Computational characterization and identification of human polycystic ovary syndrome genes

Establishment and Analysis of an Artificial Neural Network Model for Early Detection of Polycystic Ovary Syndrome Using Machine Learning Techniques

Background

To identify novel gene combinations and to develop an early diagnostic model for Polycystic Ovary Syndrome (PCOS) through the integration of artificial neural networks (ANN) and random forest (RF) methods.

Methods

We retrieved and processed gene expression datasets for PCOS from the Gene Expression Omnibus (GEO) database. Differential expression analysis of genes (DEGs) within the training set was performed using the "limma" R package. Enrichment analyses on DEGs using gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG), and immune cell infiltration. The identification of critical genes from DEGs was then performed using random forests, followed by the developing of new diagnostic models for PCOS using artificial neural networks.

Results

We identified 130 up-regulated genes and 132 down-regulated genes in PCOS compared to normal samples. Gene Ontology analysis revealed significant enrichment in myofibrils and highlighted crucial biological functions related to myofilament sliding, myofibril, and actin-binding. Compared with normal tissues, the types of immune cells expressed in PCOS samples are different. A random forest algorithm identified 10 significant genes proposed as potential PCOS-specific biomarkers. Using these genes, an artificial neural network diagnostic model accurately distinguished PCOS from normal samples. The diagnostic model underwent validation using the independent validation set, and the resulting area under the receiver operating characteristic curve (AUC) values was consistent with the anticipated outcomes.

Conclusion

Utilizing unique gene combinations, this research created a diagnostic model by merging random forest techniques with artificial neural networks. The AUC indicated a notably superior performance of the diagnostic model.

Progress of the application clinical prediction model in polycystic ovary syndrome

Clinical prediction models play an important role in the field of medicine. These can help predict the probability of an individual suffering from disease, complications, and treatment outcomes by applying specific methodologies. Polycystic ovary syndrome (PCOS) is a common disease with a high incidence rate, huge heterogeneity, short- and long-term complications, and complex treatments. In this systematic review study, we reviewed the progress of clinical prediction models in PCOS patients, including diagnosis and prediction models for PCOS complications and treatment outcomes. We aimed to provide ideas for medical researchers and clues for the management of PCOS. In the future, models with poor accuracy can be greatly improved by adding well-known parameters and validations, which will further expand our understanding of PCOS in terms of precision medicine. By developing a series of predictive models, we can make the definition of PCOS more accurate, which can improve the diagnosis of PCOS and reduce the likelihood of false positives and false negatives. It will also help discover complications earlier and treatment outcomes being known earlier, which can result in better outcomes for women with PCOS.

Weighted Gene Co-Expression Network Analysis (WGCNA) Discovered Novel Long Non-Coding RNAs for Polycystic Ovary Syndrome

Polycystic ovary syndrome (PCOS) affects reproductive-age women. This condition causes infertility, insulin resistance, obesity, and heart difficulties. The molecular basis and mechanism of PCOS might potentially generate effective treatments. Long non-coding RNAs (lncRNAs) show control over multifactorial disorders' growth and incidence. Numerous studies have emphasized its significance and alterations in PCOS. We used bioinformatic methods to find novel dysregulated lncRNAs in PCOS. To achieve this objective, the gene expression profile of GSE48301, comprising PCOS patients and normal control tissue samples, was evaluated using the R limma package with the following cut-off criterion: p-value < 0.05. Firstly, weighted gene co-expression network analysis (WGCNA) was used to determine the co-expression genes of lncRNAs; subsequently, hub gene identification and pathway enrichment analysis were used. With the defined criteria, nine novel dysregulated lncRNAs were identified. In WGCNA, different colors represent different modules. In the current study, WGCNA resulted in turquoise, gray, blue, and black co-expression modules with dysregulated lncRNAs. The pathway enrichment analysis of these co-expressed modules revealed enrichment in PCOS-associated pathways, including gene expression, signal transduction, metabolism, and apoptosis. In addition, CCT7, EFTUD2, ESR1, JUN, NDUFAB1, CTTNB1, GRB2, and CTNNB1 were identified as hub genes, and some of them have been investigated in PCOS. This study uncovered nine novel PCOS-related lncRNAs. To confirm how these lncRNAs control translational modification in PCOS, functional studies are required.

A GWAS in Idiopathic/Unexplained Infertile Men Detects a Genomic Region Determining Follicle-Stimulating Hormone Levels

CONTEXT

Approximately 70% of infertile men are diagnosed with idiopathic (abnormal semen parameters) or unexplained (normozoospermia) infertility, with the common feature of lacking etiologic factors. Follicle-stimulating hormone (FSH) is essential for initiation and maintenance of spermatogenesis. Certain single-nucleotide variations (SNVs; formerly single-nucleotide polymorphisms [SNPs]) (ie, FSHB c.-211G > T, FSHR c.2039A > G) are associated with FSH, testicular volume, and spermatogenesis. It is unknown to what extent other variants are associated with FSH levels and therewith resemble causative factors for infertility.

OBJECTIVE

We aimed to identify further genetic determinants modulating FSH levels in a cohort of men presenting with idiopathic or unexplained infertility.

METHODS

We retrospectively (2010-2018) selected 1900 men with idiopathic/unexplained infertility. In the discovery study (n = 760), a genome-wide association study (GWAS) was performed (Infinium PsychArrays) in association with FSH values (Illumina GenomeStudio, v2.0). Minor allele frequencies (MAFs) were analyzed for the discovery and an independent normozoospermic cohort. In the validation study (n = 1140), TaqMan SNV polymerase chain reaction was conducted for rs11031005 and rs10835638 in association with andrological parameters.

RESULTS

Imputation revealed 9 SNVs in high linkage disequilibrium, with genome-wide significance (P < 4.28e-07) at the FSHB locus 11p.14.1 being associated with FSH. The 9 SNVs accounted for up to a 4.65% variance in FSH level. In the oligozoospermic subgroup, this was increased up to 6.95% and the MAF was enhanced compared to an independent cohort of normozoospermic men. By validation, a significant association for rs11031005/rs10835638 with FSH (P = 4.71e-06/5.55e-07) and FSH/luteinizing hormone ratio (P = 2.08e-12/6.4e-12) was evident.

CONCLUSIONS

This GWAS delineates the polymorphic FSHB genomic region as the main determinant of FSH levels in men with unexplained or idiopathic infertility. Given the essential role of FSH, molecular detection of one of the identified SNVs that causes lowered FSH and therewith decreases spermatogenesis could resolve the idiopathic/unexplained origin by this etiologic factor.

Identification of novel candidate biomarkers and immune infiltration in polycystic ovary syndrome

Background

In this study, we aimed to identify novel biomarkers for polycystic ovary syndrome (PCOS) and analyze their potential roles in immune infiltration during PCOS pathogenesis.

Methods

Five datasets, namely GSE137684, GSE80432, GSE114419, GSE138518, and GSE155489, were obtained from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) were selected from the train datasets. The least absolute shrinkage and selection operator logistic regression model and support vector machine-recursive feature elimination algorithm were combined to screen potential biomarkers. The test datasets validated the expression levels of these biomarkers, and the area under the curve (AUC) was calculated to analyze their diagnostic value. Quantitative real-time PCR was conducted to verify biomarkers’ expression in clinical samples. CIBERSORT was used to assess differential immune infiltration, and the correlations of biomarkers with infiltrating immune cells were evaluated.

Results

Herein, 1265 DEGs were identified between PCOS and control groups. The gene sets related to immune response and adaptive immune response were differentially activated in PCOS. The two diagnostic biomarkers of PCOS identified by us were HD domain containing 3 (HDDC3) and syndecan 2 (SDC2; AUC, 0.918 and 0.816, respectively). The validation of hub biomarkers in clinical samples using RT-qPCR was consistent with bioinformatics results. Immune infiltration analysis indicated that decreased activated mast cells (P = 0.033) and increased eosinophils (P = 0.040) may be a part of the pathogenesis of PCOS. HDDC3 was positively correlated with T regulatory cells (P = 0.0064), activated mast cells (P = 0.014), and monocytes (P = 0.024) but negatively correlated with activated memory CD4 T cells (P = 0.016) in PCOS. In addition, SDC2 was positively correlated with activated mast cells (P = 0.0021), plasma cells (P = 0.0051), and M2 macrophages (P = 0.038) but negatively correlated with eosinophils (P = 0.01) and neutrophils (P = 0.031) in PCOS.

Conclusion

HDDC3 and SDC2 can serve as candidate biomarkers of PCOS and provide new insights into the molecular mechanisms of immune regulation in PCOS.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13048-022-01013-0.

Identification of Prognosis Biomarkers for High-Grade Serous Ovarian Cancer Based on Stemness

In this paper, high-grade serous ovarian cancer (HGSOC) is studied, which is the most common histological subtype of ovarian cancer. We use a new analytical procedure to combine the bulk RNA-Seq sample for ovarian cancer, mRNA expression-based stemness index (mRNAsi), and single-cell data for ovarian cancer. Through integrating bulk RNA-Seq sample of cancer samples from TCGA, UCSC Xena and single-cell RNA-Seq (scRNA-Seq) data of HGSOC from GEO, and performing a series of computational analyses on them, we identify stemness markers and survival-related markers, explore stem cell populations in ovarian cancer, and provide potential treatment recommendation. As a result, 171 key genes for capturing stem cell characteristics are screened and one vital cancer stem cell subpopulation is identified. Through further analysis of these key genes and cancer stem cell subpopulation, more critical genes can be obtained as LCP2, FCGR3A, COL1A1, COL1A2, MT-CYB, CCT5, and PAPPA, are closely associated with ovarian cancer. So these genes have the potential to be used as prognostic biomarkers for ovarian cancer.

The role of miRNA-339-5p in the function of vascular endothelial progenitor cells in patients with PCOS

RESEARCH QUESTION

miRNA-339 participates in diseases with endothelial progenitor cell (EPC) dysfunction. What is the role of miRNA-339-5p in EPC of polycystic ovary syndrome (PCOS)?

DESIGN

Clinical data were collected from 76 controls and 84 PCOS patients. Noradrenaline, asymmetric dimethylarginine (ADMA), advanced glycation end products (AGE) and silent information regulator 1 (SIRT1) in the serum were measured. The functions of EPC and the expressions of PI3K, AKT, SIRT1 and PGC-1α in EPC before and after transfection with miRNA-339-5p mimic or miRNA-339-5p inhibitor were compared.

RESULTS

Serum concentrations of noradrenaline, ADMA and AGE were significantly higher (P = 0.009, P = 0.044, P < 0.001) and the SIRT1 concentration was significantly lower (P < 0.001) in PCOS patients, especially obese ones (P = 0.034, P = 0.032, P < 0.001, P = 0.023) than in the control group. When compared with the controls, proliferation of the EPC was slightly lower (without a significant difference), the migration and tubular formation were significantly decreased (P = 0.037, P = 0.011), the expression of miRNA-339-5p in EPC was significantly higher (P = 0.035) and the expressions of PI3K, AKT, SIRT1 and PGC-1α were significantly lower in the PCOS group (mRNA: P = 0.033, P = 0.027, P = 0.027, P = 0.032; protein: P = 0.036, P = 0.028, P = 0.039, P = 0.023). After transfection, the functions of EPC from PCOS patients were best in the miRNA-339-5p inhibitor group, and weakest in the miRNA-339-5p mimic group. The miRNA-339-5p inhibitor group had higher protein expressions of PI3K, AKT and SIRT1 but lower expression of PGC-1α in PCOS patients (P < 0.001, P = 0.030, P = 0.047, P = 0.003). Similar results were obtained from the controls after transfection.

CONCLUSION

Increased sympathetic excitation and damage to EPC were observed in PCOS patients, especially obese ones. Up-regulated miRNA-339-5p could inhibit the function of EPC by inhibiting the PI3K/AKT and SIRT1/PGC-1α signalling pathways.

Establishment and Analysis of a Combined Diagnostic Model of Polycystic Ovary Syndrome with Random Forest and Artificial Neural Network

Polycystic ovary syndrome (PCOS) is one of the most common metabolic and reproductive endocrinopathies. However, few studies have tried to develop a diagnostic model based on gene biomarkers. In this study, we applied a computational method by combining two machine learning algorithms, including random forest (RF) and artificial neural network (ANN), to identify gene biomarkers and construct diagnostic model. We collected gene expression data from Gene Expression Omnibus (GEO) database containing 76 PCOS samples and 57 normal samples; five datasets were utilized, including one dataset for screening differentially expressed genes (DEGs), two training datasets, and two validation datasets. Firstly, based on RF, 12 key genes in 264 DEGs were identified to be vital for classification of PCOS and normal samples. Moreover, the weights of these key genes were calculated using ANN with microarray and RNA-seq training dataset, respectively. Furthermore, the diagnostic models for two types of datasets were developed and named neuralPCOS. Finally, two validation datasets were used to test and compare the performance of neuralPCOS with other two set of marker genes by area under curve (AUC). Our model achieved an AUC of 0.7273 in microarray dataset, and 0.6488 in RNA-seq dataset. To conclude, we uncovered gene biomarkers and developed a novel diagnostic model of PCOS, which would be helpful for diagnosis.