Exploiting microRNA Expression Data for the Diagnosis of Disease Conditions and the Discovery of Novel Biomarkers

MicroRNAs (miRNAs) play key roles in diseases and their detection in circulating biofluids makes them optimal candidate as disease biomarkers for improved diagnosis, prognosis, and therapy. Therefore, a thorough understanding of miRNAs in diseases is crucial for realizing their clinical potential. This paper introduces a pipeline of analysis aimed at predicting disease conditions of patients by leveraging the large amount of data available in ExomiRHub, a database that stores extracellular and intracellular miRNAs data from multiple clinical studies. The proposed pipeline solves the inconsistencies raised by the integration of data collected and stored following different protocols, and enables learning predictive models from a proper amount of training data. We also show how the learned models from such data can be exploited to identify novel disease biomarkers by means of explainability techniques. Our experiments show the effectiveness of predicting disease conditions using microRNA expression values, as well as the potential of such models as a tool to discover novel non-invasive disease biomarkers.