Profiling Medical Journal Articles Using a Gene Ontology Semantic Tagger

Profiling Medical Journal Articles Using a Gene Ontology Semantic Tagger Mahmoud El-Haj1, Paul Rayson1, Scott Piao1 and Jo Knight2 1School of Computing and Communications and 2Medical School, Lancaster University, UK {m.el-haj, p.rayson, s.piao, jo.knight}@lancaster.ac.uk Abstract In many areas of academic publishing, there is an explosion of literature, and sub-division of fields into subfields, leading to stove-piping where sub-communities of expertise become disconnected from each other. This is especially true in the genetics literature over the last 10 years where researchers are no longer able to maintain knowledge of previously related areas. This paper extends several approaches based on natural language processing and corpus linguistics which allow us to examine corpora derived from bodies of genetics literature and will help to make comparisons and improve retrieval methods using domain knowledge via an existing gene ontology. We derived two open access medical journal corpora from PubMed related to psychiatric genetics and immune disorder genetics. We created a novel Gene Ontology Semantic Tagger (GOST) and lexicon to annotate the corpora and are then able to compare subsets of literature to understand the relative distributions of genetic terminology, thereby enabling researchers to make improved connections between them. Keywords: semantic tagger, ontology, genetics, medical 1. Introduction The explosion of scientific literature in all fields makes it hard to keep apace of new knowledge. This is particu- larly true in the relatively new field of genomics. For ex- ample, a search in the main citation database for biomedi- cal literature (PubMed) for the term ‘genome wide associ- ation study’ results in just 5 papers from 1995, 141 from 2005 and 3,633 from 2015. We contend that the myriad of techniques developed in Information Retrieval coupled with Natural Language Processing can help address these scaling and searching issues. Such a set of techniques could help in a myriad of ways, for example, summarisation of papers or a set of papers, collocation methods to investi- gate drug-disease-gene interactions, and query expansion where terminology varies from one subfield to another. Pre- viously such techniques have been used to perform tasks such as identifying gene-gene or gene-phenotype interac- tions (Bundschus et al., 2008; Kann, 2007). In addition, by using corpus comparison methods originating in Cor- pus Linguistics, we aim to identify key words and concepts emerging from a body of literature that will provide new clues to disease aetiology. In the remainder of this paper, we describe related work on biomedical text mining and corpus comparison. Then we explain how we created an open access corpus derived from medical journal abstracts, and a novel semantic tagger to apply a lexicon derived from a standard Gene Ontology. Finally, we illustrate how these new resources allow us to profile medical journal articles using domain-specific on- tologies. 2. Related Work Over a number of years, Natural Language Processing (NLP) techniques have been widely applied to biomedical text mining to facilitate large-scale information extraction and knowledge discovery from the rapidly increasing body of biomedical literature. Substantial efforts have been ded- icated to this research …

Content truncated to 500 words