Using text mining techniques to assist gene related annotation
Date
2017
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Biomedical researchers usually describe their experimental results in research publications. With the rapid growth of biomedical publications, the information of interest needs to be extracted automatically to avoid the time consuming and labor intensive process. ☐ In this dissertation, we seek to help in the process of gene related annotations, by extracting the unstructured information buried in the literature and providing means to structure the extracted information. We start by recognizing gene names in the literature and linking them to database records. Based on this work, we develop a system which automatically selects articles that are about a specific UniProt protein entry. Next, we describe our approach for mining information related to another important bio-named entity, protein complex. Finally, we present our work on assisting the curation of gene annotation. ☐ In each of these tasks, we conduct experiments to evaluate the efficacy of our approaches. The results show that our systems achieve good performances, and can be used to assist the annotation process related to genes.
Description
Keywords
Applied sciences, Machine learning, NLP, Text mining