Using Published Pathway Figures in Enrichment Analysis and Machine Learning

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Existing pathway databases vary in scope and detail.
• There is a need for pathway resources that integrate visual, mechanistic information with literature support.
• Published pathway figures are a rich, underutilized source of biological information.

Approach:

• Pathway Figure OCR (PFOCR) extracts pathway data from published figures.
• PFOCR content is compared against established pathway databases for coverage.
• Advanced case studies explore PFOCR's utility in cancer subtype and grade prediction.

Key Points:

• PFOCR provides a novel pathway database with extensive mechanistic diagrams and literature links.
• The database approaches the breadth and depth of Gene Ontology.
• PFOCR content is extracted from published pathway figures at a high rate.
• PFOCR demonstrates competitive and in some cases superior coverage compared to existing databases.
• PFOCR offers unique advantages for advanced analyses like cancer subtype and grade prediction.

Conclusions:

• PFOCR represents a significant advancement in pathway database development.
• Its approach provides rich, mechanistic insights and supports diverse analytical applications.
• PFOCR enhances the utility of published pathway figures for biological research and precision medicine.