CRPMKB

CRPMKB provides a knowledge base that integrates and enables systematic comparison of cancer risk prediction models to support evaluation of predictive performance and development of personalized cancer risk assessment.

Key Features:

• Extensive Database: CRPMKB hosts 802 cancer risk prediction models.
• Systematic Comparison: Enables systematic evaluation of model accuracy and assessment of regional differences, cancer types, and modeling methodologies on predictive performance.
• Variable Categorization: Categorizes model variables into environmental factors, behavioral lifestyle, biological genetics, and clinical examinations.
• Pathway Enrichment Analysis: Analyzes 50 genes from lung cancer risk prediction models and identifies enrichment in p53 Signaling and Aryl Hydrocarbon Receptor Signaling.
• Biological Validation: Validates overlapping gene significance using the STRING database.
• Personalized Model Development: Supports development and stratification of models based on demographic characteristics and specific cancer types to improve prediction accuracy.

Scientific Applications:

• Comparative model evaluation: Conducts comparative studies of existing risk prediction models to assess accuracy across regions, cancer types, and methodologies.
• Variable and pathway identification: Identifies key variables and enriched pathways associated with different cancers, including p53 and Aryl Hydrocarbon Receptor signaling in lung cancer.
• Development of personalized prevention strategies: Informs development of personalized cancer prevention and risk stratification approaches based on demographic and genetic variables.

Methodology:

Integration of diverse cancer risk prediction models; analysis of model accuracy and variable distribution; pathway enrichment analyses; validation of gene significance using the STRING database.