Abstract
Background: Sarcoidosis is a complex disorder characterized by parenchymal lung involvement, intrathoracic lymphadenopathy, and extrapulmonary manifestations. Sarcoidosis can cause high mortality and morbidity.

Aims and Objectives: It is needed computational methods to estimate potential miRNA-disease-drug relationships in pulmonary sarcoidosis since experimental approaches are expensive and time consuming.

Materials and Methods: In this study, we performed principal component analysis -based unsupervised feature extraction method to GSE34608 microRNA expression profile, which was downloaded from the Gene Expression Omnibus database that consists of peripheral blood samples of eight pulmonary sarcoidosis patients and eight healthy controls.

Results: We detected a set of 100 microRNAs that could successfully discriminate pulmonary sarcoidosis patients from healthy controls with 81.2% accuracy. Among these miRNAs, we validated miR-15b by Human MicroRNA Disease Database and identified 22 pathways by miRpath and 20 approved drugs by PharmacomiR. When these pathways were integrated with drug-affected pathways detected by the database for annotation, visualization, and integrated discovery database, four overlapping pathways were determined, which were closely associated with cisplatin.

Conclusion: This study provides novel insight into pulmonary sarcoidosis pathogenesis, and we identified potential drug candidate for this disease.

Key words: Pulmonary Sarcoidosis; microRNA; Principal Component Analysis -based Unsupervised Feature Extraction; Cisplatin