DISNET: A framework for extracting phenotypic disease information from public sources

Abstract

Within the global endeavour of improving population health, one major challenge is the increasingly high cost associated with drug development. Drug repositioning, i.e. finding new uses for existing drugs, is a promising alternative; yet, its effectiveness has hitherto been hindered by our limited knowledge about diseases and their relationships. In this paper we present DISNET (Drug repositioning and disease understanding through complex networks creation and analysis), a web-based system designed to extract knowledge from signs and symptoms retrieved from medical data bases, and to enable the creation of customisable disease networks. We here present the main functionalities of the DISNET system. We describe how information on diseases and their phenotypic manifestations is extracted from Wikipedia, PubMed and MayoClinic; specifically, texts from these sources are processed through a combination of text mining and natural language processing techniques. We further present a validation of the processing performed by the system; and describe, with some simple use cases, how a user can interact with it and extract information that could be used for subsequent analyses.\n\nDatabase URL: http://disnet.ctb.upm.es