CN118186057A - Screening method of free DNA marker, DNA marker and application thereof - Google Patents

Abstract

A screening method of free DNA markers, DNA markers and application thereof. The screening method of the free DNA marker comprises the following steps: collecting a verified positive experiment sample and a verified control group sample; nucleotide sequencing of the positive test samples and control samples by a methylation sequencing method, the sequencing method following: segmenting CpG sites; and judging whether the methylation quantitative average value of the positive sample and the control sample of the CpG site segment meets a specific condition or not to form a candidate marker. The invention adopts a noninvasive mode to obtain the sample, the noise tolerance of the screened marker detection is high, and the signal of the control group is clean; the sensitivity of detection is high, the interference resistance is strong, and the accuracy of a single marker is not required; the repeatability is strong, is fit for extensive popularization.

Description

Screening method for free DNA marker, DNA marker and its application

Technical Field

The present invention relates to the technical field of gene detection, and in particular to a method for screening free DNA markers, DNA markers and applications thereof.

Background Art

Dementia is an important cause of brain health problems in the elderly in my country, and it directly endangers people's health and well-being. Alzheimer's disease (AD) is the main cause of senile dementia (more than 65%). By the end of 2021, the number of elderly people aged 60 and above in my country reached 267 million, accounting for 18.9% of the total population. It is estimated that around 2035, the number of elderly people aged 60 and above will exceed 400 million, accounting for more than 30% of the total population, entering a severe aging stage. There are currently about 47 million AD preclinical patients, 39 million MCI patients, and 15 million dementia patients in my country. Cognitive impairment cannot be reversed after the patient develops it, so "early detection, early diagnosis, and early treatment" are the key to AD prevention and treatment.

Extracellular amyloid plaques (β-amyloid (Aβ) plaques) and intracellular neurofibrillary tau tangles are the two main pathological features of AD and are also the basis for distinguishing AD dementia from non-AD dementia. At present, the more mature AD diagnostic methods are mainly to directly detect Aβ plaques and tau tangles in the living brain through positron emission tomography (PET), or indirectly by detecting the concentration of AD-related pathological proteins such as Aβ42 and p-Tau in cerebrospinal fluid (CSF). However, PET imaging is radioactive, expensive, and not all hospitals have imaging equipment. The extraction of cerebrospinal fluid is also very cumbersome and painful. In recent years, blood-based early disease diagnosis technology has begun to emerge. It is less traumatic to patients and can be collected multiple times, which has great advantages in implementation. Blood markers of diseases have important clinical value, but many diseases currently lack effective diagnostic and monitoring markers, especially in early disease diagnosis. The identification of markers includes aspects such as biological principle mining, biochemical experiments, and bioinformatics computational methods. There is a great demand for highly sensitive and flexible disease marker identification methods. In the field of AD, the current focus of blood testing technology is almost entirely on plasma protein markers, and the detection of these plasma protein markers mostly requires the use of ultra-sensitive biomarker detection systems (such as SIMOA) or mass spectrometry platforms. The consumables and prices are high, and the detection sensitivity and accuracy are not ideal; in summary, these defects have caused plasma protein biomarkers to be unable to meet the needs of large-scale early detection of AD in our communities.

Peripheral blood free DNA is a naturally occurring DNA fragment, which is mostly released into the peripheral blood after cell death in the human body. In healthy people, free DNA mainly comes from blood cells and liver, but in patients with diseases, free DNA is released by diseased tissues, triggering immune responses, causing changes in free DNA. Therefore, free DNA analysis can be used for disease diagnosis and detection, etc. In addition, the metabolic half-life of free DNA is about 6 hours, so free DNA analysis can achieve real-time monitoring of the human body. At present, free DNA analysis is mostly focused on cancer, pregnancy, infectious diseases and other fields. DNA methylation is an important DNA epigenetic modification. A common type of DNA methylation is 5-methylcytosine (5mC), that is, a methyl group is added to cytosine (C). In humans, 5mC mostly occurs on cytosine in CpG dinucleotides (i.e., cytosine and guanine connected by phosphates), which can affect the stability of the genome and regulate gene expression, and has strong cell and tissue specificity. There are tissues and organs with different functions in the human body, such as liver, lung, kidney, brain, etc. The genomes between them are almost exactly the same, but there are large differences in DNA methylation modification; when a tissue in the human body is in a pathological state, causing its cells to die and release DNA into the peripheral blood, the pathological state can be detected and diagnosed by identifying its tissue-specific DNA methylation or site. In addition, the immune response produced by the body under pathological conditions can also lead to abnormal changes in free DNA methylation. Free DNA methylation detection has been widely used in prenatal diagnosis and cancer diagnosis, and has achieved great success. However, the identification of most disease markers of free DNA methylation requires the collection of methylation data of disease-related tissues (or the collection of related tissues and obtaining them through experiments by themselves), and the identification of disease markers by comparing with peripheral blood or normal tissues, and then verifying them in peripheral blood. The different ways of obtaining methylation data have a significant impact on the subsequent marker identification.

For example, based on DNA methylation chips (common ones include Illumina HumanMethylation450BeadChip, Illumina Infinium Methylation EPIC BeadChip, etc.), it covers a fixed number of CpG sites with specific sequences. The capture depth of each CpG site in the experimental data is high, and the distance between adjacent CpG sites is far, so each CpG site is generally analyzed separately. Methods based on specific recognition enzymes for enrichment, such as MEDIP-seq, have low resolution and are rarely used in disease diagnosis. The current mainstream is sequencing-based methods, such as WGBS, RRBS, EM-seq, TAPS and other technologies, which can cover almost all CpG sites. However, due to the limitation of sequencing depth, the quantitative deviation of the methylation level of a single CpG site is large. At the same time, considering that the methylation status between adjacent CpGs is correlated, multiple adjacent CpG sites will be used as a marker, and how to segment CpG sites is very important. There are many mainstream approaches, such as treating a single element or part of its sequence as a segment based on known regulatory elements such as CpG islands, CpG island shores, and gene promoters, but these methods still have shortcomings. In addition, due to the low concentration of free DNA (only about 7 ng of free DNA per milliliter of plasma), and the great damage of sulfites in traditional WGBS experiments to DNA, free DNA methylation experiments are very difficult; disease-related CpG sites only account for a very small part of the genome, resulting in high detection costs. In summary, the development of early disease diagnosis technology based on free DNA methylation requires optimization in both experimental and computational methods.

Summary of the invention

In view of this, the main purpose of the present invention is to provide a method for screening free DNA markers, DNA markers and applications thereof, in order to at least partially solve the above technical problems.

In order to achieve the above object, as a first aspect of the present invention, a method for screening free DNA markers is provided, comprising the following steps:

Select any human gene sequence and take its first CpG site as a candidate segment; for any candidate segment, examine its next adjacent CpG site, if the interval between the CpG site and the last CpG in the candidate segment is less than or equal to a first preset threshold Di, merge the CpG site into the current candidate segment, and continue to examine the subsequent CpG sites until the interval condition is not met;

Inspect the current candidate segment, and if the number of CpG sites contained in the segment reaches a second preset threshold value Num, mark the candidate segment as a formal segment, otherwise discard the candidate segment;

After the candidate segment is examined, the next adjacent CpG site is used as a new candidate segment to start, and the above process is repeated until the preset end condition is reached;

Collect verified positive experimental samples and control group samples;

The positive experimental samples and the control group samples were sequenced by DNA methylation determination method:

For all samples, for all formal segments determined above, the methylation quantitative values of all CpG sites contained therein are averaged as the methylation level of the formal segment;

For each formal segment, if the number or proportion of samples in the control group whose methylation levels are less than a preset threshold a is not less than x, and the number or proportion of samples in the positive experimental samples whose methylation levels are greater than another preset threshold b is greater than a preset threshold y, then the formal segment is taken as a candidate marker; or

For each formal segment, if the number or proportion of samples in the control group whose methylation levels are greater than the preset threshold aa is not less than the preset threshold xx, and the number or proportion of samples in the positive experimental samples whose methylation levels are less than another preset threshold bb is greater than the preset threshold yy, then the formal segment is also taken as a candidate marker.

As a second aspect of the present invention, a candidate marker obtained by screening according to the screening method as described above is also provided;

Preferably, the candidate marker comprises the nucleotide sequence as described in SEQ ID No.1 to SEQ ID No.513.

As a third aspect of the present invention, there is also provided a primer, a primer amplification chain or a sample composition obtained by PCR amplification, replication, conversion and/or transformation based on the candidate marker as described above.

As a fourth aspect of the present invention, a methylation machine sequencing system is also provided, wherein the methylation machine sequencing system comprises a nucleic acid library of candidate markers as described above.

Based on the above technical solutions, it can be seen that the marker screening method, DNA marker and application thereof of the present invention have at least one of the following beneficial effects compared with the prior art:

1. The present invention does not require tissue samples, thus overcoming the technical difficulty of obtaining brain tissue, especially the brain tissue of early AD patients, which is almost impossible to obtain;

2. In the screening method of the present invention, whole genome sequencing is used, and the method of segmenting the genome in combination with CpG has a high degree of freedom, and the effective gene segmentation can be as short as 10 bp (i.e., 10 bases) long and contain only 3 CpG sites;

3. The noise tolerance of the marker combination screened by the present invention is high, and the control group signal is clean;

4. The detection method of the marker combination screened by the present invention has high sensitivity, which does not require a high accuracy rate of a single marker, but rather selects all potential markers;

5. The detection method of the present invention only collects peripheral blood, is non-invasive to the human body, and can be collected repeatedly;

6. Research in the field of cancer diagnosis has shown that disease-related free DNA markers in the blood appear earlier than proteins, so free DNA-related markers have better application prospects in early diagnosis; at the same time, most free DNA is double-stranded DNA, which has high stability and strong repeatability;

7. The metabolic half-life of free DNA is short, only 6 hours, so diseases can be monitored in real time; the free DNA methylation detection technology is mature and can detect multiple sites at one time, which is suitable for large-scale promotion and use;

8. The marker identification parameters of the present invention are not based on statistical tests, and sites with statistical differences are often not effective as markers; the present invention does not pursue a single marker to achieve good results, but rather identifies a certain amount of potential markers with strong anti-interference capabilities. The combined use of all or part of these markers (such as using average values, combined with machine learning technology) can achieve better diagnostic results.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a comparative dot-line graph of the free DNA methylation rate (methylation level) of a marker of Seq.21 in an early AD patient (asterisk) and a healthy elderly control (×) (each dot in the graph represents a CpG site);

Figures 2A and 2B are the arithmetic mean distribution diagrams of the methylation levels of all markers in healthy elderly people (left) and AD patients (right) based on Aβ PET imaging in two independent data sets, respectively (each point in the figure represents a patient);

3 is a performance comparison diagram (ROC curve diagram) of the markers of the present invention and other AD markers (Aβ42/Aβ40, p-Tau181, NfL, GFAP) in the classification of healthy controls and AD patients;

FIG4 is a performance comparison diagram of the free DNA marker of the present invention and other AD markers (Aβ42/Aβ40, p-Tau181, NfL, GFAP) in predicting the intensity of Aβ accumulation signals in the patient's brain.

DETAILED DESCRIPTION

In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

Unless otherwise limited herein, all technical and scientific terms used herein have the same meanings as commonly understood by those of ordinary skill in the art to which the invention belongs. Any methods and materials similar or equivalent to the methods and materials described herein can be used in the practice or testing of the present invention, but preferred methods and materials are described.

All patents and publications, including all sequences disclosed within such patents and publications, mentioned herein are expressly incorporated by reference.

Numerical ranges are inclusive of the numbers defining the range. Unless otherwise indicated, nucleic acids and amino acids are written left to right in 5' to 3' orientation and amino to carboxyl orientation, respectively.

The headings provided herein are not limitations of the various aspects or embodiments of the invention. Accordingly, the terms defined below are more fully defined in conjunction with the specification as a whole.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the invention belongs. For clarity and ease of reference, certain terms are defined below.

As used herein, the term "sample" refers to a material or mixture of materials, usually, although not necessarily, in liquid form, which contains one or more analytes of interest.

As used herein, the term "nucleic acid sample" refers to a sample comprising nucleic acids. As used herein, nucleic acid samples can be complex because they comprise a variety of different molecules comprising sequences. Mammal (e.g., mouse or human) genomic DNA is typical of complex samples. Complex samples can have more than 104, 105, 106 or 107 different nucleic acid molecules. DNA targets can be derived from any source, such as genomic DNA or artificial DNA constructs.

The term "CpG" as used herein refers to 5'-C-phosphate-G-3', i.e., two nucleotides consisting of cytosine (C) and guanine (G) adjacent to each other.

The term “human genome GRCh38” used in this article refers to the 38th version of the human genome reference sequence released by the Genome Reference Consortium (GRC) in 2013.

The term "primer" as used herein refers to a natural or synthetic oligonucleotide that can serve as a starting point for nucleic acid synthesis when forming a duplex with a polynucleotide template, and extends from its 3' end along the template to form an extended duplex. The nucleotide sequence added during the extension process is determined by the template polynucleotide sequence. Usually the primer is extended via a DNA polymerase. The primer length is usually adapted to its use in the synthesis of primer extension products, usually in the range of 8 to 100 nucleotides in length, such as 10 to 75, 15 to 60, 15 to 40, 18 to 30, 20 to 40, 21 to 50, 22 to 45, 25 to 40, etc. Typical primers can be in the range of 10-50 nucleotides in length, such as 15-45, 18-40, 20-30, 21-25, etc., and any length within the range. In some embodiments, a primer is typically no more than about 10, 12, 15, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 35, 40, 45, 50, 55, 60, 65, or 70 nucleotides in length.

As used herein, the term "plurality" includes at least 2. In some cases, the plurality can be at least 10, at least 100, at least 100, at least 10,000, at least 100,000, at least 106, 107, 108, 109 or more.

As used herein, the term "sequencing" refers to a method by which the identity of at least 10 consecutive nucleotides of a polynucleotide is obtained (e.g., the identity of at least 20, at least 50, at least 100, or at least 200 or more consecutive nucleotides).

The term "tagging" as used herein refers to the attachment of a sequence marker (which includes an identifier sequence) to a nucleic acid molecule. The sequence marker can be added to the 5' end, the 3' end, or both ends of the nucleic acid molecule. The sequence marker can be added to the fragment and the linker can be connected to the fragment by, for example, T4 DNA ligase or other ligases.

As used herein, the term "marker", also referred to as "marker", refers to a certain parameter that can distinguish between two states.

The term "free DNA", also known as "circulating cell-free DNA" (cfDNA) as used herein, refers to DNA circulating in the patient's peripheral blood. The DNA molecules in the cell-free DNA may have a median size of less than 1 kb (e.g., in the range of 50 bp to 500 bp, 80 bp to 400 bp, or 100 to 1,000 bp), although fragments with a median size outside this range may also be present. Cell-free DNA may include circulating tumor DNA (ctDNA), i.e., tumor DNA that circulates freely in the blood of cancer patients or circulating fetal DNA (if the subject is a pregnant female/female). cfDNA can be highly fragmented and can have a median fragment size of about 166 bp in some cases. cfDNA can be obtained by centrifuging whole blood to remove all cells and then isolating DNA from the remaining plasma or serum. CfDNA is mostly double-stranded, but can be denatured to obtain single strands.

As used herein, the term "amplifying" refers to producing one or more copies of a target nucleic acid using a target nucleic acid as a template.

As used herein, the term "copy of a fragment" refers to the product of amplification, wherein the copy of the fragment can be the reverse complement of the fragment strand, or have the same sequence as the fragment strand.

The present invention relates to the following methods for detecting methylated genes:

1. Whole Genome Bisulfite Sequencing (WGBS)

WGBS technology uses sodium bisulfite to treat DNA, causing unmethylated C to become U, and become T in the subsequent PCR and sequencing process, while methylated C is not affected, so that which sites are methylated can be identified during detection. Sodium bisulfite treatment of DNA is prone to damage, especially CpG islands contain a large number of unmethylated C, so coverage is easy to be low in this area. The library construction method of WGBS technology can be divided into two types. One is to first fragment the DNA and connect the adapter, and then treat the C->T with sulfite; the other is to first convert C->T, and then connect the adapter to expand. The latter requires lower DNA input. It should be noted that WGBS technology cannot distinguish between 5-hmC and 5-mC.

The commonly used WGBS alignment software is Bismark, which pre-converts the reference genome sequence into two types: C->T and G->A. During alignment, each read also undergoes two conversions: C->T and G->A. After the combination, each read is equivalent to undergoing four different alignments. The best alignment is selected from these alignments to determine the chain direction where methylation occurs and possible methylation sites.

2. Reduced Representation Bisulfite Sequencing (RRBS) technology

The principle of RRBS technology is to enrich the fragments rich in CCGG sites on genomic DNA through restriction enzyme digestion, and then perform single-base resolution methylation sequencing in the CpG-rich region of the genome through bisulfite treatment and high-throughput sequencing technology. Compared with WGBS technology, RRBS technology is a cost-effective methylation sequencing solution, and its sequencing volume is greatly reduced, which has a wide range of application value in large-scale clinical sample research.

RRBS technology uses the property of bisulfite that it can convert unmethylated cytosine (C) into thymine (T). After the genome is treated with bisulfite and sequenced, the methylation rate can be calculated based on the ratio of the number of reads that have not been converted to C or T at a single C site to the number of all covered reads. This technology has important application value for comprehensive research on epigenetic mechanisms of embryonic development, aging mechanisms, disease occurrence and development, and screening of disease-related epigenetic marker sites.

The process of RRBS technology includes: (1) Detection of DNA samples, mainly including two methods: Qubit accurately quantifies DNA concentration; agarose gel electrophoresis analyzes the degree of DNA degradation and whether there is contamination. (2) Use different extraction schemes according to different sample types to obtain high-quality genomic DNA; Qubit detects the concentration of DNA samples, and agarose gel electrophoresis detects the integrity of DNA samples; enriches CpG fragments by enzyme digestion; introduces adapter sequences and index sequences at both ends of DNA molecules; electrophoresis recovers and selects 250-500bp fragments rich in CpG; adds λDNA to the recovered genomic DNA, and then converts the non-methylated base C to U through bisulfite treatment; PCR amplifies the enriched library and purifies the PCR product to obtain the final library. (3) After the library is constructed, use Qubit2.0 for preliminary quantification, dilute the library to 1ng/μl, and then use Agilent2100 to detect the insert size of the library. If it meets expectations, use the qPCR method to accurately quantify the effective concentration of the library to ensure the quality of the library. (4) Sequencing: After the library is qualified, different libraries are pooled according to the effective concentration and target data volume requirements and then sequenced on the IlluminaNova platform.

3. Enzymatic conversion methylation detection - EM-seq sequencing

EM-seq sequencing combines enzymatic conversion methylation library preparation with a targeted methylation capture system to detect the methylation status of 3.98 million CpG sites in a 123Mb region on the human genome. EM-seq sequencing is ideal for exploring methylation levels in various applications such as cancer metastasis, human development, and functional genomics. This method has a wide coverage range, high detection sensitivity, and reduces sequencing costs. It is suitable for large cohort sample studies, especially for methylation studies of cfDNA samples.

The sample starting amount required for enzymatic conversion is relatively low, and generally 10-200ngDNA can meet the requirements. After DNA is sheared (such as ultrasonic shearing; cfDNA is naturally degraded and does not need to be sheared again), it is converted by two steps of enzymatic method to distinguish unmethylated cytosine from 5mC and 5hmC. The enzymatic treatment process is gentler on DNA and can minimize damage to DNA, so the DNA obtained after enzymatic conversion is more complete, and there are more and longer inserts in the library, which can ultimately obtain longer sequences and higher alignment rates.

The present invention discloses a method for screening free DNA markers, comprising the following steps:

Collect verified positive experimental samples (positive sample) and control group samples (negative sample).

The positions (genomic coordinates) of all CpG sites in the human genome (theoretically any version and any patch are acceptable. In this project, the 13th patch version of GRCh38, i.e., GRCh38.p13) are used. The first CpG site on the positive experimental sample is used as a candidate segment. For a candidate segment, its next adjacent CpG site is examined: if the interval between the CpG site and the last CpG in the candidate segment is within a specified range (the first preset threshold, such as 10 bases), the CpG site is merged into the current candidate segment, and the subsequent CpG sites are continued to be examined; if the interval is greater than the specified range, the current candidate segment is examined: if the number of CpG sites it contains reaches the second preset threshold (such as 3, 4, or 5), the candidate segment is marked as a formal segment, otherwise the candidate segment is discarded; after the candidate segment is examined, its next adjacent CpG site is used as a new candidate segment to start, and the process is repeated until all CpG sites are examined. The CpG segmentation method of the present invention can achieve the following: 99% of the segments are below 63 bp, with a median of 17 bp, which is very suitable for experimental verification and kit design in free DNA; each segment contains 3 or more CpG sites, which can make up for the defects in the depth of sequencing data and improve the reliability of markers; it is not based on known functional annotation elements such as CpG islands, and can identify potential markers located in enhancer regions (with better tissue specificity than gene promoters).

Among them, the first preset threshold Di is, for example, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, 40, 50, 60, 70, 80, 90 or 100 (bases); the second preset threshold Num is, for example, 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 (CpG sites).

For all experimental samples, for all CpG segments determined above, the methylation quantitative values of all CpG sites contained in them are averaged as the methylation level of the segment. There are many ways to take the average, such as the arithmetic mean. Since the sequencing depth of each methylation site in the data is different, in this project, a weighted average with sequencing depth as the weight is used. This method is the most commonly used average calculation method in methylation analysis. Other methods include the arithmetic mean, median, and so on of the methylation levels of all CpG sites.

For each CpG segment, if the number or proportion of its methylation level in all control groups is less than a preset threshold a (for example, 10%) is not less than x (for example, 90%), and the number or proportion of samples in the disease group that are greater than another preset threshold b (for example, 20%, which may be the same as a) is greater than a preset threshold y (for example, 25%), then the CpG segment is taken as a candidate marker; or the number or proportion of its methylation level in all control groups that is greater than a preset threshold aa (for example, 90%) is not less than a preset threshold xx (for example, 90%), and the number or proportion of samples in the disease group that is less than another preset threshold bb (for example, 80%, which may be the same as aa) is greater than a preset threshold yy (for example, 25%), then the CpG segment is taken as a candidate marker.

Therefore, by setting a preset threshold, several CpG segment screening results of different levels can be determined, and their detection sensitivity also varies with the set preset threshold.

For example, the preset threshold a is 10%, the preset threshold x is 90%, the preset threshold b is 20%, the preset threshold y is 25%, the preset threshold aa is 90%, the preset threshold xx is 90%, the preset threshold bb is 80%, and the preset threshold yy is 25%.

Or for example, the preset threshold a is 5%, the preset threshold x is N-2 (N is the number of samples in the control group), the preset threshold b is 20%, the preset threshold y is 33%, the preset threshold aa is 95%, the preset threshold x is N-1 (N is the number of samples in the control group), the preset threshold bb is 80%, and the preset threshold yy is 33%. Preferably, the present invention adopts a scheme in which the preset threshold a is 5%, the preset threshold x is N-1 (N is the number of samples in the control group), the preset threshold b is 5%, the preset threshold y is 25%, the preset threshold aa is 95%, the preset threshold xx is N-1 (N is the number of samples in the control group), the preset threshold bb is 95%, and the preset threshold yy is 25%.

In the above scheme, the marker identification parameters of the present invention are not based on statistical tests, which is caused by the difference in emphasis between disease diagnosis and biological research. The sites with statistical differences are often not good as markers, especially the sites where the control group is not very clean, such as the control group average 50%, the disease group average 70%, which may be statistically significant, but basically cannot be used in diagnosis, especially when the disease signal is weak and the source is unclear (such as early-stage disease). The present invention also does not pursue that a single marker can achieve a good effect (often leading to over-optimization (over-fitting)), but identifies a certain amount of potential markers to enhance anti-interference ability (such as a single marker is easily affected by primer bias, individual SNP, etc.). The present invention can also further use data processing technology, such as machine learning technology, such as random forest, support vector machine, neural network, gradient boosting decision tree, etc., or a relatively simple method, such as arithmetic mean, to combine all or part of these markers for use, so as to achieve a better diagnostic effect.

In the above scheme, the positive experimental sample is, for example, a blood or body fluid sample, such as a peripheral blood sample, whose cerebral cortex Aβ plaque pathology is positive (AD patient group) determined by AβPET; the control group sample can be a blood or body fluid sample whose cerebral cortex Aβ plaque pathology is negative determined by AβPET. Furthermore, the blood or body fluid sample is preferably obtained by non-invasive means, and is preferably tested within 6 hours of ex vivo to ensure the integrity of free DNA in the experimental sample.

In the above scheme, the specific methylation sequencing method can be, for example, the above-mentioned well-known sequencing methods, such as WGBS technology, RRBS technology, EM-seq sequencing method and TAPS technology, or digital PCR, real-time fluorescence quantitative PCR and methylated DNA specific recognition enzyme binding technology can be used to measure the methylation degree.

In the above scheme, for example, the data processing technology may also be considered:

The mean methylation level of all markers included in the analysis;

A statistical value of all markers included in the analysis, such as Q1, median, Q3, etc.

The proportion of non-zero values among all markers included in the analysis;

Develop classification methods using machine learning techniques including but not limited to decision trees, random forests, support vector machines, neural networks, gradient boosted decision trees, and more.

The method for screening free DNA markers of the present invention can screen the positions of all CpG sites in the human genome GRCh38.p13, or can screen only some of the positions.

Through screening, the present invention can obtain a series of candidate markers. Each candidate marker, as an independent gene fragment, can be used as a free DNA marker to accurately determine whether the corresponding gene fragment is methylated. Therefore, each candidate marker has its unique use value and protection value. The present invention also uses it as a composition of gene fragments for special protection.

Such a gene fragment composition includes, for example, the nucleotide sequences shown in Table 1.

Table 1 Composition of partial gene fragments screened by the method of the present invention

When it is used for free DNA markers for methylation gene detection machine sequencing as described above, it has the advantages of high resolution and good accuracy. The cfDNA methylation markers screened by the present invention combined with data processing technology can effectively distinguish healthy controls from early AD patients with an accuracy rate of more than 92%, which is significantly better than the currently commercialized plasma protein markers.

The present invention also discloses a marker obtained by screening according to the screening method as described above; and a primer, a primer amplification chain or an enriched cell-free DNA sample composition obtained by PCR amplification, replication, conversion and/or transformation based on the marker as described above. Wherein, the cell-free DNA sample composition, for example, includes a cell-free DNA molecule connected to a linker, an internal standard control composition including an amplicon, and a streptavidin carrier, and the cell-free DNA sample composition contains the candidate marker as described above.

The present invention also discloses a methylation machine sequencing system, which comprises the nucleic acid library of candidate markers as described above.

In summary, the present invention discloses and claims protection for a composition of gene fragments screened based on the above-mentioned screening method for free DNA markers, as well as primers, primer amplification chains, and sample compositions formed by PCR amplification, replication, conversion, and transformation of the composition of these gene fragments, and a nucleic acid library with unique markers formed by incorporating the above-mentioned composition of gene fragments into a nucleic acid library (library of nucleic acids), and a nucleic acid sequencing system using the same, for example, incorporating it into an Illumina nucleic acid library, and an Illumina nucleic acid sequencing system using such a nucleic acid library, etc.

The present invention will be further described below through specific examples. It should be noted that the following examples are only for illustration and are not intended to limit the present invention.

Example

1. Patient recruitment and blood sample processing

All patients included in the analysis were determined by AβPET to determine whether their cerebral cortical Aβ plaque pathology was negative (control group) or positive (AD patient group). Blood processing was routine: peripheral blood (e.g., 6 ml) was collected from the vein using a blood collection tube containing an anticoagulant (e.g., EDTA), centrifuged at low temperature (e.g., 4°C) and medium speed (e.g., 1600 g-acceleration) for a period of time (e.g., 15 minutes), then the upper layer of liquid was aspirated and transferred to a new centrifuge tube, and then centrifuged at low temperature (e.g., 4°C) and high speed (e.g., 16000 g-acceleration) for a period of time (e.g., 15 minutes) using a centrifuge, then the upper layer of liquid (i.e., plasma) was aspirated and transferred to a new centrifuge tube or cryopreservation tube. Plasma can be stored in an ultra-low temperature refrigerator (e.g., -80°C) until use.

2. EM-seq data analysis and marker identification

Free DNA is extracted using a free DNA extraction kit (commercialized), free DNA is processed and library is constructed using EM-seq technology, and then the library is sequenced using second-generation sequencing technology (such as illuminaNova Seq 6000 sequencer) to obtain the methylation level of each CpG site. The sequencing data is analyzed by bioinformatics methods, including accusation, alignment and methylation identification, to obtain the methylation level of each CpG site. The present invention uses Msuite2 software for data analysis, which includes a data quality control module, which can remove low-quality sequencing data and allow the removal of some sequences during analysis to reduce the impact of free DNA end defects (such as jagged ends) and improve the accuracy of the results.

The positions of all CpG sites in the human genome (any major version, any patch version is acceptable, the 13th patch version of GRCh38, i.e., GRCh38.p13, is used in the present invention, and older versions GRCh 37, GRCh36 or the latest versions T2T CHM13v2.0/hs1, Han1, etc. can also be used). The first CpG site is used as a candidate segment. For a candidate segment, if the interval between its adjacent next CpG and the last CpG in the segment is within a specified range (e.g., 10 bases), it is merged into the current candidate segment, and the subsequent CpG sites are continued to be investigated; if the interval is greater than the specified range, the current candidate segment is investigated, and if the number of CpG sites it contains reaches a preset threshold (e.g., 3), the candidate segment is marked as a formal segment, otherwise it is discarded, and then the next CpG site is used as a new candidate segment to start, and the process is repeated until all CpG sites are investigated.

For all experimental samples, for all CpG segments determined above, the methylation quantitative values of all CpG sites contained in them are averaged as the methylation level of the segment. There are many ways to take the average, such as the arithmetic mean. Since the sequencing depth of each methylation site in the data is different, in this project, a weighted average with sequencing depth as the weight is used, which is also the most commonly used average calculation method in methylation analysis.

For each CpG segment, if its methylation level in all control groups is less than a preset threshold a (=10%), and the proportion of samples in the disease group that are greater than another preset threshold b (=20%) is greater than a preset threshold c (=25%), then the CpG segment is taken as a candidate marker; or if its methylation level in all control groups is greater than a preset threshold aa (=90%), and the proportion of samples in the disease group that are less than another preset threshold bb (=80%) is greater than a preset threshold cc (=25%), then the CpG segment is taken as a candidate marker.

3. AD-related biomarkers

Through the experiment, 513 markers were temporarily screened as shown in Table 1, and the sequences of these markers and their position information in GRCh 38 were provided. Note that the length of each marker is different, and the number of CpGs contained is also different (all more than 3). For each marker, each CpG site contained in it can also be regarded as a marker.

4. The 513 markers measured were reused in the above-mentioned control group (healthy elderly) and AD patient group (early AD patients), and the marker sites of the present invention were determined by whole genome methylation sequencing to verify the experimental results. Among them, two sample sets were collected, and sample set 1 as the discovery set tested 19 groups of healthy elderly people and 24 groups of early AD patients for 513 markers (due to space limitations, sample set 1 was not posted); as shown in Appendix 1, sample set 2 as the validation set tested 17 groups of healthy elderly people and 13 groups of early AD patients for 513 markers (due to space limitations, sample set 2 only posted 8 groups of healthy elderly people and 8 groups of early AD patients).

5. Graphical display of some of the above test results

Figure 1 is a comparative dot-line graph of the free DNA methylation rate (methylation level) of the marker SEQ ID No. 21 for an early AD patient (asterisk) and a control healthy elderly person (×) (each dot in the figure represents a CpG site). It can be seen from the figure that the marker can significantly distinguish healthy elderly people from early AD patients, thereby playing a role in early detection and warning, improving the initial diagnosis recognition rate, and contributing to the early intervention treatment and rehabilitation of patients.

Figure 2A and Figure 2B are the distribution diagrams of the arithmetic mean of the methylation levels of all markers in healthy elderly people (left) and AD patients (right) based on AβPET imaging in two independent data sets (each point in the figure represents a patient). The definitions of sample set 1 and sample set 2 are as described above. In these two sets, the average free DNA methylation level of all sites is used as a predictive index, and it is found that it can well distinguish healthy elderly people from early AD patients, with the following accuracy rates:

Sample set 1: sensitivity = 100%, specificity = 100%;

Sample set 2: Sensitivity = 92.3%, specificity = 94.1%.

The present invention can also use any number of sites for prediction, for example, FIG1 uses one site, FIG2 uses all sites, or 2 to 512 sites can be randomly selected to calculate the average.

In addition, it should be noted that some values in the two data sets are "NA", and the sequencing depth of the corresponding samples at this site is low. This site will be ignored when calculating indicators such as the average value and will not participate in the average calculation.

Figure 3 is a performance comparison diagram (ROC curve diagram) of the markers of the present invention and other AD markers (Aβ42/Aβ40, p-Tau181, NfL, GFAP) in the classification of healthy controls and AD patients. Among them, DNA methylation-dataset 1 (AUC=1), DNA methylation-dataset 2 (AUC=0.96), plasma p-Tau181 (AUC=0.89), plasma NfL (AUC=0.86), plasma GFAP (AUC=0.76), plasma Aβ42/Aβ40 (AUC=0.80). Through ROC analysis, it can be seen that the schemes of the present invention (①, ②) have better prediction accuracy in early AD patients than existing plasma protein markers (③~⑥).

Figure 4 is a performance comparison chart of the free DNA marker of the present invention and other AD markers (Aβ42/Aβ40, p-Tau181, NfL, GFAP) in predicting the cumulative signal intensity of Aβ in the patient's brain. From the comparison between A and other B to E in Figure 4, compared with plasma protein markers, the free DNA methylation marker of the present invention has a better correlation with the brain Aβ signal intensity, indicating that the marker of the present invention has a better accuracy in dynamic monitoring.

The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A method for screening free DNA markers, characterized in that it comprises the following steps: For a gene sequence, the first CpG site is used as a candidate segment; for any candidate segment, the next adjacent CpG site is examined, and if the interval between the CpG site and the last CpG in the candidate segment is less than or equal to a first preset threshold, the CpG site is merged into the current candidate segment, and the subsequent CpG sites are continuously examined until the interval condition is not met; Inspect the current candidate segment, and if the number of CpG sites contained in the segment reaches a second preset threshold, mark the candidate segment as a formal segment, otherwise discard the candidate segment; After the candidate segment is examined, the next adjacent CpG site is used as a new candidate segment to start, and the above process is repeated until the preset end condition is reached; Collect verified positive experimental samples and control group samples; The positive experimental samples and the control group samples were measured by the DNA methylation measurement method: For all samples, for all formal segments determined above, the methylation quantitative values of all CpG sites contained therein are averaged as the methylation level of the formal segment; For each formal segment, if the number or proportion of samples in the control group whose methylation levels are less than a preset threshold a is not less than x, and the number or proportion of samples in the positive experimental samples whose methylation levels are greater than another preset threshold b is greater than a preset threshold y, then the formal segment is taken as a candidate marker; or For each formal segment, if the number or proportion of samples in the control group whose methylation levels are greater than the preset threshold aa is not less than the preset threshold xx, and the number or proportion of samples in the positive experimental samples whose methylation levels are less than another preset threshold bb is greater than the preset threshold yy, then the formal segment is also taken as a candidate marker. 2. The screening method according to claim 1, characterized in that the positive experimental sample is a blood or body fluid sample whose cerebral cortex Aβ plaque pathology is positive as determined by AβPET; preferably a peripheral blood sample; The control group samples are blood or body fluid samples whose cerebral cortex Aβ plaque pathology is determined to be negative by AβPET; Preferably, the blood or body fluid sample is obtained by a non-invasive method; further preferably, the sample is a peripheral blood sample obtained by a non-invasive method and within 6 hours of ex vivo. 3. The screening method according to claim 1, characterized in that the DNA methylation determination method is selected from WGBS technology, RRBS technology, EM-seq sequencing method, TAPS technology, digital PCR, real-time fluorescence quantitative PCR and methylated DNA specific recognition enzyme binding technology; The first preset threshold is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, 40, 50, 60, 70, 80, 90 or 100; The second preset threshold is 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10. 4. The screening method according to claim 1 is characterized in that the method for calculating the average value is selected from the group consisting of arithmetic mean, median average, and weighted average. 5. The screening method according to claim 1, characterized in that The preset threshold a is 10%, the preset threshold x is 90%, the preset threshold b is 20%, the preset threshold y is 25%, the preset threshold aa is 90%, the preset threshold xx is 90%, the preset threshold bb is 80%, and the preset threshold yy is 25%; or The preset threshold a is 5%, the preset threshold x is N-2, the preset threshold b is 20%, the preset threshold y is 33%, the preset threshold aa is 95%, the preset threshold xx is N-1, the preset threshold bb is 80%, and the preset threshold yy is 33%; wherein N is the number of samples in the control group; or The preset threshold a is 5%, the preset threshold x is N-1, the preset threshold b is 5%, the preset threshold y is 25%, the preset threshold aa is 95%, the preset threshold xx is N-1, the preset threshold bb is 95%, and the preset threshold yy is 25%; wherein N is the number of samples in the control group. 6. The screening method according to claim 1 is characterized in that, in the step of determining candidate markers based on formal segmentation, machine learning technology is also used, preferably random forest, support vector machine, neural network, gradient boosting decision tree method to optimize the screening results. 7. The screening method according to claim 1, characterized in that the positions of all or part of the CpG sites on all chromosomes in the human genome are screened; The human genome data are from GRCh38, GRCh37, GRCh36, T2TCHM13v2.0/hs1 or Han1. 8. The candidate marker obtained by screening according to any one of claims 1 to 7; Preferably, the candidate marker comprises the nucleotide sequence as described in SEQ ID No.1 to SEQ ID No.513. 9. Primers, primer amplification chains or sample compositions obtained by PCR amplification, replication, conversion and/or transformation based on the candidate marker according to claim 8. 10 . A methylation machine sequencing system, comprising the nucleic acid library of candidate markers according to claim 8 .