Discovering breast cancer drug candidates from biomedical literature

Area of Science:

• Pharmacology
• Bioinformatics
• Drug Discovery

Background:

• Traditional drug discovery is time-consuming and expensive.
• Integrating existing knowledge and literature data can accelerate the identification of potential drug candidates.
• Molecular-level evidence is crucial for understanding drug-disease relationships.

Purpose of the Study:

• To develop and implement a new paradigm for disease-specific drug candidate discovery.
• To build a literature-mining framework for extracting molecular evidence.
• To perform drug-protein association mining for breast cancer drug discovery.

Main Methods:

• Developed a novel literature-mining framework.
• Utilized a prototype system for data extraction and analysis.
• Performed drug-protein association mining focused on breast cancer.
• Validated enriched drug candidates based on molecular pharmacology.

Main Results:

• Identified 729 enriched drugs in the 'Organic Chemicals' category for breast cancer.
• Validated 79.2% of these drugs as disease-related.
• Identified 'Doxorubicin', 'Etoposide', and 'Paclitaxel' as having similar pharmacological profiles for breast cancer treatment.
• Highlighted 20.8% of identified drugs as potential candidates for future molecular therapeutics studies.

Conclusions:

• The developed paradigm enables disease-specific drug candidate discovery using molecular evidence from literature.
• The literature-mining framework is effective for identifying drug-protein associations.
• The study successfully validated existing breast cancer drugs and identified potential new therapeutic agents.