Pathway curator: an online webserver extracting genes and interactions from figures

Abstract

In the biomedical literature, gene pathways are frequently included. Many high-quality gene pathways are illustrated in the form of visuals and text, making them valuable study tools for biological processes and precision medicine. Pathway maps and literature texts provide researchers with access to a huge number of new biological treatments. For general usage, these pathway maps should be logically ordered, coordinated, and converted into a computer-readable format. Currently, keeping up with the rapid increase of the literature requires laborious extraction of information from a publication at a time. A gene pathway map recognition system is devised and implemented in this study. Based on the pathway map and relevant information supplied by users, the system extracts gene identity and gene interaction information, and the automated extraction from pathway maps is efficient. Furthermore, the tool offers users with a full view of a certain illness's pathway, which is useful for researchers and can speed up the research process in a variety of biomedical applications. This thesis first explains the project's goal and provides the background information. The project's design ideas are then presented, as well as an analysis of the system and introductions to related platforms. After that, the system's implementations are described one by one, together with the deployment and testing processes. Finally, potential improvements and future work are discussed.