 min. During this period, gently invert the reaction tube once every 2 min to ensure that the proteins in the plasma are fully denatured and cfDNA is completely released. Add 20 μL of carboxyl-modified magnetic beads with a particle size of 1 - 5 μm and a surface carboxyl density of ≥10 μmol / g. After thorough mixing by oscillation, incubate at room temperature for 15 min, and shake manually once every 5 min during the incubation to ensure the specific binding efficiency between the magnetic beads and cfDNA. Place the reaction tube on a magnetic stand and let it stand for 2 min until the magnetic beads are completely adsorbed, then slowly discard the supernatant. Add 600 μL of washing buffer containing 70% ethanol and 10 mM Tris-HCl, gently shake for 30 s, then place it on the magnetic stand again for adsorption for 2 min, and discard the supernatant. Repeat this washing step 2 times to thoroughly remove impurities such as residual proteins, salts, and anticoagulants. Add 50 μL of 10 mM Tris-HCl nuclease-free elution buffer with pH 8.0 to the magnetic beads, place the reaction tube in a 65°C constant temperature device and incubate for 10 min, gently invert it once every 3 min during this period. Immediately place it on the magnetic stand for adsorption for 3 min after incubation, and collect the supernatant, which is the extracted cfDNA solution. Subsequently, perform triple quality control: for concentration detection, the cfDNA concentration should be ≥1 ng / μL; for fragment length analysis, the main band should be concentrated at 160 - 200 bp, with no obvious degradation or contamination bands; for purity detection, the A260 / A280 ratio should be 1.8 - 2.0, with no protein or RNA contamination. Only when all three indicators meet the standards can it be considered qualified for quality control and then proceed to the subsequent experiment.

[0024] S2: Synchronously hybridize and capture cfDNA using oxidation damage-specific probes and methylation-specific probes. The oxidation damage-specific probes target and bind to the 8-hydroxydeoxyguanosine modification sites in cfDNA, and the methylation-specific probes target and bind to the CpG island methylation sites of tumor-associated genes. Nucleic acid aptamer probes targeting 8-hydroxydeoxyguanosine (8-OHdG) were selected. The probe sequences were specifically verified and showed no cross-binding with other DNA modification sites. The 5' end was labeled with a FAM fluorophore. The probes were purified by HPLC (purity ≥95%), and the concentration was adjusted to 10 μM. After aliquoting, they were stored at -20℃ protected from light, with a shelf life of 6 months. Recombinant methylation-binding domain (MBD) protein fusion probes were selected, with the N-terminus labeled with a HEX fluorophore (fluorescence quantum yield ≥0.85). These probes were also purified by HPLC (purity ≥95%), and the concentration was adjusted to 10 μM. Storage at -20℃ protected from light was avoided to prevent repeated freeze-thaw cycles that could lead to protein activity loss. Hybridization reaction systems with a total volume of 50 μL were constructed sequentially in enzyme-free reaction tubes. The precise proportions of each component were: 10 μL cfDNA template, 2 μL 10 μM oxidative damage-specific probe, 2 μL 10 μM methylation-specific probe, and 25 μL 2× hybridization buffer (containing 50 mM PBS and 1 M... After adding NaCl, 0.1% Tween-20 (pH 7.4), and 11 μL of enzyme-free water (RNase-free, DNase-free), gently pipette 10 times to mix thoroughly, avoiding the formation of air bubbles.

[0025] Place the reaction tube in a temperature-controlled device and denature at 95°C for 5 minutes to completely unwind the cfDNA double strands into single strands. Then, quickly transfer it to an ice bath for 2 minutes to maintain the single-stranded state and prevent renaturation. Avoid shaking the reaction tube during the ice bath. Transfer the reaction tube to a 42°C temperature-controlled device and hybridize at a constant temperature for 90 minutes. Keep the device sealed during hybridization to avoid temperature fluctuations and ensure that the probe binds specifically and fully to the cfDNA target site. After hybridization, add 10 μL of streptavidin magnetic beads (10 mg / mL, 2-4 μm in diameter) to the reaction tube and incubate at room temperature for 10 minutes, gently inverting the tube every 2 minutes to ensure that the magnetic beads bind specifically to the probe-cfDNA complex. Place the reaction tube on a magnetic rack for 2 minutes to adsorb the probe and discard any unbound free probe. Add 600 μL of washing buffer, gently shake for 30 seconds, then magnetically separate and discard the supernatant. Repeat the washing process 3 times, ensuring that the supernatant is completely discarded after each wash. Finally, retain the cfDNA complex bound to the double-stranded probe.

[0026] S3: Quantitatively detect the captured cfDNA and obtain the oxidative damage abundance and methylation level signals of cfDNA at each time point to form a two-dimensional time-series feature dataset; A dual-channel fluorescence synchronous detection technology is employed. The fluorescence detection unit must have independent detection capabilities for both FAM and HEX channels. The FAM channel has an excitation wavelength of 488 nm and an emission wavelength of 520 nm, while the HEX channel has an excitation wavelength of 535 nm and an emission wavelength of 556 nm. The detection sensitivity is ≤10 fM to ensure accurate capture of low-abundance signals. An oxidatively free, unmethylated β-actin housekeeping gene is introduced as an internal control. This gene is stably expressed in both normal and tumor cells, and its promoter region lacks CpG island methylation modification, making it less susceptible to oxidative damage and effectively eliminating the risk of contamination during experimental procedures or sample preparation. Systematic errors such as quality and equipment fluctuations are addressed through the following internal control calibration process: Simultaneous detection of the fluorescence signal of the β-actin gene in the subject's cfDNA (using a fluorescent label that does not cross-interfere with the target probe) is performed. The ratio of the target signal (FAM / HEX) to the internal control signal is calculated to obtain the calibrated quantitative values ​​of oxidative damage abundance and methylation level. Fluorescence detection of the cfDNA complex at each time point (T0, T1, T2, T3) is performed sequentially. Each sample has three parallel replicate wells. The average value after detection is taken as the original signal value for that time point. After internal control calibration, the final quantitative value is obtained. The calibrated FAM fluorescence values ​​(oxidative damage abundance) and HEX fluorescence values ​​(methylation level) of the four time points are organized into a two-dimensional time-series data matrix.

[0027] like Figure 3 As shown, S4: Based on the two-dimensional time-series detection data of healthy people during their physiological cycle, a benchmark model of dynamic changes in DNA oxidative damage-methylation under normal physiological conditions is established to determine the time-series feature deviation threshold. Training dataset construction: ≥1000 healthy volunteers were selected, and two-dimensional time-series data were collected according to the above-described process to form a training dataset for the healthy population. During the collection process, sampling specifications and experimental conditions consistent with those of the subjects were strictly followed to avoid dataset bias. Z-score standardization was performed on the two-dimensional time-series data of each healthy volunteer to eliminate baseline differences between individuals and obtain standardized individual two-dimensional time-series trajectories.

[0028] From the standardized time-series trajectories, 20% of healthy volunteer samples were randomly selected, and the mean values ​​of oxidative damage abundance and methylation level at each time-series node were calculated as the initial reference trajectory. The individual time-series trajectories of the remaining 80% of healthy volunteers were dynamically aligned with the initial reference trajectory. By calculating the Euclidean cumulative distance between each pair of trajectories, the correspondence between each time-series node was adjusted to ensure that the time-series trajectories of different individuals were synchronized in the time dimension. Based on the aligned individual time-series trajectories, the mean values ​​of oxidative damage abundance and methylation level at each time-series node were calculated, and the reference trajectory was updated. The trajectory alignment and update steps were repeated until the fluctuation of the reference trajectory after two adjacent updates reached the preset stability standard, and the iterative calculation was stopped. The final reference trajectory is the average dynamic change trajectory of DNA oxidative damage-methylation under normal physiological conditions.

[0029] The ROC curve analysis method was used, with the training dataset of healthy people as negative samples and the two-dimensional time series data of known early-stage tumor patients (positive samples) to draw ROC curves. The critical value corresponding to the largest area under the curve (AUC) was selected as the time series feature deviation threshold. After verification, the threshold was determined to be 1.25. The sensitivity of this threshold is ≥90% and the specificity is ≥85%, which can effectively distinguish between normal physiological fluctuations and pathological abnormalities.

[0030] S5: Compare the subject's two-dimensional time-series feature dataset with the benchmark model and calculate the time-series feature deviation value; if the deviation value exceeds the preset threshold, it is determined that the subject has abnormal DNA homeostasis, indicating the risk of early tumor development. The subject's two-dimensional time-series feature dataset is compared with the benchmark model, and the time-series feature deviation value is calculated using the following formula: ,in, It is the temporal characteristic deviation value, which characterizes the overall degree of deviation between the subject's two-dimensional temporal trajectory and the baseline trajectory of healthy individuals; This is the set total number of time-series detection nodes. It is the sequence number of the time sequence node, used to sequentially refer to each independent sampling time sequence node; For the examinee at the Quantitative values ​​of DNA oxidative damage abundance after internal reference correction at each time point. It is the first in the average dynamic baseline trajectory of healthy people. The baseline value of DNA oxidative damage abundance corresponding to each time node. The examinee was on the Quantitative values ​​of DNA methylation levels after internal control correction at each time point. It is the first in the average dynamic baseline trajectory of healthy people. The baseline values ​​of DNA methylation levels corresponding to each time node were obtained by fitting a dataset of healthy individuals using a dynamic time warping algorithm. Low risk: When If the value is ≤1.25, the examinee's DNA homeostasis is considered normal, with no risk of early tumor development. It is recommended to have a routine physical examination once a year and maintain a healthy lifestyle.

[0031] Medium risk: When 1.25 < If the S value is ≤1.8, the subject's DNA homeostasis is considered to be slightly abnormal, indicating a risk of precancerous lesions (such as polyps, dysplasia, etc.). It is recommended to have a follow-up examination after 3 months. During the follow-up examination, the sampling should be carried out at the same physiological cycle node as the first test. If S is still in this range after the follow-up examination, further targeted tumor marker testing is required.

[0032] High risk: When When the level is >1.8, the subject's DNA homeostasis is considered severely abnormal, indicating a high risk of early-stage tumors. It is recommended to undergo imaging examinations (such as CT, MRI, endoscopy, etc.) and pathological biopsy within one month to clarify the location and nature of the lesion, and to make a comprehensive diagnosis in combination with clinical symptoms.

[0033] like Figure 2 As shown, this invention provides an early tumor detection system based on DNA characteristics, the specific implementation of which is as follows: This system serves as a dedicated automated implementation platform for the aforementioned detection methods. Each module works collaboratively according to the workflow, possessing the core advantages of standardization, automation, and high precision. Its specific components and functions are as follows: Sample preprocessing module: Used for peripheral blood plasma separation, cfDNA extraction and purification, outputting standardized cfDNA samples; it includes a blood collection auxiliary unit, a centrifugation unit, a magnetic bead extraction unit, a quality control unit, and a sample transmission interface. The blood collection auxiliary unit integrates a time-series sampling reminder function, adapts to the standardized placement of anticoagulant blood collection tubes and sample labeling, ensuring precise and controllable sampling nodes; the centrifugation unit has a built-in precise temperature control and speed adjustment system, which can automatically set centrifugation parameters of 4℃, 3000g, and 15min, and automatically prompts after centrifugation to avoid human operation errors; the magnetic bead extraction unit automatically completes the quantitative addition and mixing of lysis buffer, magnetic beads, washing buffer, and elution buffer, strictly following the time and temperature parameters of incubation, magnetic separation, and washing steps, with an extraction efficiency of ≥90%; the quality control unit integrates concentration detection, fragment length analysis, and purity detection functions, automatically generates quality control reports, and issues a re-extraction prompt for unqualified samples and records the reason for the unqualified samples; the sample transmission interface and dual-dimensional probe capture module adopt a sterile sealed transmission channel to avoid sample contamination and maintain the sample low temperature during transmission.

[0034] The dual-dimensional probe capture module integrates oxidative damage-specific and methylation-specific probe reaction systems to achieve simultaneous and specific capture of cfDNA modification sites across two dimensions. It includes a probe storage unit, a reaction system preparation unit, a hybridization reaction unit, and a magnetic separation and washing unit. The probe storage unit has a built-in -20℃ low-temperature storage chamber to store both probes in the dark, supporting automatic thawing and quantitative dispensing to avoid repeated freeze-thaw cycles. The reaction system preparation unit uses a high-precision liquid workstation to automatically aspirate each reaction component according to a preset ratio, accurately preparing a 50μL hybridization reaction system with a preparation error ≤±0.5μL, and features bubble detection and removal functions. The hybridization reaction unit programmatically controls the temperature and time for denaturation (95℃, 5min), ice bath quenching (2min), and isothermal hybridization (42℃, 90min), with a temperature control accuracy ≤±0.5℃ to ensure hybridization specificity. The magnetic separation and washing unit automatically completes the addition of streptavidin magnetic beads, incubation (room temperature, 10min), magnetic separation, and three washing processes. After washing, the supernatant is automatically discarded, retaining a high-purity probe-cfDNA complex.

[0035] The time-series data acquisition module connects to the fluorescence detection unit and acquires signals of oxidative damage abundance and methylation level of cfDNA according to preset time nodes, generating a time-series feature dataset. This includes a fluorescence detection unit, a signal acquisition unit, an internal reference calibration unit, and a data storage unit. The fluorescence detection unit has dual-channel simultaneous detection capability (FAM and HEX), supports precise adjustment of excitation and emission wavelengths, has a detection sensitivity ≤10 fM, and can automatically acquire three parallel replicates for each sample and take the average. The signal acquisition unit automatically associates subject information according to the T0-T3 time nodes, acquires and stores raw fluorescence data, and supports real-time data upload and local backup. The internal reference calibration unit incorporates a β-actin housekeeping gene signal correction algorithm to automatically calculate the corrected signal value and eliminate system errors. The data storage unit uses encrypted storage, classifying and storing raw data, corrected data, and the time-series data matrix, and supports data export and traceability.

[0036] Dynamic Model Analysis Module: This module includes a built-in dual-dimensional time-series baseline model library for healthy individuals, configured with a time-series deviation calculation unit. It performs comparative analysis of subject data and the baseline model, comprising a baseline model library, a data retrieval unit, a deviation calculation unit, and a comparison analysis unit. The baseline model library contains a baseline trajectory of healthy individuals fitted using a dynamic time warping algorithm, supporting regular model updates (optimized annually based on newly added healthy population data). The data retrieval unit automatically retrieves subject data from the time-series data acquisition module and data from the baseline model library to ensure accurate data matching. The deviation calculation unit contains a built-in formula for calculating time-series feature deviation values, and performs batch data calculations through an efficient computing module. The comparison analysis unit automatically compares the deviation values ​​with preset thresholds (1.25, 1.8) to determine the DNA homeostasis.

[0037] The results output module outputs DNA homeostasis abnormality assessment results, time-series characteristic deviation curves, and early tumor risk grading reports. It includes a curve plotting unit, a report generation unit, a risk warning unit, and a report interpretation interface. The curve plotting unit automatically plots a comparison curve between the subject's two-dimensional time-series trajectory and the baseline trajectory of healthy individuals, as well as a time-series characteristic deviation curve, supporting curve zoom-in and detailed viewing. The report generation unit integrates the subject's basic information, quantitative data at each time-series node, S-value, DNA homeostasis assessment results, risk grading, and follow-up recommendations to generate a standardized PDF report, which can be printed or electronically distributed. The risk warning unit uses color coding to indicate risk levels, highlighting core conclusions and follow-up recommendations, facilitating quick access to key information for both medical staff and subjects. The report interpretation interface supports the association of medical staff with interpretation functions, allowing the addition of interpretation notes to the report to record clinical recommendations.

[0038] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for early tumor detection based on DNA characteristics, characterized in that, The method includes the following specific steps: S1: Collect peripheral venous blood from the subject according to the preset physiological cycle time nodes, separate and obtain plasma samples, and extract free cfDNA from the plasma; S2: Simultaneous hybridization capture of cfDNA was performed using oxidative damage-specific probes and methylation-specific probes. The oxidative damage-specific probes targeted and bound to the 8-hydroxydeoxyguanosine modification site in cfDNA, while the methylation-specific probes targeted and bound to the CpG island methylation site of tumor-associated genes. S3: Quantitatively detect the captured cfDNA and obtain the oxidative damage abundance and methylation level signals of cfDNA at each time point to form a two-dimensional time-series feature dataset; S4: Based on the dual-dimensional time-series detection data of healthy people during their physiological cycle, establish a benchmark model of dynamic changes in DNA oxidative damage and methylation under normal physiological conditions, and determine the time-series feature deviation threshold. S5: Compare the subject's two-dimensional time-series feature dataset with the benchmark model and calculate the time-series feature deviation value; If the deviation value exceeds the preset threshold, it is determined that the subject has abnormal DNA homeostasis, indicating an early risk of tumor development.

2. The method for early tumor detection based on DNA characteristics according to claim 1, characterized in that, In S1, four consecutive time-series detection nodes, T0, T1, T2, and T3, are set based on the subject's physiological rhythm. The interval between each node is uniformly 48 hours. Peripheral venous blood of the subject is collected at 5 mL per node using EDTA-K2 anticoagulation vacuum blood collection tubes. Within 1 hour after blood collection, the sample is centrifuged at 4℃ and 3000g for 15 minutes to complete plasma separation. The cell-free plasma layer is aspirated and transferred to an enzyme-free centrifuge tube, which is then frozen and stored at -80℃ for later use.

3. The method for early tumor detection based on DNA characteristics according to claim 1, characterized in that, In step S1, the specific steps for extracting cell-free cfDNA from plasma are as follows: Automated extraction of plasma cfDNA is performed using a magnetic bead method. 200 μL of thawed plasma sample is taken, and 400 μL of lysis buffer containing 20 mM Tris-HCl, 8 M guanidine salt, and 1% β-mercaptoethanol is added. The sample is vortexed for 30 seconds, then lysed at room temperature for 10 minutes. Subsequently, 20 μL of carboxyl-modified magnetic beads are added and thoroughly mixed. The sample is incubated at room temperature for 15 minutes to allow cfDNA to bind to the magnetic beads. After magnetic separation, the supernatant is discarded, and a solution containing 70% ethanol and 10 mM Tris-HCl is added. Magnetic separation washing was performed using s-HCl washing buffer, and the washing was repeated twice to remove impurities. Finally, 50 μL of 10 mM Tris-HCl enzyme-free elution buffer (pH 8.0) was added, and the mixture was incubated at 65°C for 10 min to complete cfDNA elution. After magnetic separation and collection of the eluent, quality control was performed using a Qubit 4.0 real-time fluorescence instrument, agarose gel electrophoresis, and a Nanodrop detector. The requirements were that the cfDNA concentration was ≥1 ng / μL, the main band of the fragment was concentrated in 160-200 bp, and the A260 / A280 ratio was 1.8-2.

0.

4. The method for early tumor detection based on DNA characteristics according to claim 1, characterized in that, In step S2, a nucleic acid aptamer probe labeled with a 5' end FAM fluorophore and targeting 8-hydroxydeoxyguanosine was selected as an oxidative damage-specific probe, while a recombinant methylation-binding domain protein fusion probe labeled with an N-terminal HEX fluorophore was selected as a methylation-specific probe. A reaction system with a total volume of 50 μL was prepared, containing 10 μL cfDNA template, 2 μL 10 μM oxidative damage probe, 2 μL 10 μM methylation probe, 25 μL 2× hybridization buffer containing 50 mM PBS, 1 M NaCl, and 0.1% Tween-20, and 11 μL enzyme-free water. After mixing, the system was denatured at 95°C for 5 min, quenched on ice for 2 min, and then hybridized at 42°C for 90 min. After hybridization, 10 μL streptavidin magnetic beads were added and incubated at room temperature for 10 min. The free probe was magnetically separated and washed three times with washing buffer, retaining the cfDNA complex bound to the dual-dimensional probe.

5. The method for early tumor detection based on DNA characteristics according to claim 1, characterized in that, In S3, a dual-channel fluorescence synchronous detection and internal reference correction technique is adopted. The abundance of oxidative damage and the level of methylation are simultaneously acquired through the FAM channel and HEX channel of the real-time fluorescence quantitative PCR instrument. The β-actin housekeeping gene without oxidative damage and methylation is introduced as an internal reference. The target detection signal is corrected for systematic error based on its fluorescence signal. The corrected FAM fluorescence value and HEX fluorescence value of the four time nodes from T0 to T3 are recorded respectively to construct a two-dimensional time series data matrix containing oxidative damage and methylation level.

6. The method for early tumor detection based on DNA characteristics according to claim 1, characterized in that, In step S4, healthy volunteers are selected, and a training dataset is constructed by acquiring two-dimensional time-series data from time-series sample collection to quantitative detection of time-series signals. The time-series data of healthy individuals is fitted using a dynamic time warping algorithm to generate the average dynamic change trajectory of DNA oxidative damage-methylation under normal physiological conditions. The time-series feature deviation threshold of 1.25 is determined by ROC curve analysis and used as a criterion to distinguish between normal physiological fluctuations and pathological abnormalities.

7. The method for early tumor detection based on DNA characteristics according to claim 6, characterized in that, In step S4, the specific steps for fitting the time series data of healthy individuals using the dynamic time warping algorithm are as follows: Select a two-dimensional time series dataset of healthy volunteers, preprocess the oxidative damage abundance and methylation level data corresponding to each time series node for each volunteer to obtain standardized individual two-dimensional time series trajectories, use the dynamic time warping algorithm to select the mean of the time series trajectories of some healthy volunteers as the initial reference trajectory, dynamically align the individual time series trajectories of the remaining healthy volunteers with the initial reference trajectory, adjust the correspondence of each time series node by calculating the cumulative distance between each pair of trajectories, and then calculate the mean of oxidative damage abundance and methylation level at each time series node based on all aligned individual time series trajectories, repeat the trajectory alignment and reference trajectory update steps until the reference trajectory tends to stabilize, that is, the trajectory fluctuation after two adjacent updates reaches the preset stability standard, stop the iterative calculation, and finally obtain the reference trajectory as the average dynamic change trajectory of DNA oxidative damage-methylation under normal physiological conditions.

8. The method for early tumor detection based on DNA characteristics according to claim 1, characterized in that, In step S5, the subject's two-dimensional temporal feature dataset is compared with the benchmark model, and the temporal feature deviation value is calculated using the following formula: ,in, It is the temporal characteristic deviation value, which represents the overall degree of deviation between the subject's two-dimensional temporal trajectory and the baseline trajectory of healthy people; This is the set total number of time-series detection nodes. It is the sequence number of the time sequence node, used to sequentially refer to each independent sampling time sequence node; For the examinee at the Quantitative values ​​of DNA oxidative damage abundance after internal reference correction at each time point. It is the first in the average dynamic baseline trajectory of healthy people. The baseline value of DNA oxidative damage abundance corresponding to each time node. The examinee was on the Quantitative values ​​of DNA methylation levels after internal control correction at each time point. It is the first in the average dynamic baseline trajectory of healthy people. The baseline values ​​of DNA methylation levels corresponding to each time node are obtained by fitting a dataset of healthy individuals using a dynamic time warping algorithm.

9. The method for early tumor detection based on DNA characteristics according to claim 8, characterized in that, In S5, It is the time series characteristic deviation value. A value ≤1.25 indicates normal DNA homeostasis and no risk of early-stage tumors; a value <1.25 indicates normal DNA homeostasis and no risk of early-stage tumors. A level ≤1.8 is considered a mild abnormality in DNA homeostasis, suggesting a risk of precancerous lesions, and a follow-up examination is recommended in 3 months. A score >1.8 indicates severe DNA homeostasis abnormalities, suggesting a high risk of early-stage tumors, and further imaging and pathological examinations are recommended.

10. A DNA-based early tumor detection system, applicable to the DNA-based early tumor detection method according to any one of claims 1-8, characterized in that, The system includes: Sample preprocessing module: used for peripheral blood plasma separation, cfDNA extraction and purification, and outputting standardized cfDNA samples; Dual-dimensional probe capture module: integrates reaction systems for oxidative damage-specific probes and methylation-specific probes to achieve simultaneous and specific capture of cfDNA dual-dimensional modification sites; Time-series data acquisition module: Connects to the fluorescence detection unit, acquires oxidative damage abundance and methylation level signals of cfDNA according to preset time nodes, and generates time-series feature dataset; Dynamic model analysis module: It has a built-in two-dimensional time-series benchmark model library for healthy people, and is equipped with a time-series deviation calculation unit to complete the comparison and analysis of the subject's data and the benchmark model; Results output module: Used to output DNA homeostasis abnormality determination results, time-series characteristic deviation curves, and early tumor risk grading reports.

Images