WO2024192928A1 - Gene combination for liver cancer detection, and related reagent and application - Google Patents

Abstract

A gene combination for liver cancer detection, and a related reagent and application, relating to the fields of biotechnology and disease detection. Provided are OSR2, TSPYL5 and SDF4 genes that can be used for liver cancer detection. By measuring the methylation content in the gene combination, whether a subject suffers from liver cancer and/or the risk of the subject suffering from liver cancer can be determined. The present invention has the advantages of high detection sensitivity and good specificity.

Description

A gene combination for liver cancer detection and related reagents and applications

CROSS-REFERENCE TO RELATED APPLICATIONS

The present disclosure claims the priority of the Chinese patent application with application number CN2023102952484 filed with the Chinese Patent Office on March 23, 2023 and entitled “A gene combination and related reagents and applications for liver cancer detection”, the entire contents of which are incorporated by reference in the present disclosure.

Technical Field

The present disclosure relates to the field of biotechnology and disease detection, and in particular, to a gene combination and related reagents and applications for liver cancer detection.

Background Art

Liver cancer mainly includes hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC), of which 70%-90% are hepatocellular carcinoma. Liver cancer is one of the most common malignant tumors in the world, and its incidence is increasing year by year. It is estimated that by 2025, the incidence rate will exceed one million cases.

High-risk groups for HCC include those with HBV infection, HCV infection, alcoholic liver disease (ALD) caused by excessive drinking, nonalcoholic fatty liver disease (NAFLD) caused by diabetes or obesity-related metabolic causes, and cirrhosis caused by other causes, as well as people with a family history of liver cancer, especially men over 40 years old. HBV infection is the main risk factor for HCC, accounting for about 85% of the number of liver cancer patients. Since 2002, the proportion of HCC caused by HBV and HCV infection has been declining due to the popularization of HBV vaccination for newborns and the improvement of HCV cure rate. However, due to changes in living standards and dietary structure, HCC cases caused by alcoholic liver disease and nonalcoholic fatty liver disease have been on the rise, accounting for 15% of the general population, becoming the second largest risk factor after HBV infection. More than 90% of HCC patients have chronic liver disease, and cirrhosis caused by any cause is an extremely high risk factor for HCC.

The prognosis of HCC depends on the tumor stage. The 5-year survival rate of early HCC is over 70%, while only 70% of patients with mid-stage HCC can survive 1 to 1.5 years after treatment. Regular screening of high-risk groups such as those with cirrhosis and chronic HBV infection can reduce the mortality rate by 37%. At present, patients with early liver cancer only account for 10% to 20% of the number of liver cancer patients. Most HCC patients are already in the middle and late stages when liver cancer is detected, and the survival rate is greatly reduced. Therefore, early detection of liver cancer and timely surgical treatment can enable patients to obtain the best survival rate.

Screening of high-risk groups for HBV or HCV infection, or cirrhosis caused by any reason, can improve the detection rate of early liver cancer and the cure rate of liver cancer. Abdominal ultrasound once every six months or combined with serum alpha-fetoprotein (AFP) is the recommended screening method for HCC. Ultrasound screening has the advantages of simplicity, low cost, non-invasiveness, real-time observation, and no radiation. However, increasing data emphasize that abdominal ultrasound depends on the level of the operator and has poor performance in screening patients with obesity and non-alcoholic fatty liver disease. A meta-analysis found that the overall sensitivity of abdominal ultrasound for detecting early HCC was only 45%. Although the sensitivity of combining AFP to detect early HCC can be increased to 63%, the specificity is reduced. In addition, most small liver cancers with a diameter of less than 1 cm are not easily detected by abdominal ultrasound and are very difficult to diagnose. Therefore, regular follow-up should be conducted for lesions with a diameter of less than 1 cm, such as once every 3 months.

Serum AFP is currently a common and important indicator for diagnosing liver cancer and monitoring therapeutic effects. Serum AFP ≥ 400ng/ml is highly suggestive of liver cancer after excluding pregnancy, chronic or active liver disease, gonadal embryonal tumors, and digestive tract tumors. When the AFP threshold is 20ng/mL, the sensitivity and specificity of AFP for detecting HCC are 41% to 65% and 80% to 90%, respectively; however, nearly 50% of HCC patients have AFP levels below 20ng/mL, so AFP cannot be used alone for HCC screening. Total AFP is not specific for liver cancer, and 10% to 42% of elevated AFP levels also occur in other liver diseases besides HCC, such as viral hepatitis.

In summary, since the sensitivity and specificity of AFP and ultrasound screening are not high, and abdominal ultrasound depends on the operator's level, most liver cancers are discovered in the late stage clinically, which greatly reduces the patient's five-year survival rate. Therefore, it is urgent to develop a product with high sensitivity and specificity for early screening and early diagnosis of people at high risk of liver cancer. The ideal detection product should be highly repeatable, independent of the operator's experience (such as abdominal ultrasound), highly accurate, and easy to implement in different clinical scenarios, and liquid biopsy meets these standards.

Summary of the invention

The purpose of the present disclosure is to provide a gene combination and related reagents and applications for liver cancer detection.

The present disclosure is implemented as follows:

The present disclosure provides the use of a reagent or a reagent combination in preparing a product for any one or more of the following uses, including (1) early screening or early diagnosis of liver cancer, (2) diagnosis or auxiliary diagnosis of liver cancer, (3) assessment or auxiliary assessment of the prognostic risk of liver cancer and (4) dynamic monitoring of liver cancer; the reagent or reagent combination comprises: a first reagent for detecting the methylation content of the OSR2 gene and a second reagent for detecting the methylation content of the TSPYL5 gene.

The present disclosure provides a reagent or a reagent combination, which includes the reagent or the reagent combination described in the aforementioned embodiments.

The present disclosure provides a kit, which includes the reagent or reagent combination described in any of the aforementioned embodiments.

The present disclosure provides a method for detecting the methylation content of a target gene, which comprises: using the reagent or reagent combination described in the aforementioned embodiment to detect the methylation content of the target gene in a sample; the target gene comprises OSR2 and TSPYL5 genes.

The present disclosure provides a method for training a liver cancer prediction model, which comprises: obtaining the reagent or reagent combination described in the aforementioned embodiment to detect the methylation content of the target gene in the training sample and annotating the result; wherein the target gene is the target gene as described in the aforementioned embodiment, and the annotated result comprises: a label representing any one of the liver cancer risk, prognosis risk and disease progression of the sample; the methylation content of the target gene of the sample to be tested is input into a pre-constructed prediction model to obtain a prediction result; the prediction model is a machine learning model that can judge any one of the liver cancer risk, prognosis risk and disease progression of the sample according to the methylation content of the target gene.

In a sixth aspect, an embodiment of the present disclosure provides an electronic device comprising: a processor and a memory; the memory is used to store a program, and when the program is executed by the processor, the processor implements the training method of the liver cancer prediction model or the liver cancer prediction method described in the aforementioned embodiment; the prediction method comprises: obtaining the methylation content of the target gene in the sample to be tested; the target gene is the target gene described in the aforementioned embodiment; the methylation content of the target gene of the sample to be tested is input into the model trained by the training method described in the aforementioned embodiment to obtain the prediction result of the sample to be tested.

The present disclosure has the following beneficial effects:

In response to the problem of low sensitivity and specificity of current ultrasound and AFP tests for early liver cancer detection, the present disclosure provides a gene combination of OSR2 and TSPYL5 genes that can be used for liver cancer detection. By detecting the methylation content of this gene combination, it can be determined whether the subject has liver cancer, the risk of liver cancer and/or the disease progression of liver cancer. It has the advantages of high detection sensitivity and good specificity.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present disclosure and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

Figure 1 is a comparison of the principles of MethyLight and HeavyMethyl detection;

Figure 2 shows the ROC curve of 21 candidate genes for distinguishing liver cancer from liver cirrhosis;

Figure 3 shows the ROC curves of OSR2 and TSPYL5 for distinguishing liver cancer from all negative samples;

Figure 4 shows the ROC curves of OSR2 and TSPYL5 for distinguishing liver cancer from cirrhosis samples;

Figure 5 shows the ROC curves of OSR2 and TSPYL5 for distinguishing liver cancer from healthy samples;

Figure 6 shows the ROC curves of OSR2 and TSPYL5 for distinguishing liver cancer from liver cirrhosis;

Figure 7 shows the ROC curves of OSR2, TSPYL5 and AFP for distinguishing liver cancer from cirrhosis.

DETAILED DESCRIPTION

In order to make the purpose, technical scheme and advantages of the embodiments of the present disclosure clearer, the technical scheme in the embodiments of the present disclosure will be described clearly and completely below. If the specific conditions are not specified in the embodiments, they are carried out according to conventional conditions or conditions recommended by the manufacturer. If the manufacturer of the reagents or instruments used is not specified, they are all conventional products that can be purchased commercially.

Methylation is a common epigenetic modification of human genomic DNA. It refers to the transfer of a methyl group (-CH ₃ ) to the 5th carbon atom of the DNA cytosine base C under the action of DNA methyltransferase. DNA methylation modification is involved in a variety of cell biological processes such as cell differentiation, genome stability, X chromosome inactivation and gene imprinting. Common methyltransferases in mammals are DNMT3a, DNMT3b and DNMT1. DNMT3a and DNMT3b realize de novo methylation of DNA, while DNMT1 is responsible for replicating the methylation pattern of the DNA template sequence.

DNA methylation mainly occurs on the base C of CpG dinucleotides. About 70% of human genes have multiple densely packed CpG dinucleotides in the promoter region upstream, forming a CpG island. Methylation of the CpG island leads to silencing of gene expression. In addition to the methylation of the CpG island in the promoter region of the gene, the first exon of the gene is similar to the promoter region and is also methylated, leading to silencing of gene expression. Gene DNA methylation also plays an important role in the occurrence and development of tumors. High methylation in the promoter region of tumor suppressor genes inhibits gene expression, causing the loss of tumor suppressor gene function, leading to uncontrolled proliferation and invasion and metastasis of tumor cells, and ultimately leading to the occurrence and development of cancer.

Studies have shown that methylation modification occurs in the early stages of tumors, earlier than gene mutations. Methylation of CpG dinucleotides often appears in clusters, that is, in a linked state, which is suitable for early screening and diagnosis of tumors. In addition, since CpG islands are related to the expression of controlled genes, the methylation of CpG islands can present a tissue-specific pattern. Therefore, DNA methylation can be used to trace the tissue origin of cancer.

At present, the detection methods of DNA methylation can be divided into chemical conversion method, enzyme conversion method, enzyme digestion method and methylated DNA immunoprecipitation method. The chemical conversion method is to convert methylated cytosine (5-mC) into uracil U by bisulfite, and then perform PCR detection or sequencing. Such methods include WGBS (Whole-Genome Bisulfite Sequencing), BSAS (Bisulfite Amplicon Sequencing), MSP (Methylation Specific PCR), BSP (Bisulfite Sequencing PCR), QMSP (Quantitative Methylation Specific PCR). The enzymatic conversion method is to convert 5-mC into 5-caC by TET enzyme, and then convert 5-caC into uracil U by APOBEC enzyme or convert 5-caC into DHU by pyridine borane, such as NEB's NEBNext Enzymatic Methyl-seq Kit and Base Genomics' TAPS method. Enzyme digestion is to cut methylated or non-methylated sites by methylation-sensitive or insensitive endonucleases, and selectively analyze DNA templates by PCR or sequencing, such as simplified genomic bisulfite sequencing RRBS. Methylated DNA Immunoprecipitation (MeDIP) is to separate methylated DNA fragments by anti-5-mC antibodies, and then perform relevant detection and analysis. Enzyme conversion method has little damage to DNA, high recovery rate of converted DNA, and relatively complete fragments, but the conversion efficiency is unstable. Enzyme digestion method has little damage to DNA and high enzyme digestion efficiency, but it needs to find specific endonucleases for one or a class of DNA sequences, which is not universal. The capture efficiency of methylated DNA immunoprecipitation method is greatly affected by the system. Although bisulfite conversion has great damage to DNA, the converted DNA is fragmented, and the conversion recovery rate is low, it has high conversion efficiency and stable conversion rate. It can identify single-base methylation and is the gold standard for DNA methylation detection. It is widely used in scientific research and clinical practice.

Fluorescence quantitative PCR based on bisulfite conversion can be divided into MethyLight and HeavyMethyl according to the principle of detecting methylation (see Figure 1). The primer design of HeavyMethyl does not distinguish between methylated and non-methylated sequences, but blocks the primer extension of the non-methylated template by designing one or two methylation-specific oligonucleotide blockers, thereby selectively amplifying the methylated template, and at the same time designs a methylation-specific probe, and realizes the selective amplification and detection of the methylated template through the oligonucleotide blocker and the methylation-specific probe. After careful design of blockers and probes and optimization of the detection system, HeavyMethyl can detect methylated DNA as low as 30-60pg, while non-methylated DNA is not amplified.

MethyLight designs methylation-specific primers and probes for DNA sequences after bisulfite conversion, so that only methylated templates are amplified and detected, while non-methylated templates are not amplified and detected. MethyLight is considered to be highly sensitive and is widely used for DNA methylation detection. For the methylation sequences of most genes, primers and probes with high sensitivity and specificity can be designed. However, for the methylation sequences of individual genes, it is difficult to design suitable primers and probes due to insufficient density of CpG dinucleotides or uneven distribution of methylated C and non-methylated C, resulting in insufficient sensitivity or specificity of the MethyLight detection system, and false positive or false negative results. Therefore, when using the MethyLight method to detect the methylation content of genes, it is necessary to fully analyze the gene sequence, select sequences suitable for designing primer probes, and carefully design, screen and fully verify feasible primer probe combinations based on the sequence characteristics suitable for designing primer probes and combined with the design elements of MethyLight primer probes.

Although fluorescent quantitative PCR based on bisulfite conversion has the above advantages, it currently has only 5 fluorescent channels at most, thus limiting the number of gene methylation targets that can be detected. Therefore, for the detection of specific cancer types, the first step is to screen methylation markers (target genes), screen out methylation markers with high sensitivity and specificity, and then perform fluorescent quantitative PCR detection. It can be seen that it is particularly critical to screen out methylation markers with good performance.

The present disclosure provides a gene combination for liver cancer detection, and a set of reagent combinations and kits for liver cancer detection based on the gene combination. By detecting the methylation content of the target gene combination, it can be determined whether the subject has liver cancer or the risk of liver cancer.

Specific technical solutions

In one aspect, the present disclosure provides the use of a reagent or a reagent combination in preparing a product for any one or more of the following uses, including (1) early screening or early diagnosis of liver cancer, (2) diagnosis or auxiliary diagnosis of liver cancer, (3) assessment or auxiliary assessment of prognostic risk of liver cancer, and (4) dynamic monitoring of liver cancer; the reagent or reagent combination comprises: a first reagent for detecting the methylation content of the OSR2 gene and a second reagent for detecting the methylation content of the TSPYL5 gene.

OSR2 is a protein-coding gene that encodes odd-skipped-related transcription factor 2.

In some embodiments, the OSR2 gene has GeneID: 116039.

In some embodiments, the detection region of the OSR2 gene is selected from: the full-length or partial sequence of the OSR2 gene (exon region and intron region) and the 5' promoter region.

In some embodiments, the detection region of the OSR2 gene includes: region 1 whose nucleotide sequence is shown in SEQ ID NO.1. Region 1 is an optional detection region, and the C in the CpG dinucleotide in its sequence is highly methylated in liver cancer, while the methylation level is low in cirrhosis, benign liver lesions and healthy people, and there is a significant difference between the two. The chromosome coordinates are chr8:99,955,670-99,961,095 (genome version number hg19), and the DNA sequence is as follows (Note: C in bold CG or all CG is methylated C):

In some embodiments, the first reagent includes: any one or more of a primer pair, a probe, and a chip.

In some embodiments, the first reagent includes: a primer pair with a nucleotide sequence as shown in SEQ ID NO. 4 to 5 and/or a probe with a nucleotide sequence as shown in SEQ ID NO. 6.

In some embodiments, TSPYL5 is a protein encoding gene predicted to activate chromatin binding activity and histone binding activity, and is involved in multiple physiological processes including cellular response to gamma radiation, positive regulation of protein kinase B signaling, and positive regulation of protein ubiquitination.

In some embodiments, the TSPYL5 gene has GeneID: 85453.

In some embodiments, the detection region of the TSPYL5 gene is selected from: the full-length or partial sequence of the TSPYL5 gene (exon region and intron region) and the 5' promoter region.

In some embodiments, the detection region of the TSPYL5 gene includes: region 2 whose nucleotide sequence is shown as SEQ ID NO.2. Region 2 is an optional detection region, and the C in the CpG dinucleotide in its sequence is highly methylated in liver cancer, while the methylation level is low in cirrhosis, benign liver lesions and healthy people, and there is a significant difference between the two. The chromosome coordinates are chr8:98,289,281-98,290,380 (genome version number hg19), and the DNA sequence is as follows (Note: C in bold CG or all CG is methylated C):

In some embodiments, the second reagent includes any one or more of a primer pair, a probe, and a chip.

In some embodiments, the second reagent includes: a primer pair with a nucleotide sequence as shown in SEQ ID NO.16-17 and/or a probe with a nucleotide sequence as shown in SEQ ID NO.18.

In some embodiments, the reagent or reagent combination also includes: a reagent or reagent combination for detecting the methylation content of any one or more genes in C1QL4 gene, CR1L gene, CYP26C1 gene, FOXG1 gene, GHSR gene, HIST1H1D gene, IRX5 gene, KCNG3 gene, LHX2 gene, MEX3A gene, NEFM gene, OTX1 gene, OXTR gene, PCDHGB6 gene, PCDHGB7 gene, PITX1 gene, PRLHR gene, PRRX1 gene, and ZIC4 gene. The detection region of the methylation content of the 19 genes is selected from the full length or partial region of the corresponding gene and its promoter region. When the detection gene is limited, those skilled in the art can obtain the corresponding detection reagent by combining conventional technical means. The 19 genes are used as a gene combination with the OSR2 gene and the TSPYL5 gene for liver cancer detection, which has good detection sensitivity and specificity, and the AUC can reach 0.92.

In some embodiments, the reagent or reagent combination further includes: a third reagent for detecting the content of an internal reference gene.

In some embodiments, the internal reference gene includes the SDF4 gene. SDF4 is a protein-coding gene that encodes a stromal cell-derived factor, which is a member of the CREC (reticulum calcium-binding protein) protein family, and the encoded protein contains six EF-hand motifs and a calcium-binding motif. This protein is located in the Golgi cavity and may be involved in regulating calcium-dependent cell activities.

In some embodiments, the SDF4 gene has GeneID: 51150.

In some embodiments, the detection region of the SDF4 gene includes: the full-length or partial sequence of the SDF4 gene.

In some embodiments, the detection region of the SDF4 gene includes: region 3 whose nucleotide sequence is shown as SEQ ID NO.3. Region 3 is an optional detection region, and the GC content of the segment sequence is 64.7%, the CpG dinucleotide density is 5.88%, and the GC content after bisulfite conversion is 32.5%. The chromosome coordinates are chr1: 1,159,192-1,159,483 (genome version number hg19), and the DNA sequence is as follows (Note: C in bold CG or all CG is methylated C):

In some embodiments, the third reagent includes: any one or more of a primer pair, a probe, and a chip.

In some embodiments, the third reagent includes: a primer pair with a nucleotide sequence as shown in SEQ ID NO. 28 to 29 and/or a probe with a nucleotide sequence as shown in SEQ ID NO. 30.

In the disclosed embodiment, OSR2 and TSPYL5 are tumor markers highly correlated with liver cancer, and the detection target is their methylated DNA sequence, and SDF4 is an internal reference gene used to control the quality of DNA samples and as a reference for the methylation content of the target gene. Specifically, by detecting the methylation content of the target genes OSR2 and TSPYL5 in the sample, the difference (△Ct) of the Ct value amplified by OSR2 and TSPYL5 minus the Ct value amplified by the internal reference gene SDF4 is used as the relative methylation content of the target gene, and after modeling with the △Ct value, it is determined whether the subject has liver cancer or the risk of liver cancer.

In some embodiments, the product includes any one or more of a reagent and a kit.

On the other hand, embodiments of the present disclosure provide a reagent or a reagent combination, which comprises the reagent or the reagent combination described in any of the preceding embodiments.

On the other hand, an embodiment of the present disclosure provides a kit, which includes the reagent or reagent combination described in any of the preceding embodiments.

The reagent or reagent combination and the kit have any one or more of the following uses: (1) early screening or early diagnosis of liver cancer; (2) diagnosis or auxiliary diagnosis of liver cancer, (3) preliminary assessment or auxiliary assessment of the prognostic risk of liver cancer and (4) dynamic monitoring of liver cancer.

On the other hand, an embodiment of the present disclosure provides a method for detecting the methylation content of a target gene, comprising: using the reagent or reagent combination described in any of the preceding embodiments to detect the methylation content of the target gene in a sample; the target genes include the OSR2 gene and the TSPYL5 gene.

In some embodiments, the use is not directly aimed at diagnosis or treatment of disease.

There are many situations where the direct purpose is not to diagnose or treat a disease. For example, when the sample to be tested is an environmental sample containing a biological sample or an artificially produced standard, the direct purpose of the test is not to diagnose or treat a disease.

In some embodiments, the method further comprises detecting an internal reference gene.

In some embodiments, the target gene also includes: for detecting any one or more of C1QL4 gene, CR1L gene, CYP26C1 gene, FOXG1 gene, GHSR gene, HIST1H1D gene, IRX5 gene, KCNG3 gene, LHX2 gene, MEX3A gene, NEFM gene, OTX1 gene, OXTR gene, PCDHGB6 gene, PCDHGB7 gene, PITX1 gene, PRLHR gene, PRRX1 gene, and ZIC4 gene.

In some embodiments, the internal reference gene includes: SDF4 gene.

In some embodiments, the sample of the detection method includes: a DNA sample or an environmental sample containing DNA. The type of the DNA sample can be any one of plasma cfDNA, serum cfDNA, tissue DNA, cell DNA and exosome DNA.

In some embodiments, the detection of the methylation content of the target gene comprises: using the reagent or reagent combination to perform qPCR amplification on the sample after bisulfite conversion;

In some embodiments, when the reagent or reagent combination comprises the first reagent to the third reagent, the qPCR reaction system comprises: the final concentration of _MgCl2 is 4-8mM, the final concentration of dNTP Mix is 200-800μM, the final concentration of Taq enzyme is 1.0-10U/reaction, the final concentration of the primer pair for OSR2 gene detection is 0.05-1.0μM, and the final concentration of the probe is 50-500nM; the final concentration of the primer pair for TSPYL5 gene detection is 0.1-1.0μM, and the final concentration of the probe is 50-500nM; the final concentration of the primer pair for SDF4 gene detection is 0.1-1.0μM, and the final concentration of the probe is 50-500nM.

In some embodiments, the qPCR reaction system includes: a final concentration of _MgCl2 of 6 mM, a final concentration of dNTP Mix of 500 μM, a final concentration of Taq enzyme of 3 U/reaction, a final concentration of the primer pair for OSR2 gene detection of 0.1 μM, a final concentration of the probe of 200 nM, a final concentration of the primer pair for TSPYL5 gene detection of 0.2 μM, a final concentration of the probe of 200 nM, a final concentration of the primer pair for SDF4 gene detection of 0.2 μM, a final concentration of the probe of 100 nM, and a total reaction volume of 25 μL.

In some embodiments, the qPCR reaction procedure includes the following:

Pre-denaturation at 95-98°C for 0.5-5 min; denaturation at 95-98°C for 5-15 sec, annealing and extension at 55-65°C for 30-60 sec, 15 cycles without collecting fluorescence; denaturation at 95-98°C for 5-15 sec, annealing and extension at 55-65°C for 30-60 sec, 30 cycles, collecting fluorescence.

In some embodiments, the qPCR reaction procedure includes the following: pre-denaturation at 95°C for 5 minutes; denaturation at 95°C for 15 seconds, annealing and extension at 62°C for 30 seconds, 15 cycles, without collecting fluorescence; denaturation at 95°C for 15 seconds, annealing and extension at 62°C for 32 seconds, 30 cycles, collecting fluorescence.

On the other hand, the present disclosure also provides a method for training a liver cancer prediction model, which includes:

Obtaining the reagent or reagent combination described in any of the foregoing embodiments to detect the methylation content of the target gene in the training sample and annotate the result; wherein the target gene is the target gene described in any of the foregoing embodiments, and the annotated result includes: a label representing any one of the liver cancer risk, prognosis risk and disease progression of the sample;

The methylation content of the target gene of the sample to be tested is input into a pre-constructed prediction model to obtain a prediction result; the prediction model is a machine learning model that can judge any of the liver cancer risk, prognosis risk and disease progression of the sample based on the methylation content of the target gene.

In some embodiments, the sample size of the training samples is greater than or equal to 10. The sample size can specifically be any one of 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900 and 1000, or a range between any two of them.

In some embodiments, the label may be a character, where a character refers to a glyph-like unit or symbol, including letters, numbers, operation symbols, punctuation marks and other symbols, as well as some functional symbols.

In some embodiments, the machine learning model can be selected from any one or a combination of support vector machine, decision tree, random forest, logistic regression, Bayesian, K nearest neighbor, K means, Markov and regression ridge algorithms.

In addition, an embodiment of the present disclosure further provides an electronic device, comprising: a processor and a memory; the memory is used to store a program, and when the program is executed by the processor, the processor implements the training method of the liver cancer prediction model or the liver cancer prediction method described in any of the above embodiments;

The prediction method includes: obtaining the methylation content of the target gene in the sample to be tested; the target gene is the target gene described in any of the aforementioned embodiments; the methylation content of the target gene of the sample to be tested is input into the model trained by the training method described in any of the aforementioned embodiments to obtain the prediction result of the sample to be tested.

The electronic device may include a memory, a processor, a bus, and a communication interface, wherein the memory, the processor, and the communication interface are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components may be electrically connected to each other via one or more buses or signal lines. The processor may process information and/or data related to target identification to perform one or more functions described in the present application.

The memory can be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable read-only memory (EEPROM), etc.

The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

In actual applications, the electronic device can be a server, a cloud platform, a mobile phone, a tablet computer, a laptop computer, an ultra-mobile personal computer (UMPC), a handheld computer, a netbook, a personal digital assistant (PDA), a wearable electronic device, a virtual reality device, etc. Therefore, the embodiments of the present application do not limit the type of electronic device.

The features and performance of the present invention are further described in detail below in conjunction with the embodiments.

Example 1 Screening of target genes

1. Methylation capture sequencing

(1) Sample

DNA was collected from 43 liver cancer tissues and 32 paired adjacent paracancerous tissues, and plasma was collected from 55 liver cancer tissues, 55 liver cirrhosis tissues, and 50 healthy subjects, with 2 mL of plasma from each tissue.

(2) Plasma cfDNA extraction

a. Thaw the plasma at room temperature or 37°C and centrifuge at 16,000g and 4°C for 10 minutes;

b. Add 80 μL of Proteinase K (Apostle MiniMax Free DNA Isolation Kit, Catalog No. A17622CN-384) to a 15 mL centrifuge tube, then add the centrifuged plasma. When transferring the plasma, do not aspirate the centrifugal precipitate or flocs, and then vortex to mix.

c. Add 200 μL of Sample Lysis Buffer to the above 15 mL centrifuge tube, vortex to mix, then place in a 60°C water bath for 20 minutes, inverting several times every 5 minutes to mix;

d. After the water bath is over, take out the 15 mL centrifuge tube, let it stand at room temperature for 10 minutes, and then centrifuge briefly to collect the plasma lysate on the tube wall to the bottom of the tube;

e. Take 6 deep-well plates and prepare for automated nucleic acid extraction. Plate 1: add 2.5 mL of cfDNA Lysis/Binding Solution, 30 μL of Magnetic Nanoparticles and all the above plasma lysate, Plate 2: add 1 mL of cfDNA Wash Solution, Plate 3: add 1 mL of cfDNA Wash Sloution, Plate 4: add 2 mL of diluted 2nd Wash Solution, Plate 5: add 500 μL of diluted 2nd Wash Solution, Plate 6: add 50 μL of cfDNA Elution Solution;

f. Place a new Deep-Well Tip Comb plate in position 7 and start the instrument self-check. After the self-check is qualified, select the nucleic acid automatic extraction program and start the automatic extraction of cfDNA. Place plates 1 to 6 in order.

g. After the instrument extraction is completed, remove plate 6 and absorb cfDNA.

(3) Methylation capture library preparation and sequencing

The cfDNA concentration was measured using Qubit dsDNA HS Assay Kit (Thermofisher Scientific, Cat. No. Q32854).

Take 40 μL of tissue DNA or cfDNA and add it to 160 μL of bisulfite conversion reagent (Zymo Research, EZ DNA Methylation Lightning Kits) for conversion, denature at 98°C for 8 minutes, convert at 54°C for 1 hour, and maintain at 4°C. The converted product is purified. The purified DNA is quantified using Qubit ssDNA Assay Kit (Thermo Fisher, Cat. No.: Q10212).

The converted DNA was used to construct the library using Hieff NGS Methyl-seq DNA Library Prep Kit for Illumina (YEASEN, catalog number: 12211ES24). The operating steps are shown in the kit manual.

The library was concentrated using Qubit dsDNA HS Assay Kit and the library fragments were quality checked using 4200 bioanalyzer (Agilent DNA 1000 Kit, Cat. No.: 5067-1504). The library that passed the quality check was captured by liquid phase hybridization. The captured library was concentrated using Qubit dsDNA HS Assay Kit and the library fragments were quality checked using 4200 bioanalyzer. After the captured library passed the quality check, it was sequenced on the Illumina sequencer NovaSeq with a read length of PE150.

2. Screening of target genes

Analyze the captured data after leaving the machine and screen the target genes according to the following conditions:

(1) Use fastp software to filter the data from the machine, including filtering the sequencing adapter sequences, removing DNA fragments with sequencing read lengths less than 50 bp, and removing DNA fragments with low average sequencing quality;

(2) Use Bismark to compare the filtered data with the hg19 reference genome (carrying the decoy bait sequence) to obtain the specific genomic location information corresponding to each DNA fragment and the methylation status information of each CpG site;

(3) Use Picard software to remove redundant data introduced during PCR amplification from the top and bottom chains, respectively, and then merge the data;

(4) Use bamUtil software to soft-clip the overlapping regions of shorter inserts to prevent duplicate statistics when detecting methylation levels;

(5) Use bamtools software to remove DNA fragments with low alignment quality, unaligned, or incompletely paired double-end reads;

(6) Screening for genes and chromosome segments that are highly methylated and linked in HCC tissue DNA and plasma cfDNA, but lowly methylated in paracancerous tissue DNA and plasma cfDNA of cirrhotic and healthy subjects.

3. Filter results

The sequencing data of tissue DNA and plasma cfDNA were analyzed, and 21 candidate genes were finally screened out. The difference levels of the median methylation percentages of the 21 candidate genes in liver cancer and cirrhosis, liver cancer and cirrhosis, and healthy people are shown in Table 1. The performance of the 21 candidate genes in distinguishing liver cancer from cirrhosis is shown in Figure 2. As can be seen from Figure 2, after integrating the 21 genes, the AUC for distinguishing liver cancer from cirrhosis reached 0.92, among which HIST1H1D, IRX5, NEFM, OSR2, OTX1, PITX1, and TSPYL5 could reach an AUC of more than 0.71 for distinguishing liver cancer from cirrhosis.

Table 1 Differences in median methylation percentages of candidate genes in liver cancer, liver cirrhosis and healthy subjects

Example 2 Screening of primer probes for OSR2, TSPYL5 and SDF4 amplification

Based on the difference in methylation percentage of candidate genes in liver cancer, liver cirrhosis, and healthy people and the AUC area of the ROC curve, combined with the characteristics of the gene sequence, two target genes, OSR2 and TSPYL5, were screened out and qPCR primer probes were designed. According to the optional sequences of OSR2 and TSPYL5 genes, the entire optional sequence was traversed to screen sequence fragments suitable for primer probe design, and several sets of primer probe combinations were carefully designed in combination with the design elements of MethyLight primer probes.

(1) Primer and probe design

Several sets of primer-probe combinations were designed in the optional DNA segments of OSR2 and TSPYL5, and the primer-probe combinations were screened.

The primer-probe combinations for OSR2 amplification are:

Primer probe combination 1:

SEQ ID NO.4: 5`-GCGTCGGTTTCGTTTTTGTACGTC-3`;

SEQ ID NO.5: 5`-CCACGCGCTACGCTACCG-3`;

SEQ ID NO.6:

5`6-FAM-ACCGCCTTTCCCGCGCTTACCCGA-3`BHQ1;

Primer probe combination 2:

SEQ ID NO.7: 5`-GTTTTTCGGAGGTAAGATTATTTGCG-3`;

SEQ ID NO.8: 5`-ACTCGATATATCCAAAACAAATTCGC-3`;

SEQ ID NO.9:

5`6-FAM-CAACCATCCTCCTTACCCGCCTCACCTATAATC-3`BHQ1;

Primer probe combination 3:

SEQ ID NO.10: 5`-TTTTGCGTCGTCGTTATTTTCG-3`;

SEQ ID NO.11：5`-AAAACCTATACGATCTCAACGCG-3`;

SEQ ID NO.12: 5`6-FAM-ATGGTGGAGCGGGTGATTTCGTGTAT-3`BHQ1;

Primer probe combination 4:

SEQ ID NO.13：5`-GGTTTTGCGTCGTCGTTATTTTC-3`;

SEQ ID NO.14: 5`-AAAACCTATACGATCTCAACGCGA-3`;

SEQ ID NO.15:5`6-FAM-ATGGTGGAGCGGGTGATTTCGTGTAT-3`BHQ1;

The primer-probe combinations for TSPYL5 amplification are:

Primer probe combination 5:

SEQ ID NO.16: 5`-CGCTCATAATAACGACGAAAACAACT-3`;

SEQ ID NO.17：5`-GGAGAGATTCGTTCGGTTTCGATC-3`;

SEQ ID NO.18:

5`Cy5-CACGCTATAACCCTACGACTCCTAACGCCA-3`BHQ-3;

Primer probe combination 6:

SEQ ID NO.19：5`-GCGGGAGGATTTTCGATTTCGA-3`;

SEQ ID NO.20: 5`-AAAACCGCAAAATCACAACGTATCTT-3`;

SEQ ID NO.21:

5`Cy5-AAACTACCTCCGCCGCCACCATAAACGAC-3`BHQ-3;

Primer probe combination 7:

SEQ ID NO.22: 5`-CCGCGATACTACAAATTTCTAAAACCTT-3`;

SEQ ID NO.23：5`-GTAGATATTGTTTTCGTGGGAATAGCG-3`;

SEQ ID NO.24:

5`Cy5-CCCGACAAAACCACGCCGATTTCCAACGC-3`BHQ-3;

Primer probe combination 8:

SEQ ID NO.25: 5`-AATACTACAATCGCAACTACCCAAA-3`;

SEQ ID NO.26：5`-GGAAGGTAGTATGGATACGTTGGA-3`;

SEQ ID NO.27:

5`Cy5-AAAACCGAAAATAAACCCTATCCGCCTAAACG-3`BHQ-3;

The primer-probe combinations for SDF4 amplification are:

Primer probe combination 9:

SEQ ID NO.28：5`-GTAGGTGGTCGTTAGATTGGG-3`;

SEQ ID NO.29：5`-TCGCTATTAATAACAAATTCACGCA-3`;

SEQ ID NO.30: 5`ROX-ACCGACGACATCCGTCTAACTATTCGA-3`BHQ2;

Primer probe combination 10:

SEQ ID NO.31：5`-TGTCGTTGTTGGGTGGTAGGTTTA-3`;

SEQ ID NO.32: 5`-CCCCAAAACCCCGAACAAC-3`;

SEQ ID NO.33：5`ROX-AACGAATACCGCCGACCCAAAACTCCA-3`BHQ2;

Primer probe combination 11:

SEQ ID NO.34：5`-GGTCGGATTAACGGGATGAAGTTTA-3`;

SEQ ID NO.35: 5`-CCCACACACGATTCGTACCTC-3`;

SEQ ID NO.36: 5`ROX-CCCGATCCCGCGTTCCCTACTATCCT-3`BHQ2.

It should be noted that the fluorescent group modification at the 5' end of the probe can be the fluorescent group used in various types of probes such as TaqMan probe, MGB probe, molecular beacon, BQ probe, such as FAM, TET, JOE, VIC, HEX, Cy3, NED, TAMRA, ROX, Cy5, AMCA, Pacific Blue, Atto 425, BODIPY FL, Oregon Green 488, R6G, Yakima Yellow, Quasar 570, AquaPhluor 593, Texas Red, Atto 590, Quasar 670, Cy5.5, Cy7 and IR Dye 750; the fluorescence quenching group modification at the 3' end of the probe can be the quenching group used in various types of probes such as TaqMan probe, MGB probe, molecular beacon, BQ probe, such as BHQ1, BHQ2, BHQ3, BBQ650, MGB, Dabcyl and DBQ1.

(2) Sample

Positive quality control DNA with a methylation content of 95% to 100% and negative quality control DNA with a methylation content of 0% (self-made).

(3) Bisulfite conversion

See Example 1 for reagents and operating procedures.

(4) Methylation-specific qPCR verification

30 ng of negative quality control product after conversion was used to verify the specificity of primers and probes, and three technical replicates were set; the converted positive quality control product DNA was gradient diluted with nuclease-free water, and 100 copies and 10 copies were taken for sensitivity verification, and 1 and 3 technical replicates were set respectively.

Prepare the qPCR reaction system according to Table 2-1.

Table 2-1 qPCR reaction system

Perform qPCR amplification according to the procedure in Table 2-2.

Table 2-2 qPCR reaction procedure

(5) 4200 Quality Inspection

The qPCR amplification product of the positive quality control with a starting amount of 100 copies was quality checked using a 4200 bioanalyzer and quality check reagents (Agilent DNA 1000 Kit, catalog number: 5067-1504).

(6) qPCR verification results

The validation results of the primer-probe combinations of OSR2, TSPYL5 and SDF4 are shown in Table 2-3. Select a primer-probe combination with high sensitivity, high specificity, and minimal primer dimer and nonspecific amplification, and determine the optimal primer-probe combination in combination with the fluorescence amplification curve. As shown in Table 2-3, the optimal primer-probe combination for OSR2 gene amplification is combination 1, the optimal primer-probe combination for TSPYL5 gene amplification is combination 5, and the optimal primer-probe combination for SDF4 gene amplification is combination 9.

Table 2-3 Primer probe screening results of target genes and internal reference genes

Note: When the test result is Undetermined, the value is set to 30. The Ct value is the test value + 15.

Example 3 Plasma sample validation of gene combination

(1) Sample

53 cases of hepatocellular carcinoma (HCC), of which 31 cases were plasma collected in 2018 and 22 cases were plasma collected in 2020; 52 cases of liver cirrhosis, of which 4 cases were plasma collected in 2019, 42 cases were plasma collected in 2020, and 6 cases were plasma collected in 2021; 50 healthy people (Health), all of which were plasma collected in 2021; benign liver lesions: 6 cases of focal nodular hyperplasia, 2 cases of hepatic adenoma, and 12 cases of hemangioma. Each sample was 2mL of plasma.

(2) Plasma cfDNA extraction

See Example 1 for reagents and operating procedures.

(3) Bisulfite conversion

See Example 1 for reagents and operating procedures.

(4) qPCR

Primer probe combination 1 for OSR2 amplification;

Primer probe combination 5 for TSPYL5 amplification;

Primer probe combination 9 for SDF4 amplification.

Prepare the qPCR reaction system according to Table 3-1.

Table 3-1 qPCR reaction system

The qPCR reaction was performed as follows

Table 3-2 qPCR reaction procedure

(5) qPCR results

Table 3-3 qPCR results

As can be seen from Figure 3, taking cirrhosis, focal nodular hyperplasia, hepatic adenoma, and hepatic hemangioma as the negative population, the performance of OSR2 and TSPYL5 in distinguishing liver cancer from the negative population is equivalent, and the AUC area of the ROC curve is 0.927. After integrating OSR2 and TSPYL5, the AUC for distinguishing liver cancer from the negative population can reach 0.955.

When cirrhosis is used as a negative population, TSPYL5 is better than OSR2 in distinguishing liver cancer from cirrhosis, with AUCs of 0.933 and 0.906, respectively. After integrating TSPYL5 and OSR2, the AUC for distinguishing liver cancer from cirrhosis can reach 0.95. (Figure 4)

When healthy people were used as negative subjects, OSR2 performed better than TSPYL5 in distinguishing liver cancer from healthy people, with AUCs of 0.935 and 0.923, respectively. After integrating OSR2 and TSPYL5, the AUC for distinguishing liver cancer from healthy people could reach 0.961 (Figure 5).

It can be seen that both OSR2 and TSPYL5 have good performance in detecting liver cancer, and the two have good complementarity.

Example 4 Application of the kit in liver cancer detection

In order to further verify the performance of OSR2 and TSPYL5 gene combination in detecting liver cancer, the plasma volume was reduced to 1 mL and a second batch of plasma samples was used for verification, some of which were also tested for AFP.

(1) Sample

There were 54 cases of liver cancer and 61 cases of liver cirrhosis, and the plasma volume of each case was 1 mL.

(2) Plasma cfDNA extraction

See Example 1 for the plasma cfDNA extraction reagent and method.

(3) Bisulfite conversion

See Example 1 for bisulfite conversion reagents and methods.

(4) qPCR amplification

The primer probe for OSR2 amplification was combination 1;

The primer probe for TSPYL5 amplification was combination 5;

The primer probe for SDF4 amplification was combination 9;

The concentration of each component of the qPCR reaction system was optimized, and the optimal reaction system is shown in Table 4-1.

Prepare the qPCR reaction system according to Table 4-1.

Table 4-1qPCR reaction system

The temperature and time of the qPCR reaction were optimized, and the optimal reaction program is shown in Table 4-2.

Set up the qPCR reaction program according to Table 4-2.

Table 4-2 qPCR reaction procedure

(5) AFP detection

In order to compare the performance with AFP, some plasma samples were tested for AFP at the same time. AFP is a self-developed reagent, and 100 μL of plasma was taken for AFP testing.

(6) Test results

Table 4-3 qPCR results

Note: NA means AFP was not detected.

As shown in Figure 6, for 1 mL of plasma, the AUC of OSR2 for distinguishing liver cancer from cirrhosis reached 0.914, and the AUC of TSPYL5 for distinguishing liver cancer from cirrhosis reached 0.915. After the integration of the two genes, the AUC for distinguishing liver cancer from cirrhosis reached 0.951. Compared with 2 mL of plasma, the AUCs of OSR2 and TSPYL5 for distinguishing liver cancer from cirrhosis were 0.906 and 0.933, respectively. After integrating OSR2 and TSPYL5, the AUC for distinguishing liver cancer from cirrhosis can reach 0.95. It can be seen that although the AUC of OSR2 is higher than that of 2 mL of plasma at the starting amount of 1 mL of plasma, and the AUC of TSPYL5 is lower than that of 2 mL of plasma, after the integration of the two genes, the AUC of 1 mL of plasma starting amount is equivalent to that of 2 mL of plasma starting amount.

In order to compare the performance of gene methylation and AFP, liver cancer and cirrhosis samples with AFP data were selected for ROC curve analysis. As shown in Figure 7, the AUC of AFP for distinguishing liver cancer from cirrhosis is 0.741, and the AUC of OSR2 and TSPYL5 for distinguishing liver cancer from cirrhosis is above 0.9, which is significantly higher than AFP. The AUC of integrating OSR2 and TSPYL5 to distinguish liver cancer from cirrhosis is 0.96. It can be seen that the performance of this target gene combination in detecting liver cancer is significantly better than AFP.

In summary, the present disclosure provides a tumor marker for liver cancer detection - OSR2 and TSPYL5 genes, the methylation content of their DNA sequences can be used to distinguish liver cancer from cirrhosis, liver cancer from healthy people, liver cancer from benign liver lesions, and the AUC is all above 0.9; the present disclosure also provides a gene combination and a kit for liver cancer detection, namely, the OSR2, TSPYL5 and SDF4 gene combination, wherein OSR2 and TSPYL5 are target genes for detecting liver cancer, and SDF4 is an internal reference gene for determining the relative methylation content. By detecting the methylation content of OSR2 and TSPYL5, the AUC for distinguishing liver cancer from cirrhosis, liver cancer from healthy people, and liver cancer from benign liver lesions can reach above 0.95, which is significantly better than the performance of the existing tumor marker protein AFP used for liver cancer screening.

The above description is only an optional embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims (10)

Use of a reagent or a reagent combination in the preparation of a product for any one or more of the following uses, the uses including (1) early screening or early diagnosis of liver cancer, (2) diagnosis or auxiliary diagnosis of liver cancer, (3) assessment or auxiliary assessment of prognostic risk of liver cancer and (4) dynamic monitoring of liver cancer; the reagent or reagent combination comprises: a first reagent for detecting the methylation content of the OSR2 gene and a second reagent for detecting the methylation content of the TSPYL5 gene. The use according to claim 1, characterized in that the GeneID of the OSR2 gene is: 116039; Preferably, the detection region of the OSR2 gene is selected from: the full-length or partial sequence of the OSR2 gene and its 5' promoter region sequence; Preferably, the detection region of the OSR2 gene includes: region 1 having a nucleotide sequence as shown in SEQ ID NO.1; Preferably, the first reagent comprises: any one or more of a primer pair, a probe and a chip; Preferably, the first reagent comprises: a primer pair having a nucleotide sequence as shown in SEQ ID NO. 4 to 5 and/or a probe having a nucleotide sequence as shown in SEQ ID NO. 6. The use according to claim 1 or 2, characterized in that the GeneID of the TSPYL5 gene is: 85453; Preferably, the detection region of the TSPYL5 gene is selected from: the full-length or partial sequence of the TSPYL5 gene and its 5' promoter region; Preferably, the detection region of the TSPYL5 gene includes: region 2 whose nucleotide sequence is shown as SEQ ID NO.2; Preferably, the second reagent comprises: any one or more of a primer pair, a probe and a chip; Preferably, the second reagent comprises: a primer pair having a nucleotide sequence as shown in SEQ ID NO.16-17 and/or a probe having a nucleotide sequence as shown in SEQ ID NO.18. The use according to claim 3, characterized in that the reagent or reagent combination further comprises: a reagent or reagent combination for detecting the methylation content of any one or more genes among C1QL4 gene, CR1L gene, CYP26C1 gene, FOXG1 gene, GHSR gene, HIST1H1D gene, IRX5 gene, KCNG3 gene, LHX2 gene, MEX3A gene, NEFM gene, OTX1 gene, OXTR gene, PCDHGB6 gene, PCDHGB7 gene, PITX1 gene, PRLHR gene, PRRX1 gene, and ZIC4 gene; Preferably, the reagent or reagent combination further comprises: a third reagent for detecting the content of an internal reference gene; Preferably, the internal reference gene includes the SDF4 gene; Preferably, the GeneID of the SDF4 gene is: 51150; Preferably, the detection region of the SDF4 gene includes: the full-length or partial sequence of the SDF4 gene; Preferably, the detection region of the SDF4 gene includes: region 3 whose nucleotide sequence is shown as SEQ ID NO.3; Preferably, the third reagent includes: any one or more of a primer pair, a probe and a chip; Preferably, the third reagent comprises: a primer pair having a nucleotide sequence as shown in SEQ ID NO. 28 to 29 and/or a probe having a nucleotide sequence as shown in SEQ ID NO. 30. The use according to claim 1 or 2, characterized in that the reagent or reagent combination further comprises: at least one of MgCl ₂ , dNTP, Taq enzyme, and buffer Buffer. A reagent or a reagent combination, characterized in that it comprises the reagent or the reagent combination according to any one of claims 1 to 5. A kit, characterized in that it comprises the reagent or reagent combination according to any one of claims 1 to 5. A method for detecting the methylation content of a target gene, characterized in that it comprises: using the reagent or reagent combination described in any one of claims 1 to 5 to detect the methylation content of the target gene in a sample; the target gene comprises the OSR2 gene and the TSPYL5 gene; Preferably, the application is not directly intended for diagnosis or treatment of a disease; Preferably, the target gene also includes: for detecting any one or more of C1QL4 gene, CR1L gene, CYP26C1 gene, FOXG1 gene, GHSR gene, HIST1H1D gene, IRX5 gene, KCNG3 gene, LHX2 gene, MEX3A gene, NEFM gene, OTX1 gene, OXTR gene, PCDHGB6 gene, PCDHGB7 gene, PITX1 gene, PRLHR gene, PRRX1 gene, and ZIC4 gene; Preferably, the method further comprises detecting an internal reference gene; Preferably, the internal reference gene includes: SDF4 gene; Preferably, the samples of the detection method include: DNA samples or environmental samples containing DNA; Preferably, the detection of the methylation content of the target gene comprises: using the reagent or the reagent combination to perform qPCR amplification on the sample after bisulfite conversion treatment; Preferably, when the reagent or reagent combination comprises the first reagent to the third reagent, the qPCR reaction system comprises: the final concentration of _MgCl2 is 4-8mM, the final concentration of dNTP Mix is 200-800μM, the final concentration of Taq enzyme is 1.0-10U/reaction, the final concentration of the primer pair for OSR2 gene detection is 0.05-1.0μM, and the final concentration of the probe is 50-500nM; the final concentration of the primer pair for TSPYL5 gene detection is 0.1-1.0μM, and the final concentration of the probe is 50-500nM; the final concentration of the primer pair for SDF4 gene detection is 0.1-1.0μM, and the final concentration of the probe is 50-500nM; Preferably, the reaction procedure of the qPCR includes: pre-denaturation at 95-98°C for 0.5-5 minutes; denaturation at 95-98°C for 5-15 seconds, annealing and extension at 55-65°C for 30-60 seconds, 15 cycles, without collecting fluorescence; denaturation at 95-98°C for 5-15 seconds, annealing and extension at 55-65°C for 30-60 seconds, 30 cycles, collecting fluorescence. A method for training a liver cancer prediction model, characterized in that it comprises: Obtain the reagent or reagent combination described in any one of claims 1 to 5 to detect the methylation content of the target gene in the training sample and annotate the result; wherein the target gene is the target gene described in claim 8, and the annotation result includes: a label representing any one of the liver cancer risk, prognosis risk and disease progression of the sample; The methylation content of the target gene of the sample to be tested is input into a pre-constructed prediction model to obtain a prediction result; the prediction model is a machine learning model that can judge any of the liver cancer risk, prognosis risk and disease progression of the sample based on the methylation content of the target gene. An electronic device, characterized in that it comprises: a processor and a memory; the memory is used to store a program, and when the program is executed by the processor, the processor implements the training method of the liver cancer prediction model or the liver cancer prediction method according to claim 9; The prediction method comprises: obtaining the methylation content of the target gene in the sample to be tested; the target gene is the target gene described in claim 8; inputting the methylation content of the target gene of the sample to be tested into the model trained by the training method described in claim 9 to obtain the prediction result of the sample to be tested.