- We present an ongoing effort on utilizing text mining methods and existing biological ontologies to help readers to access the information contained in the scientific articles. Our approach includes using multiple strategies for biological entity detection and using association analysis on extracted analysis. The entity extraction processes utilizes regular expression rules, ontologies, and keyword dictionary to get a comprehensive list of biological entities. In addition to extract list of entities, we also apply natural language processing and association analysis techniques to generate inferences among entities and comparing to known relations documented in the existing ontologies.
To address this challange, we present software developments from an ongoing project, DIVE, which features auto extraction of informational vocabulary, web based access and curation tools. The framework implements several strategies in entity extraction, including using regular expression rules, ontology and a keyword dictionary. The results of the extracted biological entities are then stored in a database and made accessible through an interactive web application for curation and evaluation by authors and other domain experts. Additional text mining and associaiton analysis can be run on extrated entities to help readers understanding of the paper. The system can benefit the entire life cycle of the digital publication, from initial manuscript submission to publishing the article and presenting information to readers. New information defined and verified by experts may also be injected to other information resources.
We implemented a rule-based approach for processing the text and structure in order to identify informational vocabulary candidates. The detection rules can be defined based on various heuristics and requirements such as publishing requirements, naming conventions, and domain ontologies. New rules can be added on demand over time. Currently, there are four types of rules implemented in the DIVE, regular expression rules, word dictionary, publishing convention, and ontology rules.
The regular expression rules utilize common naming
conventions to identify biological entities, such as gene name,
protein name, molecule structures, chemical compound, etc.
Each rule can be defined as a regular expression and used for
matching the candidate word. The word dictionary rule
consists of a pre-defined list of words that should be included
or excluded in the candidate lists. The publication content is
searched against the list at run time. The publishing
convention rules are used to identify words that are in special
format, such as in italic, or in a particular component of the
publication, such as a figure legend. The enclosing tags of the
candidates are used to define each rule. Additional rules can
be added by specifying additional tag values or by using
naming conventions to detect entities like species names. The
ontology rules utilize five biological ontology including gene
ontology [
The data association analysis can be used to generate
inferences between values from two or more fields of the data
in a given condition using FP-Growth algorithm[
Figure 2 shows top 20 inference rules based on all ontology terms extracted from the collection. Each label indicates a frequent item set found in the collection. The directional arrow indicates an inference on co-occurrence between two item sets. The shade of the directional arrow indicates the confidence level of the rule.
Such visual representations of inferred association between diverse entity types could tremendously aid a researcher in forming insights. This also has potential to be a similarity metric between articles that could help editors gauge the novelty of a new article submission. We are continuing working on evaluating the performance of the entity extraction over large data set and improving its accuracy. We are gathering feedback from domain researchers and publishing professionals for further entities candidate evaluations. We are also working on comparing the inferences from association analysis with known relationships documented in the existing ontologies.
DIVE is partially supported by CyVerse (NSF award DBI0735191 and DBI-1265383) and the Gramene, a Comparative Plant Genomics Database (NSF award IOS-1127112).