Method for diagnosing and predicting cancer types using methylated cell-free DNA

Abstract

The present invention relates to a method for cancer diagnosis and cancer type prediction using methylated cell-free nucleic acid, more specifically, to a method for cancer diagnosis and cancer type prediction using a method of extracting methylated nucleic acid from a biological sample, obtaining sequence information, generating vectorized data of nucleic acid fragments based on aligned reads, and inputting the vectorized data into a trained artificial intelligence model to analyze calculated values. The method for cancer diagnosis and cancer type prediction using methylated cell-free nucleic acid according to the present invention is useful because it generates vectorized data and analyzes it using an AI algorithm, compared to a conventional method using a step of determining chromosome amount based on read count or a detection method using a distance concept between aligned reads, which utilizes values ​​related to reads as standardized values ​​one by one, and therefore can achieve the same effect even if the read coverage is low.

Description

Technical Field

[0001] The present invention relates to a method for cancer diagnosis and cancer type prediction using methylated cell-free nucleic acids. More specifically, the method extracts nucleic acids from a biological sample, obtains and aligns sequence information including methylation information, generates vectorized data of nucleic acid fragments based on the aligned reads, and then analyzes the values calculated by inputting this into a learned artificial intelligence model. The present invention relates to a method for cancer diagnosis and cancer type prediction using such a method.

Background Art

[0002] In clinical cancer diagnosis, tissue biopsy is usually performed for confirmation after medical history investigation, physical examination, and clinical evaluation. Cancer diagnosis by clinical experiments is only possible when the number of cancer cells is 1 billion or more and the diameter of the cancer is 1 cm or more. In this case, the cancer cells already have the ability to metastasize, and at least half of them are already in a metastasized state. In addition, tissue biopsy is invasive, causing considerable discomfort to the patient, and there are also quite a few cases where tissue biopsy cannot be performed when treating cancer patients. In addition, in cancer screening, tumor markers for monitoring substances produced directly or indirectly from cancer are used. However, even when cancer is present, more than half of the results of tumor marker screening are normal, and even when there is no cancer, they are frequently positive, so there are limitations in its accuracy.

[0003] Due to the need for a cancer diagnosis method that is relatively simple, non-invasive, and has high sensitivity and specificity to complement the problems of such conventional cancer diagnosis methods, recently, liquid biopsy that utilizes a patient's body fluid for cancer diagnosis and follow-up examinations has been widely used. Liquid biopsy is a non-invasive method and is a diagnostic technique attracting attention as an alternative to conventional invasive diagnosis and examination methods.

[0004] Recently, methods have been developed to diagnose cancer and differentiate cancer types using cell-free DNA obtained from liquid biopsies (US 10975431, Zhou, Xionghui et al., bioRxiv, 2020.07.16.201350). In particular, methods for diagnosing / determining cancer type using the methylation pattern of cell-free nucleic acids are known (Li, Jiaqi et al., bioRxiv, 2021.01.12.426440, US 2020-0131582, KR 10-2148547).

[0005] On the other hand, an artificial neural network (AVM) is a computational model implemented in software or hardware that mimics the computational power of a biological system using a large number of artificial neurons connected by connecting lines. In an AVM, artificial neurons with simplified functions from biological neurons are used. These neurons are interconnected via connecting lines that have connection strengths, and they perform human cognitive functions and learning processes. Connection strength is a specific value that each connecting line possesses, and is also called connection weight. Learning in an AVM can be divided into guided learning and unguided learning. Guided learning is a method in which input data and corresponding output data are fed together into the neural network, and the connection strengths of the connecting lines are updated so that output data corresponding to the input data is output. Representative learning algorithms include the Delta Rule and Backpropagation Learning. Unguided learning is a method in which the AVM learns connection strengths on its own using only input data without target values. Unguided learning is a method in which the connection weights are updated based on the correlation between input patterns.

[0006] Many of the datasets applied in machine learning become complex, and as the dimensionality increases, the curse of dimensionality arises. This means that as the dimensionality of the required data becomes infinite, the distance between any two points diverges to infinity, and the abundance, or density, of the data becomes somewhat lower in high-dimensional space, making it difficult to adequately reflect the characteristics (features) of the data (Richard Bellman, Dynamic Programming, 2003, chapter 1). Recently, the development of deep neural networks has been reported to have significantly improved the performance of classifiers for high-dimensional data such as images, videos, and signal data by using a structure with a hidden layer between the input and output layers, and by processing the linear combination of variable values ​​transmitted from the input layer with a nonlinear function (Hinton, Geoffrey, et al., IEEE Signal Processing Magazine Vol. 29.6, pp. 82-97, 2012).

[0007] There are various patents (KR 10-2018-0124550, KR 10-2019-7038076, KR 10-2019-0003676, KR 10-2019-0001741) that utilize such artificial neural networks in the bio-field, and the inventors have filed a patent application (KR 10-2021-0067931) for a method of detecting chromosomal abnormalities through artificial neural network analysis based on sequence analysis information of cell-free DNA (cfDNA) in blood. However, there have been no cases of imaging and analyzing information on methylated cell-free nucleic acids, nor have there been any cases of representing the methylated pattern of full-length dielectric units.

[0008] Therefore, the inventors have made diligent efforts to solve the aforementioned problems and develop an artificial intelligence-based cancer diagnostic method with high sensitivity and accuracy. As a result, they have found that by generating vectorized data based on the distance or amount of methylated cell-free nucleic acid fragments and analyzing this data with a trained artificial intelligence model, cancer diagnosis and cancer type discrimination can be performed with high sensitivity and accuracy, thus completing the present invention. [Overview of the project] [Problems that the invention aims to solve]

[0009] The objective of the present invention is to provide a method for cancer diagnosis and cancer type prediction using methylated cell-free nucleic acids. Another object of the present invention is to provide a cancer diagnostic and cancer type prediction device using methylated cell-free nucleic acids.

[0010] Another object of the present invention is to provide a computer-readable storage medium that includes instructions configured to be executed by a processor that predicts cancer diagnosis and cancer type by the method described above. [Means for solving the problem]

[0011] To achieve the above objective, the present invention provides a method for providing information for cancer diagnosis and cancer type prediction, comprising the steps of: (a) extracting nucleic acids from a biological sample to obtain sequence information including methylation information; (b) aligning the obtained sequence information (reads) in a reference genome database; (c) generating vectorized data using nucleic acid fragments based on the aligned sequence information (reads); (d) inputting the generated vectorized data into a trained artificial intelligence model and comparing the analyzed output value with a cut-off value to determine the presence or absence of cancer; and (e) predicting the type of cancer through the comparison of the output value.

[0012] The present invention also provides a cancer diagnosis and cancer type prediction device that includes: a decoding unit that extracts nucleic acids from a biological sample and decodes sequence information including methylation information; an alignment unit that aligns the decoded sequence in a standard chromosome sequence database; a data generation unit that generates vectorized data using nucleic acid fragments based on the aligned sequence; a cancer diagnosis unit that inputs the generated vectorized data into a trained artificial intelligence model for analysis and determines the presence or absence of cancer by comparing it with reference values; and a cancer type prediction unit that analyzes the output result values ​​to predict the type of cancer.

[0013] The present invention also provides a computer-readable storage medium comprising instructions configured to be executed by a processor that predicts cancer diagnosis and cancer type, the instructions comprising: (a) extracting nucleic acids from a biological sample to obtain sequence information including methylation information; (b) aligning the obtained sequence information (reads) in a reference genome database; (c) generating vectorized data using nucleic acid fragments based on the aligned sequence information (reads); (d) inputting the generated vectorized data into a trained artificial intelligence model for analysis and comparing it to a cut-off value to determine the presence or absence of cancer; and (e) predicting the cancer type through comparison of the output values, thereby providing a computer-readable storage medium comprising instructions configured to be executed by a processor that predicts the presence or absence of cancer and cancer type.

[0014] The present invention also provides a method for diagnosing and predicting cancer, which includes the steps of: (a) extracting nucleic acids from a biological sample to obtain sequence information including methylation information; (b) aligning the obtained sequence information (reads) in a reference genome database; (c) generating vectorized data using nucleic acid fragments based on the aligned sequence information (reads); (d) inputting the generated vectorized data into a trained artificial intelligence model and comparing the analyzed output values ​​with a cut-off value to determine the presence or absence of cancer; and (e) predicting the type of cancer through the comparison of the output values. [Brief explanation of the drawing]

[0015] [Figure 1] This is an overall flowchart for determining chromosomal abnormalities using the artificial intelligence platform of the present invention. [Figure 2] This is an example of a GC plot generated based on methylated cfDNA according to one embodiment of the present invention, where the X-axis represents chromosomes by segment and the Y-axis represents the number of nucleic acid fragments corresponding to each segment. [Figure 3] This is a result of confirming the accuracy of neuroblastoma diagnosis using a deep learning model trained on GC plot image data generated based on the number of nucleic acid fragments using methylated cfDNA, according to one embodiment of the present invention. [Figure 4] In one embodiment of the present invention, the following results show the probability distribution for neuroblastoma diagnosis for each dataset in a deep learning model trained on GC plot image data generated based on the number of nucleic acid fragments using methylated cfDNA, where (A) represents the Training set, (B) represents the Validation set, and (C) represents the Test set. [Figure 5] This is an example of a GC plot generated based on cfDNA according to one embodiment of the present invention, where the X-axis represents chromosomes by segment and the Y-axis represents the number of nucleic acid fragments corresponding to each segment. [Figure 6] This is a result of confirming the accuracy of neuroblastoma diagnosis using a deep learning model trained on GC plot image data generated based on the number of nucleic acid fragments using cfDNA, according to one embodiment of the present invention. [Figure 7] In one embodiment of the present invention, the results show the probability distribution for neuroblastoma diagnosis for each dataset for a deep learning model trained on GC plot image data generated based on the number of nucleic acid fragments using cfDNA, where (A) represents the Training set, (B) represents the Validation set, and (C) represents the Test set. [Modes for carrying out the invention]

[0016] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by skilled experts in the art to which this invention pertains. In general, the nomenclature used herein and the experimental methods described below are well known and commonly used in the art.

[0017] The terms First, Second, A, B, etc., may be used to describe various constituent elements, but the constituent elements are not limited by such terms and are used solely for the purpose of distinguishing one constituent element from another. For example, without exceeding the scope of rights of the technology described below, the First constituent element may be named the Second constituent element, and similarly, the Second constituent element may be named the First constituent element. The terms and / or include combinations of multiple related descriptions or any of multiple related descriptions.

[0018] In the terms used in this specification, a singular expression should be understood to include plural expressions unless the context clearly dictates otherwise. Terms such as "including" mean that the recited features, numbers, steps, actions, components, parts, or combinations thereof exist, and do not exclude the possibility of the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0019] Prior to providing a detailed description of the drawings, it should be made clear that the classification of the components in this specification is merely based on the main functions each component is responsible for. That is, two or more of the components described below may be combined into one component, or one component may be further divided into two or more components according to more refined functions. And each of the components described below may further perform some or all of the functions of other components in addition to its own main function. It is of course possible that a part of the function of the main function each component is responsible for is solely performed by other components.

[0020] Also, when implementing a method or an operation method, each process constituting the method may be performed in an order different from the specified order unless the context clearly dictates a specific order. That is, each process may be performed in the same order as the specified order, may be performed substantially simultaneously, or may be performed in the reverse order.

[0021] In the present invention, when detecting cancer by aligning sequence analysis data obtained from methylated cell-free nucleic acids extracted from a sample to a reference genome, generating vectorized data based on the aligned nucleic acid fragments, calculating DPI values with a learned artificial intelligence model, and comparing them with a reference value, it was attempted to confirm that cancer can be detected with high sensitivity and accuracy.

[0022] That is, in one embodiment of the present invention, after sequencing DNA extracted from blood to include methylation information and then aligning it to a reference chromosome, the distance or amount between nucleic acid fragments is calculated for each fixed chromosomal interval, and vectorized data is generated with each genetic region on the X-axis and the distance or amount between nucleic acid fragments on the Y-axis. Then, this is learned by a deep learning model to calculate the DPI value. When the DPI value is greater than or equal to the reference value, it is determined that there is cancer, and among a large number of DPI values, the cancer type showing the highest value is determined as the actual cancer type (Figure 1).

[0023] Therefore, from one aspect, the present invention (a) A step of extracting nucleic acids from a biological sample to obtain sequence information including methylation information; (b) A step of aligning the obtained sequence information (reads) to a standard chromosome sequence database (reference genome database); (c) A step of generating vectorized data using nucleic acid fragments (fragments) based on the aligned sequence information (reads); (d) A step of inputting the generated vectorized data into a learned artificial intelligence model, analyzing, comparing the output result value with a reference value (cut-off value), and determining the presence or absence of cancer; and (e) A step of predicting the cancer type through the comparison of the output result value, relating to a method for providing information for cancer diagnosis and cancer type prediction including the above steps.

[0024] In the present invention, the nucleic acid fragment may be used without limitation as long as it is a fragment of nucleic acid extracted from a biological sample, preferably a fragment of cell-free nucleic acid or intracellular nucleic acid, but is not limited thereto.

[0025] In the present invention, the nucleic acid fragments can be obtained by any method known to the ordinary art, preferably by direct sequencing, sequencing via next-generation sequencing, sequencing via non-specific whole genome amplification, or sequencing via probe-based sequencing, but are not limited thereto. In this invention, the nucleic acid fragment may also mean a read when next-generation sequencing analysis is used.

[0026] In the present invention, the cancer may be a solid tumor or a hematological cancer, and may be preferably selected from the group consisting of non-Hodgkin lymphoma, Hodgkin lymphoma, acute myeloid leukemia, acute lymphoblastic leukemia, multiple myeloma, head and neck cancer, lung cancer, glioblastoma, colorectal cancer, pancreatic cancer, breast cancer, ovarian cancer, melanoma, prostate cancer, liver cancer, thyroid cancer, gastric cancer, gallbladder cancer, biliary tract cancer, bladder cancer, small intestine cancer, cervical cancer, cancer of unknown primary site, kidney cancer, esophageal cancer, neuroblastoma, and mesothelioma, and more preferably neuroblastoma, but is not limited thereto.

[0027] In the present invention, Step (a) above is, (ai) A step of obtaining nucleic acids containing methylation information from a biological sample; (a-ii) The step of removing proteins, fats, and other residues from the collected nucleic acids using the salting-out method, column chromatography method, or beads method to obtain purified nucleic acids; (a-iii) A step of preparing a single-end sequencing or pair-end sequencing library from purified nucleic acids or nucleic acids randomly fragmented by enzymatic cleavage, disruption, or hydroshear method; (a-iv) The step of reacting the prepared library with a next-generation sequencer; and (av) The procedure may be characterized by including the step of obtaining nucleic acid sequence information (reads) using a next-generation sequencer.

[0028] In the present invention, the step of obtaining sequence information in step (a) may be characterized by obtaining the sequence information from the isolated cell-free DNA by full-length genome sequencing at a depth of 1 million to 100 million reads, but is not limited thereto.

[0029] In the present invention, the term "biological sample" means any substance, biological fluid, tissue, or cell obtained from or derived from an individual, such as whole blood, leukocytes, peripheral blood mononuclear cells, leukocyte buffy coat, blood (including plasma and serum), sputum, tears, mucus, nasal washes, nasal aspirate, breath, urine, semen, saliva, peritoneal washings, pelvic fluids, cystic fluid, meningeal fluid, amniotic fluid, glandular fluid, and pancreatic fluid. This may include, but is not limited to, lymphatic fluid, pleural fluid, nipple aspirate, bronchial aspirate, synovial fluid, joint aspirate, organ secretions, cells, cell extract, hair, oral cells, placental cells, cerebrospinal fluid, and mixtures thereof.

[0030] In this invention, the term "reference population" refers to a reference population that can be compared, such as a standard nucleotide sequence database, and means a population of people who are currently free from a specific disease or condition. In this invention, the standard nucleotide sequence in the standard chromosome sequence database of the reference population may be a reference chromosome registered with a public health organization such as the NCBI.

[0031] In the present invention, the nucleic acid in step (a) may be cell-free DNA, and more preferably circulating tumor DNA (ctDNA), but is not limited thereto.

[0032] In the present invention, the nucleic acid containing the methylation information can be obtained by various known methods, and may preferably be characterized by being obtained by bisulfite conversion, enzyme conversion, or methylated DNA immunoprecipitation (MeDIP), but is not limited thereto.

[0033] In the present invention, there are further restriction enzyme-based detection methods that can detect DNA methylation. These methods utilize methylation restriction enzymes (MREs) to cleave unmethylated nucleic acids, or to cleave specific sequences (recognition sites) regardless of methylation, and then analyze them in combination with hybridization or PCR.

[0034] In the present invention, methods based on bisulfite substitution include whole-genome bisulfite sequencing (WGBS), reduced-representation bisulfite sequencing (RRBS), methylated CpG tandem amplification and sequencing (MCTA-seq), targeted bisulfite sequencing, and methylation array and methylation-specific PCR (MSP).

[0035] In the present invention, methods for enriching and analyzing methylated DNA include methylated DNA immunoprecipitation sequencing (MeDIP-seq) and methyl-CpG binding domain protein capture sequencing (MBD-seq).

[0036] In the present invention, other methods for analyzing methylated DNA include 5-hydroxymethylation profiling, such as 5hmC-Seal (hMe-Seal), hmC-CATCH, hydroxymethylated DNA immunoprecipitation sequencing (hMeDIP-seq), and oxidative bisulfite conversion.

[0037] In the present invention, the next-generation sequencer may be used with any sequencing method known in the art. Sequencing of nucleic acids isolated by the selection method is usually performed using next-generation sequencing (NGS). Next-generation sequencing includes any sequencing method that determines the single nucleotide sequence of individual nucleic acid molecules or, in a very similar manner, a cloned proxy for individual nucleic acid molecules (e.g., 10⁵ or more molecules are sequenced simultaneously). In one embodiment, the relative abundance of nucleic acid species in a library can be estimated by measuring the relative occurrence of their congeneral sequences in the data produced by the sequencing experiment. Next-generation sequencing methods are known in the art and are described, for example, in the literature included herein by reference (Metzker, M. (2010) Nature Biotechnology Reviews 11:31-46).

[0038] In one embodiment, next-generation sequencing is performed to determine the nucleotide sequences of individual nucleic acid molecules (e.g., HeliScope Gene Sequencing system from Helicos BioSciences and PacBio RS system from Pacific Biosciences). In another embodiment, sequencing, for example, large-scale parallel short-read sequencing (e.g., Solexa sequencer from Illumina Inc. in San Diego, California), which produces more bases per sequencing unit than other sequencing methods that produce fewer but longer reads, determines the nucleotide sequences of cloned proxies for individual nucleic acid molecules (e.g., Solexa sequencer from Illumina Inc. in San Diego, California; 454 Life Sciences (located in Branford, Connecticut) and Ion Torrent). Other methods or machines for next-generation sequencing are provided by, but are not limited to, 454 Life Sciences (located in Branford, Connecticut), Applied Biosystems (located in Foster City, California; SOLiD sequencers), Helicos Biosciences Corporation (located in Cambridge, Massachusetts), and emulsion and microflow sequencing methods, including nanoinfusions (e.g., GnuBio infusions).

[0039] Platforms for next-generation sequencing include, but are not limited to, the Roche / 454 Genome Sequencer (GS) FLX system, the Illumina / Solexa Genome Analyzer (GA), the Life / APG Support Oligonucleotide Ligation Detection (SOLiD) system, the Polonator G.007 system, the HeliScope Gene Sequencing system from Helicos BioSciences, the PromethION, GriION, and MinION systems from Oxford Nanopore Technologies, and the PacBio RS system from Pacific Biosciences.

[0040] In the present invention, the sequence sorting in step (b) includes, as a computer algorithm, a computer method or approach used for identity where read sequences in the genome (e.g., short read sequences from next-generation sequencing) may largely derive from evaluating the similarity between the read sequences and the reference sequences. Various algorithms can be applied to the sequence sorting problem. Some algorithms are relatively slow but allow for relatively high specificity. These include, for example, algorithms based on dynamic programming. Dynamic programming is a method of solving complex problems by breaking them down into simpler steps. Other approaches are relatively more efficient but generally less thorough. These include, for example, heuristic algorithms and probabilistic methods designed for large-scale database searches.

[0041] Typically, the alignment process may have two steps: candidate screening and sequence alignment. Candidate screening reduces the search space for sequence alignment from the entire genome to a shorter enumeration of possible alignment locations. As the term suggests, sequence alignment involves the step of aligning sequences with the sequences provided in the candidate screening step. This may be done using broad alignment (e.g., Needleman-Wunsch alignment) or local alignment (e.g., Smith-Waterman alignment).

[0042] Most attribute sorting algorithms can feature one of three types based on their indexing method: hash table (e.g., BLAST, ELAND, SOAP), suffix tree (e.g., Bowtie, BWA), and merge sort (e.g., Slider). Short read arrays are commonly used for sorting.

[0043] In the present invention, the sorting step of step (b) may be performed using the BWA algorithm and the Hg19 array, although this is not limited to the above. In the present invention, the BWA algorithm may include, but is not limited to, BWA-ALN, BWA-SW, or Bowtie2.

[0044] In the present invention, the length of the sequence information (reads) in step (b) is 5 to 5000 bp, and the number of sequence information entries used may be 5,000 to 5 million, but is not limited thereto. In the present invention, the vectorized data in step (c) may be any vectorized data that can be generated based on aligned nucleic acid fragments, and may preferably be characterized by being a Grand Canyon plot (GC plot), but is not limited thereto.

[0045] In the present invention, vectorized data is not limited to this, but may preferably be characterized by being image-based. Images are basically composed of pixels, and when an image composed of pixels is vectorized, it can be represented as a one-dimensional 2D vector (black and white), a three-dimensional 2D vector (color (RGB)), or a four-dimensional 2D vector (color (CMYK)), depending on the type of image.

[0046] The vectorized data of the present invention is not limited to images; for example, multiple images of n black and white can be superimposed to create an n-dimensional 2D vector (Multi-dimensional Vector) which can then be used as input data for an artificial intelligence model.

[0047] The GC plot in this invention is a plot generated by setting a specific interval (a fixed bin or bins of different sizes) as the X-axis and a numerical value that can be represented by nucleic acid fragments, such as the distance or number between nucleic acid fragments, as the Y-axis.

[0048] The present invention may further include a step of separately classifying nucleic acid fragments that satisfy the mapping quality score of the aligned nucleic acid fragments before performing step (c).

[0049] In the present invention, the mapping quality score may vary depending on the desired criteria, but is preferably 15 to 70 points, more preferably 50 to 70 points, and most preferably 60 points.

[0050] In the present invention, the GC plot in step (c) is characterized by generating vectorized data by calculating the number of nucleic acid fragments in each chromosome segment or the distance between nucleic acid fragments, representing the chromosomal segmental distribution of aligned nucleic acid fragments. The method for vectorizing the number of nucleic acid fragments or the calculated distance between nucleic acid fragments in this invention may be any known technique for vectorizing calculated values, without limitation.

[0051] In the present invention, calculating the chromosomal segment-specific distribution of the aligned sequence information in terms of the number of nucleic acid fragments can be characterized by including the following steps: i) A step of dividing the chromosome into specific intervals (bins); ii) A step of determining the number of nucleic acid fragments arranged in each interval; iii) The step of normalizing by dividing the number of nucleic acid fragments determined in each interval by the total number of nucleic acid fragments in the sample; and iv) A step to generate a GC plot, where the order of each interval is the value on the X axis and the normalized value calculated in step iii) is the value on the Y axis.

[0052] In the present invention, calculating the chromosomal segment-specific distribution of the aligned sequence information using the distance between nucleic acid fragments can be characterized by including the following steps: i) A step of dividing the chromosome into specific intervals (bins); ii) A step of calculating the distance (Fragments Distance, FD) values ​​between nucleic acid fragments aligned in each interval; iii) A step of determining a representative distance (RepFD) for each interval based on the distance values ​​calculated for each interval; iv) A step of normalizing the representative value calculated in step iii) by dividing it by the representative value of the total nucleic acid fragment distance; and iv) A step to generate a GC plot, where the order of each interval is the value on the X axis, and the normalized value calculated in step iv) is the value on the Y axis.

[0053] In the present invention, the GC plot can generate a single image by aligning the GC plots of chromosomes 1 to 22 along the Y-axis, or it can use images generated for chromosomes 1 to 22 aligned along the Z-axis.

[0054] In the present invention, the representative value (RepFD) may be characterized by one or more selected from the group consisting of the sum, difference, product, mean, median, quantile, minimum, maximum, variance, standard deviation, absolute deviation of the median, coefficient of variation, reciprocal values ​​thereof, and combinations thereof, but is not limited thereto. In the present invention, the aforementioned fixed interval (bin) may be characterized as being 1Kb to 3Gb, but is not limited thereto.

[0055] In the present invention, a step of grouping nucleic acid fragments may be further used, in which case the grouping criterion may be based on the adapter sequences of the aligned nucleic acid fragments. By separately separating the nucleic acid fragments into forward-aligned and reverse-aligned nucleic acid fragments, the distance between nucleic acid fragments can be calculated with respect to the selected sequence information.

[0056] In the present invention, the FD value can be defined as the distance between the reference value of the i-th nucleic acid fragment and any one or more nucleic acid fragments selected from the i+1 to n-th nucleic acid fragments, for the n nucleic acid fragments obtained.

[0057] In the present invention, the FD value may be calculated by determining the distance between the n obtained nucleic acid fragments and a reference value of any one or more nucleic acid fragments selected from the group consisting of the first nucleic acid fragment and the second to nth nucleic acid fragments, and using as the FD value one or more values ​​and / or one or more inverse values ​​thereof, selected from the group consisting of the sum, difference, product, mean, logarithm of the product, logarithm of the sum, median, quantile, minimum, maximum, variance, standard deviation, absolute deviation of the median, and coefficient of variation, along with weights, as a calculation result and / or statistical value including these, but is not limited to these.

[0058] In this invention, the phrase "one or more values ​​and / or one or more inverse values ​​thereof" is interpreted to mean that one or more of the aforementioned numerical values ​​may be used in combination.

[0059] In the present invention, the "reference value of nucleic acid fragments" is characterized by being a value obtained by adding or subtracting an arbitrary value from the median value of nucleic acid fragments. The aforementioned FD value can be defined as follows for the n nucleic acid fragments obtained. FD=Dist(Ri~Rj) (1 <i<j<n)

[0060] Here, the Dist function calculates statistical results including, and not limited to, one or more values ​​selected from a group consisting of the sum, difference, product, mean, logarithm of the product, logarithm of the sum, median, quantile, minimum, maximum, variance, standard deviation, absolute deviation of the median, and coefficient of variation, and / or one or more inverses thereof, as well as weights.

[0061] In other words, in this invention, the FD value (Fragment Distance Value) means the distance between arranged nucleic acid fragments. Here, the number of selection cases for nucleic acid fragments for distance calculation can be defined as follows: When there are a total of N nucleic acid fragments, JPEG0007886409000001.jpg 10170 possible combinations of distances between nucleic acid fragments. That is, if i is 1, then i+1 is 2, and the distance to any one or more nucleic acid fragments selected from the 2nd to nth nucleic acid fragments can be defined.

[0062] In the present invention, the FD value can be characterized by calculating the distance between a specific position within the i-th nucleic acid fragment and a specific position within any one or more of the i+1 to n-th nucleic acid fragments.

[0063] For example, if a nucleic acid fragment is 50 bp long and aligned at position 4,183 on chromosome 1, then the genetic position values ​​that can be used to calculate the distance of this nucleic acid fragment are from 4,183 to 4,232 on chromosome 1.

[0064] When the aforementioned nucleic acid fragment and an adjacent 50 bp long nucleic acid fragment are aligned at position 4,232 of chromosome 1, the genetic position values ​​that can be used to calculate the distance between these nucleic acid fragments are from position 4,232 to 4,281 of chromosome 1, and the FD value between the two nucleic acid fragments may be between 1 and 99.

[0065] When another adjacent 50 bp long nucleic acid fragment is aligned at position 4123 of chromosome 1, the genetic position values ​​that can be used to calculate the distance of this nucleic acid fragment are 4,123 to 4,172 on chromosome 1, the FD value between the two nucleic acid fragments is 61 to 159, and the FD value with the first example nucleic acid fragment is 12 to 110. The FD value may be used as a calculated result and / or statistical value including weights, one or more values ​​and / or one or more reciprocals thereof selected from the group consisting of the sum, difference, product, mean, logarithm of product, logarithm of sum, median, quantile, minimum, maximum, variance, standard deviation, absolute deviation of median, and coefficient of variation of either of the two FD value ranges, preferably characterized by being the reciprocal of a value in either of the two FD value ranges.

[0066] Preferably, in the present invention, the FD value is characterized by being a value obtained by adding or subtracting an arbitrary value from the median value of the nucleic acid fragment.

[0067] In this invention, the median of FD refers to the value that is furthest in the middle when the calculated FD values ​​are arranged in order of magnitude. For example, if there are three values ​​such as 1, 2, and 100, 2 is the value that is furthest in the middle, so 2 is the median. If there is an even number of FD values, the median is determined by the average of the two values ​​that are furthest in the middle. For example, if there are FD values ​​of 1, 10, 90, and 200, the median is the average of 10 and 90, which is 50.

[0068] In the present invention, the above-mentioned arbitrary values ​​may be used without limitation as long as they indicate the position of the nucleic acid fragment, but are preferably 0 to 5 kbp or 0 to 300% of the nucleic acid fragment length, 0 to 3 kbp or 0 to 200% of the nucleic acid fragment length, 0 to 1 kbp or 0 to 100% of the nucleic acid fragment length, and more preferably 0 to 500 bp or 0 to 50% of the nucleic acid fragment length, but are not limited to these.

[0069] In the present invention, the FD value is characterized in that, in the case of paired-end sequencing, it is derived based on the position values ​​of the forward and reverse sequence information (reads).

[0070] For example, in a pair of 50 bp long paired-end reads, if the forward read is aligned to position 4183 on chromosome 1 and the reverse read to position 4349, then the ends of this nucleic acid fragment will be 4183 and 4349, and the reference value that can be used for the nucleic acid fragment distance is 4183 to 4349. At this time, in another paired-end read pair adjacent to the aforementioned nucleic acid fragment, if the forward read is aligned to position 4349 on chromosome 1 and the reverse read to position 4515, then the position value of this nucleic acid fragment is 4349 to 4515. The distance between these two nucleic acid fragments may be between 0 and 333, and most preferably, it may be 166, which is the median distance between each nucleic acid fragment.

[0071] In the present invention, when obtaining sequence information by paired-end sequencing, the invention may further include a step of excluding nucleic acid fragments in the calculation process if the alignment score of the sequence information (reads) is less than a reference value.

[0072] In the present invention, the FD value is characterized in that, in the case of single-end sequencing, it is derived based on a type of position value of forward or reverse sequence information (read).

[0073] In the present invention, in the case of single-ended sequencing, when deriving a position value based on sequence information aligned in the forward direction, an arbitrary value is added, and when deriving a position value based on sequence information aligned in the reverse direction, an arbitrary value is subtracted. The arbitrary value may be any value that clearly indicates the position of the nucleic acid fragment, but is preferably 0 to 5 kbp or 0 to 300% of the nucleic acid fragment length, 0 to 3 kbp or 0 to 200% of the nucleic acid fragment length, 0 to 1 kbp or 0 to 100% of the nucleic acid fragment length, and more preferably 0 to 500 bp or 0 to 50% of the nucleic acid fragment length, but is not limited to these.

[0074] In this invention, the nucleic acid to be analyzed can be sequenced and represented in units called reads. These reads can be divided into single-end sequencing reads (SE) and paired-end sequencing reads (PE) depending on the sequencing method. An SE read means that either the 5' or 3' of the nucleic acid molecule has been sequenced in a random direction for a certain length, while a PE read means that both the 5' and 3' have been sequenced for a certain length. Due to this difference, it is a well-known fact to the average technician that when sequencing in SE mode, one read is generated from one nucleic acid fragment, while in PE mode, two reads are generated as a pair from one nucleic acid fragment.

[0075] The most ideal method for calculating the precise distance between nucleic acid fragments is to sequence the nucleic acid molecule from beginning to end, align the reads, and use the median (center) of the aligned values. However, technically, this method is currently limited by the limitations of sequencing technology and cost. Therefore, sequencing will be performed using the same methods as SE and PE. In the PE method, the start and end positions of the nucleic acid molecule can be determined, and the precise position (median) of the nucleic acid fragment can be determined through the combination of these values. However, in the SE method, information from only one end of the nucleic acid fragment can be used, which limits the calculation of the precise position (median).

[0076] Furthermore, when calculating the distance of nucleic acid molecules using end information from all reads that have been sequenced (aligned) in both the forward and reverse directions, the presence of the sequencing direction may result in inaccurate values.

[0077] Therefore, for technical reasons related to the sequencing method, the 5' end of the forward read will have a position value smaller than the center position of the nucleic acid molecule, while the 3' end of the reverse read will have a larger value. By utilizing this characteristic, it is possible to estimate a value close to the center position of the nucleic acid molecule by adding an arbitrary value (Extended bp) to the forward read and subtracting it from the reverse read.

[0078] In other words, the arbitrary value (Extended bp) may vary depending on the sample used. In the case of cell-free nucleic acids, the average length of the nucleic acid is said to be around 166 bp, so it can be set to approximately 80 bp. If the experiment is performed through fragmentation equipment (e.g., sonication), the extended bp can be set to about half of the target length set during the fragmentation process.

[0079] In the present invention, the representative value (RepFD) may be characterized by being one or more values ​​selected from the group consisting of the sum, difference, product, mean, median, quantile, minimum, maximum, variance, standard deviation, absolute deviation of the median, and coefficient of variation of the FD values, and / or one or more inverse values ​​thereof. Preferably, it may be characterized by being the median, mean, or inverse value thereof of the FD values, but is not limited thereto.

[0080] In the present invention, the artificial intelligence model in step (d) may be any model that can learn to distinguish between normal images and images with cancer, and is preferably a deep learning model.

[0081] In the present invention, the artificial intelligence model may be any artificial neural network algorithm capable of analyzing vectorized data based on an artificial neural network, but is preferably selected from the group consisting of convolutional neural networks (CNNs), deep neural networks (DNNs), recurrent neural networks (RNNs), and autoencoders, but is not limited thereto.

[0082] In the present invention, the recurrent neural network can be selected from the group consisting of LSTM (Long-short term memory) neural networks, GRU (Gated Recurrent Unit) neural networks, vanilla recurrent neural networks, and attentive recurrent neural networks.

[0083] In the present invention, when the artificial intelligence model is a CNN, the loss function for binary classification may be characterized by being expressed by the following formula 1, and the loss function for multi-class classification may be characterized by being expressed by the following formula 2.

number

number

[0084] In the present invention, the binary classification means that the artificial intelligence model learns to determine whether or not cancer is present, and the multi-class classification means that the artificial intelligence model learns to determine the type of cancer.

[0085] In the present invention, if the artificial intelligence model is a CNN, the learning process can be characterized by including the following steps: i) A step to classify the generated GC plots into training, validation, and test data; In this process, the training data is used to train the CNN model, the validation data is used to verify hyperparameter tuning, and the test data is used to evaluate the performance after the optimal model has been produced. ii) Steps to construct an optimal CNN model through hyperparameter tuning and the learning process; and iii) A step in which the performance of multiple models obtained through hyperparameter tuning is compared using validation data, and the model with the best performance on the validation data is determined to be the optimal model; In the present invention, the hyperparameter tuning process is a process of optimizing the values ​​of multiple parameters (such as the number of convolutional layers, the number of fine layers, and the number of convolutional filters) that constitute the CNN model, and the hyperparameter tuning process is characterized by using Bayesian optimization and grid search.

[0086] In the present invention, the learning process is characterized by optimizing the intrinsic parameters (weights) of the CNN model using predetermined hyperparameters, and when the validation loss begins to increase relative to the training loss, it is determined that the model has become overfitted, and the model learning is interrupted before that point.

[0087] In the present invention, the result value analyzed by the artificial intelligence model in step (d) from the input vectorized data may be any specific score or real number without limitation, and may preferably be a DPI (Deep Probability Index) value, but is not limited thereto.

[0088] In this invention, the Deep Probability Index refers to a value expressed as a probability value obtained by adjusting the output of the artificial intelligence on a scale of 0 to 1 using a sigmoid function in the case of binary classification or a softmax function in the case of multi-class classification in the last layer of the artificial intelligence model.

[0089] In the case of binary classification, a sigmoid function is used to train the model so that the DPI value for cancer becomes 1. For example, if neuroblastoma samples and normal samples are input, the model is trained so that the DPI value for the neuroblastoma sample approaches 1.

[0090] In multi-class classification, the softmax function is used to extract DPI values ​​for each class. The sum of the DPI values ​​for each class becomes 1, and the model is trained so that the DPI value for the corresponding cancer type is 1. For example, if there are three classes: neuroblastoma, liver cancer, and normal, and a neuroblastoma sample is entered, the model would be trained to make the breast cancer class close to 1.

[0091] In the present invention, the output result value of step (d) is characterized in that it is derived according to the type of cancer. In this invention, the artificial intelligence model learns to produce an output result close to 1 if cancer is present, and an output result close to 0 if cancer is not present. Using 0.5 as a baseline, it was determined that cancer is present if the result is 0.5 or higher, and that cancer is not present if the result is 0.5 or lower, and the performance was measured accordingly (Training, validation, test accuracy).

[0092] It is obvious to any engineer that the 0.5 threshold can be changed at any time. For example, to reduce false positives, a higher threshold can be set than 0.5 to tighten the criteria for determining the presence of cancer, and to reduce false negatives, a lower threshold can be measured to slightly loosen the criteria for determining the presence of cancer.

[0093] Most preferably, a trained artificial intelligence model can be used to apply unseen data (data with known answers that have not been used for training) to determine the probability of the DPI value and set a baseline value.

[0094] In the present invention, the step of predicting the type of cancer by comparing the output values ​​in step (e) can be characterized by being performed in a manner that includes determining the type of cancer showing the highest value among the output values ​​as the cancer of the sample.

[0095] From another perspective, the present invention is a decoding unit for extracting nucleic acids from a biological sample and decoding sequence information including methylation information; An alignment unit that aligns the decoded sequences into a standard chromosome sequence database; A data generation unit that generates vectorized data using aligned sequence-based nucleic acid fragments; A cancer diagnostic unit that inputs the generated vectorized data into a trained artificial intelligence model for analysis and compares it to reference values ​​to determine the presence or absence of cancer; and This invention relates to a cancer diagnosis and cancer type prediction device, including a cancer type prediction unit that analyzes output result values ​​to predict the type of cancer.

[0096] In the present invention, the decoding unit may be performed by an independent device. For example, the decoding unit of the present invention can generate sequence information including methylation information, i.e., reads, using an NGS device.

[0097] From another perspective, the present invention is a computer-readable storage medium comprising instructions configured to be executed by a processor that predicts cancer diagnosis and cancer type, (a) A step of extracting nucleic acids from a biological sample and obtaining sequence information including methylation information; (b) The step of aligning the acquired sequence information (reads) with the reference genome database; (c) A step of generating vectorized data using the aligned sequence information (reads) basis nucleic acid fragments; (d) A step of inputting the generated vectorized data into a trained artificial intelligence model and comparing the output result value obtained from the analysis with a cut-off value to determine the presence or absence of cancer; and (e) relating to a computer-readable storage medium including instructions configured to be executed by a processor that predicts the presence and type of cancer, through the step of predicting the type of cancer by comparing the output result values.

[0098] From another perspective, the present invention provides for the step of (a) extracting nucleic acids from a biological sample to obtain sequence information including methylation information; (b) The step of aligning the acquired sequence information (reads) with the reference genome database; (c) A step of generating vectorized data using the ordered sequence information (reads) based nucleic acid fragments; (d) A step of inputting the generated vectorized data into a trained artificial intelligence model and comparing the output result value obtained from the analysis with a cut-off value to determine the presence or absence of cancer; and (e) The present invention relates to a method for diagnosing and predicting cancer type, which includes the step of predicting the type of cancer through comparison of the output result values.

[0099] In other embodiments, the methods according to the present invention can be implemented using a computer. In one embodiment, the computer includes one or more processors coupled to a chipset. The chipset is also coupled to memory, a storage device, a keyboard, a graphics adapter, a pointing device, and a network adapter, etc. In one embodiment, the performance of the chipset is enabled by a memory controller hub and an I / O controller hub. In other embodiments, the memory may be used by being directly coupled to the processor instead of the chipset. The storage device is any device capable of holding data, including a hard drive, a CD-ROM (Compact Disk Read-Only Memory), a DVD, or other memory device. The memory is involved with the data and instructions used by the processor. The pointing device may be a mouse, a trackball, or other type of pointing device, and is used in combination with a keyboard to transmit input data to the computer system. The graphics adapter displays images and other information on a display. The network adapter is connected to the computer system by a short-range or long-range communication network. However, the computer used in this application is not limited to the configuration described above, and may be missing some of the configurations or may include additional configurations, and may be part of a Storage Area Network (SAN), and the computer of this application may be configured to be suitable for executing modules of a program for performing the method according to this application.

[0100] In this application, the term "module" may also mean a functional and structural combination of hardware for implementing the technical concept of this application and software for driving said hardware. For example, the term "module" may also mean a logical unit of predetermined code and hardware resources for which said code is performed, and it will be obvious to those skilled in the art that it does not necessarily mean physically connected code or a type of hardware.

[0101] The method according to this application can be implemented in hardware, firmware, or software, or a combination thereof. When implemented in software, the storage medium includes any medium that stores or transmits data in a readable form by a device such as a computer. For example, computer-readable media include ROM (Read Only Memory), RAM (Random Access Memory), magnetic disk storage media, optical storage media, flash memory devices, and other electrical, optical, or acoustic signal transmission media. [Examples]

[0102] The present invention will be described in more detail below with reference to examples. These examples are solely for illustrative purposes and it will be obvious to those with ordinary skill in the art that the scope of the present invention is not to be limited by these examples.

[0103] Example 1. Methylated cfDNA is extracted from blood and subjected to next-generation sequencing analysis. . Blood samples were collected from 185 healthy individuals and 57 neuroblastoma patients. The plasma portion was then primarily centrifuged at 3000 rpm, 25°C, and 10 minutes. The purified plasma was then further centrifuged at 16000 g, 25°C, and 10 minutes to remove precipitates and separate the plasma supernatant. Cell-free DNA was extracted from the separated plasma using the chemagen DNA kit. Using the Truseq Nano DNA HT library prep kit (Illumina), the process was first carried out up to the adapter ligation stage. Subsequently, immunoprecipitation was performed using the cfMediIP kit (diagnode) antibody at 10 rpm, 4°C, and 17 hours. After purification, PCR enrichment was performed again using the Truseq Nano DNA HT library prep kit (Illumina) to prepare the final library. The created library was sequenced using a Novaseq 6000 (Illumina) in 150-paired-end mode, producing approximately 30 million reads per sample.

[0104] Example 2. Generation of GC plots based on the number of nucleic acid fragments. The reads obtained in Example 1 were aligned to the reference genome using the bwa (version 0.7.17-r1188) alignment tool. Then, PCR duplicate nucleic acid fragments were removed using the biobambam2 bammarkduplicates (version 2.0.87) tool, and nucleic acid fragments with alignment agreement of 60 or less were removed using sambamba (version 0.6.6).

[0105] A GC plot represents the state in which NGS reads are aligned from the beginning to the end of a chromosome. After dividing all chromosomes except sex chromosomes into non-overlapping 100 kilobase bins, the number of reads assigned to each bin was counted (read count value). The read count value assigned to each bin was divided by the total number of reads in the sample to perform a normalization process. The normalized bin read count value was used as the Y value, and the order of each bin as the X value. GC plots were produced for each chromosome, and the produced GC plots were aligned from chromosome 1 to chromosome 22 to produce a single image (Figure 2).

[0106] Example 3. Construction of a GC plot-based deep learning model for neuroblastoma and calculation of DPI. 3-1. Building a Deep Learning Model The GC plot produced in Example 2 was divided into training, validation, and test data. The training data was used to train the CNN model, the validation data was used for hyperparameter tuning validation, and the test data was used for performance evaluation after producing the optimal model.

[0107] Tensorflow (version 2.4.1) was used to build and train the CNN model. The CNN model is structured in the order of convolutional layers → pooling layers → fully connected layers, with a pooling layer always inserted after each convolutional layer. The number of convolutional layers and fully connected layers were determined through hyperparameter tuning. During model training, the learning process was carried out in the direction of minimizing the loss function, which is given by equations 1 and 2.

[0108] To find the optimal model, hyperparameter tuning was performed using the scikit-optimize (version 0.7.4) Python package. The hyperparameter tuning process involves optimizing the values ​​of various parameters that make up the CNN model (number of convolutional layers, number of fine layers, number of convolutional filters, etc.). After specifying the number of convolutional layers, number of convolutional filters, convolutional patch size, number of fully connected layers, number of hidden nodes, activation function, presence or absence of dropout, and learning rate as hyperparameters, the optimal model was constructed using Bayesian optimization. When training the model with the specified hyperparameters, if the validation loss begins to increase compared to the training loss, the model is judged to be overfitted. Model training was stopped before this point, and the performance of multiple models obtained during the hyperparameter tuning process was compared using validation data. The model with the best performance was judged to be the optimal model, and its performance was evaluated using test data.

[0109] 3-2. Calculation of DPI (Deep Probability Index) When data (GC plot) is input into the optimal model obtained through hyperparameter tuning, probability values ​​are output through the model's output layer.

[0110] First, in the case of binary classification, a sigmoid function was used in the output layer of the model. The sigmoid function is given by equation 3 below.

number

number

[0111] Example 4. Construction and performance verification of a neuroblastoma deep learning model based on nucleic acid fragment number using methylated cfDNA and GC plot. The performance of the DPI values ​​was tested using normal samples (n=186) and neuroblastoma samples (n=57). All samples were divided into Train, Validation, and Test groups. After building a model using the Train samples, the performance of the model created using the Train samples was confirmed using the Validation and Test group samples. [Table 1]

[0112] As a result, as shown in Table 2, Figure 3, and Figure 4, the accuracy was confirmed to be 100%, 92%, and 94.1% for the Train, Valid, and Test groups, respectively, and the AUC values ​​from the ROC analysis were confirmed to be 1.0, 0.95, and 0.99 for the Train, Valid, and Test groups, respectively.

[0113] Figure 3 shows an analysis using the ROC (Receiver Operating Characteristic) curve as a method for measuring accuracy. A higher AUC (Area Under the Curve) value, which represents the area under the curve, is interpreted as higher accuracy. The AUC value ranges from 0 to 1, with an expected AUC value of 0.5 when predicting label values ​​randomly (baseline) and an expected AUC value of 1 when predicting with perfect accuracy.

[0114] Figure 4 shows box plots of the probability value (DPI value) of cancer calculated by the artificial intelligence model of the present invention for normal samples and neuroblastoma sample groups, with the red line representing the DPI cutoff of 0.5. [Table 2]

[0115] Example 5. Construction and performance verification of a neuroblastoma deep learning model based on nucleic acid fragment count using cfDNA and GC plot. 5-1. Extract DNA from blood and perform next-generation sequencing analysis. Blood samples (10 mL each) were collected from 186 healthy individuals and 57 neuroblastoma patients and stored in EDTA tubes. Within 2 hours of collection, the plasma portion was primarily centrifuged at 1200 g, 4°C, and 15 minutes. The purified plasma was then further centrifuged at 16000 g, 4°C, and 10 minutes to remove precipitates and separate the plasma upper layer. Cell-free DNA was extracted from the separated plasma using the Tiangen micro DNA kit (Tiangen), and the library was prepared using the MGIAsy cell-free DNA library prep set kit. Sequencing was then performed using a DNBseq G400 equipped with (MGI) in 100 base paired-end mode. As a result, approximately 170 million reads were generated per sample.

[0116] 5-2. Construction and performance verification of deep learning models The performance of the DPI value was tested using normal samples (n=186) and neuroblastoma samples (n=57). All samples were divided into Train, Validation, and Test groups. After building a model using the Train samples, the performance of the model created using the Train samples was confirmed using samples from the Validation and Test groups.

[0117] [Table 3] As a result, as shown in Table 4, Figure 6, and Figure 7, the accuracy was confirmed to be 92.9%, 97.3%, and 94.6% for the Train, Valid, and Test groups, respectively, and the AUC values ​​from the ROC analysis were confirmed to be 0.98, 0.98, and 0.95 for the Train, Valid, and Test groups, respectively.

[0118] [Table 4] Although specific aspects of the present invention have been described in detail above, it will be clear to those with ordinary skill in the art that these specific techniques are merely preferred embodiments and do not limit the scope of the invention. Therefore, the substantial scope of the invention is defined by the appended claims and their equivalents. [Industrial applicability]

[0119] The cancer diagnosis and cancer type prediction method using methylated cell-free nucleic acids according to the present invention is useful because, compared to conventional methods that use a step to determine chromosome amount based on the number of reads or detection methods that use the concept of distance between aligned reads, which utilize values ​​related to reads as individual, standardized values, this method generates vectorized data and analyzes it using an AI algorithm, thus achieving similar effects even with low read coverage.

Claims

1. (a) A step of extracting nucleic acids from a biological sample and obtaining sequence information including methylation information; (b) The step of aligning the acquired sequence information (reads) with the reference genome database; (c) A step of generating vectorized data using the ordered sequence information (reads) based nucleic acid fragments; (d) A step of determining the presence or absence of cancer by inputting the generated vectorized data into an artificial intelligence model trained to distinguish between normal vectorized data and vectorized data with cancer, and comparing the output result value obtained by the analysis with a cut-off value; and (e) A method for providing information for cancer diagnosis and cancer type prediction, which includes the step of predicting the type of cancer through comparison of the output result values, The vectorized data from step (c) is a Grand Canyon plot (GC plot), and The GC plot above shows the chromosomal segmental distribution of aligned nucleic acid fragments, as follows: i) A step of dividing the chromosome into specific intervals (bins); ii) A step of determining the number of nucleic acid fragments arranged in each interval; iii) The step of normalizing by dividing the number of nucleic acid fragments determined in each interval by the total number of nucleic acid fragments in the sample; and iv) A step to generate a GC plot, where the order of each interval is the value on the X axis and the normalized value calculated in step iii) is the value on the Y axis. A method for providing information for cancer diagnosis and cancer type prediction, characterized by generating vectorized data by calculating the number of nucleic acid fragments using a method that includes [a specific method].

2. The method according to claim 1, characterized in that step (a) is carried out by a method comprising the following steps: (ai) A step of obtaining nucleic acids containing methylation information from a biological sample; (a-ii) The step of removing proteins and fats from the collected nucleic acids using the salting-out method, column chromatography method, or beads method to obtain purified nucleic acids; (a-iii) A step of preparing a single-end sequencing or pair-end sequencing library from purified nucleic acids or nucleic acids randomly fragmented by enzymatic cleavage, disruption, or hydroshear method; (a-iv) The step of reacting the prepared library with a next-generation sequencer; and (av) Steps to obtain nucleic acid sequence information (reads) using a next-generation sequencer.

3. The method according to claim 2, characterized in that the methylation information in step (ai) is obtained by bisulfite conversion, enzyme conversion, or methylated DNA immunoprecipitation (MeDIP).

4. The method according to claim 1, characterized in that the artificial intelligence model is selected from the group consisting of a convolutional neural network (CNN), a deep neural network (DNN), a recurrent neural network (RNN), and an autoencoder.

5. The method according to claim 1, characterized in that when the artificial intelligence model is a CNN and learns binary classification, the loss function is expressed by the following formula 1, and when the artificial intelligence model is a CNN and learns multi-class classification, the loss function is expressed by the following formula 2:

6. The method according to claim 1, characterized in that the result value output by the artificial intelligence model in step (d) after analyzing the input vectorized data is a DPI (Deep Probability Index) value.

7. The method according to claim 1, characterized in that the reference value for step (d) is 0.5, and if it is 0.5 or greater, it is determined to be cancer.

8. The method according to claim 1, characterized in that the step of predicting the type of cancer by comparing the output values ​​of step (e) is performed in a manner that includes determining the type of cancer showing the highest value among the output values ​​as the cancer of the sample.

9. A decoding unit that extracts nucleic acids from biological samples and decodes sequence information including methylation information; Alignment unit that sorts the decoded sequences into a standard chromosome sequence database; A data generation unit that generates vectorized data using aligned sequence-based nucleic acid fragments; A cancer diagnostic unit that inputs the generated vectorized data into an artificial intelligence model trained to distinguish between normal vectorized data and vectorized data with cancer, analyzes it, and determines the presence or absence of cancer by comparing it to a baseline; and A cancer diagnosis and cancer type prediction device including a cancer type prediction unit that analyzes output result values ​​to predict cancer type, The aforementioned vectorized data is a Grand Canyon plot (GC plot), and The GC plot above shows the chromosomal segmental distribution of aligned nucleic acid fragments, as follows: i) A step of dividing the chromosome into specific intervals (bins); ii) A step of determining the number of nucleic acid fragments arranged in each interval; iii) The step of normalizing by dividing the number of nucleic acid fragments determined in each interval by the total number of nucleic acid fragments in the sample; and iv) A step to generate a GC plot, where the order of each interval is the value on the X axis and the normalized value calculated in step iii) is the value on the Y axis. An apparatus characterized by generating vectorized data by calculating the number of nucleic acid fragments using a method that includes [a specific method].

10. A computer-readable storage medium comprising instructions configured to be executed by a processor for cancer diagnosis and cancer type prediction, (a) A step of extracting nucleic acids from a biological sample and obtaining sequence information including methylation information; (b) The step of aligning the acquired sequence information (reads) with the reference genome database; (c) A step of generating vectorized data using the aligned sequence information (reads) basis nucleic acid fragments; (d) A step of determining the presence or absence of cancer by inputting the generated vectorized data into an artificial intelligence model trained to distinguish between normal vectorized data and vectorized data with cancer, and comparing the output result value obtained by the analysis with a cut-off value; and (e) A computer-readable storage medium comprising instructions configured to be executed by a processor that predicts the presence and type of cancer, through the step of predicting the type of cancer by comparing the output result values, The vectorized data from step (c) is a Grand Canyon plot (GC plot), and The GC plot above shows the chromosomal segmental distribution of aligned nucleic acid fragments, as follows: i) A step of dividing the chromosome into specific intervals (bins); ii) A step of determining the number of nucleic acid fragments arranged in each interval; iii) The step of normalizing by dividing the number of nucleic acid fragments determined in each interval by the total number of nucleic acid fragments in the sample; and iv) A step to generate a GC plot, where the order of each interval is the value on the X axis and the normalized value calculated in step iii) is the value on the Y axis. A computer-readable storage medium characterized by generating vectorized data by calculating the number of nucleic acid fragments using a method that includes [a specific method].

Images