DNA extraction method and application thereof in multi-chromosome copy number variation analysis

Abstract

The invention provides a DNA extraction method and application thereof in multi-chromosome copy number variation analysis, and relates to the technical field of biomedical detection. The DNA extraction method provided by the invention comprises the steps of sample pretreatment, splitting decomposition, DNA adsorption and the like, and the DNA degradation rate of the FFPE sample is controlled below 15%. The detection system based on the extraction method comprises a wet experiment detection module, a biological information analysis module and a clinical decision module, wherein the wet experiment detection module executes the DNA extraction process; the biological information analysis module performs CNV detection through a dynamic sliding window algorithm, and calculates a Z value of a chromosome arm after GC correction and baseline normalization processing; the clinical decision module integrates MaxZ (maximum Z value) and CountZ (abnormal chromosome count) indexes and clinical variables to construct a model, and generates a risk layering report. The method and the system are suitable for multi-chromosome copy number variation analysis, and detection accuracy and clinical application value are improved.

Description

Technical Field

[0001] This invention belongs to the field of biomedical detection technology and provides a unique low-degradation DNA extraction method for FFPE samples. By solving the technical bottleneck of severe DNA breakage and high degradation rate in FFPE sample processing, it provides a high-quality DNA template for subsequent multiple chromosomal copy number variation (CNV) analysis, ultimately achieving high accuracy in CNV detection of gastric cancer and precancerous lesions. It is particularly suitable for clinical molecular diagnostic scenarios that rely on FFPE samples. Specifically, it involves a DNA extraction method and its application in multiple chromosomal copy number variation analysis. Background Technology

[0002] As one of the most common malignant tumors worldwide, early diagnosis of gastric cancer is crucial for improving patient prognosis and reducing mortality. In current clinical practice, early diagnosis of gastric cancer heavily relies on molecular testing of formalin-fixed paraffin-embedded (FFPE) gastric mucosal tissue samples. However, DNA analysis of these samples often suffers from quality issues, becoming a bottleneck that limits the accuracy of subsequent multi-chromosome copy number variation (CNV) analysis.

[0003] Traditional FFPE sample DNA extraction procedures have two major drawbacks: First, the dewaxing process relies on repeated centrifugation and washing with xylene (usually requiring 3-5 operations), which is not only time-consuming but also prone to introducing exogenous contamination due to repeated opening of the cap. In addition, residual organic solvents can damage the DNA structure. Second, the decrosslinking step requires continuous treatment at a high temperature of 90-95℃ for 1-2 hours. Extreme temperatures can cause irreversible DNA strand breaks, resulting in a DNA degradation rate that generally exceeds 20%, which is far from meeting the core requirement of low-depth whole-genome sequencing for DNA integrity. This sequencing technology requires a DNA degradation rate of <15% and a concentration of ≥50ng / μL; otherwise, it will directly lead to library construction failure or a sharp drop in sequencing data quality.

[0004] CNV is one of the commonly used molecular markers for early gastric cancer. The 4q chromosome arm carries several tumor suppressor genes closely related to gastric cancer development, such as CDKN2B (cyclin-dependent kinase inhibitor 2B) and FHIT (fragile histidine triplet gene). Copy number loss on the 4q chromosome arm is a frequent event in gastric cancer and precancerous lesion samples: loss leads to downregulation or inactivation of the aforementioned tumor suppressor genes, thereby disrupting cell cycle regulation mechanisms, promoting abnormal cell proliferation, and accelerating the carcinogenesis process. The 8q chromosome arm contains key oncogenes such as MYC (proto-oncogene) and CCND1 (cyclin D1 gene). Copy number amplification on the 8q chromosome arm is extremely common in gastric cancer development. MYC gene amplification leads to overexpression of its encoded transcription factor, activating multiple downstream pro-proliferation and anti-apoptotic signaling pathways, promoting rapid tumor cell growth; CCND1 amplification accelerates the transition of cells from G1 phase to S phase, further promoting abnormal cell division. Chromosome 19 is one of the chromosomes with a high gene density in the human genome. Both arms 19p and 19q contain genes related to the pathological mechanism of gastric cancer. Although chromosome 21 is not as high-frequency as 4q and 8q in gastric cancer CNV research, it also has clinical reference value because the genes it carries are related to the tumorigenesis mechanism and can be detected in the multi-chromosome analysis system of this invention.

[0005] In summary, the reliability of CNV detection results directly determines the accuracy of clinical diagnosis. However, low-quality degraded DNA can lead to uneven distribution of sequencing reads across the genome: broken DNA fragments cannot effectively participate in library construction, resulting in abnormally low sequencing depths in some regions, which are easily misjudged as "copy number loss"; while excessive fragmentation of local segments may lead to enrichment of reads in specific regions, forming "false positive amplification".

[0006] Even though existing CNV detection algorithms attempt to improve data interpretation by optimizing window size (such as fixing the window to 100kb), the false positive rate for CNV detection of chromosome arms such as 4q and 8q, which are crucial for gastric cancer diagnosis, is still as high as 18%-25%, and the false negative rate is over 12%. This fails to provide clinicians with reliable molecular evidence and may even lead to misdiagnosis or missed diagnosis, delaying patients' treatment.

[0007] Current research in this field largely focuses on optimizing single steps: improving CNV detection algorithms (such as enhancing the noise resistance of data analysis) and optimizing library construction reagents (such as adding DNA repair enzymes). However, research on establishing a complete technical system for "FFPE sample DNA extraction-degradation control-subsequent analysis compatibility" is extremely rare. This "fragmented improvement" makes it difficult to significantly improve the accuracy of CNV analysis. Even with advanced subsequent algorithms and library construction techniques, it is impossible to compensate for errors caused by defects in the initial DNA template quality.

[0008] Although Chinese patent 201910712588.6 mentions chromosome variation detection and kit design, it does not solve the core problem of high degradation rate of DNA extracted from FFPE samples. Therefore, there is an urgent need for a more effective DNA extraction method for FFPE samples and to establish a complete technical solution for DNA extraction, degradation control, and subsequent analysis of FFPE samples. Summary of the Invention

[0009] To address the aforementioned issues, this invention provides a DNA extraction method and its application in multi-chromosome copy number variation analysis, thereby resolving technical problems such as high DNA degradation rates and impact on the accuracy of subsequent CNV analysis when using existing DNA extraction methods for samples like FFPE.

[0010] In a first aspect, the present invention provides a DNA extraction method, comprising the following steps: S1: Take a sample and add it to a hot alkaline microemulsion, then heat at 90-100℃ for 3-8 minutes; S2: Enzymatic digestion at 50-60℃ for 25-35 minutes; S3: One-step purification by adding magnetic bead-PEG binding solution; The ratio of the sample to the hot alkaline microemulsion was 5-10 mg: 250-350 μL; The hot alkaline microemulsion comprises mineral oil, Triton X-100, and NaOH; Preferably, the mineral oil comprises at least one of liquid paraffin, white oil, n-tetradecane, isohexadecane, sorbitan monooleate, and 2,6-di-tert-butyl-p-cresol.

[0011] In S1, the hot alkaline microemulsion is composed of liquid paraffin, n-tetradecane, isohexadecane, sorbitan monooleate, 2,6-di-tert-butyl-p-cresol, Triton X-100, NaOH, and ethanol.

[0012] In S1, the amounts of each component in the hot alkaline microemulsion, by volume, are: 270-330 parts liquid paraffin, 40-60 parts n-tetradecane, 40-60 parts isohexadecane, 20-30 parts sorbitan monooleate (Span 80), 5-15 parts 2,6-di-tert-butyl-p-cresol (BHT), 40-60 parts a mixture of 10% Triton X-100 and 0.5M NaOH, and finally made up to 500 parts with anhydrous ethanol.

[0013] In S1, the preferred amounts of each component in the hot alkaline microemulsion by volume are: 270-300 parts liquid paraffin, 50-60 parts n-tetradecane, 50-60 parts isohexadecane, 20-25 parts Span 80, 10-15 parts BHT, 50-60 parts a mixture of 10% Triton X-100 and 0.5M NaOH, and finally made up to 500 parts with anhydrous ethanol.

[0014] In S1, the preferred amounts of each component in the hot alkaline microemulsion by volume are: 300-330 parts liquid paraffin, 40-50 parts n-tetradecane, 40-50 parts isohexadecane, 25-30 parts Span 80, 5-10 parts BHT, 40-50 parts a mixture of 10% Triton X-100 and 0.5M NaOH, and finally made up to 500 parts with anhydrous ethanol.

[0015] In S2, the enzyme includes, but is not limited to, at least one of proteinase K, pepsin, and streptomycin.

[0016] In S3, the amount of magnetic bead-PEG added, by volume, is 1-1.5 times that of the S2 reaction system.

[0017] Secondly, the present invention provides the application of any of the above-mentioned DNA extraction methods in the analysis of multiple chromosome copy number variations.

[0018] The multi-chromosome copy number variation analysis includes, but is not limited to, chromosome arrays, low-depth whole-genome sequencing, and high-depth whole-genome sequencing.

[0019] Thirdly, the present invention provides a detection system based on multi-chromosome copy number variation analysis, characterized in that it includes: The system includes a wet laboratory testing module, a bioinformatics analysis module, and a clinical decision-making module. The wet experimental detection module is used to perform any of the above-mentioned DNA extraction methods; The bioanalysis module includes a sliding window CNV detection algorithm, which can dynamically adjust the window size and perform GC correction and baseline normalization on the detection data; The clinical decision module is used to integrate the two parameters MaxZ and CountZ, and combine them with clinical variables to construct a logistic regression model; MaxZ is the maximum absolute value of the Z value of each chromosome arm, and CountZ is the number of chromosome arms with a Z value greater than ±3.

[0020] In the bioinformatics analysis module, the window size adjustment range of the sliding window CNV detection algorithm is 50kb-200kb; GC correction is achieved by calculating the GC content of the window, fitting the LOESS curve, and calculating the corrected density; baseline normalization is achieved by constructing a baseline library and calculating the normalized copy number.

[0021] The clinical variables include, but are not limited to, age and HP infection status.

[0022] Fourthly, the present invention provides the application of the above-mentioned detection system in gastric cancer risk stratification. The bioinformatics analysis module analyzes the 4q, 8q, 19q, and 20q chromosome arms to obtain the MaxZ and CountZ dual parameters. The sample is an ex vivo sample.

[0023] Fifthly, the present invention provides a kit for implementing any of the DNA extraction methods of the claims above, characterized in that it comprises a hot alkaline microemulsion prepreg, proteinase K, magnetic bead-PEG binding solution, ethanol, and TE elution buffer.

[0024] Compared with the prior art, the present invention has the following advantages: The DNA extraction method provided by this invention, through steps such as sample pretreatment, lysis, and DNA adsorption, controls the DNA degradation rate of FFPE samples to below 15%. The detection system based on this extraction method is suitable for multi-chromosome copy number variation analysis, significantly improving detection accuracy and clinical application value. Attached Figure Description

[0025] Figure 1 This is the flowchart for Experiment Example 1.

[0026] Figure 2 This is a CNV heatmap from Experiment Example 1.

[0027] Figure 3 This is the ROC curve from Experiment Example 3.

[0028] Figure 4 This is the user interface of the detection system based on multi-chromosome copy number variation analysis described in Experiment Example 4. Detailed Implementation

[0029] The present invention will be further described in detail below with reference to specific embodiments. The following embodiments are not intended to limit the present invention, but only to illustrate the present invention. Unless otherwise specified, the experimental methods used in the following embodiments are generally performed under conventional conditions. Unless otherwise specified, the materials and reagents used in the following embodiments are commercially available.

[0030] Examples 1-3 Prepare the "hot alkaline microemulsion" as described in Table 1.

[0031] Table 1. Components of hot alkaline microemulsion

[0032] Comparative Examples 1-5 Prepare the "hot alkaline microemulsion" as described in Table 2.

[0033] Table 2 Components of Hot Alkali Microemulsion

[0034] Example 1: DNA extraction protocol for FFPE samples Proteinase K was purchased from QIAGEN (catalog number 19131), DNA binding beads from Thermo Fisher Scientific (catalog number A25562), and TE buffer from Solarbio Scientific (catalog number T1123). 1. Single-tube hot alkaline microemulsion dewaxing-decrosslinking: Take 2-8 FFPE slices and place them into a 1.5 mL tube. Add 300 µL of "hot alkaline microemulsion" and heat in a 95 °C metal bath for 5 min. The paraffin melts and is emulsified, and NaOH is used for cleavage and crosslinking. After the metal bath is complete, immediately add 30 µL of 1 M pH 7.0 Tris-HCl to neutralize and clarify the system.

[0035] 2. Synchronous enzyme digestion Add 20 µL of 10× proteinase K buffer and 10 µL of 20 mg / mL proteinase K (final concentration 0.5 mg / mL) directly to the product from step 1, and incubate in a metal bath at 56 °C for 30 min; gently shake twice during the incubation period.

[0036] 3. One-step purification with magnetic beads Add 1.2 × volume (≈480 µL) of magnetic bead-PEG binding buffer, mix for 5 min, adsorb on a magnetic rack for 1 min, discard the supernatant, wash once with 80% ethanol (500 µL, 10 s), dry for 2 min, and elute with 50 µL TE at 60℃ for 5 min to obtain supernatant DNA.

[0037] DNA purity was determined using NanoDrop (A260 / A280 should be between 1.8 and 2.0), and concentration was determined using Qubit 4.0 (ensuring ≥50 ng to meet the requirements for subsequent library preparation). The results were verified by agarose gel electrophoresis (the main DNA band was clear, with no obvious diffusion and a degradation rate of less than 15%). The experimental results are shown in Table 3.

[0038] Table 3

[0039] Note: Compared with Example 1, *** represents p < 0.001.

[0040] The A260 / A280 ratio reflects the amount of residual protein in DNA, with an ideal range of 1.8-2.0. In Examples 1-3, the A260 / A280 ranged from 1.89±0.06 to 1.92±0.05, indicating low protein complex residue. In contrast, in Comparative Examples 1-5, the A260 / A280 ranged from 1.52±0.10 to 1.65±0.08, presumably due to an imbalance in the mineral oil compound and incomplete dewaxing, leading to increased protein residue. If the A260 / A280 ratio is below 1.8, residual protein can inhibit sequencing enzyme activity, resulting in uneven sequencing depth (e.g., local read loss), which in turn affects GC correction and baseline normalization, ultimately reducing the reliability of MaxZ and CountZ parameters and failing to provide accurate molecular evidence for clinical decision-making.

[0041] Low-depth whole-genome sequencing requires a DNA concentration of ≥50 ng to complete library construction. The concentrations in Examples 1-3 reached 78.5±6.8-82.3±7.5 ng / μL, achieving efficient DNA release and adsorption. In contrast, the concentrations in Comparative Examples 1-5 were only 47.8±4.8-57.2±6.1 ng / μL. It is speculated that this may be due to the low decrosslinking efficiency of the microemulsion, resulting in insufficient DNA release. If the concentration is below 50 ng, it needs to be supplemented by nucleic acid amplification. However, amplification is prone to introducing bias, leading to "false positives" in CNV detection (such as overestimation of local copy numbers), which undermines the clinical application value of the detection system.

[0042] In Examples 1-3, liquid paraffin, n-tetradecane, and isohexadecane were compounded at a ratio of 270-330:40-60:40-60 μL, forming a stable microemulsion with Span 80, BHT, and Triton X-100 + NaOH. The mineral oil compound system, in conjunction with the surfactant, achieved efficient dewaxing and gentle decrosslinking. BHT, as a synergistic active ingredient, inhibited DNA oxidative damage, keeping the extracted DNA degradation rate below 15%. In Comparative Examples 1-5, the imbalance of the mineral oil ratio may have led to the failure of the microemulsion synergistic effect, resulting in a DNA degradation rate of 18.5±2.5%-22.5±2.6%, making it difficult to guarantee the accuracy of CNV analysis.

[0043] Experiment Example 1: Construction of an Early Diagnostic System for Gastric Cancer Based on Multi-Chromosome Copy Number Variation Analysis For details on the construction process, please refer to Figure 1 It provides a complete build process.

[0044] The system comprises three core modules: (1) Wet test detection module: The hot alkaline microemulsion prepared in Example 1 was used to extract DNA from formalin-fixed paraffin-embedded (FFPE) samples as described in Detection Example 1. (2) Bioinformatics analysis module: A CNV detection algorithm based on a sliding window was developed. The window size is dynamically adjusted from 50kb to 200kb to adapt to samples of different quality. Technical noise is significantly reduced through GC correction and baseline normalization. 1. Sliding window CNV detection algorithm Objective: To provide a copy number estimate (CN) for each window i on a chromosome. i .

[0045] 1.1 Window Division Let the total chromosome length be Lbp and the total number of reads in the whole genome be N.

[0046] If the desired resolution is 50kb–200kb, and W is the variable window size, then the number of windows m = L / W.

[0047] 1.2 Original Count Let R k For the k-th read; put R k The number of reads falling on window i is c. i =Σ k 1(R k (Located at the window).

[0048] 1.3 Density normalized to per bp per read d i =c i / (W·N) 1.4 Dynamic Window Size First, calculate the MAD (median absolute difference) for each window in 50kb increments: MAD = median(|d i -median(d)|).

[0049] If MAD < ε1 (empirical value 0.1), the sequencing fluctuation in this region is considered small, and the window is expanded to 200kb; otherwise, it remains at 50kb. Let the final window size W i ∈{50kb,200kb}.

[0050] 1.5 Copy Number Estimation Let the baseline diploid depth of this sample be μ0 (which can be determined by the median d in the whole genome). i (estimated): CN i =2·d i / μ0.

[0051] 2. GC Correction Objective: To eliminate depth bias caused by GC content.

[0052] 2.1 Calculation of GC content in the window GC i=(#G+#C) / W i .

[0053] 2.2 Establishing the LOESS Curve (GC) i ,d i Using scatter points as the data points, the LOESS (locally weighted regression) function is fitted. f(GC)≈d i .

[0054] 2.3 Corrected density d i GC =d i / f(GC i )·μ_f, Where μ_f is the mean of f across all windows, ensuring that the overall scale remains constant: μ_f=(1 / m)Σ j f(GC j ).

[0055] Therefore, the copy number after GC correction CN i GC =2·d i GC / μ0.

[0056] 3. Baseline normalization (noise reduction) Objective: To eliminate global biases caused by batch, reagent, and sequencing platform factors.

[0057] 3.1 Building a baseline library Select n normal diploid samples j=1…n from the same experimental batch, and calculate the baseline mean for window i. B i =(1 / n)Σ j CN i ,j GC .

[0058] 3.2 Normalization CN i norm =CN i GC / B i ·2 (3) Clinical decision-making module: Integrating the two parameters MaxZ (maximum Z-value) and CountZ (abnormal chromosome count), and combining clinical variables such as age and HP infection status, a logistic regression model is constructed. MaxZ reflects the CNV amplitude and is calculated as MaxZ=max(|Z1p|,|Z1q|,...,|Z22q|); CountZ reflects the range of genomic instability and is defined as the number of chromosome arms with a Z-value exceeding ±3.

[0059] The formula for the logistic regression model is as follows:

[0060] Where β0 represents the intercept term, β1 represents the MaxZ coefficient, β2 represents the CountZ coefficient, β3 represents the age coefficient, and β4 represents the HP infection status coefficient.

[0061] The sample quality control system ensures data reliability through two indicators: sequencing depth coefficient of variation (CV < 0.15) and chimerism rate (< 5%). For example... Figure 2 The CNV heatmap features are shown.

[0062] Experimental Example 2: Clinical Sample Processing Flow Take 2-10 FFPE sections of gastric mucosa tissue, use the hot alkaline microemulsion prepared in Example 1, and perform DNA extraction as described in Detection Example 1. Use Qubit quantification to ensure that the input amount is ≥50ng.

[0063] Library construction was performed using a library construction kit (NEB, catalog number: E7645S) following the manufacturer's instructions. Sequencing was conducted on the MGI platform with an average depth of 0.5X, yielding ≥5M reads / sample. Bioinformatics analysis employed a customized workflow: first, FastQC was used for quality control; then, BWA-MEM was aligned to the hg19 reference genome; copy number analysis was performed using CNVkit; and finally, the Z-scores of chromosome arms 4q, 8q, 19q, and 20q were calculated using the self-developed ZscoreCNV algorithm.

[0064] For each analysis window (dynamically adjusted to 50kb-200kb), calculate its Z-score relative to the healthy baseline:

[0065]

[0066] Z(chr): The overall Z-value of the target chromosome arm n: The number of valid windows contained in this chromosome arm (after filtering out outliers). Z(Win chr,i): The Z value of the i-th window of this chromosome arm. CN i,norm μ is the window normalized copy number. baseline σ baseline These represent the mean and standard deviation of the copy number for this window in the healthy baseline database, respectively. The critical value was determined using the Youden exponent maximization principle, and the stability of the model was confirmed in the validation set (n=75).

[0067] Experimental Example 3: Sensitivity and Specificity Analysis 150 ex vivo gastric mucosal FFPE samples were collected and divided into 3 groups according to pathological diagnosis results: healthy control group (Group 1, n=50, pathologically confirmed to have no gastric mucosal lesions), precancerous lesion group (Group 2, n=25, including low-grade intraepithelial neoplasia and high-grade intraepithelial neoplasia), and gastric cancer group (Group 3-4, n=75, including early gastric cancer and advanced gastric cancer). All samples met ethical requirements and had complete clinical data (age, H. pylori infection status).

[0068] Each group of samples was processed according to the "DNA extraction protocol for FFPE samples in Example 1". 2-8 FFPE slices (5μm thick) were taken, and DNA was obtained by hot alkaline microemulsion extraction method in Example 1. The DNA concentration was detected using Qubit4.0 (to ensure ≥50ng), purity was verified by NanoDrop (A260 / A280 1.8-2.0), and degradation rate was confirmed by agarose gel electrophoresis (<15%). Unqualified samples (concentration <50ng or degradation rate >15%) were excluded. The final number of valid samples in each group was Group1 (n=48), Group2 (n=24), and Group3-4 (n=72).

[0069] Based on the pathological diagnosis results, the samples in Group 3-4 (gastric cancer) and Group 2 that progressed to gastric cancer after 6 months of follow-up were defined as "positive samples" in the pathological results corresponding to the "high-risk" criteria, and the remaining samples (all of Group 1 and samples in Group 2 that did not progress) were defined as "negative samples". ROC curves were plotted based on the final results.

[0070] The ROC curve analysis results of this system are shown below. Figure 3 The combined model demonstrated a detection sensitivity of 89.2% (95% CI: 82.1-93.8%) and a specificity of 90.6% (95% CI: 76.2-90.3%) for gastric cancer (Group 3-4), significantly superior to existing serum biomarkers CEA, CA125, and CA199 (sensitivity <50%). For precancerous lesions (Group 2), the system identified 20% of high-risk patients whose CNV profiles were highly similar to those of early-stage gastric cancer (P<0.001).

[0071] Experimental Example 4: Clinical Laboratory Results In a 69-year-old male patient (traditional pathological diagnosis of indeterminate dysplasia), the system detected chr4q (Z=-5.32), chr8q (Z=25.52), and chr20q (Z=13.45) with a count Z=3, indicating a high risk. The system recommended a follow-up gastroscopy within 6 months. The patient was later diagnosed with early gastric cancer by pathology, demonstrating that the system can successfully identify patients who eventually progress to gastric cancer. Figure 4 The software's user interface was demonstrated, including modules for data upload, analysis progress, and analysis results.

[0072] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, and is not intended to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions made by those skilled in the art to the technical solution of the present invention do not depart from the essence and scope of the technical solution of the present invention.

Claims

1. A DNA extraction method, characterized in that, Includes the following steps: S1: Take a sample and add it to a hot alkaline microemulsion, then heat at 90-100℃ for 3-8 minutes; S2: Enzymatic digestion at 50-60℃ for 25-35 minutes; S3: One-step purification by adding magnetic bead-PEG binding solution; The ratio of the sample to the hot alkaline microemulsion was 5-10 mg: 250-350 μL; The hot alkaline microemulsion comprises mineral oil, Triton X-100, and NaOH; Preferably, the mineral oil comprises at least one of liquid paraffin, white oil, n-tetradecane, isohexadecane, sorbitan monooleate, and 2,6-di-tert-butyl-p-cresol.

2. The DNA extraction method according to claim 1, characterized in that, In S2, the enzyme includes proteinase K.

3. The DNA extraction method according to claim 1, characterized in that, In S3, the amount of magnetic bead-PEG added, by volume, is 1-1.5 times that of the S2 reaction system.

4. The application of the DNA extraction method according to any one of claims 1-3 in multichromosomal copy number variation analysis.

5. The application according to claim 4, characterized in that, The multi-chromosome copy number variation analysis includes chromosome arrays, low-depth whole-genome sequencing, and high-depth whole-genome sequencing.

6. A detection system based on multi-chromosome copy number variation analysis, characterized in that, include: The system includes a wet laboratory testing module, a bioinformatics analysis module, and a clinical decision-making module. The wet test detection module is used to perform the DNA extraction method according to any one of claims 1-3; The bioanalysis module includes a sliding window CNV detection algorithm, which can dynamically adjust the window size and perform GC correction and baseline normalization on the detection data. The clinical decision module is used to integrate the two parameters MaxZ and CountZ, and combine them with clinical variables to construct a logistic regression model; MaxZ is the maximum absolute value of the Z value of each chromosome arm, and CountZ is the number of chromosome arms with a Z value greater than ±3.

7. The detection system according to claim 6, characterized in that, In the bioinformatics analysis module, the window size adjustment range of the sliding window CNV detection algorithm is 50kb-200kb; GC correction is achieved by calculating the GC content of the window, fitting the LOESS curve, and calculating the corrected density; baseline normalization is achieved by constructing a baseline library and calculating the normalized copy number.

8. The detection system according to claim 6, characterized in that, The clinical variables include age and HP infection status.

9. The application of the detection system according to any one of claims 6-8 in gastric cancer risk stratification, wherein the bioinformatics analysis module analyzes the 4q, 8q, 19q, and 20q chromosome arms to obtain the MaxZ and CountZ dual parameters, and the sample is an ex vivo sample.

10. A kit for implementing the DNA extraction method according to any one of claims 1-3, characterized in that, It includes a hot alkaline microemulsion prepreg, proteinase K, magnetic bead-PEG binding buffer, ethanol, and TE elution buffer.

Images