Related Experiment Video
Updated: May 15, 2026
07:41
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
Mining biomarker information in biomedical literature
Erfan Younesi1, Luca Toldo, Bernd Müller
1Department of Bioinformatics, Fraunhofer Institute for Algorithms and Scientific Computing (SCAI), Schloss Birlinghoven, Sankt Augustin, 53754, Germany.
BMC Medical Informatics and Decision Making
|December 20, 2012
Summary
This study developed a specialized biomarker terminology to improve the automated retrieval of biomarker information from scientific literature. The approach enhances accuracy and aids in hypothesis generation for drug discovery.
Area of Science:
- Biomedical Informatics
- Computational Biology
- Text Mining
Background:
- Biomarker discovery relies heavily on published literature, but existing text-mining tools lack specificity for retrieving relevant information.
- Vast amounts of biomarker knowledge remain unexplored due to limitations in current named entity recognition (NER) approaches.
Purpose of the Study:
- To identify textual features that enhance biomarker information retrieval from biomedical texts.
- To improve the effectiveness of biomarker discovery for various disease areas and user needs.
Main Methods:
- A comprehensive biomarker terminology was created, organized into six concept classes and optimized for selectivity and specificity.
- Information retrieval performance was evaluated using combinations of the terminology's classes.
- Validation was performed on corpora for two neurodegenerative diseases.
Main Results:
- The terminology includes 119 entity classes and 1890 synonyms.
- Combining clinical management terms and gene/protein alteration evidence with disease/gene recognition improved retrieval rates.
- Filtering with classes like diagnostic or prognostic methods significantly reduced unspecific search results.
- The approach enables automated identification of biomarker information.
Conclusions:
- A dedicated biomarker terminology aids in the automated analysis of scientific literature for biomarker discovery.
- Extracting candidate biomarker information supports novel hypothesis generation.
- This process is valuable for early-stage decision-making in drug discovery and development.
