Method for capturing cell-free methylated DNA and its use

Abstract

This invention provides a method for capturing cell-free methylated DNA. [Solution] A method for capturing cell-free methylated DNA from a sample containing less than 100 mg of cell-free DNA, comprising the steps of: subjecting the sample to library preparation to enable subsequent sequencing of cell-free methylated DNA; adding a first amount of filler DNA to the sample, wherein at least a portion of the filler DNA is methylated; denaturing the sample; and capturing cell-free methylated DNA using a binder selective for methylated polynucleotides.

Description

[Technical Field]

[0001] Reference to related applications This application relates to the U.S. Provisional Patent Application filed on 3 May 2016, which is incorporated herein by reference in its entirety. This application asserts priority under Application No. 62 / 331,070.

[0002] This invention relates to the field of cell-free DNA, more specifically to a method for capturing cell-free methylated DNA. And relating to its use. [Background technology]

[0003] DNA methylation is a covalent modification of DNA and plays an important role in chromatin structure. It is a stable gene regulatory mechanism that performs this function. In humans, DNA methylation is a CpG gene regulatory mechanism. It mainly occurs at cytosine residues in creotides. Unlike other dinucleotides, CpG is Instead, they are not evenly distributed throughout the genome, but rather short CpG-rich DNAs called CpG islands. It is concentrated in region A. DNA methylation suppresses genes through the following two main mechanisms. Can be controlled: 1) Supplement with methyl-binding domain proteins, methyl-binding domain proteins 2) c-MY The binding sites of transcription factors (TFs) such as C Block access (Non-Patent Document 1).

[0004] Generally, most CpG sites in the genome are methylated, and most CpG islands are It remains unmethylated during normal development and in differentiated tissues (Non-Patent Document 1). Nevertheless, This allows us to elucidate tissue-specific patterns of DNA methylation in normal primary tissues. (Non-patent document 2). Furthermore, during malignant transformation, overall DNA hypomethylation occurs, and in CpG islands... Localized hypermethylation is frequently observed (Non-Patent Literature 1). In practice, DNA methylation The transformation pattern is that among cancer patients, gliablastoma (Non-Patent Literature 3), In genitalia (Non-Patent Literature 4), colorectal cancer (Non-Patent Literature 5), and breast cancer (Non-Patent Literature 6, 7) Clinically relevant subgroups with prognostic values It was used to stratify the data into categories.

[0005] For normal differentiation and its stability and role in diseases such as cancer, DNA methylation is important. It is a good biomarker that can be used to describe tumor characteristics and phenotypic states, Therefore, it has high potential in personalized medicine. Numerous sample types are DNA methylation mapping. In addition, biomeris including new FFPE tumor tissue, blood cells, urine, saliva, and stool. It is suitable for detecting cancer (Non-Patent Literature 8). More recently, genome identification has been particularly useful for cancer (somatic cell mutations). (Different) (Non-Patent Literature 9), Graft (Donor vs. Recipient DNA) (Non-Patent Literature 10), Pregnancy In situations such as pregnancy (fetal vs. maternal DNA) (Non-patent documents 11, 12), biomarkers - The use of circulating cell-free DNA (cfDNA) as a means is accelerating. The use of DNA methylation mapping of cfDNA as a biomarker is occurring. It enables the identification of source tissue and allows for minimally invasive stratification of cancer patients, thus having a significant impact. It may have. Furthermore, it may have monitoring of immune responses, neurodegenerative diseases, or myocardial infarction. In situations where genome identification is not available, it enables the use of cfDNA as a biomarker. This allows for the detection of epigenetic abnormalities in cfDNA.

[0006] Furthermore, using whole-genome DNA methylation mapping of cfDNA can overcome the important sensitivity issue in detecting circulating tumor DNA (ctDNA: circulating tumor DNA) in patients with early-stage cancer without radiological images demonstrating the disease. Existing ctDNA detection methods are based on mutation sequencing, but the sensitivity is somewhat limited because the number of recurrent mutations available for distinguishing tumor and normal circulating cfDNA is limited (Non-Patent Documents 13, 14). On the other hand, whole-genome DNA methylation mapping can utilize a large number of epigenetic changes to distinguish circulating tumor DNA (ctDNA) from normal circulating cell-free DNA (cfDNA). For example, some tumor types such as endometrial cancer can have extensive DNA methylation abnormalities without important recurrent somatic mutations (Non-Patent Document 4). Furthermore, the whole cancer data of The Cancer Genome Atlas (TCGA) shows a large number of DMRs between tumors and normal tissues across substantially all tumor types (Non-Patent Document 15). Therefore, what is emphasized by these findings is that an assay that has successfully recovered cancer-specific DNA methylation changes from ctDNA can serve as a very sensitive tool for detecting, classifying, and monitoring malignant diseases at a low sequencing-related cost.

[0007]

[0008] However, whole-genome mapping of DNA methylation in cfDNA is extremely difficult because the amount of available DNA is small and cfDNA is fragmented to less than 200 bp in length (Non-Patent Document 16). For this reason, conventional methods that require at least 50 - 100 ng of DNA ​ MeDIP-seq (Non-Patent Document 17), or requiring unfragmented DNA Reduced Representation Bisulfite Sequence It is impossible to perform ing (RRBS) (Non-Patent Document 18) in cfDNA. Another problem with mapping A methylation is that the amount of target DNA in normal cfDNA is small. (Non-patent document 19). Therefore, a sequence with sufficient depth to capture a small amount of DNA. Since the cost of making decisions is exorbitant, conducting a WGBS is impractical. On the other hand, methylation A method for selectively enriching easily obtainable CpG-rich features is available in a single read. This has the potential to maximize the amount of information used, reduce costs, and minimize DNA loss. [Prior art documents] [Non-patent literature]

[0009] [Non-Patent Document 1] Sharma, S., Kelly, TK & Jones, PA Epigenetics in cancer. Carcinogenesis 31, 27-36, doi:10.1093 / carcin / bgp220 (2010). [Non-Patent Document 2] Varley, KE et al. Dynamic DNA methylation across diverse human cell lines and tissues. Genome Res 23, 555-567, doi:10.1101 / gr.147942.112 (2013). [Non-Patent Document 3] Sturm, D. et al. Hotspot mutations in H3F3A and IDH1 define distinct epigenetic and biological subgroups of glioblastoma. Cancer Cell 22, 425-437, doi:10.1016 / j.ccr.2012.08.024 (2012).

Non-licensed Document 4

Non-licensed Document 5

Non-licensed Document 6

Non-licensed Document 7

Non-licensed Document 8

Outdoor Tools9

Outdoor Tools 10

Outdoor Content11

Outdoor Tools 12

Outdoor Tools13

Outdoor Tools 14

Outdoor Tools 15

Outdoor Tools 16

[0010] According to one embodiment, cell-free methylated DNA is obtained from a sample containing less than 100 ng of cell-free DNA. A method for capturing D, wherein the sample is subjected to library preparation, followed by cell-free methylation. Steps to enable sequencing of NA: Adding a first amount of filler DNA to the sample. In the step where at least a portion of the filler DNA is methylated, the sample The steps involve denaturation and the use of a selective binder on methylated polynucleotides in a cell-free environment. The present invention provides a method that includes the step of capturing chilled DNA.

[0011] Embodiments of the present invention can be best understood by referring to the following description and accompanying drawings. Cut. [Brief explanation of the drawing]

[0012] [Figure 1A] This figure shows that methylome analysis of cfDNA is a highly sensitive method for enriching and detecting ctDNA in small amounts of input DNA. It also shows a computer simulation of the probability of detecting at least one epimutation as a function of ctDNA concentration (column), number of DMRs tested (row), and sequencing depth (x-axis). [Figure 1B] This figure shows that methylome analysis of cfDNA is a highly sensitive technique for enriching and detecting ctDNA in small amounts of input DNA. Whole-genome Pearson correlation of DNA methylation signals of 1–100 ng of input DNA from HCT116 cell lines fragmented to mimic plasma cfDNA. Each concentration has two biological copies. [Figure 1C]This figure shows that methylome analysis of cfDNA is a highly sensitive method for enriching and detecting ctDNA in small amounts of input DNA. DNA methylation profiles (green track) obtained from cfMeDIP-seq from input DNA of different concentrations derived from HCT116 + Reduced Representation Bisulfite Sequencing (RRBS) HCT116 data obtained from ENCODE (ENCSR000DFS) and Whole-Genome Bisulfite Sequencing (WGBS) HCT116 data obtained from GEO (GSM1465024). In the heatmap (RRBS track), yellow indicates methylation, blue indicates unmethylation, and gray indicates 0 coverage. [Figure 1D] This figure shows that methylome analysis of cfDNA is a highly sensitive method for enriching and detecting ctDNA in small amounts of input DNA. Serial dilutions of the CRC cell line HCT116 into the multiple myeloma (MM) cell line MM1.S. cfMeDIP-seq was performed on pure HCT116 DNA (100% CRC), pure MM1.S DNA (100% MM), and 10%, 1%, 0.1%, 0.01%, and 0.001% CRC DNA diluted in MM DNA. All DNA was fragmented to mimic plasma cfDNA. The inventors observed a nearly perfect linear correlation (r²=0.99, p<0.0001) between (D) DMR numbers and (E) DNA methylation signals (RPKM units) within these DMRs, as predicted. [Figure 1E]This figure shows that methylome analysis of cfDNA is a highly sensitive method for enriching and detecting ctDNA in small amounts of input DNA. Serial dilutions of the CRC cell line HCT116 into the multiple myeloma (MM) cell line MM1.S. cfMeDIP-seq was performed on pure HCT116 DNA (100% CRC), pure MM1.S DNA (100% MM), and 10%, 1%, 0.1%, 0.01%, and 0.001% CRC DNA diluted in MM DNA. All DNA was fragmented to mimic plasma cfDNA. The inventors observed a nearly perfect linear correlation (r²=0.99, p<0.0001) between (D) DMR numbers and (E) DNA methylation signals (RPKM units) within these DMRs, as predicted. [Figure 1F] This figure shows that methylome analysis of cfDNA is a highly sensitive method for concentrating and detecting ctDNA in small amounts of input DNA. In the same dilution series, known somatic mutations are detectable at a rate of only 1 / 100 alleles by ultra-depth (>10,000×) targeted sequencing that exceeds the background sequencing instrument and polymerase error rate. The figure shows the percentage of reads containing each base or insertion / deletion at each mutation site in the CRC cell line. [Figure 1G] This figure shows that methylome analysis of cfDNA is a highly sensitive method for enriching and detecting ctDNA in small amounts of input DNA. The figure also shows the frequency of ctDNA (human) as a percentage of total cfDNA (human + mouse) in the plasma of mice carrying patient-derived xenografts (PDX) from two colorectal cancer patients. [Figure 2] This is a schematic diagram of the cfMeDIP-seq protocol. [Figure 3A-1] This figure shows sequencing saturation analysis and quality control. The figure shows the results of saturation analysis using the Bioconductor package MEDIPS, which analyzes cfMeDIP-seq data from each replica for each input concentration from HCT116 DNA fragmented to mimic plasma cfDNA. [Figure 3A-2]This figure shows sequencing saturation analysis and quality control. The figure shows the results of saturation analysis using the Bioconductor package MEDIPS, which analyzes cfMeDIP-seq data from each replica for each input concentration from HCT116 DNA fragmented to mimic plasma cfDNA. [Figure 3A-3] This figure shows sequencing saturation analysis and quality control. The figure shows the results of saturation analysis using the Bioconductor package MEDIPS, which analyzes cfMeDIP-seq data from each replica for each input concentration from HCT116 DNA fragmented to mimic plasma cfDNA. [Figure 3B] This figure shows sequencing saturation analysis and quality control. The procedure was tested in two replications of four start DNA concentrations (100, 10, 5, and 1 ng) of HCT116 cell lines. Reaction specificity was calculated using methylated and unmethylated Arabidopsis thaliana DNA. The enrichment ratio was calculated using genomic regions of fragmented HCT116 DNA (primers for methylated testis-specific H2B, TSH2B0, and unmethylated human DNA region (GAPDH promoter)). The horizontal dotted line indicates an enrichment ratio threshold of 25. Error bars represent ±1 s.em. [Figure 3C] This figure shows sequencing saturation analysis and quality control. The CpG enrichment score of the sequencing sample indicates robust enrichment of CpGs within the genomic region from the immunoprecipitation sample compared to the input control. The CpG enrichment score was obtained by dividing the relative frequency of CpGs in the region by the relative frequency of CpGs in the human genome. Error bars represent ±1 s.em. [Figure 4A] This figure shows quality control from cfMeDIP-seq after serial dilutions. It also shows a schematic diagram of CRC DNA (HCT116) dilution in MM DNA (MM1.S). [Figure 4B]This figure shows the quality control from cfMeDIP-seq of serial dilutions. The reaction specificity of each dilution was calculated using methylated and unmethylated Arabidopsis thaliana DNA. [Figure 4C] This figure shows quality control from cfMeDIP-seq from serial dilutions. The CpG enrichment score of the sequencing sample indicates strong enrichment of CpG within the genomic region from the immunoprecipitation sample. The CpG enrichment score was obtained by dividing the relative frequency of CpG in the region by the relative frequency of CpG in the human genome. [Figure 4D-1] This figure shows quality control from cfMeDIP-seq after serial dilutions. The figure shows the results of saturation analysis from each dilution point. [Figure 4D-2] This figure shows quality control from cfMeDIP-seq after serial dilutions. The figure shows the results of saturation analysis from each dilution point. [Figure 5A] This figure shows that the cfMeDIP-seq method can identify thousands of differentially methylated regions on circulating cfDNA obtained from pancreatic adenocarcinoma patients. Experimental design. [Figure 5B] This figure shows that the cfMeDIP-seq method can identify thousands of differentially methylated regions on circulating cfDNA obtained from pancreatic adenocarcinoma patients. The volcano plot shows circulating cfDNA from pancreatic cancer (cases, n=24) versus healthy donors (controls, n=24) using cfMeDIP-seq. The red dots indicate windows that became significant after multiple testing correction. [Figure 5C] This figure shows that the cfMeDIP-seq method can identify thousands of differential methylation regions on circulating cfDNA obtained from pancreatic adenocarcinoma patients. A heatmap of 38,085 DMRs identified in plasma DNA from healthy donors and pancreatic cancer patients. Hierarchical clustering method: Ward. [Figure 5D]This figure shows that the cfMeDIP-seq method can identify thousands of differential methylation regions on circulating cfDNA obtained from pancreatic adenocarcinoma patients. A permutation analysis estimates the frequency of predicted vs. observed overlap between DMRs identified in plasma (case vs. control) and cancer-specific DMCs identified in primary tumor tissue (primary tumor vs. normal tissue). The box plot represents the null distribution of overlaps. Diamonds represent the experimentally observed number of DNA methylation overlaps from primary tumor tissue and circulating cfDNA. Red diamonds indicate a significantly higher number of observed overlaps than expected by chance. Green diamonds indicate a significantly lower number of observed overlaps than expected by chance, and blue diamonds indicate no significant difference. The inventors calculated four possible overlaps: hypermethylation in primary tumor tissue and hypermethylation in circulating cfDNA (enriched, P value: 6.4 × 10⁻²²); hypermethylation in tumor tissue and hypomethylation in circulating cfDNA (deficient, P value: 9.43 × 10⁻¹⁷); hypomethylation in tumor tissue and hypomethylation in circulating cfDNA (enriched, P value: 1.88 × 10⁻²⁸³); and hypomethylation in tumor tissue and hypermethylation in circulating cfDNA (P value: 0.105). [Figure 5E] This figure shows that the cfMeDIP-seq method can identify thousands of differentially methylated regions on circulating cfDNA obtained from pancreatic adenocarcinoma patients. A permutation analysis is performed to estimate the frequency of predicted vs. observed overlap between DMRs identified in plasma (case vs. control) and cancer-specific DMCs identified in primary tumor tissue (primary tumor vs. normal PBMC). [Figure 6A] This figure shows the quality control of cfMeDIP-seq from circulating cfDNA derived from pancreatic adenocarcinoma patients (cases). The specificity of the reaction for each case sample was calculated using methylated and unmethylated Arabidopsis thaliana DNA. The enrichment ratio was not calculated due to the extremely limited amount of available DNA. [Figure 6B] This figure shows the quality control of cfMeDIP-seq from circulating cfDNA derived from healthy donors (controls). The specificity of the reaction for each control sample was calculated using methylated and unmethylated Arabidopsis thaliana DNA. The enrichment ratio was not calculated due to the extremely limited amount of available DNA. [Figure 6C]This figure shows the quality control of cfMeDIP-seq from circulating cfDNA derived from pancreatic adenocarcinoma patients (cases). The CpG enrichment score of the sequencing sample indicates strong enrichment of CpG within the genomic region from the immunoprecipitation sample. [Figure 6D] This figure shows the quality control of cfMeDIP-seq from circulating cfDNA derived from healthy donors (controls). The CpG enrichment score of the sequencing sample indicates strong enrichment of CpG within the genomic region from the immunoprecipitation sample. [Figure 7A] This figure shows PCA for 48 plasma cfDNA methylations from healthy donors and early-stage pancreatic adenocarcinoma patients using the top million most variable whole-genome windows. Variability was calculated for each window using the Mean Absolute Deviation (MAD) criterion, a robust measure that returns the median of absolute deviations from the median of the data. In this case, the data are RPKM values ​​across all 48 samples for a given window. PC1 vs. PC2 (left) and PC1 vs. PC3 (right) are shown. [Figure 7B] The rate of change for each principal component. [Figure 7C] Volcano plot of tumor versus normal LCM tissue from pancreatic adenocarcinoma patients using RRBS. Shows the total number of differentially methylated CpGs (DMCs) identified. Red dots indicate the window where absolute methylation difference (absolute delta-beta) > 0.25 became significant after multiple testing correction. [Figure 7D] Scatter plot showing the significance of DNA methylation differences across each overlapping window. The X-axis shows the log10q values ​​of primary pancreatic adenocarcinoma tumors versus normal tissue from RRBS data. Significance is indicated on a positive scale when a region is hypermethylated in the tumor. Hypomethylated regions are indicated on a negative scale. The Y-axis shows the log10q values ​​of plasma cfDNA methylation from pancreatic adenocarcinoma patients versus healthy donors from cfMeDIP-seq data. Blue dots indicate that both are significant. The red line indicates the trend line. [Figure 7E]Scatter plot showing differences in DNA methylation across each overlapping window. The X-axis shows the difference in DNA methylation between primary pancreatic adenocarcinoma tumors and normal tissue from RRBS data. The Y-axis shows the difference in DNA methylation of plasma cfDNA methylation between pancreatic adenocarcinoma patients and healthy donors from cfMeDIP-seq data. The blue line indicates the trend line. [Figure 7F] Volcano plot of LCM pancreatic adenocarcinoma tissue versus normal PBMCs using RRBS. Shows the total number of differentially methylated CpGs (DMCs) identified. Red dots indicate the window of absolute methylation difference (absolute delta-beta) > 0.25, which became significant after multiple testing correction. [Figure 7G] Scatter plot showing the significance of differences in DNA methylation across each overlapping window. The X-axis shows the log10q values ​​of primary pancreatic adenocarcinoma tumors versus normal PBMCs from RRBS data. Significance is indicated on a positive scale when a region is hypermethylated in the tumor. Hypomethylated regions are indicated on a negative scale. The Y-axis shows the log10q values ​​of plasma cfDNA methylation from pancreatic adenocarcinoma patients versus healthy donors from cfMeDIP-seq data. Blue dots indicate that both are significant. The red line shows the trend line. [Figure 7H] Scatter plot showing differences in DNA methylation across each overlapping window. The X-axis shows differences in DNA methylation between primary pancreatic adenocarcinoma tumors and normal PBMCs from RRBS data. The Y-axis shows differences in DNA methylation of plasma cfDNA methylation between pancreatic adenocarcinoma patients and healthy donors from cfMeDIP-seq data. [Figure 8A]This figure shows that the transcription factor (TF) footprint can be identified and the active transcription network in the origin tissue can be inferred using circulating cfDNA methylation profiles. Expression profiles of all TFs (n=33) whose motifs were enriched (using HOMER20 software) in hypomethylated regions (hypomethylation footprint in control) in cfDNA from healthy donors across multiple human tissues. Expression data were obtained from the Genotype-Tissue Expression (GTEx) project.21 Several TFs that are preferentially expressed in the hematopoietic system were identified (PU.1, Fli1, STAT5B, KLF1). [Figure 8B] This figure shows that circulating cfDNA methylation profiles can be used to identify transcription factor (TF) footprints and infer active transcriptional networks in the origin tissue. Expression profiles of all TFs with hypomethylated motifs in the control (n=33) versus expression profiles of 1,000 random sets of 33 TFs in whole blood (GTEx data). [Figure 8C] This figure shows that the transcription factor (TF) footprint can be identified using circulating cfDNA methylation profiles, and the active transcriptional network in the origin tissue can be inferred. Expression profiles of all TFs (n=85) whose motifs were enriched in hypomethylated regions (case hypomethylation footprint) in cfDNA from pancreatic adenocarcinoma patients. Several pancreas-specific or pancreatic cancer-associated TFs were identified. Furthermore, Hallmark TFs driving molecular subtypes of pancreatic cancer were also identified. [Figure 8D] This figure shows that the circulating cfDNA methylation profile can be used to identify the transcription factor (TF) footprint and infer the active transcription network in the origin tissue. Expression profiles of all TFs with hypomethylated motifs in cases (n=85) vs. expression profiles of 1,000 random sets of 85 TFs in normal pancreas (GTEx data). [Figure 8E] This figure shows that the circulating cfDNA methylation profile can be used to identify the transcription factor (TF) footprint and infer the active transcription network in the origin tissue. Expression profiles of all TFs with hypomethylated motifs in cases (n=85) vs. expression profiles of 1,000 random sets of 85 TFs in pancreatic adenocarcinoma tissue (TCGA data). [Figure 9] This figure shows the recovery percentage of added unmethylated Arabidopsis thaliana DNA after cfMeDIP-seq using 10 ng, 5 ng, and 1 ng of initial cancer cell-free DNA (n=3) combined with 90 ng, 95 ng, and 99 ng of filler DNA, respectively, before immunoprecipitation, or without filler DNA. The filler DNA used had different compositions of artificial methylation and present unmethylated lambda DNA to increase the final amount before immunoprecipitation to 100 ng. The recovery percentage of the desired added unmethylated DNA was <1.0%, and lower recovery rates indicate higher reaction specificity. [Figure 10] This figure shows the recovery percentage of added methylated Arabidopsis thaliana DNA after cfMeDIP-seq using initial tumor cell-free DNA amounts of 10 ng, 5 ng, and 1 ng (n=3) combined with 90 ng, 95 ng, and 99 ng of filler DNA, respectively, before immunoprecipitation, or without filler DNA. The filler DNA used had different compositions of artificial methylation and present unmethylated lambda DNA to increase the final amount before immunoprecipitation to 100 ng. The minimum recovery percentage of the desired added methylated DNA is 20%. [Modes for carrying out the invention]

[0013] The inventors have developed a bioinformatics method for mixing ctDNA in different proportions of 0.001% to 10% with other substances. This was simulated (Figure 1A, row plane). The inventors found that ctDNA is normal cfDNA Compared to 1, 10, 100, 1,000, or 10,000 DMRs (Differential Methylation Regions) We also simulated scenarios with (Figure 1A, row plane). Then, we read the reads in various sequences. Samples were taken at each locus at the determination depth (10×, 100×, 1000× and 1 (0000×) (Figure 1A, x-axis). The inventors believe that the amount of cancer ctDNA present is low and covers Even at a low severity level, an increase in the number of DMRs increases the probability of detecting at least one cancer-specific event. We found that it becomes more difficult (Figure 1A).

[0014] To overcome these challenges, the inventors have developed a method for using cell-free DNA to develop a method for developing whole-genome DNA cfMeDIP-seq (cell-free methylated DNA immunoprecipitation and) is used for methylation mapping. We developed a new method called high-throughput sequencing. The cfMe method described here is... The DIP-seq method is robust to input DNA up to 100 ng, unlike existing low-input MeDIP- Seq procedure 17 It was developed through modifications. However, most plasma samples are 100 ng or It generates far less DNA. To overcome this challenge, the inventors of this invention introduced D To artificially increase the amount of NA to 100 ng, foreign λDNA (filler DNA) is used. It was added to an adapter-linked cfDNA library (Figure 2). This is nonspecific due to the antibody. Minimize the amount of binding, and minimize the amount of DNA lost due to binding with plastic products. The filler DNA is an amplifier similar in size to the adapter-linked cfDNA library. It consists of lycon, unmethylated DNA and in vitro methylated DNA with different methylation levels. It was composed of A (Figures 9 and 10). Different patients produce different amounts of cfDNA. The addition of this filler DNA is also useful in practice, as it standardizes the input DNA amount to 100 ng. This allows downstream procedures to proceed regardless of the amount of cfDNA available. It remains exactly the same in all samples.

[0015] According to one embodiment, cell-free methylated DNA is obtained from a sample containing less than 100 ng of cell-free DNA. A method for capturing, a. The sample is subjected to library preparation, followed by cell-free sequencing of methylated DNA. Steps to make it possible, b. A step of adding a first amount of filler DNA to the sample, wherein the filler DNA Steps in which at least a portion is methylated, c. A step of denaturing the sample, d. Capture cell-free methylated DNA using a selective binder for methylated polynucleotides. Steps to take This document presents methods that include the following:

[0016] In some embodiments, this method amplifies the captured cell-free methylated DNA, This then includes a step to determine the sequence.

[0017] Polymerase chain reaction (PCR) n) Subsequently, various sequencing techniques, such as the Sanger sequencing method, are known to those skilled in the art. Ilumina (Solexa) sequencing method, Roche 454 sequencing method, Ion t orrent: Various methods including Proton / PGM sequencing and SOLiD sequencing. Next-generation sequencing (N), also known as high-throughput sequencing, includes sequencing technology. GS (next-generation sequencing) technology is also available. NGS allows for DNA and RNA sequencing that is more efficient than previously used Sanger sequencing methods. It can be done much faster and less expensively. In some embodiments, the sequence determination is It is optimized for short-read sequencing.

[0018] Cell-free methylated DNA is DNA that circulates freely in the bloodstream, and various aspects of DNA It is methylated in the known region. Samples, such as plasma samples, separate cell-free methylated DNA. It can be collected for analysis.

[0019] In this specification, "library preparation" refers to list end-repair. r) A tailing, adapter linking, or to enable subsequent DNA sequencing This includes any other preparations performed on cellular DNA.

[0020] In this specification, "filler DNA" may be non-coding DNA, or Ampli It can consist of components.

[0021] DNA samples can be denatured, for example, by using sufficient heat.

[0022] In some embodiments, the sample contains less than 50 ng of cell-free DNA.

[0023] In some embodiments, the first amount of filler DNA is about 10%, 20%, or 30%. , 40%, 50%, 60%, 70%, 80%, 90%, or 100% methylated filler DN Contains A. In a preferred embodiment, the first amount of filler DNA is about 50% methyl Contains chemical filler DNA.

[0024] In some embodiments, the first amount of filler DNA is 20 ng to 100 ng. In preferred embodiments, 30 ng to 100 ng of filler DNA. More preferred In the embodiment, 50 ng to 100 ng of filler DNA. Cell-free DNA from the sample. When the first amount of filler DNA is used in combination, at least 50 ng of total DNA is preferred. Each sample contains at least 100 ng of total DNA.

[0025] In some embodiments, the filler DNA is 50 bp to 800 bp long. In one embodiment, the length is 100bp to 600bp, and in a more preferred embodiment, 20 0bp~600bp length.

[0026] Filler DNA is double-stranded. For example, filler DNA can be treated as junk DNA. This is possible. Filler DNA can be either endogenous or exogenous. For example, filler The DNA is non-human DNA, preferably λDNA in a preferred embodiment. "λDNA" refers to enteric bacterial phage λDNA. In some embodiments, fila —The DNA does not align with human DNA.

[0027] In some embodiments, the binder is a protein containing a methyl CpG binding domain. Yes, there is. One such exemplary protein is the MBD2 protein. Methyl-CpG-binding domain (MBD) "ain)" refers to certain domains of proteins and enzymes that are approximately 70 residues long, It binds to DNA containing one or more symmetrically methylated CpGs. MeCP2, MBD1, MBD 2. MBD4 and BAZ2 mediate the binding of MBD to DNA, MeCP2, and MBD1 In the case of MBD2, it preferentially mediates binding to methylated CpG. Human protein MECP2, MBD1, MBD2, MBD3, and MBD4 are methyl CpG binding domains. This includes nuclear protein families associated with the presence of each of the (MBD) molecules. Each of these proteins, with the exception of MBD3, specifically binds to methylated DNA. It is possible.

[0028] In another embodiment, the binder is an antibody, and cell-free capture of methylated DNA is performed. This method includes immunoprecipitation of methylated DNA using an antibody. In this specification, "immunoprecipitation" is used. This involves using antibodies that specifically bind to certain antigens (polypeptides, nucleotides, etc.). This refers to a technique for precipitating the antigen from a solution. This process involves a specific protein or DN. A can be used to isolate and concentrate from the sample, and the antibody becomes a solid at some point in the procedure. It needs to bond to a substance. Examples of solid substrates include beads such as magnetic beads. Other types of beads and solid substrates are known in the art.

[0029] An exemplary antibody is a 5-MeC antibody. In the immunoprecipitation procedure, in some embodiments At least 0.05 μg of antibody is added to the sample, and in a more preferred embodiment, a small amount At the very least, add 0.16 μg of antibody to the sample. To confirm the immunoprecipitation reaction, In embodiments, the method described herein involves adding a second amount of control D after step (b). The process further includes the step of adding NA to the sample.

[0030] Another exemplary antibody is the 5-hydroxymethylcytosine antibody.

[0031] In another embodiment, the method described herein ensures the capture of cell-free methylated DNA. To confirm, the step of adding a second amount of control DNA to the sample after step (b) is performed. It also includes.

[0032] In this specification, “contrast” may include both positive and negative contrasts, or at least a positive contrast. can.

[0033] According to a further embodiment, this specification is for measuring the DNA methylation profile in a sample. This document provides instructions on how to use the methods described in the book.

[0034] In a further aspect, the profile is correlated with the known methylation profile of tumor tissue. This specification allows for the confirmation of the presence of cell-free DNA derived from cancer cells in a sample. The method described above provides a way to use it.

[0035] In a further embodiment, the profile is compared with the known methylation profile of a particular tissue. This specification describes how to identify the tissue of origin of cell-free DNA in a sample by involving it. This provides the use of DNA methylation profiles.

[0036] In some embodiments, the origin tissue of cancer cells in cell-free DNA within a sample is identified. Uses further including the uses described herein.

[0037] In a further embodiment, the use described herein for monitoring immunotherapy is provided. ru.

[0038] In a further embodiment, the use described herein for the diagnosis of autoimmune conditions is provided. .

[0039] According to a further embodiment, the present invention is for measuring cell turnover in the subject from which the sample was taken. Provides the use described in the details.

[0040] The following examples illustrate various aspects of the present invention, which are disclosed herein. This invention is not intended to limit the broad range of embodiments. [Examples]

[0041] method Donor replenishment and sample acquisition Pancreatic adenocarcinoma (PDAC) patient samples are being collected by the University Health Network. The biodata was obtained from the k BioBank, and healthy controls were collected from Mount Sinai in Toronto, Canada. Supplementation from the Family Medicine Center of the Hospital (MSH) Completed. All samples collected with the patient's consent were submitted to Research Ethics Board. rd, University Health Network and M in Toronto, Canada This was obtained with facility approval from Ount Sinai Hospital.

[0042] Specimen processing - Purified tumor and normal cells In the case of primary PDAC samples, the specimen should be processed immediately after excision, and a representative section should be used for examination. Disintegration was confirmed. Laser capture microdisexcision of new liquid nitrogen-frozen tissue samples. Laser capture microdissection (LCM) This was performed on an ICA LMD7000 instrument. In short, it was maintained in a gas-phase liquid nitrogen environment. The frozen tissue was embedded in OCT cutting medium and cut into 8 μm thick sections using a cryotome. Place the sample on a PEN membrane slide (Leica), lightly stain with hematoxylin, and then examine the tumor area. Microscopic identification was facilitated. LCM was performed on the same day that the section was cut to minimize nucleic acid degradation. Ta.

[0043] Microscopically dissected tumor cells are transferred by gravity to the cap of a sterile, ribonuclease-free microcentrifuge tube. Approximately 150,000 to 200,000 tumor cells were collected for DNA sampling, and further... The cells were stored at -80°C until processing. LCM generally collects a sufficient amount of purified tumor cells. It took 1-2 days per case. Qiagen Cell Lysis Buff Genomic DNA was extracted using er. Unstained frozen sections on glass slides were then prepared appropriately. Histologically examined corresponding normal reference tissue by rubbing it into a DNA extraction buffer. These were collected from the frozen duodenal or gastric mucosa of each patient.

[0044] Sample processing - cfDNA EDTA and ACD plasma samples were collected at BioBank and Mount S in Toronto, Canada. Family Medicine Center at inai Hospital (MSH) Obtained from e. All samples were stored at -80°C or in gas-phase liquid nitrogen until use. Cell-free DN A is extracted from 0.5-3.5 ml of plasma using QIAamp Circulating Nucle Extraction was performed using the ic Acid Kit (Qiagen). Before use, the extracted DNA was quenched. Quantified using bits.

[0045] Sample processing - PDX cfDNA University Health Network Research Ethi As approved by the CS Board, University Health Network Human colorectal tumor tissue obtained from the twork Biobank with the patient's consent is used in the Co The cells were digested into single cells using lagene A. The single cells were from 4-6 week old NOD / SCID male mau. The injection was administered subcutaneously to the mice. After euthanizing the mice by CO2 inhalation, blood was collected by cardiac puncture. The blood samples were then stored in EDTA tubes. Plasma was isolated from the collected blood samples and stored at -80°C. Cellular DNA is extracted from 0.3-0.7 ml of plasma using QIAamp Circulating N Extraction was performed using the ucleic acid kit (Qiagen). All animals The business is part of University Health Network's Animal Care program. It was carried out in accordance with the ethical guidelines approved by the Committee.

[0046] RRBS Cell-free DNA was obtained from LCM-enriched tumor and normal samples from the same patient. Gu and others made minor changes to Nomu's DNA in 2011. 18 Follow the procedure to submit to RRBS In short, 10 ng of genomic DNA measured by Qubit was treated with restriction enzymes. Digestion with MspI, followed by terminal repair, A-tailing, and Illumina TruSe It was used for adapter coupling with the q-methylation adapter. Then, the prepared library was Z Using the YMO EZ DNA methylation kit, follow the manufacturer's instructions to perform bisulfite conversion. Each purified library was subjected to gel size selection of fragments ranging from 160 bp to 300 bp. The optimal number of cycles for amplifying the compound was determined using qPCR, and then the sample was subjected to KAPA H Using iFi Uracil + Mastermix (Kapa Biosystems) The solution was amplified and purified using AMPure beads (Beckman Coulter). After subjecting the final library to BioAnalyzer analysis, UHN Princess Illumina HiSe at Margaret Genomic Centre The sequence was determined using q 2000.

[0047] Preparation of λPCR products of exogenous enterobacteriaceae phages Intestinal bacterial phage λDNA (ThermoFischer Scientific) Amplification was performed using the primers shown in Table 1 to generate six different PCR amplicon products. The PCR reaction was performed using KAPA HiFi Hotstart ReadyMix. The following conditions were used: enzyme activation at 95°C for 3 minutes, followed by 30 cycles at 98°C for 20 seconds, 60 Final extension at 72°C for 15 seconds, 72°C for 30 seconds, and 72°C for 1 minute. PCR amplicon QI Purify using the AQuick PCR purification kit (Qiagen), then run the mixture onto a gel, and... Idleness and amplification were verified at 1 CpG, 5 CpG, 10 CpG, 15 CpG, and 20 CpG. L amplicon is CpG Methyltransferase (M.SssI)(T Methylation using hermoFischer Scientific, QIAQui Purified using a ck PCR purification kit. Methylation of the PCR amplicon was performed using restriction enzyme H. Using pyCH4IV (New England Biolabs Canada) The test was conducted, and the mixture was flowed onto a gel to ensure methylation. Non-methylation (20CpGS) and methylation were also performed. Chilled (1CpG, 5CpG, 10CpG, 15CpG, 20CpGL) amplifier After measuring DNA concentration using PicoGreen, 50% methylated and 50% unmethylated DNA were measured. It was stored together with the λPCR product.

[0048] cfMeDIP-seq A schematic diagram of the cfMeDIP-seq procedure is shown in Figure 2. Before cfMeDIP, the DNA sample is prepared Using the Kapa Hyper Prep Kit (Kapa Biosystems) The library was prepared according to the manufacturer's procedures, with some modifications. In short, the target DN A was added to 0.2 mL PCR tubes and subjected to end repair and A tailing. Adapter chain The connection is to the NEBNext adapter (NEBNext Multi for Illumina kit). iplex Oligos (New England Biolabs) is used to determine the final concentration. Prepare the solution at 0.181 μM and incubate at 20°C for 20 minutes, then add the AMPure XP bean. The solution was purified using USER enzyme (New England Bio). The eluted library was extracted using USER enzyme (New England Bio). Digestion at (labs Canada), followed by Qiagen MinElu before MeDIP. Purification was performed using the te PCR Purification Kit.

[0049] Combine the prepared library with the storage methylated / unmethylated λPCR product to create the final DNA The amount was set to 100 ng, with some modifications. Taiwo et al., 2012. 17 Using the procedure, MeDI Provided to P, the Diagenode MagMeDIP kit (Cat# It was used according to the manufacturer's procedure, with some modifications to C02010021. 0.3 ng control Methylated and 0.3 ng of control unmethylated Arabidopsis thaliana DNA, (total amount of DNA [cf (To make 100ng of DNA + filler + control) filler DNA and buffer A After adding the sample to a PCR tube containing dapter-linked DNA, the sample is heated at 95°C for 10 minutes, and then... The samples were then placed in an ice bath for 10 minutes. Each sample was then divided into two 0.2 mL PCR tubes. One is for 10% input control, and the other is the sample to be subjected to immunoprecipitation. MagMeDIP 5-mC monoclonal antibody 33D3 (Cat#C15200081) contained in the kit ) was diluted 1:15 before the production of the diluted antibody mixture and added to the sample. Washing magnetic beads (manufacturing (Following the instructions of the person in charge) The sample was also added before incubation at 4°C for 17 hours. Diag The enode was purified using the iPure Kit and eluted in 50 μl of buffer C. Before proceeding to the next step, confirm the success of reaction (QC1) and the presence of added Arabidopsis thaliana DNA. The detection of unmethylated DNA is verified by qPCR, and the recovery percentage of unmethylated DNA < 1% and the specificity of the reaction are determined. Sex %>99% (1-[Recovery of added unmethylated control DNA / Recovery of added methylated control DNA The calculation (by [ ]) was confirmed. The optimal number of cycles to amplify each library was used with qPCR. The decision was made, and then the sample was sent to the KAPA HiFi Hotstart Mastermi Amplify using x, and administer NEBNext Multiplex Oligos at a final concentration of 0. The solution was added up to 3 μM. The PCR settings used for library amplification were as follows: Activate at 95°C for 3 minutes, followed by a predetermined cycle of 98°C for 20 seconds, 65°C for 15 seconds, and 7 Final extension at 2°C for 30 seconds, followed by 72°C for 1 minute. The amplified library is then processed using MinElut. Purified using an ePCR purification column, then gel-sized using 3% Nusieve GTG The adapter dimer was removed by selection using an agarose gel. , methylated human DNA regions (testis-specific H2B, TSH2B) and unmethylated human DNA The enrichment ratio of the region (GAPDH promoter) is obtained from cell-free DNA (fine DNA obtained from ATCC). HCT116 cell line DNA that has been sheared to mimic cell lines (mycoplasma-free) Measurements were taken on the generated MeDIP-seq and cfMeDIP-seq libraries. After the final library was subjected to BioAnalyzer analysis, UHN Princes At the Margaret Genomic Centre, Illumina Hi The sequence was determined using Seq 2000.

[0050] Different methylation percentages in filler DNA cfMeDIP-seq methylation and demethylation of lambda D in the filler components of the procedure The following was performed using different percentages of NA.

[0051] [Table 1]

[0052] As shown in Figures 9 and 10, it was used to increase the final amount before immunoprecipitation to 100 ng. The filler DNA (lambda DNA) still yields good results in terms of methylated DNA recovery. While obtaining it as (Figure 10), it is preferable to have minimal recoverable unmethylated DNA (Figure 9). The composition should contain some artificially methylated DNA (100%~15%) The inventors have identified filler DNA that is 100% unmethylated filler DNA or is present in filler DNA. In samples without NA, the recovery of methylated DNA is very high, but the recovery of unmethylated DNA is... The recovery rate is also high. This is because additional methylated DNA in the filler DNA is present during the reaction. Helps to occupy excess antibodies and the amount of nonspecific binding with unmethylated DNA present in the sample. This indicates minimizing the amount of methylated DNA present throughout the sample. This is clear, and this can vary greatly from sample to sample, so use different cell-free DNA samples. In some cases, optimizing antibody dosage may not be very economical, or even possible. This filler DNA helps normalize different starting amounts and different cell-free DNA samples. It can be treated similarly (i.e., using equal amounts of antibody), and still, good methyaluronic acid The converted data can then be retrieved.

[0053] Ultra-deep target sequencing for point mutation detection The inventors of the present invention have developed QIAgen Circulating Nucleic Acid Using the ticket, go to Princess Margaret Cancer Center In this case, the corresponding tumor tissue molecular profiling data was created before registration in the initial clinical trial. Cell-free DNA was collected from approximately 20 mL of patient-derived plasma (4-5 x 10 mL EDTA blood collection tubes). The DNA was isolated from cell lines (diluted CRC and MM cell lines) using PureGene G Extracted using the entra kit and processed to approximately 180 bp using a Covaris sonicator. The DNA is fragmented, and larger fragments are removed using Ampure beads to create cell-free DNA. The fragment size was mimicked. The DNA sequencing library was created from 83 ng of fragmented DNA. NEXTflex-96 DNA Barcode Adapter (Bio Scientist KAPA Hyper Prep using fic, Austin, TX) adapter The system was constructed using a kit (Kapa Biosystems, Wilmington, MA). To isolate DNA fragments containing known mutations, the inventors of the present invention used Illumina Tr Tests performed in clinical laboratories using the uSeq Amplicon Cancer Panel. A biotinylated DNA capture probe targeting mutation hotspots was developed from 48 genes tested. (xGen Lockdown Custom Probes Mini Pool, Integrated DNA Technologies, Coralville, I A) was designed. A barcode library was pooled, and then a custom hybrid capture ra Ibrari according to the manufacturer's instructions (IDT xGEN Lockdown Procedure Version 2.1) The method was applied accordingly. These fragments were analyzed using an Illumina HiSeq 2000 instrument. The sequence was determined up to >10,000 × lead coverage. The obtained reads were then processed using bwa-mem. Alignment is performed using SAMtools and muTect version 1.1.4. It was detected.

[0054] Modeling the relationship between the number of tumor-specific features and the detection probability based on sequencing depth. The inventors created 145,000 simulated genomes. The percentage of chilled DMR was set to 0.001%, 0.01%, 0.1%, 1%, and 10%. These consisted of 1, 10, 100, 1000, and 10000 independent DMRs, respectively. The inventors have created 14,500 diploid genomes (equivalent to 100 ng of DNA) From these initial mixtures, 10, 100, 1,000, and 10,000 reads / genes were extracted. Further loci were collected to represent the degree of sequencing coverage at these depths. This process was used The experiment was repeated 100 times for each combination of bar degree, abundance, and number of features. The inventors considered the parameters For each combination, estimate the frequency of successful detection of at least one DMR and plot the probability curve. (Figure 1A) Visual evaluation of the effect of the number of features on the detection success probability conditioned on sequencing depth. It was worth it.

[0055] Calculation and visualization of differential methylation regions from cfDNA of pancreatic cancer patients and healthy donors. 24 pancreatic cancer (PC) patients and 24 healthy individuals Differential methylation regions (DMRs) between cfDNA samples from the naar are processed using MEDIPS R packages. Calculated using the page25 For each sample, BAM alignment (against the human genome hg) 19) Using the file, MEDIPS R subjects were prepared. Next, two sets of samples were used. The DMR was calculated by comparing the RPKMs of the two groups using a t-test. The raw p-values ​​from the t-test were obtained. The adjustment was performed using the Benjamini-Hochberg procedure. Then, the DMR was adjusted. We defined it as all windows with a p-value less than 0.1. A total of 38,085 DMRs were found. Of these, 6,651 were hyper-pancreatic cancer patients and 31,544 were hypo-pancreatic cancer patients. The scaled RPKM values ​​from these DMRs are shown as a heatmap (Figure 5C). This heatmap uses the distance function "Euclidean" and is a clustering method for column-direction clustering. Created using the styling function "ward" and row-direction clustering "average". It was done.

[0056] Comparison of RRBS samples from 24 pancreatic cancer tissues and 5 normal PBMCs. Download 5 normal PBMC samples profiled by RRBS from GEO. (All control samples from contract ID GSE89473) were then analyzed, and their methylation profiles were obtained. It was compared with 24 pancreatic cancer tissue RRBS samples. The downloaded BED file was R me The syntax was parsed and processed using the thylKit package. 26 Next, these five samples This was compared with similarly processed RRBS samples from 24 pancreatic cancer patients. (Custom function used) Then, at least 18 of the 24 PC samples and 4 of the 5 PBMC samples CpGs present were extracted, and only CpGs in autosomes were retained, totaling 1,806,808 cells. A background set of CpG was created. From these, Benjamini-Ho Using the chberg-adjusted criteria of p-value < 0.01 and delta-beta > 0.25, D MCs were obtained. 134,021 DMCs were hyper-effective compared to PBMCs in pancreatic cancer. It was found that, similarly, using the same q-value cutoff and delta-beta < -0.25 The inventors obtained 179,662 Hypo DMCs. A total of 313,683 The DMC is shown as a red dot in the corresponding volcano plot (Figure 7F), and the negative lo of the q value g10 was plotted against delta-beta (the horizontal line at negative log10q value = 2 is The q-value cutoff for calling DMC is shown, and the vertical dotted line is the delta-beta cutoff. (That is.)

[0057] Differences between primary tumors and normal PBMCs, as well as differences between cfDNA from pancreatic cancer patients and healthy donors. Evaluation of overlapping secondary methylation signals Permutation analysis was performed to identify DMRs in plasma (using the inventors' cfMeDIP-seq Circulating cfDNA subjected to the procedure and cancer-specific DMC (RRB) identified in the primary tumor tissue The frequency of predicted versus observed overlaps in S) was compared. The inventors examined four possible examples. :Hyper DMC overlaps with Hyper DMR, Hy overlaps with Hypo DMR per DMC, Hypo DMC which overlaps with Hypo DMR, and finally Hyper A Hypo-DMC that overlaps with a DMR. In each example, Hyper or Hypo-DMC is H The number of "biological crossovers" was obtained by overlaying it with yper or Hypo DMR. Then, the DMC Each set is left blank across the entire background set of 1,806,808 CpGs. For this purpose, the mixture was mixed 1000 times and duplicated again with each set of DMR. These random biological The intersections are placed on the same scale using Z-scores and shown as box plots and diamonds, respectively. (Figure 5E). The horizontal dotted lines in these plots are derived from the Bonferroni correction. This is the cutoff Z score associated with a q value of 0.05.

[0058] Comparison of RRBS samples from 24 pancreatic cancer tissues and 24 normal tissues, from these tissues Furthermore, evaluation of the overlap of differential methylation signals from cfDNA of pancreatic cancer patients and healthy donors. 24 PC samples were compared with 5 normal PBMC samples, and 24 normal sets of samples were derived from the same patient. We compared the weaving separately. Background set (763,874 CpGs) and PC The DMC Hyper&Hypo inside (34,013 and 11,160 respectively) can be used in the same way. These are calculated and used to create a volcano plot (Figure 7C) and a box plot (Figure 5C). D) was created in the same way.

[0059] PCA plots on 24 PCs and 24 healthy cfDNA samples. The inventors have developed 24 PCs and 24 PCs using the most variable whole genome window of the top million. We performed PCA-based clustering analysis without a dependent variable on individual healthy cfDNA samples. (Figures 7A-B). For each window, the variability was measured using the mean absolute deviation (MAD) measurement criterion. I calculated it. This is a robust measure that returns the median of absolute deviations from the median of the data, and the data 'T' represents the RPKM value across these 48 samples for a given window.

[0060] Regarding the low-methylated motifs in 24 PC and 24 healthy cfDNA samples Heatmap showing GTEx expression profiles of related TFs RNA-Seq data from GTEx databases for tissue analysis of the entire human gene spectrum. The result was obtained in the form of median RPKM (file GTEx_Analysis_v6p_RNA-se q_RNA-SeQCv1.1.8_gene_median_rpkm.gct.gz under https: / / gtexportal.org / home / datas (Obtained from ets). The target TF was matched to their gene names, and a heat map was created (Figure 8A, 8). C) was created using the median RPKM of each TF scaled across the entire organization. The distance function "manhattan" and the clustering function "average" are used in the row direction. It was used for both rasterization and columnar clustering.

[0061] Regarding the low-methylated motifs in 24 PC and 24 healthy cfDNA samples Violin plot showing GTEx expression profiles of related TFs The inventors have identified a motif that is significantly enriched in the hypomethylated region in case-versus-control groups. To estimate whether the detected TF is significantly upregulated in pancreatic cancer samples, The researchers used a randomized test with the ssGSEA score as the test statistic. The inventors identified 85 TFs that were found to be significantly associated with low methylation motifs. The scores were calculated using 1,000 random sets of 85 TFs (of all human TFs). The list is at http: / / www.tfcheckpoint.org / data / Obtained from the file TFCheckpoint_download_180515.txt Expression levels from 178 pancreatic adenocarcinoma patients on TCGA were used.

[0062] The distribution of these scores can be seen in the related violin plot (Figure 8E).

[0063] Next, we use the Wilcoxon rank-sum test to compare the irregular distribution and the observed distribution, p A value of <2.2e-16 was obtained.

[0064] The same analysis was performed on GTEx data from normal pancreases (Figure 8D). Whole blood GTEx data In contrast, the motif is hypomethylated footprint in plasma cfDNA from healthy donors. The analysis was repeated for TF (n=33) identified as lint (Figure 8B).

[0065] Results / Discussion Suitable whole-genome methods for cfDNA methylation mapping The cfMeDIP-seq method described here is robust up to 100 ng of input DNA. Existing low-input MeDIP-seq procedures 17 It was developed through modification. However, plasma samples The majority produce far less than 100 ng of DNA. To overcome this challenge... Therefore, the inventors have artificially expanded the amount of initiation DNA to 100 ng by using an external λD NA (filler DNA) was added to the adapter-linked cfDNA library (Figure 2). This minimizes the amount of nonspecific binding by antibodies and prevents loss due to binding to plastic products. The amount of DNA used is also minimized. Filler DNA is an adapter-linked cfDNA library. It consists of amplicons of similar size, with different densities of unmethylated DNA and CpG. It is composed of in vitromethylated DNA. Different patients produce different amounts of cfDNA. Therefore, the addition of this filler DNA is also useful in practice, as it standardizes the input DNA amount to 100 ng. This allows downstream procedures to proceed regardless of the amount of cfDNA available. However, this remains exactly the same in all samples.

[0066] The inventors first obtained a fragment of similar size to that observed in cfDNA. Using DNA from the colorectal cancer cell line HCT116, the cfMeDIP-seq procedure was performed. Confirmed. I chose HCT116 because public DNA methylation data is available. Yes, it was. The inventors used 100 ng of shear cell line DNA to determine the most standard MeDI P-seq procedure 17 and cf using 10 ng, 5 ng, and 1 ng of the same shear cell line DNA The MeDIP-seq procedure was performed simultaneously. This was done in two biological replications. Under all conditions, the inventors have found that the reaction has a specificity of over 99% (1-[added non-methylated control). DNA recovery / recovery of added methylated control DNA), and unmethylated region (GAPDH) We obtained an extremely high enrichment of the known methylation region (TSH2B0) exceeding [a certain value] (Figure 3B).

[0067] The library is sequenced until saturated with approximately 30 to 70 million reads per library (Table 2). (Figure 3A). The raw reads were aligned with both the human genome and the lambda genome, and the lambda genome and the parenchyma were aligned. It was found that there was no alignment (Tables 3A and 3B). Therefore, filler D The addition of exogenous λDNA as NA did not interfere with the generation of sequencing data. Finally, The inventors calculate the CpG concentration score as a quality control measure for the immunoprecipitation step. 25 All libraries showed similar enrichment for CpG, but the input control, as expected, The inventors confirmed the effectiveness of their immunoprecipitation method even with extremely low input (1 ng), without showing concentration (Figure 3C). did.

[0068] According to whole-genome correlation estimation, which compares different input DNA levels, MeDIP-seq( Both the 100ng method and the cfMeDIP-seq (10, 5, and 1ng) method are extremely robust. The Pearson correlation was at least 0.94 between any two biological copies (Figure 1B). Further analysis shows that cfMe in 5 and 10 ng of input DNA DIP-seq is a method that obtains methylated particles at 100 ng using conventional MeDIP-seq. It can robustly reproduce the rofile (at least 0.9 pairwise-Pearson correlation). (Figure 1B). The performance of cfMeDIP-seq with 1 ng of input DNA is 100 ng. Although lower than MeDIP-seq in g, it still shows a strong Pearson correlation of >0.7. This is shown (Figure 1B). The inventors believe that the cfMeDIP-seq procedure is the most standard Re duced Representation Bisulfite Sequencin g(RRBS) and Whole-Genome Bisulfite Sequence It is also possible to reproduce the DNA methylation profile of HCT116 using ng(WGBS). They acknowledged it (Figure 1C). In short, the inventors' data suggests that cfMeD IP-seq is a whole-genome methylation technique for fragmented, low-input DNA materials such as circulating cfDNA. This is a robust procedure for mapping.

[0069] cfMeDIP-seq shows high sensitivity in detecting tumor-derived ctDNA. To evaluate the sensitivity of the cfMeDIP-seq procedure, we used colorectal cancer (C RC)HCT116 cell line DNA to multiple myeloma (MM)MM1.S cell line DNA in stages Diluted. Both were sheared to mimic cfDNA size. The inventors of CR Change C DNA from 100%, 10%, 1%, 0.1%, 0.01%, 0.001% to 0%. The solutions were diluted, and cfMeDIP-seq was performed for each of these dilutions (Figures 4A-D). The inventors have developed a method for detecting three-point variations in the same sample using ultra-depth (10,000 × median) Target sequencing was also performed (coverage). 5% false positive rate (FDR: False Discovery) D identified in pure MM DNA at each CRC dilution point using the ry rate threshold The number of MR observations was almost perfectly linear with the predicted number of DMRs based on the dilution factor up to a 0.001% dilution (r 2 = 0.99, p < 0.0001) (Figure 1D). Furthermore, the DNA methylation signals within these DMRs also showed almost perfect linearity between the observed and predicted signals (r = 2 0.99, p < 0.0001) (Figure 1E). In contrast, beyond a 1% dilution, ultra-deep target sequencing was unable to accurately distinguish CRC-specific mutants from false mutants due to PCR or sequencing errors (Figure 1F). Therefore, cfMeDIP-seq has excellent detection sensitivity for cancer-derived DNA and outperforms mutant detection by ultra-deep target sequencing using standard procedures.

[0070] Cancer DNA is frequently hypermethylated in CpG-rich regions. 1 Since cfMeDIP- seq specifically targets methylated CpG-rich sequences, the inventors hypothesized that ctDNA would be preferentially enriched during the immunoprecipitation procedure. To test this, the inventors generated patient-derived xenografts (PDXs) from two colorectal cancer patients and collected mouse plasma. Tumor-derived human c fDNA was present at a frequency of less than 1% within the total cfDNA pool in the input samples and was present at a two-fold abundance after immunoprecipitation (Figure 1G). According to what these results suggest, the cfMeDIP procedure was able to further increase ctDNA detection sensitivity through biased sequencing of ctDNA.

[0071] Methylome analysis of plasma cfDNA can identify patients with early-stage pancreatic adenocarcinoma from healthy donors. The inventors performed methylome analysis of plasma cfDNA for the detection of ctDNA in early-stage cancer We decided to investigate whether it could be used. The inventors studied 24 patients (cases) with early-stage pancreatic cancer. Furthermore, methylome analysis was performed on preoperative plasma from 24 age- and sex-appropriated healthy donors (controls). (Tables 4A, 4B, and 5). For each patient, high tumor purity laser capture microscopy was performed. Dissection (LCM: laser-capture microdissected) tumor specimens And normal tissue samples were examined. cfMeDIP-seq was used to analyze circulating cFD in tumor and normal tissue. The tests were performed on NA and RRBS (Figures 5A and 6, Tables 6A and 6B). t-tests and Be-tests were also performed. By using the njamini-Hochberg correction in multiple testing, the inventors compared cases with 38,085 DMRs were obtained between the diffracted cfDNAs (p<0.01, q<0.1) (Figure 5) B~C).

[0072] Is the difference in cfDNA methylation profiles between cases and controls due to the presence of ctDNA? To evaluate this, DNA from the primary tumor and normal tissue obtained from the same patient after surgical resection was used. The methylation patterns were mapped using RRBS. The inventors identified tumors (n=24) and positive 45,173 differentially methylated CpG (DMC) samples were found among normal tissues (n=24). We identified retinally methylated CpG (Figure 7A-C).

[0073] cfDNA methylation profile when reproducing the methylation profile of those initial tumors The usefulness of the file is demonstrated by the combination of DMC in tumors and DMR in cfDNA (both of which are highly methylated). Regarding the methylation (both being hypomethylated, one being hypermethylated and the other hypomethylated), the background The test was conducted by examining the concentration of cfDNA. The inventors investigated the concentration of cfDNA. Significant enrichment of tumor-specific hypermethylation and hypomethylation sites was observed, but tumor-specific Highly methylated sites were poorly presented in cfDNA hypomethylated DMRs (Figure 5D). In fact, DNA methylation status of a given region in a tumor and methylation profile in plasma cfDNA The two factors are correlated (Figures 7D-E).

[0074] Finally, especially in the early stages of cancer, the majority of plasma cfDNA molecules in cancer patients are of non-tumor origin. Therefore, it may be released from blood cells. 14 The inventors have found that pancreatic adenocarcinoma tumor tissue and normal Peripheral blood mononuclear cells (PBMC) The differences in DNA methylation of cells were evaluated. The inventors compared tumors (n=24) and PB. Among MCs (n=5), 313,683 DMCs were identified (Figure 7F). The inventors stated, Significant enrichment of tumor-specific hypermethylation and hypomethylation sites in the same direction in cfDNA However, tumor-specific hypermethylation sites were not adequately presented in cfDNA hypomethylation DMR. (Figure 5E). Here again, the DNA methylation status of a given region in the tumor and plasma cfDNA are compared. The methylation profiles are correlated (Figure 7G-H).

[0075] In short, these results suggest that the circulating cfDNA methylation of cases and controls is The differences in the circulatory profile are primarily due to the presence of tumor-derived DNA in the circulatory system. This was the case (Figures 5D-E and 7C-H).

[0076] Plasma cfDNA methylomes enable the estimation of tumor-associated active transcription factor networks. The DMR between the case and the control is highly concentrated in tumor-derived DMR (Figure 5D-E). The inventors have found that cfDNA methylomes are associated with tumor-specific or tissue-related active transcription factors. It was hypothesized that enrichment of the following motifs was revealed. Using these cfDNA methylomes it was possible to infer the active transcriptional networks in the origin tissues of these DNA molecules. Since most TFs show variable binding based on the DNA methylation status of the target sequences 28 , to infer the active transcriptional network, the inventors examined whether DMRs in cfDNA could reveal enrichment of transcription factor (TF) footprints. Motif analysis was performed separately on hypomethylated DMRs using H OMER software for healthy donors (Figure 8A) and 20 pancreatic cancer patients (Figure 8C) to reveal potential TF footprints. 20 The inventors identified 33 motifs as hypomethylated footprints in healthy donors compared to pancreatic adenocarcinoma cases, and 85 motifs as hypomethylated footprints in pancreatic adenocarcinoma cases compared to healthy donors.

[0077] Among the 33 motifs identified as hypomethylated footprints in healthy donors, the inventors identified several TFs that are preferentially expressed in the hematopoietic system, including PU.1, Fli1, STAT5B, and KLF1 (Figure 8A - B). Similarly, among the 85 motifs identified as hypomethylated footprints in pancreatic adenocarcinoma cases, the inventors identified several TFs that are preferentially expressed in the pancreas, including RBPJL, PTF1a, Onecut1 (HNF6), and NR5A2 (Figure 8C - D). The TF motifs identified as hypomethylated footprints in pancreatic adenocarcinoma cases

[0078] Among the 33 motifs identified as hypomethylated footprints in healthy donors the inventors identified several TFs, including PU.1, Fli1, STAT5B, and KLF1, that are preferentially expressed in the hematopoietic system (Figure 8A - B). Similarly, among the 85 motifs identified as hypomethylated footprints in pancreatic adenocarcinoma cases, the inventors identified several TFs, including RBPJL, PTF1a, Onecut1 (HNF6), and

[0079] the inventors identified several TFs, including RBPJL, PTF1a, Onecut1 (HNF6), and NR5A2, that are preferentially expressed in the pancreas (Figure 8C - D). The TF motifs identified as hypomethylated footprints in pancreatic adenocarcinoma cases including RBPJL, PTF1a, Onecut1 (HNF6), and NR5A2, that are preferentially expressed in the pancreas (Figure 8C - D). The TF motifs identified as hypomethylated footprints in pancreatic adenocarcinoma cases The gene was frequently overexpressed in pancreatic adenocarcinoma patients with TCGA (Figure 8E). The inventors have identified TF as a driver for each molecular subtype of pancreatic cancer, as previously identified. Accordingly, we were able to identify several hypomethylation footprints in pancreatic adenocarcinoma cases. Ta 24 These include c-MYC and HIF1a (squamous epithelium subtype driver), N R5A2, MAFA, RBPJL and NEUROD1 (ADEX driver) and last This included FOXA2 and HNF4A (pancreatic progenitor cell subtypes).

[0080] In short, these results suggest that methylome analysis of circulating cfDNA is possible. Using differential methylation TF footprints, the active transcription network within tumors This allows us to infer and identify systemic shifts in immune cell populations between healthy donors and cancer patients. There is a possibility that it will happen.

[0081] Here, the inventors have developed a novel whole DNA system suitable for ultra-low input and fragmented DNA, such as circulating cell-free DNA. This describes a method for genomic DNA methylation. The inventors have shown that cfMeDIP-seq can be used to methylate low levels of DNA. This demonstrates extreme robustness in DNA, enabling rapid library preparation. This was achieved. Furthermore, since the inventors' method relies on the enrichment of methylated DNA, live To determine the sequence of a Lari until it is saturated, only about 30 to 70 million reads per library are needed. It eliminates the need for whole-genome sequencing and significantly reduces associated costs. Relatively low cost. In addition, the short processing time allows for the rapid deployment of cfMeDIP-seq to clinical settings. can.

[0082] Furthermore, cfMeDIP-seq relies on epigenetic information rather than genomic information. Therefore, to non-invasively monitor tissue damage in a broad set of non-malignant diseases It may be usable. For example, it could be used for infection or after cancer immunotherapy. It can monitor the immune response, and the circulating cardiac DNA or neurodegeneration after myocardial infarction. It can monitor brain DNA in the early stages of disease.

[0083] Finally, in relation to oncology, multiple cancer types have clinically distinct subgroups. It was found that these subgroups are among the many cancer types, including gliablastoma. 3 , ependymomas 4 , colorectal 5 ,breast 6、7 and pancreatic cancer 24 Different D with prognostic predictors in Stratification can be performed based on the NA methylation profile. Recent data suggests that According to some sources, pancreatic cancer patients were stratified into four subgroups driven by several mechanisms. It is possible 24 : Squamous epithelium, pancreatic progenitor cells, immunogenicity, and endocrine and exocrine abnormal differentiation (ADEX:aberrantly differentiated endocrin e exocrine). In the circulating cfDNA methylome of pancreatic cancer patients, the inventors This involves identifying the low methylation footprint from TFs that drive these subtypes. For example, the inventors found that two pathways M were concentrated in squamous epithelial subtypes. YC and HIF1 alpha (hypoxia-inducible factor 1-alpha) were identified. 24 The inventors This involves the enrichment of two TFs, HNF4A and FOXA2, in the progenitor cell subtype. It was also possible to determine this.24 Finally, the inventors were also able to identify three TFs, NR5A2, RBPJL, and MAFA, that were enriched in the ADEX subtype. 24 This suggests that cfMeDIP-seq could also be used as a biomarker to stratify cancer patients with a minimally invasive approach.

[0084] The present invention has been described with respect to specific embodiments. Modifications and changes within the spirit and scope of the present invention will be apparent to those skilled in the art. The specific embodiments disclosed herein are not intended to limit the scope of protection, which should be determined solely by the claims. All publications and references disclosed herein are incorporated by reference in their entirety.

[0085]

Table 2

[0086]

Table 3

[0087]

Table 4

[0088]

Table 5

[0089]

Table 6

[0090]

Table 7

[0091] Table 8

[0092] Table 9

[0093] Table 10

[0094] References

[0095] Table 11

[0096] Table 12

[0097] Table 13

Claims

1. A method for capturing cell-free methylated DNA from a sample containing less than 100 ng of cell-free DNA. There is, a. The sample is subjected to library preparation, and the sequence of the cell-free methylated DNA thereafter Steps that enable decision-making b. A step of adding a first amount of filler DNA to the sample, wherein the filler Steps in which at least a portion of the DNA is methylated, c. The step of denaturing the sample, and d. Capture cell-free methylated DNA using a selective binder for methylated polynucleotides. Steps to take Methods that include...

2. The step further includes amplifying the captured cell-free methylated DNA and subsequently sequencing it. The method according to claim 1.

3. The method according to claim 1, wherein the sample contains less than 50 ng of cell-free DNA.

4. The aforementioned first amount of filler DNA is approximately 5%, 10%, 15%, 20%, 30%, and 40%. 50%, 60%, 70%, 80%, 90%, or 100% methylated filler DNA, preferred Alternatively, it may contain 5% to 50%, 10% to 40%, or 15% to 30% methylated filler DNA. The method according to claim 1, wherein the remainder is unmethylated filler DNA.

5. The first amount of filler DNA is 20 ng to 100 ng, preferably 30 ng to 100 The method according to claim 1, wherein the amount is ng, more preferably 50 ng to 100 ng.

6. The cell-free DNA from the sample and the first amount of filler DNA together, at least The present invention also contains 50 ng of total DNA, preferably at least 100 ng of total DNA, according to claim 1. Method of description.

7. The filler DNA is 50 bp to 800 bp long, preferably 100 bp to 600 bp long. The method according to claim 1, more preferably having a length of 200 bp to 600 bp.

8. The method according to claim 1, wherein the filler DNA is double-stranded.

9. The method according to claim 1, wherein the filler DNA is junk DNA.

10. The method according to claim 1, wherein the filler DNA is endogenous or exogenous DNA.

11. The filler DNA is non-human DNA, preferably λDNA, according to claim 10. method.

12. The method according to claim 1, wherein the filler DNA does not align with human DNA.

13. The binder is a protein containing a methyl CpG binding domain, according to claim 1. method.

14. The method according to claim 13, wherein the protein is the MBD2 protein.

15. Step (d) is the step of immunoprecipitation of the cell-free methylated DNA using an antibody. The method according to claim 1, including the method described in claim 1.

16. At least 0.05 μg, preferably at least 0.16 μg, of the antibody is subjected to immunoprecipitation. The method according to claim 15, comprising the step of adding to the sample for that purpose.

17. Claim 15, wherein the antibody is a 5-MeC antibody or a 5-hydroxymethylcytosine antibody. Methods used.

18. To confirm the immunoprecipitation reaction, a second amount of control DNA is added to the sample after step (b). The method according to claim 15, further comprising the step of adding to

19. To confirm cell-free methylated DNA capture, a second amount of control is added after step (b). The method according to claim 1, further comprising the step of adding DNA to the sample.

20. Any one of claims 1 to 19 for measuring the DNA methylation profile in the aforementioned sample Use of the method described in item 1.

21. By correlating the aforementioned profile with known methylation profiles of tumor tissue, To confirm the presence of cell-free DNA derived from cancer cells in the aforementioned sample, D according to claim 20 Use of NA methylation profile.

22. By correlating the aforementioned profile with a known methylation profile of a specific organization, The DN according to claim 20 for identifying the tissue of origin of the cell-free DNA in the sample. Use of A methylation profile.

23. Claim 22 for identifying the tissue origin of the cancer cells in the cell-free DNA within the sample. The use according to claim 21, further comprising the use described therein.

24. The use according to any one of claims 20 to 23 for monitoring immunotherapy.

25. The use according to any one of claims 20 to 23 for the diagnosis of an autoimmune state.

26. The use according to claim 22 for measuring cell turnover in the subject from which the sample was collected. 。

Images