SNP (Single Nucleotide Polymorphism) marker for predicting risk of polycystic ovarian syndrome and application thereof

Abstract

The invention belongs to the field of molecular biology and reproductive medicine, and particularly relates to an SNP marker for risk prediction of polycystic ovarian syndrome and application of the SNP marker. According to the invention, through large-scale cross-racial GWAS research, 94 independent gene loci are accurately locked from massive samples in the global range, and 73 gene loci are newly found and comprise 250 genetic variation sites (SNP). A PRS model constructed and optimized based on the loci can more stably, economically and accurately predict PCOS risks in an earlier stage, accurately distinguish PCOS high-risk individuals from PCOS low-risk individuals, facilitate layered prevention and control and save medical resources, in addition, the loci can guide targeted drug development such as PPARG agonists, and a new thought is provided for accurate treatment of PCOS.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of molecular biology and reproductive medicine, and particularly relates to SNP markers for predicting the risk of polycystic ovary syndrome and application thereof. BACKGROUND

[0002] The information disclosed in the background of the present application is only intended to increase the understanding of the overall background of the present application and should not necessarily be regarded as acknowledging or implicitly suggesting that this information constitutes prior art known to those of ordinary skill in the art.

[0003] Polycystic ovary syndrome (PCOS) is a common gynecological endocrine disorder in women of reproductive age. It has complex clinical manifestations, including menstrual irregularity, hirsutism, obesity, and infertility. This disease not only poses a serious threat to the reproductive health of patients, but is also closely linked to long-term complications such as metabolic syndrome, cardiovascular disease, and cancer. Currently, the diagnosis of PCOS is mainly based on the comprehensive evaluation of clinical symptoms, biochemical indicators, and ultrasound examination. However, the existing diagnostic methods still have obvious limitations. In terms of clinical manifestations, the characteristic symptoms of PCOS such as menstrual irregularity and hirsutism lack specificity. Menstrual irregularity can be caused by various gynecological diseases, endocrine abnormalities, or environmental factors; while the assessment of hirsutism is affected by racial differences and subjective judgment, making it difficult to achieve accurate diagnosis solely relying on clinical symptoms. In terms of biochemical indicators, although elevated androgen levels and abnormal LH / FSH ratio are important diagnostic indicators, the hormone levels of some patients may only be at the critical value or fluctuate within the normal range, leading to diagnostic uncertainty. In terms of ultrasound examination, transvaginal ultrasound can detect polycystic-like changes in the ovary, but for patients in the early stages of the disease or with atypical clinical manifestations, their morphological changes in the ovary may not yet meet the diagnostic criteria, affecting the accuracy of diagnosis.

[0004] SNP is a single nucleotide variation on the genome, including transition, transversion, insertion, and deletion, and is the most common DNA polymorphism site in the human genome, widely existing in chromosomes and mitochondrial DNA. These small genetic differences not only affect the phenotypic diversity between individuals, but also are closely related to disease susceptibility, drug response, and other factors, and may become an important genetic basis for the occurrence, development, and prognosis of complex diseases.

[0005] Disease prediction based on key SNP markers has the advantages of high sensitivity, strong specificity and rapid detection, and through the construction of SNP prediction spectrum, it makes a prospective diagnosis of the disease. As the third generation of genetic markers, SNPs have been widely used in high-risk population screening, disease-related gene mapping and pharmacogenomics research. Therefore, screening the association between genotype and disease using single nucleotide polymorphism (SNP) and performing whole genome association analysis (GWAS) is expected to further deepen the understanding of the pathogenesis of PCOS. In addition, genetic variation is not only an important basis for disease risk prediction, but also a potential drug target. The genetic loci related to PCOS discovered by whole genome association study (GWAS) often point to key genes and pathways involved in the development of the disease, providing a clear direction for targeted drug development.

[0006] In the risk prediction of PCOS, the existing methods mainly rely on clinical feature analysis and limited genetic polymorphism detection. The prediction model based on clinical indicators has significant limitations, and its prediction performance is easily disturbed by environmental factors, resulting in insufficient stability and accuracy. The risk prediction based on disease genetic association sites is currently limited to a small number of sites (such as the previous GWAS reporting more than 20 sites), and is mostly based on a single population (such as European or Asian), which is difficult to fully reflect the complex genetic background of PCOS. For example, there are significant differences in gene frequency and disease genetic susceptibility between European and Asian populations, and the prediction performance of some prediction models developed for European populations will be significantly reduced when applied to Asian populations. SUMMARY

[0007] Therefore, the present application provides a SNP marker for predicting the risk of polycystic ovary syndrome and its application. The present application aims to provide a set of SNP markers for predicting the risk of PCOS with high sensitivity and cross-racial applicability, and to construct a multi-gene risk score (PRS) system to address the problems of low prediction accuracy and poor population applicability in the prior art. At the same time, based on the discovery of key genetic variation sites, new intervention targets closely related to the development of PCOS and their corresponding drugs are mined through pharmacogenomics research. Through the present application, stable, economical, precise and early prediction of the risk of polycystic ovary syndrome can be achieved, which helps early intervention and targeted treatment, and significantly improves the health level of patients.

[0008] In order to achieve the above-mentioned purpose, the present application is realized by the following technical solutions: The technical scheme of the present application is that 12419 cases of Chinese PCOS and 34235 cases of controls are subjected to whole genome association research, then up to 13773 cases of European cases and 411088 cases of controls are subjected to multi-ethnic meta-analysis, 94 PCOS-related independent loci are identified, 73 of which are newly found, and 250 SNPs are contained. Although there is different local selection pressure, there is high genetic overlap between Chinese and European populations, and a PRS scoring system is constructed based on these sites to achieve accurate prediction of PCOS risk, and combined with pharmacogenomic research, a series of new targets and corresponding drugs are found.

[0009] In a first aspect, the present application provides a SNP marker for predicting the risk of polycystic ovary syndrome, the marker being: rs10737472, rs10604385, rs4648892, rs2742963, rs12038623, rs61780052, rs6673300, rs35320516, rs10922273, rs12046808, rs4915135, rs1338285, rs144834966, rs13028479, rs12470971, rs13021112, rs10205592, rs199865706, rs1375194, rs13035011, rs7563201, rs6753358, rs6710610, rs7578655, rs1464727, rs3811516, rs7567846, rs138311200, rs7606567, rs4384834, rs10183851, rs1922476, rs4953651, rs12994034, rs6708637, rs1504187, rs1504165, rs6733807, rs140140896, rs55687579, rs75775350, rs72827480, rs1837294, rs1830321, rs957293, rs10206585, rs2857534, rs6712883, rs13411274, rs3771066, rs7574706, rs76172579, rs56009971, rs7564590, rs1531788, rs6735267, rs6710653, rs7599312, rs10191896, rs6785890, rs13061415, rs2067819, rs7615916, rs146753246, rs4456810, rs3924105, rs61262534, rs10516202, rs55904482, rs191505728, rs35232147, rs62306364, rs11372697, rs9312937, rs199669078, rs4705862, rs10072700, rs2070721, rs11242115, rs41525648, rs891988, rs2879271, rs9324872, rs4624820, rs34779343, rs9460408, rs9465511, rs11349096, rs6926381, rs6926229, rs7757706,rs9378247, rs4947328, rs486416, rs3807020, rs2185710, rs2502403, rs35869530, rs6912459, rs12527757, rs60793153, rs5886708, rs10228327, rs11777247, rs17807624, rs2740434, rs745379, rs17153782, rs2740433, rs2686208, rs804260, rs4840581, rs804256, rs17053711, rs11775358, rs4733407, rs10808334, rs6994022, rs12681434, rs12681435, rs12544187, rs4460442, rs2183958, rs1123534, rs7864171, rs7865239, rs10818820, rs12347448, rs77719717, rs200979118, rs117889661, rs75304159, rs79375299, rs3945628, rs4620350, rs2799465, rs10986104, rs5900588, rs7848238, rs4314746, rs7042632, rs6606575, rs57994832, rs11188578, rs200584089, rs2756100, rs1926029, rs76383974, rs11592408, rs16932777, rs10741694, rs10835634, rs11031005, rs11031034, rs11031040, rs556622, rs11040598, rs10082560, rs76735130, rs11229231, rs3740664, rs55965748, rs11234902, rs1426394, rs11225154, rs10501995, rs11225161, rs11436634, rs1784692, rs202185443, rs3217844, rs3217856, rs35481774, rs76624470, rs1131017, rs7971751, rs2272046, rs113923326, rs17115231, rs1275468, rs1148006, rs11180609, rs2255591, rs2855530, rs2761887, rs798791,rs2075773, rs77938952, rs2445768, rs2445773, rs1008815, rs10852606, rs12931034, rs7200621, rs8043701, rs9939609, rs7190396, rs250150, rs16950653, rs30403, rs30404, rs56738967, rs9928324, rs7203756, rs76862947, rs73575083, rs7358, rs78978329, rs1642799, rs72829444, rs78378222, rs2680703, rs2680698, rs2632513, rs78911886, rs12978702, rs11666480, rs11669774, rs10407022, rs4325676, rs11670032, rs1676465, rs7260162, rs2301595, rs8105605, rs11672660, rs10412726, rs734518, rs3817996, rs853854, rs6019857, rs6022786, rs6068688, rs171008, rs237048, rs1038023, rs9613564, rs62237617, rs17879961, rs56119508, rs544626416, rs112158446, rs7288513, rs5762906, rs34293193, rs756630, rs138219, rs2015414, rs73636638, rs5951636.

[0010] SNP marker related information is shown in Table 1: Table 1

[0011] In a second aspect, the present application provides a detection method of the SNP marker, which is selected from at least one of a gene chip method, a GSA method, a restriction fragment length polymorphism analysis, a single strand conformation polymorphism analysis, a PCR-ASO probe method, a PCR-SSO method, a PCR-SSP method, a PCR-fluorescence method, a PCR-DNA sequencing method, a direct sequencing method, a PCR fingerprint method, an AELP, a DGGE, a RAPD or a single base extension termination method.

[0012] Preferably, the detection method of the SNP marker is a gene chip method.

[0013] Preferably, the SNP marker itself and / or a product of the SNP marker, such as a metabolic product of the SNP marker and / or other products directly associated with the biomarker, can be detected.

[0014] In a third aspect, the present application provides an application of a substance for detecting the SNP marker of the first aspect in the preparation of a product for diagnosing, detecting, monitoring or predicting polycystic ovary syndrome.

[0015] Preferably, the product can exist in any of the existing known forms of primers, probes, nucleic acid membrane strips, (gene or protein) chips, preparations, kits, instruments, detection devices and equipment, etc. A person skilled in the art can realize the above products according to the actual situation without creative labor, and therefore the above products all belong to the protection scope of the present application.

[0016] In a fourth aspect, the present application provides a multi-gene risk score system, which comprises: (1) a detection unit, which comprises: a SNP marker detection for a to-be-tested sample of a subject; (2) an analysis unit, which comprises: an analysis of the SNP marker detected in (1); (3) an evaluation unit, which comprises: a judgment of the SNP marker analyzed in (2) to determine the risk of the subject.

[0017] The risk of the subject includes a diagnosis, detection, monitoring or prediction evaluation of polycystic ovary syndrome of the subject.

[0018] Preferably, the system is constructed based on the PRSice-2 (version 2.3.5) software platform, and the SNP site of the first aspect of the present application is incorporated into the system.

[0019] Preferably, the system is trained after being constructed; the training method is that three combination methods are set respectively in the training stage, namely using Chinese data alone, using European data alone, and training together after merging the two types of data; at the same time, we adopt the "one by one leave queue" cross-validation strategy: in each iteration, one of the Chinese queues is completely excluded and used as an independent test set, and the remaining three population queues are used for model training.

[0020] In a fifth aspect, the application provides a use of a modulator of PPARG activity in the preparation of a medicament for preventing or treating polycystic ovary syndrome.

[0021] Preferably, the medicament is: agonist pioglitazone, antagonist GW9662, metformin, telmisartan and balsalazide.

[0022] In a sixth aspect, the application provides a use of a modulator of GATA4 activity in the preparation of a medicament for preventing or treating polycystic ovary syndrome.

[0023] Preferably, the medicament comprises siRNA, and the nucleotide sequence of the siRNA is shown in SEQ ID NO. 1-SEQ ID NO. 8.

[0024] GATA4-human-1200 Sense strand: 5'-CAGAGAGUGUGUCAACUGUTT-3' (SEQ ID NO. 1) Antisense strand: 5'-ACAGUUGACACACUCUCUGTT-3' (SEQ ID NO. 2) GATA4-human-1516 Sense strand: 5'-AACGGAAGCCCAAGAACCUTT-3' (SEQ ID NO. 3) Antisense strand: 5'-AGGUUCUUGGGCUUCCGUUTT-3' (SEQ ID NO. 4) GATA4-human-1821 Sense strand: 5'-AGCUCCAAGCAGGACUCUUTT-3' (SEQ ID NO. 5) Antisense strand: 5'-AAGAGUCCUGCUUGGAGCUTT-3' (SEQ ID NO. 6) GATA4-human-1708 Sense strand: 5'-CGUUCUCAGUCAGUGCGAUTT-3' (SEQ ID NO. 7) Antisense strand: 5'-AUCGCACUGACUGAGAACGTT-3' (SEQ ID NO. 8).

[0025] In a seventh aspect, the use of a BHMT agonist in the manufacture of a medicament for preventing or treating polycystic ovary syndrome.

[0026] Preferably, the medicament is betaine and acceptable salts thereof.

[0027] Preferably, the medicament can further comprise a pharmaceutically acceptable carrier. The pharmaceutically acceptable carrier can be a buffer, an emulsifier, a suspending agent, a stabilizer, a preservative, an excipient, a bulking agent, a coagulating agent and a harmonizing agent, a surfactant, a diffusing agent or an antifoaming agent. The medicament can further comprise a pharmaceutically acceptable carrier. The pharmaceutically acceptable carrier can be a virus, a microcapsule, a liposome, a nanoparticle or a polymer and any combination thereof.

[0028] Preferably, the medicament can be used in combination with other prophylactic and / or therapeutic compounds for polycystic ovary syndrome. The other prophylactic and / or therapeutic compounds can be administered simultaneously with the main active ingredient, even in the same composition.

[0029] Compared with the prior art, the present application has the following beneficial effects: (1) The prediction accuracy is significantly improved: The present application relies on comprehensive and large-scale GWAS research, successfully discovers 250 SNP sites closely related to PCOS. These sites widely cover multiple key physiological pathways related to the pathogenesis of PCOS, including but not limited to hormone regulation, metabolic processes, and inflammatory responses, etc., thereby being able to more comprehensively and accurately capture the genetic signals of the disease. Compared with the prediction methods in the prior art which only rely on a small number of sites or a single genetic factor, the PRS model constructed by the present application exhibits high accuracy in assessing the risk of individuals suffering from PCOS and risk stratification. Through verification on a large number of samples, the model effectively reduces the occurrence of misdiagnosis and missed diagnosis, and provides more reliable and solid basis for early intervention and treatment strategy for patients.

[0030] (2) Strong cross-racial applicability: Given that the research samples are widely derived from different ethnic groups around the world, the SNP sites discovered by the present application and the carefully constructed PRS model fully take into account the significant genetic differences among different ethnic groups. This outstanding feature enables the model to be widely and effectively applied to different populations around the world, successfully solving the problem that the prediction efficiency of existing prediction methods significantly decreases when applied across races, and providing a unified, reliable and universally applicable PCOS risk prediction tool for women of childbearing age around the world.

[0031] (3) Realize early prediction: PCOS often lacks obvious clinical symptoms in the early stage of the disease, and traditional prediction methods based on symptoms and partial biochemical indicators are difficult to effectively identify high-risk individuals at this stage. However, the prediction model constructed based on stable genetic information can accurately assess the risk of disease before the disease has clinical manifestations, just by detecting the genetic information of individuals. This advantage helps to take appropriate intervention measures and early drug treatment at the early stage of the disease.

[0032] (4) Economical and efficient: Compared with some complex and costly detection technologies, the PRS model constructed based on more than 200 SNP sites has significant economic efficiency. The detection technology it relies on is relatively mature and widely available in the market, and the cost is relatively low. Only one genotyping detection is needed to obtain the key site information required to construct the model, without the need for repeated detection or the use of expensive special detection equipment, thereby greatly reducing the detection cost. At the same time, based on the PRS model analysis, it is efficient and fast, and can quickly give accurate prediction results in a short time.

[0033] (5) Assist drug research and development and precision treatment: Through in-depth functional analysis of the discovered SNP sites and key genes, such as research on PPARG, GATA4, BHMT, etc., the potential therapeutic targets of PCOS are further clarified, providing a clear and precise direction for the development of targeted drugs, such as the development of new PPARG modulators. Identifying gene intervention targets not only helps to accelerate the development process of new drugs, but also enables precision treatment based on the genetic characteristics of individual patients. BRIEF DESCRIPTION OF DRAWINGS

[0034] The drawings accompanying the specification of the present invention are used to provide a further understanding of the present invention, and the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, and do not constitute an improper limitation on the present invention.

[0035] Figure 1 is a Manhattan plot of PCOS GWAS for Chinese and European cross-racial analysis; Figure 2 is an independent locus identified by PCOS GWAS for Chinese and European cross-racial analysis; (A) number of sites in each analysis; (B) a Venn diagram showing the overlap between various ethnic groups; Figure 3 is a Circos plot of sites related to PCOS and their priority genes and pathways; Figure 4 is a performance plot of a polygenic risk score model under different P-value thresholds; Figure 5is the plot of different ethnic data source multi-gene risk score model; Figure 6 is the analysis of multi-group PRS quantile and disease risk odds ratio relationship; Figure 7 is the prediction effect verification AUC curve of the optimal PRS model of PCOS; Figure 8 is the gene-drug interaction network diagram of PCOS associated loci; Figure 9 is the PPARG target drug function verification diagram; (A) differential gene expression analysis of GCs treated with four kinds of PPARG modulating drugs; (B) gene set enrichment analysis of granulosa cells treated with pioglitazone (PPARG agonist); (C) gene set enrichment analysis of granulosa cells treated with GW9662 (PPARG antagonist); (D) gene set enrichment analysis of granulosa cells treated with telmisartan; (E) gene set enrichment analysis of granulosa cells treated with balsalazide; (F) deconvolution analysis shows the proportion of five GC subgroups under different treatment conditions; Figure 10 is the GATT target intervention function verification diagram; (A) volcano plot of the transcriptome between the GATA4 intervention group and the control group in the RNA sequencing data; (B) gene set enrichment analysis between the GATA4 intervention group and the control group in the RNA sequencing data; (C) pathway enrichment analysis of up-regulated DEGs and down-regulated DEGs; (D) the proportion of granulosa cell (GC) samples at different follicular development stages in GATA4 knockdown (siGATA4) and control (siNC); Figure 11 is the BHMT target drug function verification diagram; (A) volcano plot of the transcriptome between the betaine treatment group and the control group in the RNA sequencing data; (B) pathway enrichment analysis of up-regulated DEGs and down-regulated DEGs; (C) the proportion of granulosa cell samples at different follicular development stages in the betaine treatment group and the control group. DETAILED DESCRIPTION

[0036] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0037] Explanation of terms: "PCOS" refers to polycystic ovary syndrome, which is a common gynecological endocrine disorder disease in women of childbearing age.

[0038] "SNP or SNP site" refers to DNA sequence polymorphism caused by single amino acid variation at the genomic level, including single base conversion, transversion, insertion and deletion, etc. The SNP site of the present application is named in the "rs-" manner, and the skilled in the art can determine its exact position according to the rs- naming above from the relevant database and relevant information system such as single nucleotide polymorphism database (dbSNP), nucleotide sequence.

[0039] "Marker" refers to a biochemical indicator that can mark the change or possible change of system, organ, tissue, cell and subcellular structure or function, which can be used for disease diagnosis, disease staging or to evaluate the safety and effectiveness of new drugs or new therapies in the target population.

[0040] "PRS" refers to polygenic risk score, which is a method for assessing the risk of an individual suffering from a certain disease.

[0041] "Allele" refers to a gene located at the same position of a pair of homologous chromosomes that controls the same shape but different forms.

[0042] The technical solutions of the present application are further described below in conjunction with specific examples.

[0043] Example 1 Sample data source Chinese samples, consisting of four cohorts, each of which was subjected to genotyping detection. All Chinese PCOS cases were diagnosed according to the Rotterdam criteria. The control group included healthy women from hospitals and the general population. In total, the Chinese data set contained genotype information of 46654 individuals, of which 12419 were PCOS and 34235 were controls.

[0044] European samples, including four GWAS meta-analysis datasets: 2018Felix, FinnGen (R10), EstBB (Estonian Biobank), and 23andMe. In the 2018Felix dataset, PCOS cases were diagnosed according to the National Institutes of Health (NIH) or Rotterdam criteria. In the FinnGen (R10) dataset, PCOS cases were defined as females with International Classification of Diseases (ICD) code ICD-10 E28.2, ICD-9 256.4, or ICD-8 256.90, and controls included all females without a diagnosis of PCOS. Similarly, in the EstBB dataset, PCOS cases were identified using ICD-10 E28.2, and all female participants without this diagnosis were included as controls. In the 23andMe dataset, both PCOS cases and controls were identified by self-report. Since the 23andMe research team was unable to share the full meta-analysis statistics for the PCOS dataset, only the top 10,000 variants were provided for analysis. Overall, the European-wide genome-wide meta-analysis dataset included 8,589 cases and 328,329 controls. When the dataset from 23andMe was included, the total was 13,773 cases and 411,088 controls.

[0045] Data quality control and analysis Quality control was performed separately for each cohort in the Chinese samples, including filtering out SNPs with missing rate lower than 5% before sample removal, and excluding subjects with missing rate higher than 2%. Autosomal heterozygosity bias was assessed, and samples with inbreeding coefficient (|F_het| < 0.2) were retained; after sample removal, SNPs with missing rate lower than 2% were retained. In addition, SNPs also had to satisfy Hardy-Weinberg equilibrium. Progressive pre-phasing and imputation were performed using Eagle2 and minimac3, with the 3rd phase 1000 Genomes Project reference panel as the imputation reference, and autosomes and X chromosome were imputed separately. Principal component analysis (PCA) was performed for each cohort using high-quality, high-frequency autosomal SNPs. First-degree relatives (PI_HAT > 0.2) and PCA outliers were identified and removed. Principal component scores were calculated and overlaid with 1000 Genomes Project samples to help describe the ancestral characteristics in the analysis samples. In each cohort, association tests were performed using additive logistic regression models including the top ten principal components as covariates to adjust for potential population structure.

[0046] Discovery of new genetic loci associated with PCOS: GWAS was performed in 12419 Chinese PCOS women and 34235 controls, with 7580757 variants after imputation against the 3rd phase reference panel of the 1000 Genomes Project, and 70 loci associated with PCOS were identified, containing 96 significantly associated SNPs (P<5x10 -8 ).

[0047] Meta-analysis was performed in 8589 European PCOS cases and 328329 controls, and 16 loci were identified, containing 22 significantly associated SNPs. After further integrating the 23andMe data (5184 cases and 82759 controls), 9 new loci were discovered.

[0048] By combining the data from Chinese and European populations, meta-analysis was performed across ethnicities, and 77 loci were identified, containing 118 genome-wide significant SNPs. After further integrating the 23andMe data, 8 new loci were discovered.

[0049] In summary, GWAS analysis incorporating all data found 94 independent loci associated with PCOS, as shown in Figure 1 , Figure 2 , in which: CHN: GWAS performed in Chinese cohort, containing 12419 cases and 34235 controls.

[0050] EUR: Meta-analysis performed in European cohort, containing 8589 cases and 328329 controls.

[0051] EUR2: Meta-analysis performed in European cohort focusing on 10000 variants, containing 13773 cases and 411,088 controls.

[0052] MIX: Meta-analysis combining Chinese and European cohorts, containing 21008 cases and 362564 controls.

[0053] MIX2: Meta-analysis combining Chinese and European cohorts and focusing on 10000 variants, containing 26192 cases and 445323 controls.

[0054] 21 loci have been previously reported to be associated with PCOS or its symptoms, and the remaining 73 loci are novel discoveries, with a total of 250 significantly associated SNPs (P<5x10 -8rs10737472, rs10604385, rs4648892, rs2742963, rs12038623, rs61780052, rs6673300, rs35320516, rs10922273, rs12046808, rs4915135, rs1338285, rs144834966, rs13028479, rs12470971, rs13021112, rs10205592, rs199865706, rs1375194, rs13035011, rs7563201, rs6753358, rs6710610, rs7578655, rs1464727, rs3811516, rs7567846, rs138311200, rs7606567, rs4384834, rs10183851, rs1922476, rs4953651, rs12994034, rs6708637, rs1504187, rs1504165, rs6733807, rs140140896, rs55687579, rs75775350, rs72827480, rs1837294, rs1830321, rs957293, rs10206585, rs2857534, rs6712883, rs13411274, rs3771066, rs7574706, rs76172579, rs56009971, rs7564590, rs1531788, rs6735267, rs6710653, rs7599312, rs10191896, rs6785890, rs13061415, rs2067819, rs7615916, rs146753246, rs4456810, rs3924105, rs61262534, rs10516202, rs55904482, rs191505728, rs35232147, rs62306364, rs11372697, rs9312937, rs199669078, rs4705862, rs10072700, rs2070721, rs11242115, rs41525648, rs891988, rs2879271, rs9324872, rs4624820, rs34779343, rs9460408, rs9465511, rs11349096, rs6926381, rs6926229, rs7757706, rs9378247, rs4947328, rs486416, rs3807020,rs2185710, rs2502403, rs35869530, rs6912459, rs12527757, rs60793153, rs5886708, rs10228327, rs11777247, rs17807624, rs2740434, rs745379, rs17153782, rs2740433, rs2686208, rs804260, rs4840581, rs804256, rs17053711, rs11775358, rs4733407, rs10808334, rs6994022, rs12681434, rs12681435, rs12544187, rs4460442, rs2183958, rs1123534, rs7864171, rs7865239, rs10818820, rs12347448, rs77719717, rs200979118, rs117889661, rs75304159, rs79375299, rs3945628, rs4620350, rs2799465, rs10986104, rs5900588, rs7848238, rs4314746, rs7042632, rs6606575, rs57994832, rs11188578, rs200584089, rs2756100, rs1926029, rs76383974, rs11592408, rs16932777, rs10741694, rs10835634, rs11031005, rs11031034, rs11031040, rs556622, rs11040598, rs10082560, rs76735130, rs11229231, rs3740664, rs55965748, rs11234902, rs1426394, rs11225154, rs10501995, rs11225161, rs11436634, rs1784692, rs202185443, rs3217844, rs3217856, rs35481774, rs76624470, rs1131017, rs7971751, rs2272046, rs113923326, rs17115231, rs1275468, rs1148006, rs11180609, rs2255591, rs2855530, rs2761887, rs798791, rs2075773, rs77938952, rs2445768, rs2445773,rs1008815, rs10852606, rs12931034, rs7200621, rs8043701, rs9939609, rs7190396, rs250150, rs16950653, rs30403, rs30404, rs56738967, rs9928324, rs7203756, rs76862947, rs73575083, rs7358, rs78978329, rs1642799, rs72829444, rs78378222, rs2680703, rs2680698, rs2632513, rs78911886, rs12978702, rs11666480, rs11669774, rs10407022, rs4325676, rs11670032, rs1676465, rs7260162, rs2301595, rs8105605, rs11672660, rs10412726, rs734518, rs3817996, rs853854, rs6019857, rs6022786, rs6068688, rs171008, rs237048, rs1038023, rs9613564, rs62237617, rs17879961, rs56119508, rs544626416, rs112158446, rs7288513, rs5762906, rs34293193, rs756630, rs138219, rs2015414, rs73636638, rs5951636, Characterization of PCOS-associated loci To identify potential functional variants and causal genes for PCOS, we used RegulomeDB and Open Targets Genetics methods. RegulomeDB assesses the regulatory potential of key variants by cross-analyzing them with functional genomics experimental data and computational predictions, and assigns a rank score and a probability score. These two scores annotate and rank variants from complementary perspectives: a higher rank score indicates a stronger regulatory context, while a higher probability score suggests a greater functional impact. Open Targets Genetics integrates GWAS results with functional genomics data - including expression quantitative trait loci (eQTLs), protein quantitative trait loci (pQTLs), and chromatin interaction data - and combines gene proximity and computational-based functional predictions to prioritize causal genes for each locus. The overall V2G score (ranging from 0 to 1) integrates multiple genomic evidence datasets, such as variant proximity, chromatin interactions, and expression correlation. The data used were pre-filtered to remove low-confidence evidence, ensuring that each non-zero score represents at least one robust supporting evidence, with higher scores indicating stronger evidence of functional association. For each genomic evidence dataset, a score (ranging from 0 to 1) was generated in a consistent manner, reflecting the level of evidence for functionality.

[0055] According to the RegulomeDB evaluation results, the scores of 136 SNPs among the above-mentioned SNPs are less than 3 (may affect binding), and the probability scores of 19 SNPs are greater than 0.8, supporting their function as regulatory variants. It is worth noting that SNP rs12544187 located at 8p11.22, with a RegulomeDB rank score of 1a and a probability score of 0.95, highlights it as a strong candidate for further functional regulatory research. This SNP is located in the intronic region of the FGFR1 gene, which is crucial for the development of the gonadotropin-releasing hormone (GnRH) nervous system, suggesting that it may play an important role in PCOS by affecting GnRH.

[0056] Further analysis found that the above SNPs have diverse tissue-specific regulatory scores, reflecting their different effects in the biological system. rs2680703 (17q22) showed a high regulatory score in multiple tissues, indicating its wide physiological impact. The gene SUPT4H1 involved in rs2680703 is involved in mRNA processing and transcription elongation mediated by RNA polymerase II. rs4840581 (corresponding to the gene GATA4) has a significant score in endocrine and adrenal tissues, indicating that this locus or gene is mainly involved in the regulation of endocrine hormones affecting the occurrence and development of PCOS. In addition, rs78378222 (corresponding to the gene TP53 / SHBG) showed a significant score in endocrine glands, exocrine glands, thyroid glands, and ovaries, suggesting that this locus or gene plays a key role in PCOS reproduction and hormone regulation.

[0057] Several key genes associated with PCOS risk were identified using Open Targets Genetics analysis, including AMH, HLA-C, KHNYN, MLC1, ANGEL1, SUOX, RAB5B, and HEATR3, all of which had overall V2G scores exceeding 0.4. Among the newly identified loci, SNP rs10407022 located at 19p13.3 is a missense mutation (p.Ser49Ile) in the AMH gene, and its risk allele T is associated with increased mRNA levels of AMH in multiple tissues such as the ovary and adipose tissue. Based on evidence from ovarian eQTLs, four genes (RPS26, CBLN3, HEATR3, and MLC1) were prioritized. Based on pQTL and eQTL analysis, the FUT10 gene located at 8p12 was found to be very important. Based on analysis of epigenetic data (including promoter capture Hi-C, enhancer-TSS interaction, and DHS-promoter correlation), a series of key genes were identified, including but not limited to YAP1, DENND1A, TMEM150B, and HLA-C.

[0058] PCOS-associated gene and gene set analysis After uniform treatment of individual genotypes, we performed secondary gene- and gene-set-based association analyses. Using the Phase 3 East Asian panel (EAS) of the Human Genome Project as a linkage disequilibrium reference, gene-based tests were performed using MAGMA (v1.10) with a multiple linear principal component regression model. We tested a total of 19,049 genes, including SNPs within 35 kb upstream and 10 kb downstream of each gene. For gene-set analysis, we tested gene sets from the Molecular Signatures Database (MSigDB) collection (v2023.1.Hs), including Gene Ontology (GO), BioCarta, WikiPathways, Human Phenotype Ontology (HPO), and Kyoto Encyclopedia of Genes and Genomes (KEGG). As a supplementary method, we also performed gene-based tests using VEGAS2 (v2.01.17) with a multivariate normal simulation. Using default parameters, pathway analysis was performed by VEGAS2Pathway to assess the enrichment of gene P values in Biosystems pathways. We defined gene-level significance as P = 0.05 / 19,049 = 2.62 x 10 -6 . Pathway-level significance thresholds were set at 1.49 x 10 -6 for MAGMA and 5.14 x 10 -6 for VEGAS2.

[0059] Gene analysis across the genome using MAGMA found 219 significantly associated genes (P < 2.62 x 10 -6), in particular: HLA-DQB1, DENND1A, HLA-B, FSHR, SUOX, THADA, ERBB3, ZNF436, CDK2, INSR, ZFP36L2, FAM98A, STON1-GTF2A1L, SUPT4H1, KRR1, HEATR3, BZRAP1, HLA-DQA1, LOC101927250, ATP8B3, LHCGR, SPRY4, MICA, CNEP1R1, RAB5B, TCEA3, LTB, GTF2A1L, KLF16, GREB1, C9orf3, SOX6, HSF5, ZBTB12, ASAP3, HLA-DRB1, GLDN, ZNF728, SMOC2, CDCA2, YAP1, FOXP2, KCTD9, C2, PPARG, AIF1, TEX14, ZNRF3, GNRH1, NCR3, UBD, CYP19A1, TNXB, HNRNPR, HLA-A, ID4, LHX2, ARL14EP, PMEL, IKZF4, RPS26, PA2G4, RNF43, REXO1, RAD51C, HOXD1, EHMT2, C4A, HOXD3, RPL41, OR2H2, CDSN, ATF6B, ZC3H10, VARS, LSM2, AMH, TRIM37, HSPA1B, BHMT, CFB, LHX9, TCF19, CCHCR1, C6orf48, RNF39, PLEKHJ1, MYO10, MAS1L, ANGEL1, SCAMP4, FKBPL, HMGA2, SF3A2, NELFE, HSPA1A, ZBED9, HSPA1L, PRRC2A, C17orf47, BTN2A2, C6orf15, EIF4A1, HIST1H4I, POU5F1, ABHD17A PARGC1A, CD68, PSORS1C2, HIST1H2AG, FAM13A, EMILIN1, AGBL5, SLC44A4, BTN3A1, OST4, MPDU1, LOC100129940, HLA-DRA, KIF15, ADAT3, MTMR4, KHK, E2F2, VWA7, SENP3, TMEM42, HIST1H2AH, LOC554223, BTN1A1, DXO, HAUS5, 4-Sep, OR4A16, OR4C11, ZNRD1, PPP1R11, PPT2, HIST1H2BJ, SOX15, ZNF518A, GLIPR1, FOLH1, OR4A15, ABHD1, LRRC74A, PRR30, SHBG, JMY, FZD4, PSORS1C1,PLEKHH2, BTN2A1, LST1, EGFL8, C4B, PREB, MAPRE3, TRIM48, CGREF1, KIAA1143, CYP21A2, LOC101928592, STK19, NEIL2, MICB, PPM1E, RASGRP3, TTC28, RBM42, ZNF502, FUT10, CCNJ, PROX1, SKIV2L, DDAH2, NEU1, CCNE1, TCF23, GDPD1, GATA4, INHBB, AKR1C4, PRR11, ESR1, TNFSF13, DOCK5, HLA-G, SMG8, PAPD5, MAK16, LOC101060256, MAST4, MSH5, LY6G6C, HIST1H2BK, MAP1A, ERBB4, TP53BP1, SKA2, BTN3A2, PEX6, PDGFC, PRRT1, HOXD4, PPP2R5D, PLEKHM3, GRK5, CLIC1, BMP4, AGPAT1, ZBTB16, TRIM31, SAPCD1, HLA-DQA2, ZNF501, TNF, CAPS2. Gene set analysis further revealed six significantly enriched gene sets, including ligand binding domain (LBD) binding and peroxisome proliferator-activated receptor gamma (PPARG) pathway ( Figure 3). Meanwhile, VEGAS2 identified 154 significantly associated genes, specifically: THADA, DENND1A, STON1-GTF2A1L, ARL14EP, LHCGR, RPS26, SUOX, IKZF4, ERBB3, RAB5B, YAP1, C9orf3, MIR548H3, LINC01126, ZFP36L2, CDK2, RNF43, LOC100506368, HMGA2, FSHR, LOC100129940, RAD51C, PPM1E, BZRAP1-AS1, BZRAP1, HLA-DQB1, TRIM37, KLF16, HLA-DQA1, REXO1, FAM98A, C2, ZBTB12, HSF5, ABHD17A, INSR, AMH, FZD4, MTMR4, HEATR3, TEX14, MICA, TCEA3, HLA-B, SMOC2, CNEP1R1, SPRY4, SF3A2, LST1, MYO10, CFB, LHX2, C17orf47, CCHCR1, SOX6, PRR11, ASAP3, LINC01015, HCG4B, KCTD9, ZNRF3, VARS, CDCA2, ZNF728, HLA-A, MAS1L, LOC101929164, KRR1, GLIPR1, C1orf213, GREB1, ZNF436, HOXD-AS1, LOC102800310, BTN2A2, GNRH1, SKIV2L, HOXD3, BTN2A1, MOG, HCG22, LOC101929144, HCG27, LOC285819, HCG9, PPARG, DOCK5, GLDN, ANGEL1, TMEM214, C8orf49, LOC440040, MAPRE3, E2F2, KIF15, FXR2, BHMT, KHK, CCNJ, VASH1, FOXP2, KIAA1143, SHBG, CCND2, ZNF518A, ZNF502, EMILIN1, TMEM42, JMY, AGBL5, LOC102723617, UBD, LOC554223, PLEKHM3, MPDU1, SOX15, ETV2, FAM13A, ZNRD1-AS1, FUT10, LOC100288123, EHMT2, CGREF1, ZNRD1, FOLH1, HOXD-AS2, HLA-J, HLA-E, HCG4, HNRNPR, PSORS1C1, CD68, SENP3, SENP3-EIF4A1, PPT2, EIF4A1, BMP4, ZBED9, CCNE1, OR4A16, RNU6-28P,TP53BP1, HSPA1L, OR4C15, GNMT, LOC100270746, MAST4, LOC441601, ABHD1, CYP21A1P, LY6G6F, OR5T3, FKBPL, HSPA1A, of which 116 overlap with the MAGMA analysis, specifically: HLA-DQB1, DENND1A, HLA-B, FSHR, SUOX, THADA, ERBB3, ZNF436, CDK2, INSR, ZFP36L2, FAM98A, STON1-GTF2A1L, KRR1, HEATR3, BZRAP1, HLA-DQA1, LHCGR, SPRY4, MICA, CNEP1R1, RAB5B, TCEA3, KLF16, GREB1, C9orf3, SOX6, HSF5, ZBTB12, ASAP3, GLDN, ZNF728, SMOC2, CDCA2, YAP1, FOXP2, KCTD9, C2, PPARG, TEX14, ZNRF3, GNRH1, UBD, HNRNPR, HLA-A, LHX2, ARL14EP, IKZF4, RPS26, RNF43, REXO1, RAD51C, EHMT2, HOXD3, VARS, AMH, TRIM37, BHMT, CFB, CCHCR1, MYO10, MAS1L, ANGEL1, FKBPL, HMGA2, SF3A2, HSPA1A, ZBED9, HSPA1L, C17orf47, BTN2A2, EIF4A1, ABHD17A, CD68, FAM13A, EMILIN1, AGBL5, MPDU1, TMEM214, LOC100129940, KIF15, MTMR4, KHK, E2F2, SENP3, TMEM42, LOC554223, OR4A16, ZNRD1, PPT2, SOX15, ZNF518A, GLIPR1, FOLH1, ABHD1, SHBG, JMY, FZD4, PSORS1C1, BTN2A2, LST1, MAPRE3, CGREF1, KIAA1143, PPM1E, ZNF502, FUT10, CCNJ, SKIV2L, CCNE1, PRR11, DOCK5, MAST4, TP53BP1, PLEKHM3, BMP4. VEGAS2 analysis identified mainly eight pathways, such as cell cycle G1 / S transition, primordial germ cell development, gonad development, ovarian follicle development and beta 2 microglobulin binding, among others. Figure 3 .

[0060] Cross-ethnic fine mapping analysis To identify causal variants in the genome-wide significant loci, we performed fine-mapping analysis using the SuSiE model. For single-population analysis, we used the echolocatoR method combined with the SuSiE tool. Cross-ancestry fine-mapping used SuSiEx. In the cross-ancestry analysis, we generated 95% credible sets and calculated the posterior inclusion probability (PIP) values for both cross-ancestry and single-population analyses.

[0061] A total of 109 95% credible sets were identified by using SuSiEx for cross-ancestry population fine-mapping. In the cross-ancestry population analysis, most SNPs, such as rs9312937 (MYO10), rs2740433 (GATA4), rs1784692 (ZBTB16), and rs2091788 (FSHR), had PIP values greater than 0.95 in the cross-ancestry population fine-mapping, which were significantly higher than the PIP values in the single-population analysis. The SNP rs250157 (MAF) located at 16q23.2 showed the highest PIP value of 0.894 in the cross-ancestry population analysis, and this SNP was predicted to be associated with the motif targeting estrogen receptors ESRRA and ESRRG.

[0062] PRS model construction and validation The construction of PRS was completed based on the PRSice-2 (version 2.3.5) software platform. The research team set 10 GWAS P-value thresholds (1, 0.5, 0.4, 0.3, 0.2, 0.1, 0.05, 0.01, 1x10 -4 , and 5x10 -8 ) according to the significance of the association signal, and generated a set of candidate gene risk scores under each threshold. The model performance was first calculated on the observation scale Nagelkerke R 2 , and was converted to the susceptibility scale according to the prevalence of PCOS in the general population of 7.8%. At the same time, by comparing the odds ratio (OR) of individuals in the top 10% and the bottom 10%, we further quantified the risk difference between the high and low groups (ΔOR) Figure 4 .

[0063] To evaluate the impact of different population sources on model performance, three combination methods were set in the training stage, namely using Chinese data alone, using European data alone, and training together after merging Chinese and European data. This method maximizes the risk of overfitting and can fully test the generalization ability of the model. At the same time, we used the "one-by-one leave queue" cross-validation strategy: in each iteration, one Chinese cohort was completely excluded as an independent test set, and the remaining three population cohorts were used for model training.

[0064] When the Chinese cohort was used as the test set, the model training results showed that when only Chinese data was used and the P-value threshold was set to 5 10 -8 and 0.05, PRS could explain 8.4% and 15.6% of the PCOS phenotype variation, respectively; at this time, the risk of the top 10% of the population was 13.63 times (95% confidence interval: 10.37-17.91) that of the bottom 10% of the population. When the European samples were additionally included in the training set, the model's explanation at the P-value threshold of 5 10 -8 12.19 (95% confidence interval: 9.25-16.06). In contrast, the model trained only using European data performed the worst, with a phenotype explanation of 5.1% and 3.6% at the two thresholds, and an OR of only 1.46 (95% confidence interval: 1.17-1.81) (( Figure 5 , Figure 6 ).

[0065] Taking into account the model's explanation, risk ratio, and test feasibility, the study ultimately determined that the "cross-racial training data and P < 5 10 -8 " scheme was the optimal PRS model, and the SNPs included are shown in Table 1. This model can explain about 10% of the genetic variation of PCOS in Chinese women and can effectively distinguish between high- and low-risk populations of PCOS.

[0066] To further test the stability of the model, a sample of 760 PCOS patients and 870 healthy controls was used as a validation cohort, and the optimal model was directly applied to compare the model's prediction results with the actual disease status to evaluate the model's sensitivity, specificity, and accuracy. The validation results confirmed that the model exhibited good performance indicators (( Figure 7 ). The AUC value reached 0.872, demonstrating excellent overall discrimination ability, with a sensitivity of 0.899 and an overall accuracy of 0.749, fully demonstrating the model's stability and generalizability in real-world scenarios. When the subjects were further risk stratified by PRS quartiles (Q1-Q4), it was observed that the sensitivity increased from 0.418 in Q1 to 0.979 in Q4, while the specificity decreased from 0.906 to 0.570, showing a typical gradient change; the positive predictive value of the highest risk quartile was as high as 0.991, almost serving as a "clinical diagnosis" warning signal, while the high specificity (0.906) of the lowest risk quartile helped reduce unnecessary follow-up and intervention. This stepwise performance indicates that the model not only can perform binary classification diagnosis, but also can achieve fine genetic risk stratification, providing reliable quantitative evidence for early screening, individualized management, and precise intervention of PCOS, and has the potential to continue to expand and optimize in multi-ethnic cohorts.

[0067] Genomics-driven drug discovery for polycystic ovary syndrome We employed two complementary approaches to leverage genomic data to identify potential drugs for PCOS prevention or treatment Figure 8 ). First, we performed an enrichment analysis of genes associated with PCOS using GREP (gene repositioning drugs), excluding genes within the MHC region. These genes were determined by genome-wide summary statistics of our core dataset and analyzed using MAGMA and VEGAS2. We performed Fisher’s exact test to measure the enrichment of disease-associated genes among drug target genes classified according to the Anatomical Therapeutic Chemical (ATC) classification system codes. Second, we retrieved known or potentially druggable genes from the Drug- Gene Interaction Database (DGIdb 5.0). This included genes associated with PCOS identified by MAGMA or VEGAS2, as well as prioritized genes from all identified loci with an overall V2G score greater than 0.05 from Open Targets Genetics. DGIdb 5.0 aggregates drug-gene interactions from various studies and databases, categorizing genes into 43 potential druggable groups and defining over 30 types of interactions according to the source data.

[0068] The results of the GREP-based analysis showed significant enrichment of PCOS-associated genes among drug target genes in the “reproductive urinary and sex hormones” and “antineoplastic and immunomodulating agents” categories, identifying key endocrine therapy targets such as CYP19A1 and ESR1, FSHR, and related drugs such as aromatase inhibitors, selective estrogen receptor modulators, sex hormone modulators, and insulin sensitizers such as pioglitazone and metformin, which act through PPARG, and telmisartan, betaine.

[0069] Based on the drug-gene interaction analysis of the DGIdb 5.0 database, this study successfully identified several drugs that promote fertility, such as trofupimil alpha and urinary folliculin, which can interact with FSHR. In addition, the combination of lutupimil alpha and LHCGR also confirms their established and critical role in PCOS management. At the same time, this study also found that metformin has the potential to target the PPARG site and can be used as a candidate drug to further explore its application value in PCOS treatment. The newly discovered interaction between betaine hydrochloride and BHMT further highlights its broad prospects in the field of PCOS treatment, especially in improving PCOS-related symptoms. It is worth noting that the interaction between TNFSF13 and the unapproved anti-inflammatory drug atacicept brings potential hope and progress in addressing PCOS-related inflammation issues, and is expected to open up new treatment ideas and strategies. In addition, this analysis also reveals some research targets, such as GATA4 and HOXD3, for which compounds have not yet been approved, but their existence indicates significant preclinical research interest in improving PCOS cardiovascular symptoms, and effective drugs targeting these targets are expected to be developed in the future, providing more treatment options for PCOS patients. At the same time, research compounds involving SHBG and other genes also suggest that hormone regulation pathways have not been fully explored in PCOS treatment, providing new clues and directions for a deeper understanding of the pathogenesis of PCOS and the search for new treatment targets.

[0070] Potential target intervention and candidate drug effect verification PPARG as a target ( Figure 9 ), the reason for prioritizing PPARG over other potential targets is its unique position at the intersection of PCOS metabolic and reproductive dysfunction. We treated human granulosa cells (GCs) with a variety of PPARG modulators, including the classic agonist pioglitazone, the antagonist GW9662, and two newly discovered modulators - telmisartan and balsalazide. Transcriptional profiling showed that pioglitazone activation of PPARG significantly enhanced steroidogenic gene expression and improved multiple metabolic pathways, including amino acid metabolism, fructose metabolism, fatty acid elongation, and citric acid cycle. Importantly, pioglitazone also exhibited strong anti-inflammatory effects in GCs, which are closely related because the role of chronic low-grade inflammation in the pathogenesis of PCOS has been established. The results of GW9662, a PPARG antagonist, demonstrated that it significantly inhibited the Hedgehog pathway, which is crucial for follicular development. The dual impact on metabolic and follicular development pathways makes PPARG a promising therapeutic target.

[0071] Telmisartan is an angiotensin II receptor blocker with partial PPARG agonist activity (equivalent to 20-30% potency of pioglitazone). Our findings with telmisartan are particularly interesting as telmisartan treatment of human granulosa cells activated ALKBH3, an RNA demethylase that protects against oxidative damage, while stimulating GNAS, a key component of FSH / cAMP / PKA signaling in granulosa cells, for improved ovarian function markers, including enhanced steroidogenesis, vitamin uptake, and fatty acid metabolic pathways. That is, telmisartan can provide additional benefits in preventing ovarian remodeling in PCOS.

[0072] Balsalazide is an anti-inflammatory drug that, through our drug-gene interaction analysis, was predicted to interact with PPARG. Our study showed a significant effect on granulosa cell survival. As an FDA-approved drug for inflammatory bowel disease, it promoted cell cycle progression and DNA replication in granulosa cells and inhibited pathways related to circadian rhythm synchronization, possibly reflecting a shift from a differentiated to a proliferative state in granulosa cells. These multi-modal effects suggest that balsalazide is a promising therapeutic candidate for PCOS, capable of addressing the functional defects in PCOS granulosa cell development.

[0073] The above findings collectively reinforce the position of PPARG as an attractive therapeutic target for PCOS, with effects encompassing metabolic regulation, steroidogenesis, inflammation control, and follicular development, providing a solid foundation for translational research into PCOS treatment targeting PPARG.

[0074] Targeting GATA4 Figure 10 GATA4 was knocked down in human granulosa cells using siRNA, with four siRNAs mixed for transfection. The specific sequences of the four siRNAs are as follows: GATA4-Human-1200 Sense strand: 5'-CAGAGAGUGUGUCAACUGUTT-3' Antisense strand: 5'-ACAGUUGACACACUCUCUGTT-3' GATA4-Human-1516 Sense strand: 5'-AACGGAAGCCCAAGAACCUTT-3' Antisense strand: 5'-AGGUUCUUGGGCUUCCGUUTT-3' GATA4-Human-1821 Sense strand: 5'-AGCUCCAAGCAGGACUCUUTT-3' Sense: 5'- AAGAGUCCUGCUUGGAGCUTT -3' GATA4 - Human - 1708 Sense: 5'- CGUUCUCAGUCAGUGCGAUTT -3' Antisense: 5'- AUCGCACUGACUGAGAACGTT -3' Knocking down GATA4 in human granulosa cells, transcriptome sequencing analysis identified VWCE, CNFN, TGFBI and CREB3L1 as significantly upregulated genes, which are enriched in collagen synthesis, extracellular matrix (ECM) remodeling and hormone signaling pathways (e.g. relaxin and neuroactive ligand-receptor interaction). In contrast, downregulated genes (e.g. DCN, RAB23, LAMB4, etc.) are associated with DNA replication, mitotic progression and immune regulation (e.g. graft versus host disease and Th1 / Th2 differentiation). That is, GATA4-targeted intervention can significantly alter the transcriptional program associated with polycystic ovary syndrome extracellular matrix remodeling fibrosis abnormalities, etc.

[0075] Targeting BHMT (GSHMT1) as a target, Figure 11 GSEA showed that pathways related to follicle maturation (including neuroactive ligand-receptor interaction, TGF-beta signaling and protein digestion / absorption) were significantly activated, while pathways related to metabolic dysfunction (type 2 diabetes) and inflammation (leukocyte transendothelial migration) were inhibited. Single-cell deconvolution analysis further showed that betaine-driven primitive and pre-ovulatory granulosa cell populations increased, promoting follicle development, consistent with its antioxidant, anti-apoptotic and improved hormone synthesis function, showing its potential as a PCOS treatment drug.

[0076] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A SNP biomarker for predicting the risk of polycystic ovary syndrome, characterized in that, The markers are: rs10737472, rs10604385, rs4648892, rs2742963, rs12038623, rs61780052, rs6673300, rs35320516, rs10922273, rs12046808, rs4915135, rs1338285, rs144834966, rs13028479, rs12470971, rs13021112, rs10205592, rs199865706, rs1375194, rs13035011, rs7563201, rs6753358, rs671 0610, rs7578655, rs1464727, rs3811516, rs7567846, rs138311200, rs7606567, rs4384834, rs10183851, rs1922476, rs4953651, rs12994034, rs670 8637, rs1504187, rs1504165, rs6733807, rs140140896, rs55687579, rs75775350, rs72827480, rs1837294, rs1830321, rs957293, rs10206585, rs28 57534、rs6712883、rs13411274、rs3771066、rs7574706、rs76172579、rs56009971、rs7564590、rs1531788、rs6735267、rs6710653、rs7599312、rs101 91896, rs6785890, rs13061415, rs2067819, rs7615916, rs146753246, rs4456810, rs3924105, rs61262534, rs10516202, rs55904482, rs191505728, rs35232147, rs62306364, rs11372697, rs9312937, rs199669078, rs4705862, rs10072700, rs2070721, rs11242115, rs41525648, rs891988, rs28792 71. rs9324872, rs4624820, rs34779343, rs9460408, rs9465511, rs11349096, rs6926381, rs6926229, rs7757706, rs9378247, rs4947328, rs486416,rs3807020、rs2185710、rs2502403、rs35869530、rs6912459、rs12527757、rs60793153、rs5886708、rs10228327、rs11777247、rs17807624、rs2740434、rs745379、rs17153782、rs2740433、rs2686208、rs804260、rs4840581、rs804256、rs17053711、rs11775358、rs4733407、rs10808334、rs6994022、rs12681434、rs12681435、rs12544187、rs4460442、rs2183958、rs1123534、rs7864171、rs7865239、rs10818820、rs12347448、rs77719717、rs200979118、rs117889661、rs75304159、rs79375299、rs3945628、rs4620350、rs2799465、rs10986104、rs5900588、rs7848238、rs4314746、rs7042632、rs6606575、rs57994832、rs11188578、rs200584089、rs2756100、rs1926029、rs76383974、rs11592408、rs16932777、rs10741694、rs10835634、rs11031005、rs11031034、rs11031040、rs556622、rs11040598、rs10082560、rs76735130、rs11229231、rs3740664、rs55965748、rs11234902、rs1426394、rs11225154、rs10501995、rs11225161、rs11436634、rs1784692、rs202185443、rs3217844、rs3217856、rs35481774、rs76624470、rs1131017、rs7971751、rs2272046、rs113923326、rs17115231、rs1275468、rs1148006、rs11180609、rs2255591、rs2855530、rs2761887、rs798791、rs2075773、rs77938952、rs2445768、rs2445773、rs1008815、rs10852606、rs12931034、rs7200621、rs8043701、rs9939609、rs7190396、rs250150、rs16950653、rs30403、rs30404、rs56738967、rs9928324、rs7203756、rs76862947、rs73575083、rs7358、rs78978329、rs1642799、rs72829444、rs78378222、rs2680703、rs2680698、rs2632513、rs78911886、rs12978702、rs11666480、rs11669774、rs10407022、rs4325676、rs11670032、rs1676465、rs7260162、rs2301595、rs8105605、rs11672660、rs10412726、rs734518、rs3817996、rs853854、rs6019857、rs6022786、rs6068688、rs171008、rs237048、rs1038023、rs9613564、rs62237617、rs17879961、rs56119508、rs544626416、rs112158446、rs7288513、rs5762906、rs34293193、rs756630、rs138219、rs2015414、rs73636638、rs5951636。、 2. The use of a substance that detects the SNP marker of claim 1 in the preparation of products for the diagnosis, detection, monitoring or prediction of polycystic ovary syndrome.

3. The application as described in claim 2, characterized in that, The products include primers, probes, nucleic acid membrane strips, gene or protein chips, formulations, reagent kits, instruments, detection devices, and equipment.

4. A multi-gene risk scoring system, characterized in that, The system includes: (1) A detection unit, the detection unit comprising: a means for determining the subject's sample to be tested for SNP marker detection; (2) An analysis unit, the analysis unit comprising: analyzing the SNP markers detected in (1); (3) Assessment unit, the assessment unit comprising: assessing the subject’s disease risk based on the SNP markers analyzed in (2); The risk profile of the subjects includes the diagnosis, detection, monitoring, or predictive assessment of polycystic ovary syndrome.

5. The system as described in claim 4, characterized in that, The system was constructed based on the PRSice-2 software platform, and the SNP sites described in claim 1 were incorporated into the system. Preferably, the system was trained after it was built; Preferably, the training method is as follows: three combination methods are set in the training phase, namely, using Chinese data alone, using European data alone, and merging Chinese and European data for joint training; at the same time, we adopt the "one-by-one queue" cross-validation strategy: in each iteration, one of the Chinese queues is completely excluded and used as an independent test set, and the remaining three population queues are used for model training.

6. The use of PPARG activity modifiers in the preparation of drugs for the prevention or treatment of polycystic ovary syndrome, preferably, the drugs are: agonist pioglitazone, antagonist GW9662, metformin, telmisartan and balsalazine.

7. The use of GATA4 activity modulators in the preparation of drugs for the prevention or treatment of polycystic ovary syndrome, preferably, the drug comprises siRNA, the nucleotide sequence of which is shown in SEQ ID NO.1-SEQ ID NO.

8.

8. The use of BHMT agonists in the preparation of drugs for the prevention or treatment of polycystic ovary syndrome, preferably, the drug being betaine or an acceptable salt thereof.

9. The application according to any one of claims 6-8, characterized in that, The drug also includes a pharmaceutically acceptable carrier, which may be a buffer, emulsifier, suspending agent, stabilizer, preservative, excipient, filler, coagulant and blending agent, surfactant, dispersant or defoamer.

10. The application according to any one of claims 6-8, characterized in that, The drug can also be used in combination with other drugs for the prevention and / or treatment of polycystic ovary syndrome.

Images