DNA methylation markers and their use for non-invasive detection of cancer

Abstract

To provide DNA methylation markers for noninvasive detection of cancer and uses thereof.SOLUTION: Provided is a "binary categorical discrimination" method for finding a small number of exquisite DNA methylation position combinations in the human genomes (CG IDs) to detect cancer in DNA in patient-derived biological material and to distinguish it from cell-free DNA and blood cell DNA of other tissues. Another method for determining the tissue of origin of tumor DNA uses a combination of unique DNA methylation positions in the human genomes (CG IDs). Combinations of CG IDs derived from tumor DNA have been disclosed to accurately detect cancer by measuring DNA methylation of specific CG ID combinations and deriving a "methylation score". Also provided are kits for predicting cancer using CG IDs from a small volume of plasma using a multiplexed next generation sequence methylation assay, a pyrosequencing assay, and a methylation specific PCR.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This invention relates to DNA methylation signatures in the field of human DNA, particularly in molecular diagnostics. [Background technology]

[0002] Cancer is a leading cause of death in humans. Early detection of cancer can significantly improve cure rates and reduce the enormous personal and economic costs for patients, their families, and the healthcare system. For example, hepatocellular carcinoma (HCC) is the fifth most common cancer in the world (El-Serag, 2011). It is particularly prevalent in Asia, with its incidence being highest in areas where hepatitis B is prevalent, suggesting a possible causal relationship (Flores & Marrero, 2014). Follow-up of high-risk populations, such as patients with chronic hepatitis, and early diagnosis of the transition from chronic hepatitis to HCC would improve cure rates. Survival rates for hepatocellular carcinoma are currently very low because it is almost always diagnosed at an advanced stage. If diagnosed early, liver cancer can be effectively treated with a cure rate of over 80%. Advances in imaging have improved the non-invasive detection of HCC (Tan, Low, &Thng, 2011; Valente et al., 2014). However, current diagnostic methods, including imaging and immunoassays using single proteins such as alpha-fetoprotein, often fail to diagnose HCC early (Flores & Marrero, 2014). These challenges are not limited to HCC but are common to other cancers as well. For example, early detection of breast and colorectal cancer could dramatically reduce morbidity and mortality, as well as costs for public health systems and insurance companies. Furthermore, certain cancers, such as pancreatic cancer, are almost always detected late, resulting in virtually constant mortality. While advances in imaging have improved early cancer detection, high-resolution imaging such as MRI is expensive, requires highly trained personnel, and is not available in many locations. It has not yet evolved into a method for screening large populations. To impact the reduction of cancer morbidity and mortality, there is a need to develop non-invasive, powerful, yet low-cost methods that can be used across a wide geographical area for routine screening of populations. The main challenge is that solid tumors hide within visceral structures and evolve long before they show clinical symptoms. However, it is possible to obtain tumor material non-invasively.

[0003] It is now widely established that tumor DNA is released into the system and can be found in plasma (Warton & Samimi, 2015), and may also be found in other secreted bodily fluids such as urine and saliva, as well as in feces. Measuring the molecular properties of tumor DNA can confirm that DNA found in bodily fluids originates from a tumor (Zhai et al., 2012). Tumor cells produce mutations that can distinguish tumor DNA from normal cell DNA, but the number of possible mutations is enormous, and not all tumors produce common mutations (Dominguez-Vigil, Moreno-Martinez, Wang, Roehrl, & Barrera-Saldana, 2018).

[0004] DNA methylation, a covalent modification of DNA, is a major mechanism of epigenetic regulation of genomic function and is ubiquitously altered within tumors (Aguirre-Ghiso, 2007; Baylin et al., 2001; Ehrlich, 2002; Issa et al., 1993). Tumor DNA methylation profiles can be a powerful tool for tumor classification, prognosis, and predicting response to chemotherapy (Stefansson et al., 2014). The main drawback of using tumor DNA methylation in early diagnosis is the need for invasive procedures and anatomical visualization of the suspected tumor. Circulating tumor cells are a non-invasive source of tumor DNA and are used to measure DNA methylation of tumor suppressor genes (Radpour et al., 2011). HCC DNA hypomethylation is detectable in patient blood (Ramzy, Omran, Hamad, Shaker, & Abboud, 2011), and recently, genome-wide bisulfite sequencing has been applied to detect hypomethylated DNA in the plasma of HCC patients (Chan et al., 2013). However, this source of information is limited, especially in the early stages of cancer, and DNA methylation profiles are confused with host DNA methylation profiles. Genome-wide bisulfite sequencing is a relatively expensive procedure and requires significant bioinformatics analysis, making it impractical as a screening tool. Therefore, the challenge is to develop a low-cost, high-throughput assay that can reliably distinguish tumor DNA from non-tumor DNA, depicting a small number of CGs and enabling screening of broad populations across extensive and diverse geographical regions. Recently, several groups have conducted comparative analyses of genome-wide DNA methylation maps of cancer, normal DNA, and blood DNA (Zhai et al., 2012). However, a major challenge with these approaches is that they do not account for cell-free DNA from other tissues present in the blood at varying levels that are not a priori anticipated. Contamination with DNA from other tissues that have a similar methylation profile to cancer tissue can lead to false positives. Furthermore, past approaches quantitatively compare DNA methylation in normal and cancerous tissues. This quantitative difference can be diluted when tumor DNA is mixed with different, unknown amounts of DNA from other non-transformed tissues, potentially leading to false negatives. These shortcomings in current methods necessitate the different approaches disclosed in the subject matter of this invention.

[0005] Further publications relating to the use of systems and methods for cancer detection include: Grigg G, Clark S. Sequencing of 5-methylcytosine residues in genomic DNA. Bioessays. June 1994;16(6):431-6, 431; Zeschnigk M, Schmitz B, Dittrich B, Buiting K, Horsthemke B, Doerfler W. Segments imprinted in the human genome: Different DNA methylation patterns of Prader-Willi / Angelman syndrome regions determined by genome sequencing. Hum Mol Genet. March 1997;6(3):387-95; Feil R, Charlton J, Bird AP, Walter J, Reik W. Methylation analysis of individual chromosomes: An improved protocol for bisulfite genome sequencing. Nucleic Acids Res. February 25, 1994;22(4):695-6;Martin V, Ribieras S, Song-Wang X, Rio MC, Dante R. Genome sequencing shows a correlation between DNA hypomethylation in the 5' region of the pS2 gene and its expression in human breast cancer cell lines. Gene. May 19, 1995;l57(l-2):26l-4;WO 97 46705, WO 95 15373, and WO 45560. Aguirre-Ghiso, JA (2007). Models, mechanisms, and clinical evidence of cancer dormancy. Nat Rev Cancer, 7(11), 834-846. doi:l0.l038 / nrc2256 Baylin, SB; Esteller, M.; Rountree, MR; Bachman, KE; Schuebel, K.; & Herman, JG (2001). Abnormal patterns of DNA methylation, chromatin formation, and gene expression in cancer. Hum Mol Genet, 10(7), 687-692. Breitbach, S., Tug, S., Helmig, S., Zahn, D., Kubiak, T., Michal, M., ...Simon, P. (2014). Direct quantification of cell-free circulating DNA from crude plasma. PLoS One, 9(3), e87838. doi:l0.l37l / journal.pone.0087838 Chan, KC, Jiang, P., Chan, CW, Sun, K., Wong, J., Hui, EP, ...Lo, YM (2013). Non-invasive sequencing of cancer-related genome-wide hypomethylation and copy number abnormalities using plasma DNA bisulfite sequencing. detection. Proc Natl Acad Sci USA, 110(47), 18761-18768. doi:l0.l073 / pnas.l3l3995110 Dominguez-Vigil, IG, Moreno-Martinez, AK, Wang, JY, Roehrl, MHA, & Barrera-Saldana, HA (2018). The dawn of liquid biopsy in the fight against cancer. Oncotarget, 9(2), 2912-2922. doi:l0.18632 / oncotarget.23l3l Ehrlich, M. (2002). DNA methylation in cancer: too much, too little. Oncogene, 21(35), 5400-5413. El-Serag, HB (2011). Hepatocellular carcinoma. N Engl J Med, 365(12), 1118-1127. doi:10.1056 / NEJMra1001683 Flores, A., & Marrero, JA (2014). Latest trends in hepatocellular carcinoma: Focusing on diagnosis and treatment. Clin Med Insights Oncol, 8, 71-76. doi:l0.4l37 / CMO.S9926 Issa, JP, Vertino, PM, Wu, J., Sazawal, S., Celano, P., Nelkin, BD, ...Baylin, SB (1993). Increased cytosine DNA-methyltransferase activity in progressive colon cancer. J Natl Cancer Inst, 85(15), 1235-1240. Luczak, MW, & Jagodzinski, PP (2006). The role of DNA methylation in cancer progression. Folia Histochem Cytobiol, 44(3), 143-154. Radpour, R., Barekati, Z., Kohler, C., Lv, Q., Burki, N., Diesch, C., ... Zhong, XY (2011). Hypermethylation of tumor suppressor genes involved in key regulatory pathways for the development of blood tests for breast cancer. PLoS One, 6(1), el6080. doi:l0.l37l / journal.pone.00l6080 Ramzy, II, Omran, D., Hamad, O., Shaker, O., & Abboud, A. (2011). Evaluation of serum LINE-1 hypomethylation as a prognostic marker for hepatocellular carcinoma. Arab J Gastroenterol, 12(3), 139-142. doi:l0.l0l6 / j.ajg.20l1.07.002 Stefansson, OA; Moran, S.; Gomez, A.; Sayols, S.; Arribas-Jorba, C.; Sandoval, J.; ... Esteller, M. (2014). Definitions of biologically distinct breast cancer subtypes based on DNA methylation. Mol Oncol. doi:l0.l0l6 / j.molonc.20l4.l0.0l2 Tan, CH, Low, SC, & Thng, CH (2011). APASL and AASLD Consensus Guidelines for Imaging Diagnosis of Hepatocellular Carcinoma: A Review. Int J Hepatol, 2011, 519783. doi:10.4061 / 2011 / 519783 Valente, S., Liu, Y., Schnekenburger, M., Zwergel, C., Cosconati, S., Gros, C., ...Mai, A. (2014). Selective non-nucleoside inhibitor of human DNA methyltransferase activated in cancer, including cancer stem cells. J Med Chem, 57(3), 701-713. doi:l0.l02l / jm40l2627 Warton, K., & Samimi, G. (2015). Methylation of cell-free circulating DNA in cancer diagnosis. Front Mol Biosci, 2, 13. doi:l0.3389 / fmolb.20l5.000l3 Xu, RH, Wei, W., Krawczyk, M., Wang, W., Luo, H., Flagg, K., ... Zhang, K. (2017). Diagnosis and Prediction of Hepatocellular Carcinoma. Circulating tumor DNA methylation markers for later detection. Nat Mater, 16(11), 1155-1161. doi:l0.l038 / nmat4997 Zhai, R., Zhao, Y., Su, L., Cassidy, L., Liu, G., & Christiani, DC (2012). Genome-wide DNA methylation profiling of cell-free serum DNA in esophageal adenocarcinoma and Barrett's esophagus. Neoplasia, 14(1), 29-33. [Overview of the Initiative]

[0006] Embodiments of the claimed subject matter demonstrate that cancer is associated with a set of “categorically” distinct DNA methylation signatures that differ from the DNA methylation profiles of any normal tissue and blood cells. These sites create a binary differentiation between cancer and other tissues, so that these sites are methylated only in cancer and not at all in other cancers. Therefore, it is possible to detect even a few molecules of cancer cells in the background of a normal cellular DNA profile of methylation using deep next-generation sequencing. Embodiments of the subject matter of the present invention enable the detection of cell-free tumor DNA even in the background of cell-free DNA from other tissues, and are therefore particularly suitable for the early detection of cancer using cell-free (CF) DNA extracted from bodily fluids such as saliva, plasma, urine, and feces. Embodiments also enable the early detection of cancer in tissue smears such as pap tests, as well as in biopsies and needle biopsies. Previous analyses in the prior art have only compared normal cells and cancer cells from the same tissue and blood, and derived sites with quantitatively different DNA methylation levels (Xu et al., 2017). However, sites discovered through such prior art analysis may be subject to the CF of other organizations. When mixed with DNA, CF tumor DNA cannot be detected (see Figure 2 for the HCC ctDNA marker from Sun Yat-sen University Cancer Hospital). One embodiment of the subject matter of the claims reveals a unique set of sites that are not methylated in all tissues but are methylated in certain cancers. Another embodiment reveals a method for discovering categorically differentiated methylation sites in cancer, other tissues and other diseases, called the "binary-categorical differentiation method," using various sources of genome-wide DNA methylation data obtained by next-generation sequencing, MeDIP arrays, MeDIP sequencing, etc. One embodiment reveals “category” DNA methylation site combinations for detecting profiles of a. hepatocellular carcinoma (HCC), b. lung cancer, c. prostate cancer, d. breast cancer, e. colorectal cancer, f. head and neck squamous cell carcinoma (HNSC), g. pancreatic cancer, h. brain cancer (glioblastoma), i. stomach cancer, j. ovarian cancer, k. cervical cancer, l. esophageal cancer, m. bladder cancer, n. kidney cancer, o. testicular cancer, p. common solid tumors, and q. hematological cancers in genome-wide discovery sets. Another embodiment also reveals “category” DNA methylation site combinations that differentiate tumors by their tissue of origin. This embodiment distinguishes prior art methods and assays for detecting methylated CF DNA with low tissue specificity. The embodiment validates the Polygene DNA methylation assay for detecting cancer in DNA methylation data from hundreds of patients and in the tissues of tumor origin with high sensitivity and specificity. The present invention discloses a method for simultaneously and accurately measuring DNA methylation in a polygene set of CG IDs from hundreds of individuals by sequential amplification using target-specific primers followed by barcode primers, multiplex sequencing in a single next-generation Miseq sequencing reaction, and data extraction and methylation quantification from small amounts of bodily fluids such as plasma, saliva, and urine. Another embodiment of the subject matter of the present invention also discloses the measurement of methylation of the aforementioned DNA-methylated CG IDs using a pyrosequencing assay or methylation-specific PCR.Another embodiment discloses the calculation of either a “category” or a polygene weighted methylation score to distinguish individuals with cancer from healthy individuals. Another embodiment discloses a novel process to predict cancer in individuals without other clinical evidence of cancer from plasma, urine, feces, tissue biopsy, or tissue swab. Another embodiment may be used by those skilled in the art to detect cancer, as well as other diseases involving cell death and the release of CF DNA, such as Alzheimer's disease and other neurodegenerative diseases of neurons, and cardiac diseases of cardiomyocytes. The DNA methylation markers (CG IDs) described in the embodiment are used for: a. Non-invasive early detection of cancer in "healthy" people through routine "health checkups". b. Monitoring of "high-risk" individuals, such as patients with chronic hepatitis at high risk of HCC or smokers at high risk of lung cancer. c. Monitoring the response to treatment in patients undergoing cancer treatment and detecting recurrence or metastasis.

[0007] The embodiments demonstrate the usefulness of detecting cancer in unknown samples using polygene or category scores based on the DNA methylation assay methods disclosed herein. The disclosed embodiments also demonstrate the use of next-generation bisulfite sequencing (bisulfite) to detect cancer in bodily fluids, feces, urine, and any cancerous or affected tissue. It can be used by those skilled in the art by using methylation analysis methods available to those skilled in the art, such as sequencing, Illumina Epic microarrays, capture sequencing, methylated DNA immunoprecipitation (MeDIP), methylation-specific PCR, and methylation assay methods that will become available.

[0008] The embodiments also disclose the potential for discovering novel “polygene” category DNA methylation markers for other cancers and diseases using any method available to those skilled in the art for genome-wide sequencing, such as analysis of binary category distinction (BCD) methods for discovering specific highly sensitive markers used for non-invasive detection of disease, following next-generation bisulfite sequencing, MeDip sequencing, ion torrent sequencing, Epic microarrays, etc. Embodiments also disclose a method for discovering regions that binaryly and categorically distinguish cancer from all normal tissues by detecting cancer from tumor-derived DNA in cell-free DNA using a "binary category" DNA methylation marker for cancer, which includes the step of deriving a "binary category" DNA methylation marker from a genome-wide DNA methylation map using a "binary category distinction (BCD)" method. The genome-wide DNA methylation map is one or more of cancer cells, normal tissue, and blood DNA. The aforementioned "binary category discrimination (BCD)" method includes the use of one or more genome-wide DNA methylation arrays, DNA microarrays, genome-wide bisulfite sequencing, methylated DNA immunoprecipitation (MeDIP) sequencing, and hybridization with oligonucleotide arrays.

[0009] Embodiments of the subject matter of the present invention include the following: In a first embodiment, the embodiment provides polygene DNA methylation markers for cancer in cell-free DNA in bodily fluids such as plasma for the early detection of cancer, the set of polygene DNA methylation markers obtained from genome-wide DNA methylation obtained by mapping methods such as Illumina 450K or EPIC arrays, genome-wide bisulfite sequencing, methylated DNA immunoprecipitation (MeDIP) sequencing, or hybridization with oligonucleotide arrays, using the "Binary Category Discrimination (BCD) analysis" disclosed herein.

[0010] In other embodiments, for the early detection of liver cancer hepatocellular carcinoma (HCC) using plasma cfDNA or other body fluid cfDNA, the polygenic DNA methylation markers are combinations of CG IDs from the following list (or a short subset of this list shown as an example below Table 1).

[0011]

Table 1

[0012] Subset for detection: cg02012576,cg03768777,cg24804544,cg05739190

[0013] In other embodiments, to identify the origin of cancer as HCC using plasma cfDNA or other body fluid DNA and to distinguish it from 10 other common solid tumor cancers, the polygenic DNA methylation markers are combinations of CG IDs from the following list or a short subset of this list (shown as an example below Table 2).

[0014]

Table 2

[0015] Subset for specificity: cg14126493

[0016] In other embodiments, for the early detection of lung cancer using plasma cfDNA or other body fluid cfDNA, the polygenic DNA methylation markers are combinations of CG IDs from the following list or a short subset of this list (shown as an example below Table 3).

[0017]

Table 3

[0018] Subset for detection: cg04223424,cg23141355

[0019] In other embodiments, to identify the origin of the cancer as lung cancer using plasma CF DNA or other bodily fluid DNA, and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 4).

[0020] [Table 4]

[0021] Subset for specifications: cg05917732, cg25470077

[0022] In other embodiments, to detect prostate cancer early using plasma CF DNA or other bodily fluid CF DNA, and to identify the origin of the cancer as prostate cancer, as well as to distinguish it from 16 other common solid tumor cancers, the polygene DNA methylation markers are combinations of CG IDs from the list below (or shorter subsets of this list shown as examples below Table 5).

[0023] [Table 5]

[0024] Subset for detection specifications: cg14283569 [A subset of the four combinations shown in the table above]

[0025] In other embodiments, to detect breast cancer early using plasma CF DNA or other bodily fluid CF DNA, the polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (such as the examples shown below Table 6).

[0026] [Table 6]

[0027] Subset for detection: cg13031251,cg09734791,cg09695735,cg03637878

[0028] In other embodiments, to identify the origin of the cancer as breast cancer using plasma CF DNA or other bodily fluid DNA, and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 7).

[0029] [Table 7]

[0030] Subset for specifications: cg03113878,cg20180843

[0031] In other embodiments, to detect colorectal cancer (CRC) early using plasma CF DNA or other bodily fluid CF DNA, and to identify the origin of the cancer as colorectal cancer, as well as to distinguish it from 16 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 8).

[0032] [Table 8]

[0033] Detection - Subset for Specification: cg09854653, cg01566242

[0034] In other embodiments, to detect pancreatic cancer early using plasma CF DNA or other bodily fluid CF DNA, polygene DNA methylation markers are used, as listed below. This is a combination of IDs or a short subset of this list (shown as an example below Table 9).

[0035] [Table 9]

[0036] Subset for detection: cg25024074,cg15386964,cg16232979

[0037] In other embodiments, to identify the origin of the cancer as pancreatic cancer using plasma CF DNA or other bodily fluid DNA and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 10).

[0038] [Table 10]

[0039] Subset for specifications: cg01237565,cg08182975,cg20983577,cg25591377

[0040] In other embodiments, to detect brain cancer (glioblastoma) early and identify the origin of the cancer as brain cancer (glioblastoma), and to distinguish it from 10 other common solid tumor cancers, the polygene DNA methylation markers are combinations of CG IDs from the list below or short subsets of this list (shown as examples below Table 11).

[0041] [Table 11]

[0042] Detection - Subset for Specification: cg19929355

[0043] In other embodiments, to detect stomach cancer early using plasma CF DNA or other bodily fluid CF DNA, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 12).

[0044] [Table 12]

[0045] Subset for detection: cg05611779,cg09734791,cg15760257

[0046] In other embodiments, to identify the origin of the cancer as gastric cancer using plasma CF DNA or other bodily fluid DNA, and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 13).

[0047] [Table 13]

[0048] Subset for specifications: cg0511779, cg19235339

[0049] In other embodiments, to detect ovarian cancer early using plasma CF DNA or other bodily fluid CF DNA, polygene DNA methylation markers are used, as listed below. This is a combination of IDs or a short subset of this list (as shown as an example below Table 14).

[0050] [Table 14]

[0051] Subset for detection: cg24339193,cg22694l53,cg11252337,cg21210985

[0052] In other embodiments, to identify the origin of the cancer as ovarian cancer using plasma CF DNA or other bodily fluid CF DNA and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or short subsets of this list (shown as examples below Table 15).

[0053] [Table 15]

[0054] Subset for specifications: cg07068768,cg19846609

[0055] In other embodiments, plasma CF DNA or other bodily fluid CF DNA is used to treat the cervix. To detect cancer early, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of these IDs (shown as examples below Table 16).

[0056] [Table 16]

[0057] Subset for detection: cg00757182, cg01601746

[0058] In other embodiments, to identify the origin of the cancer as cervical cancer using plasma CF DNA or other bodily fluid DNA, and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or short subsets of this list (shown as examples below Table 17).

[0059] [Table 17]

[0060] Subset for specifications: cg07066594,cg09260640,cg1296l842

[0061] In other embodiments, to detect head and neck squamous cell carcinoma (HNSC) early using plasma CF DNA or other bodily fluid CF DNA, the polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 18).

[0062] [Table 18]

[0063] Subset for detection: cg07900968,cg20334243,cg27420520

[0064] In other embodiments, to identify the origin of the cancer as head and neck squamous cell carcinoma (HNSC) using plasma CF DNA or other bodily fluid DNA, and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 19).

[0065] [Table 19]

[0066] Subset for specifications: cg18006328, cg19287220

[0067] In other embodiments, to detect esophageal cancer early using plasma CF DNA or other bodily fluid CF DNA, polygene DNA methylation markers are used, as listed below. This is a combination of IDs or a short subset of this list (as shown as an example below Table 20).

[0068] [Table 20]

[0069] Subset for detection: cg03280624,cg03735888,cg09734791,cg27420520

[0070] In one embodiment, to identify the origin of the cancer as esophageal cancer using plasma CF DNA or other bodily fluid DNA, and to distinguish it from 10 other common solid tumor cancers, the polygene DNA methylation markers are combinations of CG IDs from the list below or short subsets of this list (shown as examples below Table 21).

[0071] [Table 21]

[0072] Subset for specifications: Cg09556952,cg12473285

[0073] In other embodiments, to detect bladder cancer early using plasma CF DNA or other bodily fluid CF DNA, polygene DNA methylation markers are used, as listed below. This is a combination of IDs or a short subset of this list (as shown as an example below Table 22).

[0074] [Table 22]

[0075] Subset for detection: cg04223424,cg10723962,cg25024074

[0076] In other embodiments, to identify the origin of the cancer as bladder cancer using plasma CF DNA or other bodily fluid DNA, and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 23).

[0077] [Table 23]

[0078] Subset for specifications: cg13544006

[0079] In other embodiments, to detect renal (kidney) cancer early using plasma CF DNA or other bodily fluid DNA, to identify the origin of the cancer as renal cancer, and to distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 24).

[0080] [Table 24]

[0081] A subset for detection specifications: cg08884571,cg00011225,cg23946709

[0082] In other embodiments, to detect testicular cancer early using plasma CF DNA or other bodily fluid CF DNA, identify the origin of the cancer as testicular cancer, and distinguish it from 10 other common solid tumor cancers, polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 25).

[0083] [Table 25]

[0084] Subsets for detection and specification: cg14531093, cg25159927

[0085] In other embodiments, to detect one of the 13 most common solid tumors early using plasma CF DNA or other bodily fluid CF DNA, the polygene DNA methylation markers are combinations of CG IDs from the list below or shorter subsets of this list (shown as examples below Table 26).

[0086] [Table 26]

[0087] Subset for detection: cg10723962, cg15759056, cg24427504, cg25024074

[0088] In other embodiments, to detect hematological cancers such as AML and CLL early using leukocytes, plasma CF DNA, or other bodily fluid CF DNA, the polygene DNA methylation marker is a combination of CG IDs (or a shorter subset of these combinations shown below Table 27) drawn by the BCD method against genome-wide DNA methylation data as shown in Table 27.

[0089] [Table 27]

[0090] Detection - Subset for Specification: cg18658397,cg18780412,cg20439288,cg22828045,cg25375340

[0091] In other embodiments, to detect melanoma early using plasma CF DNA or other bodily fluid CF DNA, identify the origin of the cancer as melanoma, and distinguish it from 16 other common solid tumor cancers, the polygene DNA methylation markers are combinations of CG IDs shown in the list below (or a shorter subset of this list shown as examples below Table 28).

[0092] [Table 28]

[0093] Detection - Subset for Specification: cg15307891, cg18866529, cg27084903

[0094] In another aspect of the subject matter of the present invention, a kit and process for detecting cancer is provided, comprising means and reagents for detecting DNA methylation measurements of polygene DNA methylation markers.

[0095] In one embodiment, a kit for detecting hepatocellular carcinoma is provided, comprising means and reagents for measuring DNA methylation of CG IDs in Tables 1 and 2.

[0096] In another embodiment, a kit for detecting lung cancer is provided, comprising means and reagents for DNA methylation measurements of CG IDs as shown in Tables 3 and 4.

[0097] In another embodiment, a kit for detecting prostate cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs listed in Table 5.

[0098] In another embodiment, a kit for detecting breast cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 6 and 7.

[0099] In another embodiment, a kit for detecting colorectal cancer is provided, comprising means and reagents for DNA methylation measurements of CG IDs as shown in Table 8.

[0100] In another embodiment, a kit for detecting pancreatic cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 9 and 10.

[0101] In yet another embodiment, a kit for detecting brain cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Table 11.

[0102] In another embodiment, a kit for detecting gastric cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 12 and 13.

[0103] In another embodiment, a kit for detecting ovarian cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 14 and 15.

[0104] In another embodiment, a kit for detecting cervical cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 16 and 17.

[0105] In another embodiment, a kit for detecting head and neck squamous cell carcinoma (HNSC) is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 18 and 19.

[0106] In another embodiment, a kit for detecting esophageal cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 20 and 21.

[0107] In another embodiment, a kit for detecting bladder cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Tables 22 and 23.

[0108] In another embodiment, a kit for detecting kidney cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Table 24.

[0109] In another embodiment, a kit for detecting testicular cancer is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs in Table 25.

[0110] In other embodiments, means and reagents for detecting DNA methylation measurements of CG IDs in Table 26 are used to detect one of 13 common cancers (bladder, brain, breast, cervix, colon, esophagus, HNSC, HCC (liver), lung, ovary, pancreas, prostate, stomach). A kit will be provided.

[0111] In another embodiment, a kit for detecting hematological cancers such as AML and CLL is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs detected by the BCD method, which are specific to different subtypes of hematological cancers listed in Table 27.

[0112] In another embodiment, a kit for detecting melanoma is provided, comprising means and reagents for detecting DNA methylation measurements of CG IDs listed in Table 28.

[0113] In another embodiment, a DNA pyrosequencing methylation assay is used to predict HCC in body fluids, such as plasma CF DNA, by using the CG IDs listed above and, for example, by using the primers and standard conditions for the pyrosequencing reaction disclosed below. cg02012576 Forward: GGTAGTTAGGAAGTTTAGAGGTTGTAGTA Reverse (biotin-labeled): ACCACTACCCCAACCCAACCCTA Array: GGTTTTAGGATGTTTG cg03768777(VASH2) Forward: AGAATAATATTAGAGAATGGGATATGGAA Reverse (biotin-labeled): ACAACTCCAAAATCCTACCT Array: GAATGGGATATGGAATGA cg05739190(CCNJ) Forward: GTTTAGGAGTTGGGTTTTAGTTGAG Reverse (biotin-labeled): ACCCCACCCTAACTCCCTTACC Array: TGGGTTTTAGTTGAGG cg24804544(GRID2IP) Forward (biotin-labeled): GGGTAGGGGAGGGTTTTGAAATA Reverse: TAACCCCCCCTCCAACCTCATTC Array: CACCCAACTTCTCAAT The specificity of the tissue originating from cancer is determined by measuring the DNA methylation of the following CGID cg02012576(HPX). Forward (biotin-labeled): ATTTTTATGGGTATTAGTTTTAGGGAGAA Reverse (biotin-labeled): CCAAAACTATCCTATAACCTCTACAACTCA Array: ACCATTACCACCCCT

[0114] In another embodiment, a polygenic multiplexed amplicon bisulfite sequencing DNA methylation assay is used to predict cancer in bodily fluids, such as plasma CF DNA, by using the CG ID described above. For example, prostate cancer is predicted using the primers and standard conditions disclosed below. These standard conditions include bisulfite conversion, sequential amplification with target-specific primers (PCR1) followed by barcode primers (PCR2), and multiplex sequencing on a single next-generation Miseq sequencer (Illumina), demultiplexing using Illumina software, data extraction and methylation quantification using standard methods of methylation analysis such as Methylkit, subsequent calculation of a weighted DNA methylation score, and prediction of cancer from small amounts of bodily fluids such as plasma, saliva, or urine.

[0115] The first step in detecting prostate cancer by PCR is as follows: In the case of CGID cg02879662 Forward primer: 5'ACACTCTTTCCCTACACgACgCTCTTCCgATCTNNNNNGGTAGGAGTTTTGGG AATTGG3' Reverse primer: 5'gTgACTggAgTTCAgACgTgTgCTCTTCCgATCTCCACCCCTACAATCCCTAA3' In the case of CGID cg16232979 Forward primer: 5'ACACTCTTTCCCTACACgACgCTCTTCCgATCT NNNNNYGGTTTYGGGTTTYGTATT3' Reverse primer: 5' gTgACTggAgTTCAgACgTgTgCTCTTCCgATCTACRCAAAAATATAAATCRACR ATC3'

[0116] To test whether the cancer is specifically occurring in the prostate, the first PCR is performed as follows: For CGID:cg14041701 and cg14498227 Forward primer: 5'ACACTCTTTCCCTACACgACgCTCTTCCgATCTNNNNNGTTTTGYGTTTYGGA TTTGGGTT3' Reverse primer: 5' gTgACTggAgTTCAgACgTgTgCTCTTCCgATCTCATAAACAACACCTTTAAATAA ACACTAAA3'

[0117] To barcode the sample, use the following primers for the second PCR reaction. Forward primer: 5'AATgATACggCgACCACCgAgATCTACACTCTTTCCCTACACgAC3' Barcode primer (reverse): 5'CAAgCAgAAgACggCATACgAgATAGTCATCGgTgACTggAgTTCAgACgTg3' (The red bases are indices, and 200 variations of this index are used.)

[0118] In other embodiments, receiver operating characteristic (ROC) assays are used to detect cancer by defining a threshold between cancer and normal using weighted DNA methylation measurements of CG IDs. Samples above / below the threshold are classified as cancer. For example, the above CG IDs for detecting HCC are: JPEG0007850764000029.jpg20168

[0119] In another embodiment, a hierarchical clustering analysis assay is used to predict cancer by using methylation measurements of the CG IDs listed above.

[0120] In another aspect of the subject matter of the present invention, a method for identifying DNA methylation markers for detecting cancer and other diseases includes the step of performing a statistical analysis with respect to DNA methylation measurements obtained from clinical samples using the previously disclosed "binary category distinction (BCD)" method.

[0121] In another embodiment, the method involves performing statistical analysis and a "binary category distinction (BCD)" method on DNA methylation measurements obtained from a sample, with DNA methylation measurements obtained by running an EPIC array of DNA extracted from an Illumina Beadchip 450K or at least one sample.

[0122] In another embodiment, DNA methylation measurements are obtained by DNA pyrosequencing of DNA extracted from the sample, followed by mass spectrometry-based (Epityper) analysis. TM This is obtained by performing targeted amplification of a region spanning the target CG ID, as disclosed herein, from a PCR-based methylation assay and barcoding in a second set of amplifications, followed by indexed multiplex sequencing on an Illumina next-generation sequencer, after the bisulfite-converted DNA.

[0123] In other embodiments, the statistical analysis includes a receiver operating characteristic (ROC) assay.

[0124] In other embodiments, the statistical analysis includes a hierarchical clustering analysis assay.

[0125] definition As used herein, the term "CG" refers to dinucleotide sequences in DNA containing cytosine and guanosine bases. These dinucleotide sequences can be methylated in human and other animal DNA. CG IDs are provided by Illumina. As defined in the 450K manifest, its location in the human genome is revealed (the CG annotations listed here are publicly available at https: / / bioconductor.org / packages / release / data / annotation / html / IlluminaHumanMethylation450k.db.html and are installed as the R package IlluminaHumanMethylation450k.db:IlluminaHumanMethylation450k annotation data. R package version 2.0.9.).

[0126] As used herein, the term diagnostic instrument or instrument means any instrument known to those skilled in the art that can be used in conjunction with one or more reagents to perform a diagnostic test used, for example, to derive a DNA methylation measurement of a DNA methylation score according to an embodiment of the claimed subject matter.

[0127] As used herein, the term “beta value” refers to the estimated methylation level at CG ID locations derived by normalization and quantification of the Illumina 450K array using the intensity ratio of methylated and unmethylated probes, with the formula beta value = methylated C intensity / (methylated C intensity + unmethylated C intensity). The beta value is between 0 and 1, where 0 is not fully methylated and 1 is fully methylated.

[0128] As used herein, the term “penalty-based regression” refers to a statistical method aimed at identifying the minimum number of predictors necessary to predict an outcome from a larger list of biomarkers being performed, as described in Goeman JJ, L1 Penalty-Based Estimation in Cox Proportional Hazards Models, Biometrical Journal 52(1), 70-84, where “penalty-based” is, for example, in the R statistical package.

[0129] As used herein, the term "clustering" refers to the phenomenon where objects within the same group (called a cluster) are (in a sense) more relative to each other than to other groups (clusters). This refers to grouping a set of subjects, similar to the previous example.

[0130] As used herein, the term “hierarchical clustering” refers to a statistical method of constructing a hierarchy of “clusters” based on how similar (close) or dissimilar (far) the clusters are to each other, as described, for example, in Kaufman, L.; Rousseeuw, PJ (1990) Searching for Groups in Data: An Overview of Cluster Analysis (1st Edition) New York: John Wiley. ISBN 0-471-87876-6.

[0131] As used herein, the term “Receiver Operating Characteristic (ROC) assay” refers to a statistical method for creating graphical plots illustrating the performance of predictors. For example, as described in Hanley, James A. and McNeil, Barbara J. (1982) “The Meaning and Use of Area Under Receiver Operating Characteristic (ROC) Curves,” Radiology 143(1):29-36, the true positive rate of a predictor is plotted against the false positive rate at different threshold settings of the predictor (i.e., different %) of methylation.

[0132] As used herein, the term “multivariate or polygene linear regression” refers to a statistical method for estimating the relationship between multiple “independent variables” or “predictions,” such as the methylation rate of CG IDs, and a “dependent variable,” such as cancer. This method determines the “weight” or coefficient of each CG ID when predicting the “outcome” (the dependent variable, such as cancer) when several “independent variables,” such as CG IDs, are included in the model. [Brief explanation of the drawing]

[0133] [Figure 1]Figure 1 shows a candidate list of completely unmethylated sites across hundreds of blood samples and normal tissues. Figure A shows that CG IDs across 17 tissues that were unmethylated in all individuals (<0.l) (GSE50192) of the Illumina 450K genome-wide methylation array overlapped with unmethylated CG IDs from a genome-wide DNA methylation array of blood samples from 312 individuals (GSE61496), generating a list of 33,477 CG IDs. B shows the strongest candidate list of unmethylated CG IDs, and the list of 33,477 CG IDs in A overlapped with unmethylated CG IDs from a DNA methylation array of blood samples from 656 individuals (females and males) aged 19 to 101 years (GSE40279). For the combined portion, these analyses generated a list of 28,754 highly reliable unmethylated CG IDs across tissues and blood samples from many individuals across all age groups. These 28,754 locations were used to identify sites that are categorically methylated in cancer but not in other tissues, using the "Binary Categorical Discrimination (BCD)" method disclosed in the subject matter of the present invention. [Figure 2] Figure 2 illustrates the lack of tissue specificity among current circulating DNA markers for HCC. The illustrated heatmap shows 10 CG IDs listed as candidate biomarkers for HCC and methylation levels at these sites in other normal tissues by Xu et al. (Xu et al., 2017). Some of the CG IDs proposed as specific biomarkers for HCC are also methylated in other tissues, and the methylation levels of blood DNA vary. (Blue represents 0 methylation, dark red represents 100% methylation) [Figure 2A] Figure 2A is an exploded view of a portion of Figure 2 according to an embodiment of the subject of the present invention. [Figure 2B] Figure 2B is an exploded view of a portion of Figure 2 according to an embodiment of the subject of the present invention. [Figure 2C] Figure 2C is an exploded view of a portion of Figure 2 according to an embodiment of the subject of the present invention. [Figure 3]Figure 3 shows the specificity of HCC DNA methylation markers discovered using the BCD method against cancer DNA. The illustrated heatmap shows the four CG IDs selected as HCC DNA methylation markers by the BCD method described here. Methylation levels differ categorically between cancer (HCC) and normal tissues and blood, resulting in the site being unmethylated in all individuals with blood and other tissues, while being measurably methylated in HCC. [Figure 3A] Figure 3A is an exploded view of a portion of Figure 3 according to an embodiment of the subject matter of the present invention. [Figure 3B] Figure 3B is an exploded view of a portion of Figure 3 according to an embodiment of the subject of the present invention. [Figure 3C] Figure 3C is an exploded view of a portion of Figure 3 according to an embodiment of the subject of the present invention. [Figure 4] Figure 4 shows the lack of cancer tissue origin specificity of current DNA methylation markers for colorectal cancer and a comparison with the “detection-speck” method according to embodiments of the subject matter of the present invention. Figure A shows the CG site of the Sept9 gene contained in the colorectal cancer CF DNA methylation marker “Epi-Colorectal Cancer” (commercialized by Epigenomics), which lacks specificity for colorectal cancer because the marker can be used to detect many other cancers using methylation data from the TCGA collection of cancer DNA methylation data (HKG-Colorectal Cancer (HKG-epiCRC), blue). Markers disclosed in the subject matter of the present invention (Table 8) for the detection of colorectal cancer found using the BCD method (HKG-Colorectal Cancer, orange) (Table 8) are highly specific to colorectal cancer when tested against other common solid tumor cancers. Figures B and C are scatter plots of DNA methylation values ​​of tumor DNA from different individuals of various cancers using either the DNA methylation marker HKG-Colorectal Cancer (HKG-epiCRC) (B) or Epi-Colorectal Cancer (C). Of particular note is the strict categorical difference in DNA methylation between colorectal cancer and other cancers, as demonstrated by the scattered heterogeneous profiles of the HKG-epiCRC marker (B) versus the Epi-colorectal cancer marker (C). [Figure 4A] Figure 4A is an exploded view of a portion of Figure 4 according to an embodiment of the subject matter of the present invention. [Figure 4B] Figure 4B is an exploded view of a portion of Figure 4 according to an embodiment of the subject of the present invention. [Figure 5] Figure 5 illustrates the discovery of polygene DNA methylation markers for the early detection of hepatic cancer (HCC). Figure A shows a table listing the sources and number of patients, and the methylation data from these patients was used to discover four sets of CGIDs for HCC detection by embodiments using the BCD method (Table 1) and CG IDs (Table 2) to determine the origin of the cancer. Figure B (Detection) in the lower left panel of Figure 5 shows the sum of the methylation scores of these CG IDs (Table 1) for each test subject listed from 1 to 145 (79 normal and 66 HCC). The polygene score categorically distinguishes between people with HCC and those with normal liver tissue. Figure C in the lower right panel shows the methylation score of one CGID (Table 2) to detect the origin of a specific tumor using data from eight different tumors (Table 2). The marker categorically distinguishes HCC from cancers of other origins. [Figure 6] Figure 6 shows the validation of a polygene DNA methylation marker for HCC (speck) using DNA methylation data from GSE76269 (n=227). Figure A is an ROC plot showing the region under the curve for the HCC DNA methylation marker using DNA methylation data from 227 liver cancer patients and 10 healthy individuals. Figure B in Figure 6 shows the sensitivity, specificity, and accuracy of HCC detection. Figure C shows the predictive rate of HCC detection in the validation dataset. [Figure 7]Figure 7 shows the validation of the accuracy and specificity of Polygene HKG-epi liver cancer detection and spec marker for HCC versus other cancers in TCGA methylation data (n=4166). Figure A in Figure 7 shows the detection rate of HKG-liver detection / spec marker DNA methylation data from patients with different cancers. Note the nearly perfect specificity for HCC. Figure B is an ROC plot of the specificity and sensitivity of the HKG-liver detection marker for HCC in DNA methylation data from 4166 patients in TCGA. Figure C shows the sensitivity and specificity for HCC versus cancers of other origins. [Figure 8] Figure 8 illustrates the discovery of polygene DNA methylation markers for lung cancer. Figure A in Figure 8 is a table listing the sources and number of patients, and the methylation data from these patients was used to discover a series of CGIDs for lung cancer detection, disclosed in embodiments using the BCD method (Table 3) and CG IDs (Table 4) to determine the origin of the cancer tissue. Figure B (Detection) in the lower left panel of Figure 8 shows the sum of the methylation scores of these CG IDs (Table 3) for each test subject listed from 1 to 20 individuals (10 normal and 10 lung cancer). The polygene score categorically distinguishes between individuals with lung cancer and those with normal tissue. Figure C in the lower right panel of Figure 8 shows the methylation scores of CGIDs (Table 4) for detecting the origin of a specific tumor using data from individuals with eight different tumors (n=80). In these embodiments, the markers categorically distinguish lung cancer from cancers of other origins. [Figure 9] Figure 9 shows the validation of the accuracy and specificity of Polygene HKG-epi lung cancer detection and spec marker for lung cancer versus other cancers in TCGA methylation data (n=4166). Figure A in Figure 9 shows the detection rate of HKG-epi lung cancer detection / spec marker using DNA methylation data from patients with different cancers. Note the specificity for lung cancer. Figure B in Figure 9 shows the ROC plot of the specificity and sensitivity of the HKG-lung cancer detection marker for lung cancer in DNA methylation data from 4166 patients from TCGA. Figure C in Figure 9 shows the sensitivity and specificity for lung cancer versus cancers of other origins. [Figure 10] Figure 10 illustrates the discovery of polygene DNA methylation markers for prostate cancer. Figure A in Figure 10 is a table listing the sources and number of patients, and the methylation data from these patients was used to discover a set of CGIDs for the detection of prostate cancer, disclosed in embodiments using the BCD method (Table 5) and CG IDs (Table 5) to determine the origin-specific cancer tissue. Figure B (Detection) in the lower left panel of Figure 10 shows the sum of the methylation scores of these CG IDs (Table 5) for each test subject listed from 1 to 15 (5 normal and 10 prostate cancer). The polygene score categorically distinguishes between people with prostate cancer and normal people. Figure C in the lower right panel of Figure 10 shows the methylation scores of CG (Table 5) for detecting the origin-specific tumor tissue using data from people with eight different tumors (n=80). In these embodiments, the marker categorically distinguishes prostate cancer from cancers of other origins. [Figure 11] Figure 11 shows the validation of the accuracy and specificity of Polygene HKG-epi prostate cancer detection and spec marker for prostate cancer versus other cancers in TCGA methylation data (n=4166). Figure A of Figure 11 shows the detection rate of HKG-prostate cancer detection / spec marker using DNA methylation data from patients with different cancers. The specificity for prostate cancer should be noted. Figure B of Figure 11 is an ROC plot of the specificity and sensitivity of the HKG-prostate cancer detection marker for lung cancer using DNA methylation data from 4166 patients in TCGA. Figure C of Figure 11 shows the sensitivity and specificity for prostate versus cancers of other origins. [Figure 12]Figure 12 illustrates the discovery of polygene DNA methylation markers for breast cancer. Figure A in Figure 12 is a table listing the sources and number of patients, and the methylation data from these patients was used to discover a set of CGs for breast cancer detection, disclosed in embodiments using the BCD method (Table 6) and CG IDs (Table 7) to determine the origin of cancer. Figure B (Detection) in the lower left panel of Figure 12 shows the sum of the methylation scores of these CG IDs (Table 6) for each test subject listed from 1 to 27 (17 normal and 10 breast cancer). The polygene score categorically distinguishes between people with breast cancer and people with normal breast tissue. Figure C in the lower right panel of Figure 12 shows the methylation scores of CG IDs (Table 7) for detecting the origin of a specific tumor using data from people with eight different tumors (n=80). In these embodiments, the marker categorically distinguishes breast cancer from cancers of other origins. [Figure 13] Figure 13 shows the detection of non-invasive and invasive breast cancer by the HKG-epi breast cancer-detection polygene DNA methylation marker in the validation cohort GSE60185 (n=285). Figure A of Figure 13 is an ROC plot showing the area under the curve for the breast cancer polygene DNA methylation marker using DNA methylation data from 239 breast cancer patients, 17 non-breast cancer mastectomy patients, and 29 adjacent tissue samples. Sensitivity, specificity, and precision for all breast cancers are shown in B, and the predictive rates for DCIS (ductal carcinoma in situ), invasive breast cancer, and mixed breast cancer samples are shown in Figure C of Figure 13. Notably, the breast cancer marker detects very early-stage breast cancer (DCIS). [Figure 14]Figure 14 illustrates the validation of the accuracy and specificity of Polygene HKG-epi breast cancer detection and spec marker for breast cancer versus other cancers in TCGA methylation data (n=4166). Figure A of Figure 14 shows the detection rate of HKG-epi breast cancer detection / spec marker in DNA methylation data from patients with different cancers. Note the specificity for breast cancer. Figure B of Figure 14 is an ROC plot of the specificity and sensitivity of the HKG-breast cancer detection marker for detecting breast cancer using DNA methylation data from 4166 patients in TCGA. Figure C of Figure 14 shows the sensitivity and specificity for breast cancer versus cancers of other origins. [Figure 15] Figure 15 illustrates the discovery of polygene DNA methylation markers for colorectal cancer (CRC). Figure A in Figure 15 is a table listing the sources and number of patients, and the methylation data from these patients was used to discover a series of CGIDs for the detection of colorectal cancer, as disclosed in embodiments using the BCD method (Table 8) and CG IDs (Table 8) for determining the origin of cancer. Figure B (Detection) in the lower left panel of Figure 15 shows the sum of the methylation scores of these CG IDs for each test subject listed from 1 to 75 (25 normal and 50 colorectal cancer). The polygene score categorically distinguishes between people with cancer and people with normal tissue. Figure C in the lower right panel of Figure 15 shows the methylation scores of CGIDs that detect the specific origin of tumors using DNA methylation data from people with eight different tumors (n=80). In these embodiments, the markers categorically distinguish colorectal cancer from cancers of other origins. [Figure 16]Figure 16 shows the validation of the accuracy and specificity of Polygene HKG-epiCRC detection and speckle marker for colorectal cancer versus other cancers using the TCGA DNA methylation dataset (n=4166). Figure A of Figure 16 shows the detection rates of HKG-epiCRC detection / speckle marker using DNA methylation data from patients with different cancers. Note the specificity for colorectal cancer. Figure B of Figure 16 is an ROC plot of the specificity and sensitivity of the HKG-epi colorectal cancer detection marker for colorectal cancer using DNA methylation data from 4166 patients in TCGA. Figure C of Figure 16 shows the sensitivity and specificity for colorectal cancer versus cancers of other origins. [Figure 17] Figure 17 illustrates the discovery of polygene DNA methylation markers for pancreatic cancer. Figure A in Figure 17 is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of pancreatic cancer disclosed in the present invention, using the BCD method (Table 9) and CG IDs (Table 10) to determine the origin of the cancer. Figure B (Detection) in the lower left panel of Figure 17 shows the sum of the methylation scores of these CG IDs (Table 9) for each test subject listed from 1 to 32 (12 normal and 20 pancreatic cancer). The polygene score categorically distinguishes between people with pancreatic cancer and people with normal tissue. Figure C in the lower right panel of Figure 17 shows the methylation scores of CGIDs (Table 10) for detecting the specific origin of the tumor using data from people with 10 different tumors (n=100). In these embodiments, the markers categorically distinguish pancreatic cancer from cancers of other origins. [Figure 18]Figure 18 shows the validation of the accuracy and specificity of Polygene HKG-epi pancreatic cancer detection and spec marker for pancreatic cancer versus other cancers in TCGA methylation data (n=4854). Figure A of Figure 18 shows the detection rate of HKG-epi pancreatic cancer detection / spec marker using DNA methylation data from patients with different cancers. Note the specificity for pancreatic cancer. Figure B of Figure 18 is an ROC plot of the specificity and sensitivity of the HKG-epi pancreatic cancer detection marker for pancreatic cancer using DNA methylation data from 4854 patients in TCGA. Figure C shows the sensitivity and specificity for pancreatic cancer versus cancers of other origins. [Figure 19] Figure 19 illustrates the discovery of polygene DNA methylation markers for brain cancer (glioblastoma). Figure A of Figure 19 is a table listing the sources and number of patients, and the methylation data from these patients were used in the discovery of a series of CGIDs for the detection of brain cancer, disclosed in the present invention, using the BCD method (Table 11) and CG IDs (Table 11) to determine the specific origin of the cancer. Figure B (Detection / Spec) in the lower left panel shows the sum of the methylation scores of these CG IDs (Table 11) for each test subject listed from 1 to 16 (6 normal and 10 brain cancer). The polygene score categorically distinguishes between people with brain cancer, 110 people with other cancers, and people with normal tissue. [Figure 20] Figure 20 shows the validation of the accuracy and specificity of polygene HKG-epi brain cancer detection and spec marker for brain cancer versus other cancers in TCGA methylation data (n=4854). Figure A shows the detection rate of HKG-epi brain cancer detection / spec marker using DNA methylation data from patients with different cancers. The specificity for brain cancer should be noted. Figure B is an ROC plot of the specificity and sensitivity of the HKG-epi brain cancer detection marker for brain cancer using DNA methylation data from 4854 patients in TCGA. Figure C shows the sensitivity and specificity for brain cancer in the TCGA dataset (n=695). [Figure 21]Figure 21 illustrates the discovery of polygene DNA methylation markers for gastric [stomach] cancer. Figure A is a table listing the sources and number of patients, and the methylation data of these patients were used to discover a series of CGIDs for the detection of gastric cancer disclosed in the present invention, using the BCD method (Table 12) and CG IDs (Table 13) to determine the specific origin of the cancer. Figure B (Detection) in the lower left panel of Figure 21 shows the sum of the methylation scores of these CG IDs (Table 12) for each test subject listed from 1 to 28 (14 normal and 20 gastric cancer). The polygene score categorically distinguishes between people with gastric cancer and people with normal tissue. Figure C (Spec) in the lower right panel of Figure 21 shows the polygene methylation scores of people with 10 different tumors (n=100). In these embodiments, the marker categorically distinguishes gastric cancer from cancers of other origins. [Figure 22] Figure 22 shows the validation of the accuracy and specificity of polygene HKG-gastric detection and speck markers for gastric cancer versus other cancers in TCGA methylation data (n=4817). Figure A shows the detection rates of HKG-epi gastric cancer detection / speck markers using DNA methylation data from patients with different cancers. The specificity for gastric cancer should be noted. Figure B is an ROC plot of the specificity and sensitivity of the HKG-epi gastric cancer detection speck 1 marker for gastric cancer using DNA methylation data from 4420 patients in TCGA. Figure C is an ROC plot of the specificity and sensitivity of the HKG-epi gastric cancer-speck 1 marker for gastric cancer using DNA methylation data from 4854 patients in TCGA. Notably, there is significant cross-reactivity with colorectal cancer and esophageal cancer, demonstrating a common origin. [Figure 23]Figure 23 illustrates the discovery of polygene DNA methylation markers for ovarian cancer. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of ovarian cancer disclosed in the present invention, using the BCD method (Table 14) and CG IDs (Table 15) to determine the origin of the cancer. Figure B (Detection) in the lower left panel of Figure 23 shows the sum of the methylation scores of these CG IDs for each test subject listed from 1 to 15 (5 normal and 10 ovarian cancer). The polygene score categorically distinguishes between people with ovarian cancer and those with normal tissue. Figure C in the lower right panel of Figure 23 shows the methylation scores of CGIDs that detect specific tumor origins using data from people with 11 different tumors (n=110). In these embodiments, the markers categorically distinguish ovarian cancer from cancers of other origins. [Figure 24] Figure 24 illustrates the validation of the accuracy and specificity of Polygene HKG-epi ovarian cancer detection and spec marker for ovarian cancer versus other cancers in TCGA methylation data (n=6522). Figure A shows the detection rates of HKG-epi ovarian cancer detection / spec marker using DNA methylation data from patients with different cancers. Note the specificity for ovarian cancer. Figure B is an ROC plot of the specificity and sensitivity of HKG-epi ovarian cancer detection and spec marker for ovarian cancer in DNA methylation data from 4723 patients in TCGA. Figure C shows the sensitivity and specificity for ovarian cancer. [Figure 25]Figure 25 illustrates the discovery of polygene DNA methylation markers for cervical cancer. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of cervical cancer disclosed in the present invention, using the BCD method (Table 16) and CG (Table 17) to determine the origin of the cancer. Figure B (Detection) in the lower left panel of Figure 25 shows the sum of the methylation scores of these CG IDs (Table 16) for each test subject listed from 1 to 30 (20 normal and 10 with cervical cancer). The polygene score categorically distinguishes between people with cervical cancer and those with normal tissue. Figure C in the lower right panel of Figure 25 shows the methylation scores of CGIDs (Table 17) for detecting the specific origin of the tumor using data from people with eight different tumors (n=80). In these embodiments, the markers categorically distinguish cervical cancer from cancers of other origins, but note the measurable detection of some colorectal cancers. [Figure 26] Figure 26 shows the validation of the accuracy and specificity of Polygene HKG-Cervical Cancer Detection and Spec Marker for cervical cancer versus other cancers in TCGA methylation data (n=6522). Figure A shows the detection rate of HKG-Cervical Cancer Detection / Spec Marker using DNA methylation data from patients with different cancers. Note the specificity for cervical cancer. Figure B is an ROC plot of the specificity and sensitivity of the HKG-Cervical Cancer Detection Spec Marker for cervical cancer using DNA methylation data from 4420 patients in TCGA. Figure C shows the sensitivity and specificity for cervical cancer. [Figure 27]Figure 27 illustrates the discovery of polygene DNA methylation markers for head and neck squamous cell carcinoma (HNSC). Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for HNSC detection disclosed in the present invention, using the BCD method (Table 18) and CG (Table 19) to determine the origin of the cancer. Figure B in the lower left panel of Figure 27 shows the sum of the methylation scores for these CG IDs (Table 18) for each test subject listed from 1 to 140 (10 cancer, 10 normal, and 120 other cancers). Figure C shows polygene scores that categorically distinguish HNSC from normal tissue samples and categorically distinguish HNSC from cancers of other origins in the embodiment. [Figure 28] Figure 28 illustrates the validation of the accuracy and specificity of the Polygene HKG-epiHNSC detection / spec marker for HNSC versus other cancers in TCGA methylation data (n=4166). Figure A shows the detection rate of the HKG-epiHNSC detection / spec marker using DNA methylation data from patients with different cancers. Note the specificity of HNSC. Figure B is an ROC plot of the specificity and sensitivity of the HKG-epiHNSC detection marker for HNSC in DNA methylation data from 4166 patients in TCGA. Figure C shows the sensitivity and specificity for HNSC versus cancers of other origins. [Figure 29]Figure 29 illustrates the discovery of polygene DNA methylation markers for esophageal cancer. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of esophageal cancer, as disclosed in the embodiments, using the BCD method (Table 20) and CGID (Table 21) to determine the specific origin of the cancer. Figure B in the lower left panel of Figure 29 shows the sum of the methylation scores for these CGIDs (Table 20) for each test subject listed from 1 to 22 (8 normal and 14 cancer). Figure C shows the polygene scores that categorically distinguish esophageal cancer from normal tissue and categorically distinguish esophageal cancer from other origins from esophageal cancer listed from 1 to 220 (20 cancer, 190 other cancers, and 10 normal blood). [Figure 30] Figure 30 shows the validation of the accuracy and specificity of the polygene HKG-epi esophageal cancer detection / spec marker for esophageal cancer versus other cancers in TCGA methylation data (n=7102). Figure A shows the detection rate of the HKG-epi esophageal cancer detection / spec marker using DNA methylation data from patients with different cancers. Note the specificity for esophageal cancer. Figure B is an ROC plot of the specificity and sensitivity of the HKG-epi esophageal cancer detection marker for esophageal cancer in DNA methylation data from 4166 patients in TCGA. Figure C shows the sensitivity and specificity for esophageal cancer versus cancers of other origins. [Figure 31]Figure 31 illustrates the discovery of polygene DNA methylation markers for bladder cancer. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used in the discovery of a series of CGIDs for the detection of bladder cancer disclosed in the embodiments, using the BCD method (Table 22) and CGID (Table 23) to determine the specific origin of the cancer. Figure B (Detection) in the lower left panel of Figure 31 shows the sum of the methylation scores of these CGIDs (Table 22) for each test subject listed from 1 to 15 (5 normal, 10 bladder cancer). Figure C in the lower right panel of Figure 31 shows the methylation scores of CGIDs (Table 23) for detecting the specific origin of the tumor using data from people with 13 different tumors (n=130). In these embodiments, the markers distinguish bladder cancer from cancers of other origins. Measurable detection of colorectal cancer by these markers should also be noted. [Figure 32] Figure 32 shows the validation of the accuracy and specificity of Polygene HKG-epiBladder detection and spec markers for bladder cancer versus other cancers in TCGA (n=4723). Figure A shows the detection rates of HKG-epi bladder cancer spec marker (A) and detection marker (B) in DNA methylation data from patients with different cancers (A) and bladder cancer (B). Figure C is an ROC plot of the specificity and sensitivity of the HKG-epi bladder cancer spec marker for bladder cancer using DNA methylation data from 4420 patients in TCGA. Figure D is an ROC plot of the specificity and sensitivity of the HKG-epi bladder cancer detection marker for bladder cancer (n=440). [Figure 33]Figure 33 illustrates the discovery of polygene DNA methylation markers for kidney cancer. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of kidney cancer and determination of the specific origin of the cancer (Table 24) as disclosed in the embodiments using the BCD(hypo) method. Figure B (Detection / Spec) in the lower left panel of Figure 33 shows the sum of the methylation scores of these CGIDs (Table 24) for each test subject listed from 1 to 226 (180 other cancers, 10 healthy blood, 6 normal kidneys, and 30 kidney cancers). In these embodiments, the polygene score categorically distinguishes kidney cancer, other cancers, and normal blood. [Figure 34] Figure 34 shows the validation of the accuracy and specificity of Polygene HKG-epi kidney cancer detection and spec marker against other cancers and normal tissues using TCGA DNA methylation data (n=7102). Figure A shows the detection rate of HKG-epi kidney cancer detection / spec marker using DNA methylation data from different cancers. The specificity for kidney cancer should be noted. Figure B is an ROC plot of the specificity and sensitivity of the HKG-cervical cancer detection spec marker for kidney cancer using DNA methylation data from 6367 cancers in TCGA. Figure C shows the sensitivity and specificity for renal (kidney) cancer. Of further note are the crossovers with brain, HCC, and testicular cancers. [Figure 35] Figure 35 illustrates the discovery of polygene DNA methylation markers for testicular cancer. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of testicular cancer and the determination of origin-specific cancers (Table 25) as disclosed in embodiments using the BCD(hypo) method. Figure B (Detection / Spec) in the lower left panel of Figure 35 shows the sum of the methylation scores for these CG IDs (Table 25) for each test subject listed from 1 to 190 (10 for testicular cancer, 180 for other cancers and blood). In these embodiments, the polygene score categorically distinguishes testicular cancer, normal blood, and other cancers. [Figure 36] Figure 36 illustrates the validation of the accuracy and specificity of Polygene HKG-epi testicular cancer detection and spec markers for testicular cancer versus other normal tissues and cancers in TCGA methylation data (n=7102). Figure A shows the detection rate of HKG-epi testicular cancer detection / spec markers using DNA methylation data from patients with different cancers. Note the specificity for testicular cancer. Figure B is an ROC plot of the specificity and sensitivity of the HKG-epi testicular cancer detection spec marker for testicular cancer using DNA methylation data from 6367 patients in TCGA. Figure C shows the sensitivity and specificity for testicular cancer. [Figure 37] Figure 37 illustrates the discovery of pan-cancer polygene DNA methylation markers for 13 common cancers. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of 13 common cancers (Table 26) (bladder cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer CRC, esophageal cancer, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, and stomach cancer) disclosed in embodiments using the BCD method (Table 26). Figure B shows the sum of the methylation scores of these CGIDs for each test subject listed from 1 to 180 (170 cancer patients, 10 healthy blood samples). In these embodiments, the polygene score distinguishes cancer from normal tissue. [Figure 38]Figure 38 illustrates the validation of the accuracy and specificity of the polygene HKG epi pan-cancer marker in TCGA methylation data (n=7102). Figure A shows the methylation scores calculated using the epi pan-cancer polygene DNA methylation marker from 13 different cancer patients using TCGA data. Figure B is an ROC plot of the specificity and sensitivity of the HKG-pan-cancer detection and spec marker using DNA methylation data for all cancers from 4878 patients in TCGA. Figure C shows the overall sensitivity and specificity of the pan-cancer marker for detecting cancer. In these embodiments, one or more colors are used, e.g., orange (weighted methylation score) and blue (detection of one BCD marker per sample is scored as positive cancer). [Figure 39] Figure 39 illustrates the discovery of polygene DNA methylation markers for melanoma. Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for melanoma detection (Table 28) as disclosed in embodiments using the BCD method (Table 45). Figure B shows the sum of the methylation scores of these CGIDs for each test subject listed from 1 to 220 (other cancers and healthy blood) and 10 patients with melanoma. In these embodiments, the polygene score distinguishes melanoma, other cancers, and normal tissue. [Figure 40] Figure 40 illustrates the validation of the accuracy and specificity of Polygene HKG-epi melanoma detection and spec markers for melanoma versus other normal tissues and cancers in TCGA methylation data (n=7102). Figure A shows the detection rate of HKG-epi melanoma detection / spec markers using DNA methylation data from patients with different cancers. Note the specificity for melanoma (over-detection with liver cancer, brain cancer, and prostate cancer). Figure B is an ROC plot of the specificity and sensitivity of the HKG-melanoma detection spec marker for melanoma using DNA methylation data from 6367 patients in TCGA. Figure C shows the sensitivity and specificity for melanoma. [Figure 41]Figure 41 illustrates the discovery of polygene DNA methylation markers for hematological cancer (acute myeloid leukemia (AML)). Figure A is a table listing the sources and number of patients, and the methylation data from these patients were used to discover a series of CGIDs for the detection of hematological cancer AML (Table 27) as disclosed in embodiments using the BCD method (Table 27). Figure B shows the sum of the methylation scores of these CGIDs for each test subject listed from 1 to 10 (healthy blood) and 10 patients with AML. In these embodiments, the polygene score distinguishes AML from normal blood. [Figure 42] Figure 42 illustrates the validation of the accuracy and specificity of Polygene HKG-epiAML detection and spec markers against normal blood in GSE40279 and GSE61496 (n=968) compared to AML in GSE86409 (n=79) and TCGA (n=140). Figure A shows the detection rate of HKG-epiAML detection / spec markers using DNA methylation data from patients with AML and individuals with healthy blood. Figure B is an ROC plot of the specificity and sensitivity of the HKG-AML detection spec marker for AML using DNA methylation data from GSE86409 (n=79), TCGA (n=140), GSE40279, and GSE61496 (n=968). Figure C shows the sensitivity and specificity for AML. [Figure 43] Figure 43 shows the validation of primers selected to detect different cancers exhibiting BCD characteristic ~0 methylation in plasma derived from normal individuals (each sample is a mixture of plasma from normal patients). A first PCR1 reaction targeting specific CGs was performed using sequence-targeted primers. After the second PCR, the amplified fragments were purified and subjected to next-generation sequencing. DNA methylation was quantified at the respective CG ID positions shown. [Figure 44] Figure 44 shows the validation of the primers shown, selected to detect different cancers exhibiting BCD characteristic ~0 methylation in plasma derived from normal individuals (each sample is a mixture of plasma from normal patients). [Figure 45] Figure 45 shows the primer design for multiple amplification and sequencing. Note that the first PCR reaction targets a specific region of interest, while the PCR1 primer has a complementary sequence to the second PCR2 primer. The second set of primers introduces the index for each patient, as well as reverse and forward sequencing primers. [Figure 46] Figure 46 shows the optimization of PCR conditions for detecting prostate cancer. The right panel shows multiplex PCR1 reactions using various primer concentrations, such as the DNA shown, for three prostate cancer markers: HIF3A 232 bp, TPM4 213 bp, and CTTN 199 bp. [Figure 47] Figure 47 shows a bioinformatics workflow for determining DNA methylation levels. PCR2 products are combined, quantified, and purified, then next-generation sequencing is performed using a Miseq Illumina sequencer. The sequences are demultiplexed, and FASTQ files are generated for each patient and analyzed using the workflow shown in the scheme. DNA methylation scores are calculated for each patient. [Modes for carrying out the invention]

[0134] Detailed explanation All descriptions in the drawings are for illustrative purposes only and are not intended to limit the scope of the claimed subject matter.

[0135] Embodiment 1: Discovery of categorically unmethylated CGIDs in normal tissue and blood DNA across hundreds of individuals.

[0136] Cell-free DNA derived from tumors is known to be found in bodily fluids such as plasma and urine, as well as in feces. It has also been established that the DNA methylation profile of CF tumor DNA is similar to that of tumor DNA (Dominguez-Vigil et al., 2018). Extensive data have established that tumor DNA is distinctly methylated compared to normal tissue (Luczak & Jagodzinski, 2006). Therefore, many groups have attempted to describe the CGID location (CG ID in the Illumina 450K manifest) of DNA that is distinctly methylated between cancerous tissue and tissue of normal origin, e.g., liver cancer and adjacent liver tissue, using logistic regression. However, because these methods measure quantitative differences between cancer and non-transformed tissue rather than categorical qualitative differences, the quantitative differences between tumor and normal tissue are diluted and erased by CF DNA from normal tissue, leading to false negatives and reduced sensitivity. Furthermore, other tissues not included in the analysis may have similar DNA methylation profiles to tumor DNA, and this can lead to false positives because most studies compare tumor DNA only to its untransformed counterparts, rather than to other tissues. Because unpredictable amounts of DNA have been detected in CF DNA (Breitbach et al., 2014), measured DNA methylation reflects a complex of unknown and unpredictable mixtures of tissue DNA and tumor DNA from different sources. Thousands of tumor samples have been subjected to genome-wide DNA methylation analysis using Illumina 450K arrays and discovered in the public domain (TCGA). Examining the methylation profiles of many normal and cancerous tissues, the inventors found that there is a significant group of CGs in the genome that are not completely methylated in all normal tissues but are methylated in tumor DNA. A subset of these sites is not methylated across a large number of individuals whose DNA methylation has been profiled in the public domain. The inventors also noticed that in many cancers, these strongly unmethylated sites are methylated in cancer. Thus, this creates a qualitative "categorical difference" between tumor DNA and all other DNA that may be found in the blood. Using deep next-generation sequencing, even minute methylated molecules can be easily identified against a background of completely unmethylated copies.

[0137] Database; Illumina 450K DNA methylation data We used publicly available databases of normalized beta values ​​for methylation of ~450,000 CGs across the entire human genome from numerous individuals, deposited in either the Gene Expression Omnibus (GEO) https: / / www.ncbi.nlm.nih.gov / geo / or The Cancer Genome Atlas (TCGA) https: / / cancergenome.nih.gov / . We used the following databases to collect data from many normal tissues and blood DNA: GSE50192, GSE50192, We obtained a list of strong non-methylated CG IDs in GSE40279.

[0138] DNA from leukocytes is one of the main sources of CF DNA in plasma. The inventors first used Illumina 450K data of GSE50192 and the logical formulas COUNTIF and IF functions in Excel to generate a list of 47,981 unmethylated CGIDs in all individuals of 17 different somatic cell human tissues: NmCGID x =COUNTIF (betaCGID x n1:n i ,">0.1") umCGID x =IF(NmCGID x (=0, TRUE, FALSE) NmCGIDx = Number of normal subjects with methylated CGIDx umCGIDx = Unmethylated CGIDx in all subjects betaCGIDx = methylation value of the given CGIDX x = Any CGID on the Illumina 450k array n1 = the first subject in the array n i = The last subject in the array.

[0139] Next, the inventors used the same criteria to generate a list of 68,260 unmethylated CGIDs (UMCGIDs) in blood DNA from 312 individuals. Then, the inventors superimposed the lists of 47,981 and 68,260 CG IDs to obtain a list of 33,477 CG IDs that were not methylated in both blood and somatic tissues of all individuals (Figure 1A). To increase the robustness of this list of unmethylated CG IDs, the inventors plotted a list of 60,379 CG IDs in an Illumina 450K array of whole blood DNA from 656 men and women (GSE40279) aged 19 to 101 years. These are strongly unmethylated sites in the blood of hundreds of individuals, regardless of sex or age. This list of 60,379 CG IDs overlapped with a list of 33,477 CG IDs that were not methylated in both body tissues and blood, generating a final list of 28,754 CG IDs used to identify categorical methylation markers for cancer. This list includes the positions of CG IDs that are not strongly methylated between tissues and individuals. This includes placement.

[0140] To identify categorically distinct DNA methylation sites between cancer and normal tissue, the inventors investigated whether any of these 28,754 CG IDs were methylated in different cancers. Following their investigation of tumor DNA methylation data, the inventors found that methylation of a subset of these 28,754 CG IDs was common in tumor DNA from individual patients. However, not all individuals are methylated at the same sites. Therefore, a combination of CG IDs is necessary to detect cancer with high specificity. Accordingly, the inventors discovered polygeneic combinations of CG IDs for cancer detection.

[0141] The inventors used 10 to 50 DNA methylation profiles from the public domain of TCGA or GEO as a “discovery set” to discover a polygene set of CGIDs whose methylation status differs “categorically” between tumor and normal tissue and which can detect cancer with the highest sensitivity and specificity. These CGIDs were then tested with hundreds of TCGA and GEO tumor DNA methylation array data as a “validation set” to validate the sensitivity and specificity of the polygene DNA methylation markers for detecting cancer, as disclosed in Embodiment 2.

[0142] Embodiment 2: Binary Category Discrimination (BCD) method for detecting cancer in cell-free DNA. Using the following publicly available databases containing normalized beta values ​​for methylation of approximately 450,000 CGs (CG IDs) across the entire human genome, we derived a list of cancer-specific DNA methylation markers.

[0143] [Table 29] [Table 30] [Table 31] [Table 32] [Table 33] [Table 34] [Table 35] [Table 36] [Table 37] [Table 38] [Table 39] [Table 40] [Table 41] [Table 42] [Table 43] [Table 44] [Table 45] [Table 46]

[0144] BCD method The following are steps of a binary category classification (BCD) method used in embodiments to discover polygene DNA methylation markers for the early prediction of different cancers.

[0145] 28,754 CGIDs that were not strongly methylated in normal tissue were filtered out. For the discovery cohort, we used Microsoft Excel's COUNTIF and IF functions to plot 28,754 CG IDs that were not strongly methylated in normal tissue. These CG IDs were categorically methylated in certain cancers but not in unaffected or normal tissue. NmcCGIDx=COUNTIF (betaCGIDxCancer n1:n i ,">0.2") NmnCGIDx=COUNTIF (betaCGIDxNormal n1:ni ," >0.1") DMCGIDx = IF((AND(NmcCGIDx > 0, NmnCGIDx = 0)), "TRUE", "FALSE") DM CGIDx was sorted from the highest number to the lowest number. The positions of the top 20 TRUE DM CGIDx were selected. NmcCGIDx = the number of cancer patients having methylated CGIDx Nmn = the number of normal adjacent or similar tissue samples having methylated CGIDx betaCGIDx = the methylation level of CGIDx n = patients from 1 to i DM = differentially methylated CGIDx

[0146] The inventors noticed that testicular cancer and kidney cancer showed a widespread lack of methylation in CG IDs that were highly methylated in all tissues. Thus, for what is not categorically methylated in cancer but is methylated in normal tissue, an improved BCD method called "BCDhypo" was used to discover the positions of categorically differentially methylated CG IDs in testicular cancer and kidney (renal) cancer. The following steps were used to discover the positions of CGIDs that were differentially hypomethylated in testicular cancer and kidney cancer.

[0147] In the discovery cohort, the Excel COUNTIF and IF functions were used to draw hypomethylated CGIDs in testis or kidney that were fully methylated in normal tissue. NucCGIDx = COUNTIF (betaCGIDxCancer n1:ni, "<0.2") NunCGIDx = COUNTIF (betaCGIDxNormal n1:ni, "<0.9") DHMCGIDx = IF((AND(NucCGIDx > 0, NunCGIDx = 0)), "TRUE", "FALSE") The positions of the DHM CGIDs were sorted from the highest number to the lowest number. The top 20 TRUE DHM sites were selected and a penalized regression analysis was performed. NucCGIDx = the number of cancer patients with unmethylated CGID X NunCGIDx = the number of normal tissue samples with unmethylated CGID X n = patients from 1 to i DHM = differentially hypomethylated CGID

[0148] Next, the inventors performed penalized regression using the top 20 DM (or DHM) CGIDx with the package that penalizes in R, and mapped the minimum combination of CGIDx that predicts cancer with the highest sensitivity and specificity. To determine the regression coefficients between the combination of polygenes of CGIDx and the methylation levels of these CGIDs in cancer, the combination of polygenes of CGIDx was further tested with a multivariate linear regression equation. Using this model, the methylation score for each patient with typical cancer was calculated. JPEG0007850764000048.jpg1359Ms = methylation score, α = intercept, β i = CG ID i of the coefficient, CG i = combination of methylation levels per CG. 1 to i = number of combined CGs.

[0149] Embodiment 3: Discovery of polygenic DNA methylation markers for liver cancer (HCC). The inventors used normalized Illumina 450K DNA methylation data from GSE61258 (normal liver) and 66 randomly selected samples from the TCGA HCC collection of HCC DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGIDx, which in Embodiment 1 was identified as strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a set of binary categorically distinctly methylated CGID polygenes that detected HCC with high sensitivity and specificity in the training cohort (Figure 5B, Table 1) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGID as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 randomly selected DNA methylation samples from TCGA representing eight different tumor types. Using this training cohort, the inventors discovered a polygene set of CGID that was distinctly methylated between HCC and other tumors (Figure 5C, Table 2) (spec). A weighted DNA methylation score was developed for CGID as described in Embodiment 2.

[0150] Embodiment 4: Usefulness of HCC polygene DNA methylation marker for detecting HCC. Next, the inventors demonstrated that the weighted HCC DNA methylation score detected HCC in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from GSE76269 of 227 HCC patients from CGID, Table 1. Using this method, 95% of HCC samples were detected as HCC (Figure 6C). The ROC curve shown in Figure 6A reveals the specificity (1) and sensitivity (0.96) of this methylation score for detecting cancer. Next, the inventors used the “validation cohort” along with methylation data from GSE75041 and TCGA of HCC and eight other cancers to demonstrate the usefulness of combining the speculative and detecting DNA methylation score for detecting HCC and distinguishing HCC from other cancers. The ROC curve shown in Figure 7B reveals the specificity (0.97) and sensitivity (0.95) of this methylation score for distinguishing HCC from other normal tissues and other cancers. These DNA methylation markers and the calculated methylation scores can be used to screen for and detect cancer early in individuals at risk and in generally healthy individuals using non-human biomaterials such as tissue, feces, saliva, plasma, and urine.

[0151] Embodiment 5: Discovery of polygene DNA methylation markers for lung cancer. The inventors have obtained data from 10 individuals with GSE61258 (normal lungs) and lung cancer DNA. Normalized Illumina 450K DNA methylation data from 10 randomly selected samples from the chilled data TCGA lung cancer collection were used as a “training” cohort. The inventors first listed the “training cohort” dataset of 28754 CGIDs, which were found in Embodiment 1 as sites strongly unmethylated across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically methylated CGIDs that detected lung cancer (including both adenocarcinoma and squamous cell carcinoma samples) in the training cohort with high sensitivity and specificity (Figure 8B, Table 3) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 DNA methylation samples randomly selected from TCGA representing eight different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between lung cancer and other tumors (Figure 8C, Table 4) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0152] Embodiment 6: Usefulness of a lung cancer polygene DNA methylation marker for detecting lung cancer. Next, the inventors demonstrated that the weighted lung cancer DNA methylation score and threshold developed in Embodiment 5 (Detection) could detect lung cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from GSE66836, GSE63704, GSE76269, and TCGA from 919 lung cancer patients. Using this method, 96% of lung cancer samples were detected as lung cancer (Figure 9A). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for lung cancer and eight other cancers to demonstrate the usefulness of combining the speckled and detected DNA methylation scores to distinguish lung cancer from other cancers (Figure 9A). The ROC curve shown in Figure 9B reveals the specificity (0.96) and sensitivity (0.84) of this methylation score for detecting lung cancer from other normal tissues and other cancers (Figure 9C). These DNA methylation markers and the calculated methylation scores can be used to detect cancer early in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0153] Embodiment 7: Discovery of polygene DNA methylation markers for prostate cancer. The inventors used normalized Illumina 450K DNA methylation data from five individuals from GSE52955 (normal prostate) and from 10 randomly selected samples from the TCGA prostate cancer collection of prostate cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGID, which in Embodiment 1 was identified as strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected prostate cancer with high sensitivity and specificity in the training cohort (Figure 10B, Table 5) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 randomly selected DNA methylation samples from TCGA representing eight different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between prostate cancer and other tumors (Figure 10C, Table 5) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0154] Embodiment 8: Usefulness of a prostate cancer polygene DNA methylation marker for detecting prostate cancer. Next, the inventors demonstrated that the weighted prostate cancer DNA methylation score and threshold developed in Embodiment 7 (Detection) could detect prostate cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from GSE73549, GSE2955, and TCGA from 430 prostate cancer patients. Using this method, 99% of prostate cancer samples were detected as prostate cancer (Figure 11A). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for prostate cancer and eight other cancers to demonstrate the usefulness of combining the spec and detection DNA methylation scores to distinguish prostate cancer from other cancers (Figure 11A). The ROC curve shown in Figure 11B reveals the specificity (0.99) and sensitivity (0.98) of this methylation score for detecting prostate cancer from other normal tissues and other cancers (Figure 11C). These DNA methylation markers and the calculated methylation scores can be used to detect cancer early in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0155] Embodiment 9: Discovery of polygene DNA methylation markers for breast cancer. The inventors used normalized Illumina 450K DNA methylation data from 17 individuals from GSE60185 (normal breast) and from 10 randomly selected samples from the TCGA breast cancer collection of breast cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGIDs, which were found in Embodiment 1 as strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected breast cancer with high sensitivity and specificity in the training cohort (Figure 12B, Table 6) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 randomly selected DNA methylation samples from TCGA representing eight different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between breast cancer and other tumors (Figure 12C, Table 7) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0156] Embodiment 10: Usefulness of breast cancer polygene DNA methylation markers for detecting breast cancer. Next, the inventors demonstrated that the weighted breast cancer DNA methylation score and threshold developed in Embodiment 9 (Detection) could detect breast cancer in a “validation cohort” containing 891 breast cancer patients using normalized Illumina 450K DNA methylation beta values ​​from GSE60185, GSE75067, and TCGA. Using this method, 91% of breast cancer samples were detected as breast cancer (Figure 13A), with both DCIS and invasive cancer being detected. Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for breast cancer and eight other cancers to demonstrate the usefulness of combining the speckled and detection DNA methylation scores for distinguishing breast cancer from other cancers (Figure 14A). The ROC curve shown in Figure 14B reveals the specificity (0.89) and sensitivity (0.87) of this methylation score for distinguishing breast cancer from other normal tissues and other cancers (Figure 14C). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of breast cancer in women at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0157] Embodiment 11: Discovery of polygene DNA methylation markers for colorectal cancer (CRC). The inventors used normalized Illumina 450K DNA methylation data from 25 individuals with GSE(32146) (normal) and 50 randomly selected samples from the TCGA colorectal cancer collection of colorectal cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGIDs, which were found in Embodiment 1 as strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically methylated CGIDs that detected colorectal cancer with high sensitivity and specificity in the training cohort (Figure 15B, Table 8) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 randomly selected DNA methylation samples from TCGA representing eight different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between colorectal cancer and other tumors (Figure 15C, Table 8) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0158] Embodiment 12: Usefulness of a colorectal cancer polygene DNA methylation marker for detecting colorectal cancer. Next, the inventors demonstrated that the weighted colorectal cancer DNA methylation score and threshold developed in Embodiment 11 (Detection) could detect colorectal cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from GSE69550 and TCGA from 459 colorectal cancer patients. Using this method, 98% of colorectal cancer samples were detected as colorectal cancer (Figure 16A). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for colorectal cancer and eight other cancers to demonstrate the usefulness of combining the speckled and detected DNA methylation scores for distinguishing colorectal cancer from other cancers (Figure 16A). The ROC curve shown in Figure 16B reveals the specificity (0.96) and sensitivity (0.98) of this methylation score for distinguishing colorectal cancer from other normal tissues and other cancers (Figure 16C). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of cancer in individuals at risk of CRC and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0159] Embodiment 13: Discovery of polygene DNA methylation markers for pancreatic cancer. The inventors used normalized Illumina 450K DNA methylation data from 12 individuals from GSE53051 (normal) and 20 randomly selected samples from the TCGA collection of pancreatic cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGID, which in Embodiment 1 was identified as strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected pancreatic cancer with high sensitivity and specificity in the training cohort (Figure 17B, Table 9) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 100 randomly selected DNA methylation samples from the TCGA representing 10 different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between pancreatic cancer and other tumors (Figure 17C, Table 10) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0160] Embodiment 14: Usefulness of a pancreatic cancer polygene DNA methylation marker for detecting pancreatic cancer. Next, the inventors demonstrated that the weighted pancreatic cancer DNA methylation score and threshold developed in Embodiment 13 (Detection) could detect pancreatic cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from 891 pancreatic cancer patients from TCGA. Using this method, 86% of pancreatic cancer samples were detected as pancreatic cancer (Figure 18A). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for pancreatic cancer and nine other cancers to demonstrate the usefulness of combining the speckled and detection DNA methylation scores for distinguishing pancreatic cancer from other cancers (Figure 18A). The ROC curve shown in Figure 18B reveals the specificity (0.93) and sensitivity (0.86) of this methylation score for detecting pancreatic cancer and distinguishing pancreatic cancer from other normal tissues and other cancers (Figure 18C). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of cancer in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0161] Embodiment 15: Discovery of polygene DNA methylation markers for brain cancer. The inventors used normalized Illumina 450K DNA methylation data from 10 individuals from GSE65820 (normal) and 10 samples randomly selected from the TCGA collection of brain cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset of 28754 CGIDs, which were found in Embodiment 1 as sites strongly unmethylated across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a set of binary categorically methylated CGIDs that detected brain cancer with high sensitivity and specificity in the training cohort (Figure 19B, Table 11) (detection), using the BCD method described in Embodiment 2. Weighted DNA methylation scores and cancer thresholds were developed for the CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 110 DNA methylation samples randomly selected from the TCGA representing 11 different tumor types. Using this training cohort, the inventors found that detection of CGID also distinguishes brain cancer from other tumors (Figure 19B, Table 11) (Detection-Spec). As described in Embodiment 2, a weighted DNA methylation score was developed for CGID.

[0162] Embodiment 16: Usefulness of a brain cancer polygene DNA methylation marker for detecting brain cancer. Next, the inventors demonstrated that the weighted brain cancer DNA methylation score and threshold developed in Embodiment 15 (Detection) detected brain cancer in a "validation cohort" that included normalized Illumina 450K DNA methylation beta values from 689 brain cancer patients from TCGA, 40 patients from GSE58298, and 136 patients from GSE36278. Using this method, 91% - 97% of brain cancer samples were detected as brain cancer (Figure 20A). Next, the inventors demonstrated the usefulness of the same CGIDs for distinguishing brain cancer from other cancers using the "validation cohort" along with methylation data from GSE and TCGA of brain cancer and nine other types of cancer (Figure 20A). The ROC curve shown in Figure 20B reveals the specificity (1) and sensitivity (0.97) of this methylation score for detecting brain cancer from other normal tissues and other cancers (Figure 20C). These DNA methylation markers and methylation scores calculated from methylation values can be used for early cancer detection in at-risk and generally healthy individuals using different biological materials from tissue, feces, saliva, plasma, and urine from patients.

[0163] Embodiment 17: Discovery of a polygenic DNA methylation marker for gastric cancer. The inventors used normalized Illumina 450K DNA methylation data from 18 individuals from GSE99553 (normal) and 10 randomly selected samples from the TCGA collection of gastric cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGID, which in Embodiment 1 was found to be strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected gastric cancer with high sensitivity and specificity in the training cohort (Figure 21B, Table 12) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 100 randomly selected DNA methylation samples from the TCGA representing 11 different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between gastric cancer and other tumors (Figure 21C, Table 13) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0164] Embodiment 18: Usefulness of a gastric cancer polygene DNA methylation marker for detecting gastric cancer. Next, the inventors demonstrated that the weighted gastric cancer DNA methylation score and threshold developed in Embodiment 17 (Detection) could detect gastric cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from 397 gastric cancer patients from TCGA. Using this method, 88% of gastric cancer samples were detected as gastric cancer (Figure 23A). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for gastric cancer and 10 other cancers to demonstrate the usefulness of combining the speckled and detected DNA methylation scores to distinguish gastric cancer from other cancers (Figure 23A). The ROC curve shown in Figure 22B reveals the specificity (0.9) and sensitivity (0.9) of this methylation score for detecting gastric cancer from other normal tissues and other cancers (Figure 22C). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for cancer screening and early detection in at-risk individuals and generally healthy individuals using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0165] Embodiment 19: Discovery of polygene DNA methylation markers for ovarian cancer. The inventors used normalized Illumina 450K DNA methylation data from five individuals with GSE65820 (normal) and from 10 randomly selected samples from the TCGA collection of ovarian cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGIDs, which were found in Embodiment 1 as sites strongly unmethylated across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected ovarian cancer with high sensitivity and specificity in the training cohort (Figure 23B, Table 14) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 100 randomly selected DNA methylation samples from the TCGA representing 10 different tumor types and blood. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between ovarian cancer and other tumors (Figure 23C, Table 15) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0166] Embodiment 20: Usefulness of ovarian cancer polygene DNA methylation markers for detecting ovarian cancer. Next, the inventors demonstrated that the weighted ovarian cancer DNA methylation score and threshold developed in Embodiment 19 (Detection) could detect ovarian cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from 114 ovarian cancer patients from TCGA. Using this method, 86% of ovarian cancer samples were detected as ovarian cancer (Figure 24A). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for ovarian cancer and nine other cancers to demonstrate the usefulness of the speck DNA methylation score for distinguishing ovarian cancer from other cancers (Figure 24A). The ROC curve shown in Figure 24B reveals the specificity (0.99) and sensitivity (1) of this methylation score for distinguishing ovarian cancer from other normal tissues and other cancers (Figure 24C). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of cancer in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0167] Embodiment 21: Discovery of polygene DNA methylation markers for cervical cancer. The inventors used normalized Illumina 450K DNA methylation data from 20 individuals from GSE46306 (normal) and 10 randomly selected samples from the TCGA collection of cervical cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGIDs, which were found in Embodiment 1 as sites strongly unmethylated across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected cervical cancer with high sensitivity and specificity in the training cohort (Figure 25B, Table 16) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 randomly selected DNA methylation samples from the TCGA representing eight different tumor types and blood types. Using this training cohort, the inventors discovered a set of polygenes for CGID that were distinctly methylated between cervical cancer and other tumors (Figure 25C, Table 17) (spec). A weighted DNA methylation score was developed for CGID as described in Embodiment 2.

[0168] Embodiment 22: Usefulness of a cervical cancer polygene DNA methylation marker for detecting cervical cancer. Next, the inventors demonstrated that the weighted cervical cancer DNA methylation score and threshold developed in Embodiment 21 (Detection) could detect cervical cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from 313 cervical cancer patients from TCGA. Using this method, 91% of the cervical cancer samples were detected as cervical cancer (Figure 26A). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for cervical cancer and nine other cancers to demonstrate the usefulness of the speck DNA methylation score for distinguishing cervical cancer from other cancers (Figure 26A). The ROC curve shown in Figure 26B reveals the specificity (0.9) and sensitivity (0.9) of this methylation score for detecting cervical cancer and distinguishing cervical cancer from other normal tissues and other cancers (Figure 26C). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of cancer in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0169] Embodiment 23: Discovery of polygene DNA methylation markers for head and neck squamous cell carcinoma (HNSC). The inventors used normalized Illumina 450K DNA methylation data from 10 individuals with GSE(52068) (normal) and 10 samples randomly selected from the TCGA cancer collection of HNSC DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGIDs, which were found in Embodiment 1 as strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected HNSCs with high sensitivity and specificity in the training cohort (Figure 27B, Table 18) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 DNA methylation samples randomly selected from the TCGA representing 12 different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between HNSCs and other tumors (Figure 27C, Table 19) (speck).

[0170] Embodiment 24: Usefulness of a polygene DNA methylation marker for detecting HNSCs in head and neck squamous cell carcinoma (HNSC). Next, the inventors demonstrated that the weighted HNSC DNA methylation score and threshold developed in Embodiment 23 (Detection) are normalized from GSE52068 in Illumina 450K The inventors demonstrated the detection of HNSCs in a "validation cohort" containing DNA methylation beta values. Using this method, 88%–96% of HNSC samples were detected (Figure 28A). Next, the inventors used the "validation cohort" along with methylation data from GSE and TCGA of HNSCs and 12 other cancers to demonstrate the usefulness of the DNA methylation detection score for distinguishing HNSCs from other cancers (Figure 28A). The ROC curve shown in Figure 28B reveals the specificity (0.86) and sensitivity (0.88) of this methylation score for distinguishing HNSCs from other normal tissues and other cancers (Figure 28C). The marker also detects several other cancers (its specificity for these cancers is limited due to its relatively high sensitivity). These DNA methylation markers and the calculated methylation scores from methylation values ​​can be used for early cancer detection in at-risk individuals and generally healthy individuals using different biomaterials from patient tissues, feces, saliva, plasma, and urine.

[0171] Embodiment 25: Discovery of polygene DNA methylation markers for esophageal cancer. The inventors used normalized Illumina 450K DNA methylation data from 10 individuals with GSE(52068) (normal) and 10 randomly selected samples from the TCGA cancer collection of esophageal cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGIDs, which were found in Embodiment 1 as sites strongly unmethylated across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically methylated CGIDs that detected esophageal cancer with high sensitivity and specificity in the training cohort (Figure 29B, Table 20) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 randomly selected DNA methylation samples from TCGA representing 12 different tumor types. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between esophageal cancer and other tumors (Figure 29C, Table 21) (spec). .

[0172] Embodiment 26: Usefulness of an esophageal cancer polygene DNA methylation marker for detecting esophageal cancer. Next, the inventors demonstrated that the weighted esophageal cancer DNA methylation score and threshold developed in Embodiment 25 (Detection) detect esophageal cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from GSE52068. Using this method, 88%–96% of esophageal cancer samples were detected (Figure 30A). Next, the inventors demonstrated the usefulness of the detection DNA methylation score for distinguishing esophageal cancer from other cancers using the “validation cohort” along with methylation data from GSE and TCGA for esophageal cancer and 12 other cancers (Figure 30A). The ROC curve shown in Figure 30B reveals the specificity (0.86) and sensitivity (0.88) of this methylation score for distinguishing esophageal cancer from other normal tissues and other cancers (Figure 30C). The marker also detects several other cancers (its specificity for these cancers is limited due to its relatively high sensitivity). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of cancer in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0173] Embodiment 27: Discovery of polygene DNA methylation markers for bladder cancer. The inventors used normalized Illumina 450K DNA methylation data from five individuals from GSE52955 (normal) and from 10 randomly selected samples from the TCGA cancer collection of bladder cancer DNA methylation data as a “training” cohort. The inventors first listed the “training cohort” dataset 28754 CGID, which in Embodiment 1 was found to be strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors discovered a polygene set of binary categorically distinctly methylated CGIDs that detected bladder cancer with high sensitivity and specificity in the training cohort (Figure 31B, Table 22) (detection). Weighted DNA methylation scores and cancer thresholds were developed for CGIDs as described in Embodiment 2. Next, the inventors generated a “training cohort” from 80 randomly selected DNA methylation samples from TCGA representing 13 different tumor types and normal blood. Using this training cohort, the inventors discovered a polygene set of CGIDs that were distinctly methylated between bladder cancer and other tumors (Figure 31C, Table 23) (spec). A weighted DNA methylation score was developed for CGIDs, as described in Embodiment 2.

[0174] Embodiment 28: Usefulness of a bladder cancer polygene DNA methylation marker for detecting bladder cancer. Next, the inventors demonstrated that the weighted bladder cancer DNA methylation score and threshold developed in Embodiment 27 (Detection) could detect bladder cancer in a “validation cohort” containing normalized Illumina 450K DNA methylation beta values ​​from 439 bladder cancer patients from TCGA. Using this method, 96% of bladder cancer samples were detected as bladder cancer (Figure 32B). Next, the inventors used the “validation cohort” along with methylation data from GSE and TCGA for bladder cancer and 13 other cancers to demonstrate the usefulness of the speck DNA methylation score for distinguishing bladder cancer from other cancers (Figure 32B). The ROC curve shown in Figure 32C reveals the specificity (0.86) and sensitivity (0.88) of this methylation score for distinguishing bladder cancer from other normal tissues and other cancers (Figure 32C). However, there was a fairly high proportion of cross-detection of gastric, pancreatic, esophageal, and colorectal cancers. These DNA methylation markers and methylation values The calculated methylation score can be used for early detection of cancer in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0175] Embodiment 29: Discovery of polygene DNA methylation markers for kidney cancer. The inventors used normalized Illumina 450K DNA methylation data for kidney cancer from 10 individuals from GSE52955 (normal) and 10 samples randomly selected for each cancer from 13 cancers in the TCGA dataset as a “training” cohort, and used normal tissue and blood (GSE40279, GSE52955). The inventors first listed the “training cohort” dataset 28754 CGIDs, which were found in Embodiment 1 as strongly unmethylated sites across normal tissue and blood samples, as a candidate list. Next, the inventors used the BCDhypo method described in Embodiment 2 to detect kidney cancer with high sensitivity and specificity in the training cohort and discovered a polygene set of binary categorically methylated CGIDs that are “detect-spec” (Figure 33B, Table 24) (detect-spec), which are specific to kidney cancer over other cancers. As described in Embodiment 2, weighted DNA methylation scores and cancer thresholds were developed for CGIDs.

[0176] Embodiment 30: Usefulness of a renal cancer polygene DNA methylation marker for detecting renal cancer. The inventors demonstrated that the weighted renal cancer DNA methylation score and threshold developed in Embodiment 29 (Detect-Speck) detected renal cancer and distinguished it from other cancers in a “Validation Cohort” containing normalized Illumina 450K DNA methylation beta values ​​from 871 renal cancer patients from TCGA. Using this method, 90% of renal cancer samples were detected as renal cancer (Figure 34A). Next, the inventors demonstrated the usefulness of the “Detect-Speck” DNA methylation score for distinguishing renal cancer from other cancers using the “Validation Cohort” along with methylation data from GSE and TCGA for renal cancer and 13 other cancers (Figure 34A). The ROC curve shown in Figure 34B reveals the specificity (0.87) and sensitivity (0.91) of this methylation score for detecting renal cancer from other normal tissues and other cancers (Figure 34C) (high crossover with HCC, brain cancer, and testicular cancer). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of kidney cancer in individuals at risk and generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0177] Embodiment 31: Discovery of polygene DNA methylation markers for testicular cancer. The inventors used normalized Illumina 450K DNA methylation data from 13 individuals from GSE46306 (normal) and 10 samples randomly selected for each cancer from 13 cancers in the TCGA dataset as a “training” cohort, and also used normal tissue and blood (GSE40279, GSE61496). The inventors first listed 28754 CGIDs from the “training cohort” dataset as candidate sites, which were found in Embodiment 1 as strongly unmethylated sites across normal tissue and blood samples. Next, the inventors used the BCDhypo method described in Embodiment 2 to detect testicular cancer with high sensitivity and specificity in the training cohort and discovered a polygene set of binary categorically methylated CGIDs that are “detect-spec” (Figure 35B, Table 25) (detect-spec), which are specific to testicular cancer compared to other cancers. As described in Embodiment 2, weighted DNA methylation scores and cancer thresholds were developed for CGIDs.

[0178] Embodiment 32: Usefulness of a testicular cancer polygene DNA methylation marker for detecting testicular cancer. Next, the inventors demonstrated that the weighted testicular cancer DNA methylation score and threshold developed in Embodiment 31 (Detect-Speck) detected testicular cancer and distinguished it from other cancers in a “Validation Cohort” containing normalized Illumina 450K DNA methylation beta values ​​from 156 testicular cancer patients from TCGA. Using this method, 96% of testicular cancer samples were detected as testicular cancer (Figure 36A). Next, the inventors demonstrated the usefulness of the “Detect-Speck” DNA methylation score for distinguishing testicular cancer from other cancers using the “Validation Cohort” along with methylation data from GSE and TCGA for testicular cancer and 13 other cancers (Figure 36A). The ROC curve shown in Figure 36B reveals the specificity (0.97) and sensitivity (0.96) of this methylation score for detecting testicular cancer from other normal tissues and other cancers (Figure 36C). These DNA methylation markers and the calculated methylation scores from methylation values ​​can be used for early detection of cancer in at-risk individuals and generally healthy individuals using different biomaterials from patient tissue, feces, saliva, plasma, and urine.

[0179] Embodiment 33: Discovery of polygene pan-oncological DNA methylation markers for 13 common solid tumors. The inventors used normalized Illumina 450K DNA methylation data from 10 randomly selected samples for each of the 13 cancers (bladder cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer, esophageal cancer, HNSC, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, and stomach cancer) in the TCGA dataset as a “training” cohort, and normal tissue and blood from TCGA and GEO. Next, the inventors performed penalized regression on CGIDs listed in a combination list of CGIDs that detected the 10 cancers listed in Table xy and a candidate list of CGIDs that detected any of the 10 common cancers with high sensitivity and specificity (Figure 37B, Table 26) (detection). As described in Embodiment 2, weighted DNA methylation scores and cancer thresholds were developed for CGIDs.

[0180] Embodiment 34: Usefulness of a pan-cancer polygene DNA methylation marker for detecting cancer. Next, the inventors demonstrated that the weighted cancer DNA methylation score and threshold developed in Embodiment 33 ("Detection") could detect 13 common cancers (bladder cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer, esophageal cancer, HNSC, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, and stomach cancer) in a "validation cohort" containing normalized Illumina 450K DNA methylation beta values ​​from 3644 cancer patients from TCGA with other normal tissues. Using this method, 90%–95% of cancer samples were detected (Figure 38A). The ROC curve shown in Figure 38B reveals the specificity (0.99) and sensitivity (0.95) of this methylation score for detecting the 13 cancers from other normal tissues (Figure 38C). These DNA methylation markers and the methylation scores calculated from methylation values ​​can be used for the early detection of cancer in individuals at risk and in generally healthy individuals, using different biomaterials from patient tissue, stool, saliva, plasma, and urine.

[0181] Embodiment 35. Discovery of polygene DNA methylation markers for detecting melanoma. The inventors used normalized Illumina 450K DNA methylation data from 10 randomly selected melanoma samples and 220 samples from other cancers (bladder cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer, esophageal cancer, HNSC, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, stomach cancer) and normal blood as a "training" cohort. Next, the inventors used Illumina 450K DNA methylation data from 10 randomly selected melanoma samples in the TCGA and GEO datasets and 220 samples from other cancers (bladder cancer, brain cancer, breast cancer, cervical cancer, colorectal cancer, esophageal cancer, HNSC, liver cancer, lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, stomach cancer) as a "training" cohort. Penalized regression was performed on the list of CGID combinations for detecting chromoma and the list of candidate CGIDs for detecting melanoma with high sensitivity and specificity (Figure 39, Table 28) (Detection-Spec). As described in Embodiment 2, weighted DNA methylation scores and melanoma thresholds were developed for CGIDs.

[0182] Embodiment 36: Usefulness of a melanoma polygene DNA methylation marker for detecting melanoma. Next, the inventors demonstrated that the weighted melanoma DNA methylation score and threshold developed in Embodiment 35 ("Detection-Spec") could detect melanoma in a "validation cohort" containing normalized Illumina 450K DNA methylation beta values ​​from 475 melanoma patients from TCGA with other cancers and normal tissues. Using this method, 98% of melanoma samples were detected (Figure 40A). The ROC curve shown in Figure 40B reveals the specificity (0.98) and sensitivity (0.95) of this methylation score for detecting melanoma from other normal tissues and other cancers (Figure 40C). The calculated methylation scores from these DNA methylation markers and methylation values ​​can be used for the early detection of melanoma in at-risk individuals and generally healthy individuals using different biomaterials from patient tissues, feces, saliva, plasma, and urine.

[0183] Embodiment 37: Discovery of polygene DNA methylation markers for detecting acute myeloid leukemia (AML). The inventors used normalized Illumina 450K DNA methylation data from 10 randomly selected AML samples and 10 normal blood samples within a GEO dataset as a “training” cohort. Next, the inventors performed penalized regressions against a list of CGID combinations for AML detection and a list of candidate CGIDs for detecting AML with high sensitivity and specificity (Figure 41, Table 27) (Detection-Spec). As described in Embodiment 2, weighted DNA methylation scores and AML thresholds were developed for the CGIDs.

[0184] Embodiment 38: Usefulness of an acute myeloid leukemia (AML) polygene DNA methylation marker for detecting AML in blood DNA. Next, the inventors demonstrated that the weighted AML DNA methylation score and threshold developed in Embodiment 37 ("Detection-Spec") could detect AML in a "validation cohort" including 79 AML patients from GEO and 140 patients from TCGA, as well as normalized Illumina 450K DNA methylation beta values ​​from normal blood. Using this method, 100% of AML samples were detected (Figure 42A). The ROC curve shown in Figure 42B reveals the specificity (1) and sensitivity (1) of this methylation score for detecting AML from blood (Figure 42C). These DNA methylation markers and the calculated methylation scores from methylation values ​​can be used with blood DNA for the early detection of AML in at-risk individuals and generally healthy individuals.

[0185] Embodiment 39: Bisulfite conversion, multiple amplification, and next-generation sequencing and methylation score calculation for predicting prostate cancer. Blood was collected in 9 ml tubes containing K3-EDTA and processed within 1 hour. Fresh blood samples were centrifuged at 1000 g for 10 minutes at 4°C. The supernatant was carefully transferred to a Falcon tube without disturbing the cell layer, centrifuged again for 10 minutes to completely remove any remaining cells, and frozen at -80°C. Plasma samples were thawed, and DNA was extracted using several available methods and commercially available kits for plasma DNA extraction, such as the Qiagen kit for plasma DNA or the EZ DNA Direct Extraction method. The DNA was purified using commercially available methods such as AMPure XP magnetic beads, and the purified DNA was treated with sodium bisulfite, for example, using the EZ DNA Bisulfite Treatment Kit. A library of target sequences was then prepared. It is generated by a two-step PCR reaction (Figure 40). The first PCR reaction targets specific CGIDs in Tables 5 and 6. Note that the PCR1 primer has a sequence complementary to the second PCR2 primer (Figure 45). Using human bisulfite-converted genomic DNA from HEK293 cells, we simultaneously amplified three sequences of DNA containing CGIDs for detecting prostate cancer from HIF3A (232 base pair region), TPM4 (213 base pair region), and CTTN (199 base pair region) in a multiplex PCR reaction using the following primers in a standard Taq polymerase reaction: For CGID cg02879662, forward primer: 5'ACACTCTTTCCCTACACgACgCTCTTCCgATCTNNNNNGGTAGGAGTTTTGGG AATTGG3' and reverse primer: 5'gTgACTggAgTTCAgACgTgTgCTCTTCCgATCTCCACCCCTACAATCCCTAA3' For CGID cg16232979, forward primer: 5'ACACTCTTTCCCTACACgACgCTCTTCCgATCT NNNNNYGGTTTYGGGTTTYGTATT3' and reverse primer: 5'gTgACTggAgTTCAgACgTgTgCTCTTCCgATCTACRCAAAAATATAAATCRACR ATC3' For CGID:cg14041701 and cg14498227, forward primer: 5'ACACTCTTTCCCTACACgACgCTCTTCCgATCTNNNNNGTTTTGYGTTTYGGA TTTGGGTT3' and reverse primer: 5'gTgACTggAgTTCAgToACgTgTgCTCTTCCgATCTCATAAACAACACCTTTAAA TAAACACTAAA3'. The amplified fragment was fractionated on an agarose gel.

[0186] To barcode the sample, use a second PCR reaction with the following primers: Forward primer: 5'AATgATACggCgACCACCgAgATCTACACTCTTTCCCTACACgAC3' Barcode Primer (Reverse): 5'CAAgCAgAAgACggCATACgAgATAGTCATCGgTgACTggAgTTCAgACgTg3' (The bolded base is the index. 200 variations of this index are used. The second set of primers introduces the index for each patient, along with reverse and forward sequencing primers. Prostate cancer HIF3A 232 bp, TPM4) Multiplex PCR1 reactions for three markers, 213 bp and CTTN 199 bp, are shown in the right-hand panel using various primer concentrations, as shown in Figure 46.

[0187] Embodiment 40: Usefulness of bisulfite conversion, multiplex amplification and next-generation sequencing methods, and calculation of methylation scores for predicting cancer. The inventors demonstrate that Embodiment 39 can be used to simultaneously predict prostate cancer and other cancers with high throughput using plasma samples from hundreds of patients. The highly predictable indexed amplification of CG IDs and the streamlined method for calculating cancer-indicating methylation scores can be used for the early detection of prostate cancer and other cancers.

[0188] Embodiment 41: Demonstration that the selected biomarker exhibits true BCD characteristics is that it is completely hypomethylated in the plasma of healthy individuals. Plasma DNA was extracted from plasma prepared from 40 healthy individuals and subjected to targeted amplification with cancer-specific primers for the following cancers: liver cancer, prostate cancer, lung cancer (Figure 43), and gastric cancer, pan-cancer, and CRC (Figure 44), subsequently in Embodiments 39 and 4. As described in section 0, barcoding was performed using a second amplification set (PCR2) and next-generation sequencing. All CGs showed very low levels of methylation in plasma from healthy individuals (Figures 43 and 44).

[0189] Embodiment 42: Bioinformatics workflow for determining DNA methylation levels. The PCR2 product is subjected to next-generation sequencing using a Miseq Illumina sequencer, after quantification and purification. The sequences are demultiplexed using Illumina software for indexed sequencing, generating a FASTQ file for each patient. Using a Perl text editing script (https: / / www.activestate.com / activeperl / downloads), T and C are counted in the FASTQ file for each patient's CG ID, and the percentage of methylated C within a single CG ID for that patient is quantified by dividing by the number of C / C+T (see scheme in Figure 42). The output CSV file is used to calculate the methylation score (Ms) for each patient using the following equation: JPEG0007850764000049.jpg1371, α=Intercept, β=CG ID i The coefficient is CG = the methylation level per CG for combinations of CG IDs from 1 to n. n = the number of CGs combined. MS = methylation score.

[0190] Uses of the subject matter of this invention The applications of the subject matter of the present invention are generally in the fields of molecular diagnostics and early cancer prediction. Those skilled in the art can use the subject matter of the present invention to derive similar non-invasive biomarkers for the early prediction of other cancers and other diseases involving cell death and loss of cell-free DNA systems, such as neurological disorders, diabetes, cirrhosis, and cardiovascular diseases such as damage to cardiac tissue. The subject matter of the present invention provides a method for finding sophisticated methylation markers of specific cell types and tissues using BCD and BCDhypo methods. Also disclosed are a wide range of methods and biomarkers for early cancer prediction that can be used by those skilled in the art to detect cancer early, dramatically increase survival rates, and cure cancer. The methods disclosed by the present invention may be used by those skilled in the art to perform regular annual screenings of healthy individuals, identify and immediately treat individuals who are beginning to develop cancer, prevent the disastrous personal, social, and economic impacts of cancer mortality and morbidity, and to monitor “high-risk” individuals and monitor the response to treatment of patients being treated to detect recurrence or metastasis. The adoption of the present invention described herein by healthcare providers and health checkup facilities for routine medical management will have a significant impact on reducing the burden of cancer and healthcare costs.

[0191] The fact that the subject matter of the present invention comprises numerous different dependent claims does not mean that combinations of these claims cannot be used to predict cancer. Embodiments disclosed herein for measuring, statistically analyzing, and predicting cancer should not be considered limiting. Various other modifications for measuring DNA methylation in cancer patients are apparent to those skilled in the art, including Illumina EPIC arrays, capture array sequencing, next-generation sequencing, methylation-specific PCR, epitypers, restriction enzyme-based analysis, and other methods in the public domain. Similarly, in addition to those listed herein, there are numerous statistical methods in the public domain for using the subject matter of the present invention for predicting cancer in patient samples.

[0192] While the subject matter of the present invention has been described in relation to its embodiments, including one or more preferred embodiments, it should be understood that many other possible modifications and changes can be made without departing from the spirit and scope of the claimed subject matter.

Claims

1. This includes the step of deriving binary DNA methylation markers from a genome-wide DNA methylation map using the "binary category distinction (BCD)" method. A method for using binary DNA methylation markers for cancer as an auxiliary means for binary detection of cancer, The aforementioned BCD method, (a) A step of preparing beta values ​​of multiple categories of samples, including cancer, normal tissue, and blood, as input data. (b) From all CGIDs, select CGIDs that are universally unmethylated in normal tissue and blood by having a beta value of 0.1 or less in each sample of normal tissue and blood, and extract them as a candidate set of binary DNA methylation markers for cancer. (c) For each CG ID included in the candidate set, compare the beta values ​​of the cancer group and the normal tissue group, and use a beta value of 0.2 as the threshold. (i) Methylation occurs only in the cancer group and not in the normal tissue group, or (ii) Methylation occurs only in the normal tissue group, while hypomethylation occurs in the cancer group. The following steps are taken: setting a binary pattern as a condition, and determining the CG ID that satisfies the condition through spreadsheet processing, (d) A step of extracting the determined CG ID as a differentially methylated CG ID (DM) or a differentially hypomethylated CG ID (DHM), and (e) A step of combining the obtained DM or DHM to construct a cancer-specific methylation marker set. A method that includes this.

2. The method according to claim 1, wherein the genome-wide DNA methylation map is one or more of those created based on cancer cell-derived DNA, normal tissue-derived DNA, and blood-derived DNA.

3. In step (a) of the "Binary Category Discrimination (BCD)" method, the generation of the beta value involves a genome-wide DNA methylation array, a DNA microarray, and a genome-wide DNA methylation array. The method according to claim 1, comprising the use of one or more idiovisulfite sequences, methylated DNA immunoprecipitation (MeDIP) sequences, and hybridization with an oligonucleotide array.

Images