University of British Columbia
Building and inferring knowledge bases using biomedical text mining
Abstract
dc:descriptionBiomedical researchers have the overwhelming task of keeping abreast of the latest research. This is especially true in the field of personalized cancer medicine where knowledge from different areas such as clinical trials, preclinical studies, and basic science research needs to be combined. We propose that automated text mining methods should become a commonplace tool for researchers to help them locate relevant research, assimilate it quickly and collate for hypothesis generation. To move towards this goal, we focus on extracting relations from published abstracts and full-text papers. We first explore the use of co-occurrences in sentences and develop a method for inferring new co-occurrences that can be used for hypothesis generation. We next explore more advanced relation extraction methods by developing a supervised learning method, VERSE, which won part of the BioNLP 2016 Shared Task. Our classical method outperforms a deep learning method showing its applicability to text mining problems with limited training data. We develop it further into the Kindred Python package which integrates with other biomedical text mining resources and is easily applied to other biomedical problems. Finally, we examine the applicability of these methods in personalized cancer research. The specific role of genes in different cancer types as drivers, oncogenes, and tumor suppressors is essential information when interpreting an individual cancer genome. We built CancerMine, a high-quality knowledgebase, using the Kindred classifier and annotations from a team of annotators. This allows for quantifiable comparisons of different cancer types based on the importance of different genes. The clinical relevance of cancer mutations is generally locked in the raw text of literature and was the focus of the CIViCmine project. As a collaboration with the Clinical Interpretation of Variants in Cancer (CIViC) project team, we built methods to prioritise relevant papers for curation. Through this work, we have focussed on different ways to extract structured knowledge from individual sentences in biomedical publications. The methods, guidelines, and results developed will aid biomedical text mining research and the personalized cancer treatment community.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy - PhD
- Level thesis:degree_level
- doctoral
- Discipline thesis:degree_discipline
- Bioinformatics
- Grantor dc:publisher
- University of British Columbia
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Lever, Jake
Rights
dc:rights- Statement dc:rights
-
- Attribution-ShareAlike 4.0 International
- Language dc:language
- eng
Identifiers
dc:identifier.*- Handle dc:identifier
- http://hdl.handle.net/2429/67285