A rapid methylation typing method and system for central nervous system tumors based on nanopore sequencing

Abstract

This invention discloses a rapid methylation typing method and system for central nervous system tumors based on nanopore sequencing. The method includes: DNA extraction and ligation library preparation from FFPE or frozen samples; real-time sequencing using a nanopore sequencing platform, with adaptive sampling to enrich diagnostically relevant CpG target regions for FFPE samples; real-time base and 5mC identification to construct methylation features, and integrated output of methylation typing and confidence scores through a multi-classifier; simultaneous CNV analysis based on sequencing depth to extract key diagnostic features; when the methylation typing confidence score is insufficient, combining CNV features to form an integrated interpretation; and optional detection of SNVs / Indels when coverage meets the target. The method is cost-effective and provides integrated output of methylation typing, CNV, and optional mutation information, offering effective support for the rapid and accurate diagnosis of central nervous system tumors.

Description

Technical Field

[0001] This invention relates to the field of detection methods, specifically to a rapid methylation typing method and system for central nervous system tumors based on nanopore sequencing. Background Technology

[0002] The pathological diagnosis of central nervous system (CNS) tumors has gradually shifted from purely morphological methods to an integrated diagnosis combining histology and molecular characteristics. The WHO CNS classification system now uses DNA methylation typing as important molecular evidence for many complex CNS tumors, especially applicable to cases where the subtype remains unclear after routine pathology, immunohistochemistry, and limited gene testing.

[0003] Currently, the most widely used methylation typing platform in clinical practice is the Illumina MethylationEPIC chip. This technology detects the methylation status of CpG sites across the entire genome, and combines a reference database with a machine learning classifier to achieve molecular tumor typing. It has advantages such as mature technology, high stability, and a comprehensive reference database.

[0004] However, EPIC chip technology has significant drawbacks in real-world clinical workflows: First, the testing cycle is long, requiring batch processing with turnaround times often lasting weeks to months, making it difficult to meet the needs of rapid clinical decision-making; second, the high equipment and maintenance costs make it difficult for many non-ultra-large centers to routinely implement the technology; third, FFPE samples require a remediation step, placing high demands on sample quantity and quality, and degraded samples may not yield reliable results; fourth, the fixed workflow makes it difficult to achieve dynamic control such as "sequencing and analysis simultaneously, stopping once confidence level is reached"; fifth, the information dimension is limited, mainly providing methylation information, with limited resolution for copy number variations, and it cannot directly obtain single nucleotide variation information, often requiring separate testing, leading to extended cycles and increased costs.

[0005] On the other hand, nanopore sequencing technology features real-time output, no need for fixed batches, and the ability to directly identify 5mC modifications from sequencing signals, theoretically making it suitable for "rapid methylation typing." However, applying it to FFPE, the most common material in routine pathology, still faces challenges: FFPE DNA fragments are short and prone to damage, the proportion of effective information is unstable, classification confidence is insufficient under low coverage, and the cost of repeated experiments increases. Currently, there is no integrated solution optimized for FFPE samples that can stably obtain methylation typing under low coverage and simultaneously output multidimensional molecular information.

[0006] Therefore, there is a need for a rapid detection method and system that can be used for real-world pathological samples, obtain stable methylation typing even with low coverage, and simultaneously output CNV and optional SNV information, generating a structured report within 24 hours.

[0007] To address the aforementioned issues, the applicant proposes a rapid methylation typing method and system for central nervous system tumors based on nanopore sequencing. Summary of the Invention

[0008] The purpose of this invention is to provide a rapid methylation typing method and system for central nervous system tumors based on nanopore sequencing, so as to solve the problems in the prior art.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a rapid methylation typing method for central nervous system tumors based on nanopore sequencing, comprising the following steps:

[0010] Sample reception and preparation steps: Receive central nervous system tumor samples, perform pathological evaluation and tumor cell enrichment;

[0011] DNA extraction and quality control steps: Extract sample DNA and perform quality testing;

[0012] Library construction steps: Nanopore sequencing libraries were constructed using ligation methods;

[0013] Nanopore sequencing and adaptive sampling steps: Real-time sequencing was performed using a nanopore sequencing platform, and adaptive sampling was enabled for FFPE samples to enrich a predefined set of diagnostically relevant CpG target regions;

[0014] Real-time analysis and typing steps: Real-time identification of bases and 5mC modifications, construction of methylation feature maps, and output of methylation typing categories and their confidence levels through multi-classifier integration;

[0015] Report generation steps: Generate a structured report containing at least methylation typing results.

[0016] Optionally, the real-time analysis and classification step further includes:

[0017] CNV analysis sub-steps: Copy number variation analysis is performed based on the coverage depth of sequencing reads to extract key diagnostic features;

[0018] Integrated interpretation sub-step: When the confidence level of methylation typing does not reach the preset threshold, an integrated diagnostic suggestion is formed by combining the CNV analysis results and pathological information, and the source of evidence and confidence level are marked in the report.

[0019] Optionally, the real-time analysis and typing step also includes an SNV / Indel detection sub-step: when the sequencing coverage reaches a preset threshold, variant detection and annotation are initiated, and a list of potential pathogenic or actionable mutations is output.

[0020] Optionally, the nanopore sequencing and adaptive sampling steps are equipped with dynamic termination conditions: sequencing is automatically terminated when the minimum amount of data generated reaches a preset value, or when the confidence level output by any one or more classifiers reaches a preset confidence threshold.

[0021] Optionally, the set of diagnostic-related CpG target regions is described in BED format and includes the genomic regions required for molecular subtyping of central nervous system tumors; the adaptive sampling rapidly determines the read front signal during sequencing, and if the target region is hit, sequencing continues and the read is retained; if not hit, the molecule is ejected by reverse voltage.

[0022] A rapid methylation typing system for central nervous system tumors based on nanopore sequencing, comprising:

[0023] Sample and DNA preparation unit: used for pretreatment, DNA extraction and quality control of central nervous system tumor samples;

[0024] Nanopore sequencing unit: Supports real-time base recognition and 5mC modification recognition, and has adaptive sampling function for enrichment sequencing of predefined diagnostic-related CpG target regions;

[0025] The computation and analysis unit communicates with the sequencing unit in real time to receive sequencing data and perform real-time analysis, including methylation feature construction, multi-classifier genotyping, CNV analysis, and integrated interpretation.

[0026] Report output unit: Used to generate structured reports containing methylation typing results, CNV features, and quality control indicators.

[0027] Optionally, the calculation and analysis unit further includes:

[0028] Methylation typing module: Constructs features based on the methylation level of CpG sites or regions, and calls one or more machine learning classifiers to output tumor category and confidence score;

[0029] CNV analysis module: Performs copy number variation analysis based on read depth and extracts key diagnostic features;

[0030] Integrated interpretation module: When the confidence level of methylation typing is insufficient, it outputs remedial interpretation suggestions by combining CNV features and pathological information;

[0031] It also includes an SNV detection module: when the coverage reaches a threshold, variant detection and annotation are enabled.

[0032] Optionally, the nanopore sequencing unit supports single-sample independent operation or barcode-reused mixed-sample operation, and supports flow cell washing and library reloading to supplement coverage.

[0033] As described above, the central nervous system tumor samples include FFPE paraffin tissue, frozen tissue, or pre-extracted DNA; the structured report also includes CNV spectra, coverage and quality control indicators, and optionally includes SNV / Indel results and MGMT predictions.

[0034] Beneficial effects: First, turnaround time is significantly shortened, supporting on-demand testing of single samples. Utilizing nanopore real-time sequencing and analysis, the entire process from sample to report can be completed within 24 hours, and sequencing can be terminated early upon reaching a confidence threshold. No batch sample collection is required, making it suitable for medium-sized pathology centers and emergency consultation scenarios, significantly reducing waiting and logistics costs.

[0035] Second, it is compatible with common FFPE materials to improve the success rate of detection. Addressing the fragmented and damaged nature of FFPE samples, adaptive sampling is used to enrich diagnostically relevant CpG target regions. Even with low coverage, reportable methylation typing results can still be obtained, effectively improving the usability and data utilization of FFPE samples and reducing the cost of repeated testing.

[0036] Third, it provides integrated output of multidimensional molecular information. A single sequencing run can simultaneously obtain methylation typing, diagnostic CNV features, and SNV / Indel detection results when coverage meets the standard, without the need for separate chip hybridization or targeted sequencing, reducing sample consumption, simplifying the process, and achieving comprehensive diagnostic efficiency of "one test for multiple uses".

[0037] Fourth, it possesses clear interpretation rules and cost control mechanisms. Confidence assessment is established through multi-classifier integration and threshold strategies. When methylation typing is insufficient, CNV features are introduced as remedial evidence to improve diagnostic output. Simultaneously, mechanisms such as dynamic termination, barcode reuse, and flow cell washing control the cost per sample, facilitating the technology's widespread adoption. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the overall process for rapid nanopore methylation spectroscopy-based accurate detection and typing (CNS tumors) according to an embodiment of the present invention;

[0039] Figure 2 This is a schematic diagram of the integrated analysis pipeline structure of nanopore methylation spectroscopy + CNV + SNV according to an embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the adaptive sampling targeted enrichment principle in an embodiment of the present invention (for low-quality / low-yield FFPE samples). Detailed Implementation

[0041] The preferred embodiments of the present invention are described below with reference to the accompanying drawings to make the technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0042] This invention provides a rapid methylation typing method and system for central nervous system tumors based on nanopore sequencing, aiming to solve the technical problems of existing EPIC chip methylation typing technology, such as long turnaround time, high cost, poor adaptability to FFPE samples, and limited information dimensions. It should be noted that the following embodiments are only for illustrative purposes and are not intended to limit the technical solution of this invention. Other embodiments obtained by those skilled in the art without departing from the concept of this invention are all within the scope of protection of this invention.

[0043] This invention provides a rapid methylation typing method for central nervous system tumors based on nanopore sequencing, comprising the following core steps: sample reception and preparation, DNA extraction and quality control, ligation-based library preparation, nanopore sequencing and adaptive sampling targeted enrichment, real-time analysis and quality monitoring, and report generation and output. This method relies on an integrated detection system, which includes at least a sample and DNA preparation unit, a nanopore sequencing unit, a computation and analysis unit, and a report output unit. The nanopore sequencing unit supports real-time base recognition and 5mC modification recognition, while the computation and analysis unit incorporates real-time processing software and algorithms, enabling simultaneous sequencing and analysis to dynamically determine whether the data volume meets the typing requirements.

[0044] In the sample receiving and preparation steps, the sample types applicable to this invention include, but are not limited to, formalin-fixed paraffin-embedded tissues, frozen tissues, and pre-extracted DNA samples. For FFPE samples, hematoxylin-eosin staining verification and tumor cell proportion assessment are required first. Preferably, the tumor cell proportion is not less than 50%, more preferably not less than 70%, to ensure that the subsequent methylation typing results accurately reflect tumor characteristics rather than interference from the background microenvironment or normal brain tissue. For samples with a low tumor cell proportion, tumor enrichment is required, which may include macro-slicing of the FFPE section or manual scraping of the tumor-enriched area under a microscope, removing as much necrotic tissue, hemorrhage areas, and areas with obvious inflammatory reactions as possible. This step is crucial for ensuring the accuracy of methylation typing because the DNA methylation profile of non-tumor cells may interfere with the overall detection results, especially when the tumor cell proportion is low, which may lead to decreased typing confidence or even misclassification.

[0045] In the DNA extraction and quality control steps, this invention is not limited to a specific DNA extraction kit or method; appropriate solutions can be adopted according to the sample type. For FFPE samples, it is recommended to use a dedicated extraction kit that can effectively remove cross-links and recover fragmented DNA; for frozen tissues, conventional phenol-chloroform extraction or commercial column extraction kits can be used. After extraction, strict quality control is required, mainly including concentration detection, purity assessment, and fragmentation degree analysis. Concentration detection can be performed using fluorescent dye methods or ultraviolet spectrophotometry; purity assessment is mainly achieved by measuring the A260 / A280 ratio and the A260 / A230 ratio. Ideally, the A260 / A280 ratio for purified DNA samples should be between 1.8 and 2.0, and the A260 / A230 ratio should be greater than 2.0; fragmentation degree analysis can be performed using agarose gel electrophoresis or a microfluidic fragment analyzer, focusing on the size of the main DNA band and the distribution range of fragments. For FFPE samples, due to cross-linking during formaldehyde fixation and oxidative degradation during long-term storage, the DNA is often highly fragmented, with the main band potentially less than 500 base pairs or even lower. This invention sets a certain range for the amount of DNA input, typically ranging from several hundred nanograms to several micrograms, depending on the requirements of the library preparation kit and the desired sequencing coverage. For severely degraded FFPE DNA, shorter fragments are allowed in the library preparation process, but they must meet the minimum molecular length requirements of the library preparation kit; for example, the main fragment should not be shorter than 200 base pairs, otherwise it may lead to decreased ligation efficiency and insufficient library yield.

[0046] In the ligation-based library preparation step, this invention employs a ligase-based library construction strategy, mainly including DNA end repair, dA tail addition, adapter ligation, and purification. Compared to transposase-based library construction, ligation can more completely preserve the methylation modification information of DNA molecules, as transposase-based methods may exhibit bias or introduce bias in certain cases regarding modified bases. During library construction, barcodes can be added as needed to enable mixed sequencing of multiple samples. For emergency samples requiring rapid results, it is recommended to use independent library construction and sequencing for each sample to avoid waiting for sample collection; for routine samples or large sample volumes, a barcode reuse and mixing strategy can be used to improve sequencing throughput and reduce per-sample cost. After library construction, it is recommended to retain a portion of the library as a backup library for subsequent flow cell washing and re-insertion if needed. For example, if the amount of data obtained from the initial sequencing is insufficient to reach the genotyping confidence threshold, the used flow cells can be washed to remove residual molecules before re-inserting the backup library to supplement coverage and improve genotyping confidence. This mechanism improves the efficiency of single-use consumables without increasing flow tank consumption, which helps control testing costs.

[0047] In the nanopore sequencing and adaptive sampling targeted enrichment steps, this invention employs a nanopore sequencing platform for real-time sequencing. After sequencing begins, the system performs real-time base recognition and simultaneously identifies 5mC modifications. For FFPE samples, due to the high degree of DNA fragmentation and the often unstable proportion of effective information, this invention preferably enables the adaptive sampling function. The specific implementation is as follows: A set of "diagnosis-related CpG target regions" is provided in advance. This set is described in BED format and contains the genomic regions required for molecular subtyping of central nervous system tumors. These regions can be determined based on existing literature, public databases, or self-built reference sets, covering key methylation differential sites or regions for each major CNS tumor type. During sequencing, the nanopore sequencing device rapidly determines the leading signal or sequence of each DNA molecule entering the pore and compares it with the target region set in real time. If the molecule is determined to originate from a target region, sequencing continues and the read is retained; if the molecule is determined to originate from a non-target region, it is ejected from the pore by applying a reverse voltage, and the next molecule is then read. Through this mechanism, this invention can significantly increase the proportion of effective data in the target region under the same sequencing time and throughput, concentrating limited sequencing data on key diagnostic regions, thereby obtaining reportable methylation typing results even under conditions of lower coverage. This is particularly beneficial for FFPE samples, effectively overcoming the instability of their effective information ratio and improving detection success rate. The set of target regions for adaptive sampling can be dynamically adjusted as needed, for example, expanding or optimizing the region list as new research evidence emerges, keeping the detection content up-to-date.

[0048] During sequencing, this invention also supports setting dynamic termination conditions. Dynamic termination conditions can include reaching a minimum data volume requirement, such as generating a certain amount of BAM or FASTQ format data, or automatically terminating sequencing when the methylation typing confidence level output by any one or more classifiers reaches a preset threshold. The confidence threshold can be set according to actual needs, for example, to 90% or 95%. The dynamic termination mechanism helps to minimize sequencing time and control sequencing costs while ensuring typing accuracy. For samples with higher data volume requirements, such as samples requiring SNV detection, sequencing can continue until a preset coverage threshold is reached, and coverage can be improved by combining flow cell washing and library reloading operations.

[0049] In the real-time analysis and quality monitoring steps, this invention constructs an integrated analysis pipeline. This pipeline runs synchronously with the sequencing process, continuously receiving newly generated sequencing data and processing it in real time. Real-time processing includes base identification, 5mC modification identification, and sequence alignment with a reference genome. After alignment, methylation features are summarized by CpG sites or regions. The summarized methylation features are kept consistent with the target region set to facilitate subsequent analysis. Based on the summarized methylation features, this invention calls one or more machine learning classifiers to output tumor categories and confidence scores. The machine learning classifiers can be built based on algorithms such as random forests, support vector machines, and neural networks. The training set can use a publicly available methylation reference database or a self-built local dataset. The results of multiple classifiers can be integrated through threshold determination, voting mechanisms, or the highest score rule to form the final classification conclusion. Concurrently, the analysis pipeline performs copy number variation analysis, segmenting sequencing reads based on their coverage depth to generate genome-wide copy number profiles and extracting key diagnostic features, such as chromosome 7 gain and chromosome 10 deletion associated with glioblastoma, and co-deletion of the short arm of chromosome 10 and the long arm of chromosome 19 associated with oligodendroglioma. When sequencing coverage reaches a preset threshold, the analysis pipeline can also selectively initiate single nucleotide variant and small fragment insertion / deletion detection, annotating detected variants and outputting a list of potentially pathogenic or actionable mutations, such as BRAF gene V600E mutations, PTEN gene mutations, NF1 gene mutations, and TP53 gene mutations. Furthermore, prediction of the methylation status of the MGMT promoter based on the methylation level of relevant regions can be performed as an optional output.

[0050] The analytical pipeline of this invention also incorporates consistency interpretation and remedial rules. Specifically, when the methylation typing result reaches the confidence threshold, the system directly outputs the methylation category as the primary conclusion; when the methylation typing result does not reach the confidence threshold, the system automatically combines key features extracted from copy number variation analysis with original pathological information to form an integrated diagnostic suggestion or remedial interpretation, and clearly indicates the source of evidence and confidence level in the final report. For example, if the methylation typing confidence is insufficient but the copy number profile shows typical co-deletion features on the short arm of chromosome 1 and the long arm of chromosome 19, the system can indicate "Methylation typing did not reach the threshold, but copy number features support the diagnosis of oligodendroglioma," for clinicians' reference. This remedial mechanism helps avoid "unreportable" situations due to insufficient data or sample quality issues, improving the overall diagnostic output and result consistency.

[0051] In the report generation and output steps, this invention generates a structured report. The report includes at least the following: methylation typing results, including classification category, score or confidence level, model version used, and threshold description; copy number variation spectrum and summary of key diagnostic features; coverage and quality control indicators, such as total data volume, average coverage of the target region, and alignment rate; optionally, it includes single nucleotide variant and small fragment insertion / deletion detection results and annotations, as well as MGMT promoter methylation prediction. The report's conclusion section should contain clear and constructive language, such as "reportable," "requires retesting," or "requires comprehensive judgment in conjunction with other evidence," facilitating direct use by clinicians for diagnostic and treatment decisions. The report can be output in PDF format and can also generate a graphical summary for intuitive understanding.

[0052] The technical solution of the present invention will be further described below through specific embodiments.

[0053] Example 1: Detection of FFPE Samples from Glioblastoma

[0054] We received a pathologically diagnosed high-grade glioblastoma, an FFPE sample with wild-type IDH to be excluded, and the tumor cell proportion was assessed to be approximately 80%. DNA was extracted using a commercially available FFPE DNA extraction kit, and the concentration met requirements, with an A260 / A280 ratio of 1.85. Fragment analysis showed approximately 300 base pairs in the main band. Library construction was performed using ligation, and barcode labeling was enabled. Before sequencing, a pre-defined set of CNS tumor diagnostic-related CpG target regions was loaded. This set contains approximately 20,000 target regions, covering key sites for known CNS tumor classification. Adaptive sampling was enabled. After approximately four hours of sequencing, real-time analysis showed a methylation typing confidence level of 95%, clearly classifying it as glioblastoma. Simultaneously, copy number variation analysis showed typical chromosome 7 gain and chromosome 10 deletion features. The system automatically terminated sequencing, generated a report, and the total time was approximately six hours. The report output glioblastoma methylation typing results, copy number spectra, and key feature summaries. Coverage statistics showed an average target region coverage of approximately 15-fold, meeting quality control requirements.

[0055] Example 2: Detection of FFPE samples from oligodendrogliomas

[0056] A FFPE sample suspected of being oligodendroglioma was received, with tumor cells accounting for approximately 60%. DNA was extracted, a library was constructed, and sequencing was performed using adaptive sampling. After three hours of sequencing, the methylation typing confidence level reached 88%, but this did not reach the preset 90% threshold. The system continued sequencing and initiated a salvage interpretation process. Copy number variation analysis showed a clear co-deletion feature on the short arm of chromosome 1 and the long arm of chromosome 19, consistent with typical molecular alterations in oligodendroglioma. After five hours of sequencing, the methylation typing confidence level rose to 92%, reaching the threshold, and the final classification was oligodendroglioma. The report simultaneously output the methylation typing results and copy number characteristics, indicating their consistency. The total time was approximately seven hours.

[0057] Example 3: Detection of frozen samples from medulloblastoma

[0058] We received a frozen tissue sample from a child with medulloblastoma; the tumor cell percentage exceeded 90%. After DNA extraction and library construction, adaptive sampling was not used due to the high sample quality. Two hours after sequencing, the methylation typing confidence level reached 96%, clearly classifying it as WNT-type medulloblastoma. Copy number variation analysis showed no characteristic changes. The system terminated sequencing and generated a report. Because of the high quality of the frozen sample DNA, the data output efficiency was significantly better than that of FFPE samples, further shortening the detection time.

[0059] Example 4: Remedial interpretation of samples with insufficient confidence in methylation typing

[0060] A FFPE sample was received, with a low proportion of tumor cells, approximately 40%. After DNA extraction and library construction, sequencing was performed. Six hours after sequencing, the methylation typing confidence level only reached 75%, consistently failing to reach the threshold. The system automatically activated salvage interpretation rules, extracting features based on copy number variation analysis. This revealed a gain on chromosome 7 and a deletion on chromosome 10, suggesting a possible glioblastoma. The report explicitly stated, "Methylation typing confidence is insufficient, but copy number features support a glioblastoma diagnosis; a comprehensive assessment combining pathological morphology is recommended." This report provided valuable reference information for clinicians, avoiding situations where reports could not be submitted due to insufficient typing confidence.

[0061] Example 5: Multi-classifier ensemble and voting mechanism

[0062] An ensemble system of three classifiers based on random forest, support vector machine, and neural network was constructed. For the same test sample, the three classifiers output classification results: random forest classifies it as IDH wild-type glioblastoma with a confidence level of 0.93; support vector machine classifies it as IDH wild-type glioblastoma with a confidence level of 0.91; and neural network classifies it as IDH wild-type glioblastoma with a confidence level of 0.95. The system employs a voting mechanism; if the three classifiers produce consistent results, it directly outputs IDH wild-type glioblastoma. If discrepancies exist, the highest confidence rule or a weighted voting mechanism can be used to determine the final classification. The classifier ensemble method is indicated in the report to improve the reliability of the results.

[0063] Example 6: Dynamic Termination Conditions and Cost Control

[0064] The dynamic termination conditions are set as follows: methylation genotyping confidence reaches 90% or sequencing time reaches 12 hours. For a high-confidence sample, the confidence level reaches 90% after three hours of sequencing, and the system automatically terminates, effectively avoiding unnecessary sequencing time. For a difficult sample, if the confidence level is still only 85% after 12 hours of sequencing, the system terminates according to the preset maximum sequencing time and initiates salvage readout rules, outputting integrated suggestions based on copy number characteristics. This mechanism ensures detection quality while minimizing sequencing costs.

[0065] Example 7: Flow tank washing and reloading

[0066] In one case, six hours after the initial sequencing, insufficient library loading resulted in lower-than-expected data output, with methylation typing confidence reaching only 80%. The reserved backup library was washed and reused in the flow cell for another four hours of sequencing, increasing the total data volume by approximately 50% and raising the methylation typing confidence to 93%, meeting the reportable standard. This mechanism replenished the data volume without increasing flow cell consumption, thus improving the detection success rate.

[0067] Example 8: Enabling SNV Detection Function

[0068] Eight hours after sequencing a sample, the methylation typing confidence level reached 95%, and the average target region coverage reached 50-fold, meeting the preset threshold for SNV detection. The system automatically initiated the variant detection process, detecting the BRAF gene V600E mutation, which was confirmed as a known pathogenic mutation through annotation. The report, based on the methylation typing and copy number results, supplements the output with this SNV detection result, providing clinical reference information for targeted therapy.

[0069] Example 9: MGMT promoter methylation prediction

[0070] Based on the methylation level of CpG sites associated with the MGMT promoter in the target region, the average methylation score was calculated, and methylation positivity or negativity was determined according to a preset threshold. For a glioblastoma sample, the predicted result was MGMT promoter methylation positivity, suggesting potential sensitivity to temozolomide treatment. This predicted result was included as optional content in the report for clinical reference.

[0071] Example 10: System Composition and Workflow

[0072] The system described in this invention includes a sample and DNA preparation unit, a nanopore sequencing unit, a computing and analysis unit, and a report output unit. The sample and DNA preparation unit includes a pathological assessment area, DNA extraction equipment, and quality control instruments; the nanopore sequencing unit includes a nanopore sequencer and related consumables; the computing and analysis unit includes a high-performance server or workstation equipped with real-time analysis software and a classifier model; the report output unit includes report generation software and a printing device. The workflow is as follows: After receiving the sample, the pathologist completes pathological assessment and tumor enrichment in the sample and DNA preparation unit; technicians extract DNA and perform quality control; qualified DNA is transferred to the nanopore sequencing unit for library preparation and sequencing; during sequencing, the computing and analysis unit receives and analyzes data in real time; after reaching the dynamic termination conditions, the report output unit automatically generates a structured report and pushes it to the clinician. The entire process can be completed within 24 hours, supporting on-demand testing of single samples.

[0073] In summary, this invention constructs a rapid methylation typing method and system for central nervous system tumors by integrating nanopore real-time sequencing, adaptive sampling targeted enrichment, multi-classifier integrated analysis, simultaneous CNV detection, and salvage interpretation rules. This method significantly shortens the detection cycle, improves adaptability to routine FFPE pathological materials, achieves integrated output of methylation typing, copy number variation, and optional mutation information, and establishes a clear confidence level interpretation and cost control mechanism, demonstrating good clinical practical value and promising prospects for wider application.

[0074] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A rapid methylation typing method for central nervous system tumors based on nanopore sequencing, characterized in that, Includes the following steps: Sample reception and preparation steps: Receive central nervous system tumor samples, perform pathological evaluation and tumor cell enrichment; DNA extraction and quality control steps: Extract sample DNA and perform quality testing; Library construction steps: Nanopore sequencing libraries were constructed using ligation methods; Nanopore sequencing and adaptive sampling steps: Real-time sequencing was performed using a nanopore sequencing platform, and adaptive sampling was enabled for FFPE samples to enrich a predefined set of diagnostically relevant CpG target regions; Real-time analysis and typing steps: Real-time identification of bases and 5mC modifications, construction of methylation feature maps, and output of methylation typing categories and their confidence levels through multi-classifier integration; Report generation steps: Generate a structured report containing at least methylation typing results.

2. The method according to claim 1, characterized in that, The real-time analysis and classification steps further include: CNV analysis sub-steps: Copy number variation analysis is performed based on the coverage depth of sequencing reads to extract key diagnostic features; Integrated interpretation sub-step: When the confidence level of methylation typing does not reach the preset threshold, an integrated diagnostic suggestion is formed by combining the CNV analysis results and pathological information, and the source of evidence and confidence level are marked in the report.

3. The method according to claim 2, characterized in that, The real-time analysis and typing steps also include an SNV / Indel detection sub-step: when the sequencing coverage reaches a preset threshold, variant detection and annotation are initiated, and a list of potential pathogenic or actionable mutations is output.

4. The method according to claim 1, characterized in that, The nanopore sequencing and adaptive sampling steps are equipped with dynamic termination conditions: sequencing is automatically terminated when the minimum amount of data generated reaches a preset value, or when the confidence level output by any one or more classifiers reaches a preset confidence threshold.

5. The method according to claim 1, characterized in that, The diagnostic-related CpG target region set is described in BED format and includes the genomic regions required for molecular subtyping of central nervous system tumors. The adaptive sampling rapidly determines the read front signal during sequencing. If the target region is hit, sequencing continues and the read is retained. If the target region is not hit, the molecule is ejected by reverse voltage.

6. A rapid methylation typing system for central nervous system tumors based on nanopore sequencing, characterized in that, include: Sample and DNA preparation unit: used for pretreatment, DNA extraction and quality control of central nervous system tumor samples; Nanopore sequencing unit: Supports real-time base recognition and 5mC modification recognition, and has adaptive sampling function for enrichment sequencing of predefined diagnostic-related CpG target regions; The computation and analysis unit communicates with the sequencing unit in real time to receive sequencing data and perform real-time analysis, including methylation feature construction, multi-classifier genotyping, CNV analysis, and integrated interpretation. Report output unit: Used to generate structured reports containing methylation typing results, CNV features, and quality control indicators.

7. The system according to claim 6, characterized in that, The calculation and analysis unit further includes: Methylation typing module: Constructs features based on the methylation level of CpG sites or regions, and calls one or more machine learning classifiers to output tumor category and confidence score; CNV analysis module: Performs copy number variation analysis based on read depth and extracts key diagnostic features; Integrated interpretation module: When the confidence level of methylation typing is insufficient, it outputs remedial interpretation suggestions by combining CNV features and pathological information; It also includes an SNV detection module: when the coverage reaches a threshold, variant detection and annotation are enabled.

8. The system according to claim 6, characterized in that, The nanopore sequencing unit supports single-sample independent operation or barcode-reused mixed-sample operation, and supports flow cell washing and library reloading to supplement coverage.

9. The method according to any one of claims 1 to 5, or the system according to any one of claims 6 to 8, characterized in that, The central nervous system tumor samples include FFPE paraffin tissue, frozen tissue, or pre-extracted DNA; the structured report also includes CNV spectra, coverage and quality control indicators, and optionally includes SNV / Indel results and MGMT predictions.

Images