Identifying miRNA biomarkers of polycystic ovary syndrome through text mining

Abstract

Objective: Polycystic ovary syndrome (PCOS) is an endocrine disorder with diverse clinical manifestations that often occurs in women of childbearing age. However, its molecular pathogenesis remains unclear, and this study aimed to identify miRNA targets in PCOS through text mining and database analysis. Methods: First, three different sets of text mining genes (TMGs) associated with “polycystic ovary syndrome”, “obesity/adiposis”, and “anovulation” keywords were retrieved from the GenCLiP3 database, and overlapping genes were selected. Second, Gene ontology annotation and biological pathway enrichment analyses of these overlapping TMGs were performed, followed by protein–protein interaction (PPI) network analysis. Third, genes in the gene module clustered in the PPI were selected to predict potential miRNAs for PCOS via miRNA-mRNA analysis. Results: A total of 4291 TMGs related to three different keywords were obtained through text mining; 72 intersect TMGs were retained among the three gene sets, and 62 TMGs participated in the establishment of the PPI network, of which 18 were aggregated in the gene module. Finally, 11 miRNAs that simultaneously bound to two TMGs (IGF1, ESR1, MAPK1, NAMPT, PIK3CA, and SERPINE1) could be prioritized as targets to study PCOS. Conclusion(s): The discovery of 11 miRNAs (miR-301a-3p, miR-301b-3p, miR-3666, miR-454-3p, miR-130a-3p, miR-130b-3p, miR-4295, miR-190a-3p, miR-5011-5p, miR-548c-3p, and miR-4799-5p) and 6 TMGs, which are associated with the HIF-1 signaling pathway (P = 4.799E-08), could be used as potential targets for PCOS.