Mining microRNA information from biomedical literature

Abstract

MicroRNAs (miRNAs) are essential gene regulators whose dysregulation often leads to diseases. Since their discovery, the number of related publications has been growing exponentially. Easy access to such information is crucial for researchers to efficiently exploit the existing knowledge and connect different facts about miRNAs. To keep up with the growing literature, text mining methods are increasingly being used to automatically extract relations between a miRNA and a target, a process, or a disease, to assist in scaling up database curation and knowledge discovery. However, to efficiently connect the different facts, it is important to integrate the results of these relation extraction methods and build a knowledge map, where various aspects of miRNAs are visible simultaneously. Here, we present an integrative text mining approach to collect miRNA information from the literature. ☐ We have developed a web-based resource that integrates different miRNA information by mining the entire miRNA research literature. We developed an interface that allows the users to visualize and navigate all this extracted information, where we normalize the detected entities using standard ontologies to allow for easy querying. A second part of the work is concerned with the development of new relation extraction tools, with the focus on relations that are highly important in miRNA research. The final part of this work focuses on a novel way of summarizing the miRNAs, with an emphasis on the role of miRNA in gene regulation and their subsequent impact on diseases, cellular processes and signaling pathways. Each part in this work has been implemented with easy access to extracted information and they have been evaluated and shown to achieve high precision and recall.