CrcBiomeScreen

CrcBiomeScreen implements a reproducible machine learning framework for microbiome-based colorectal cancer screening through systematic evaluation of normalization strategies, taxonomic resolutions, and class imbalance handling.

Key Features:

• Modular Pipeline Execution: Users can apply the full pipeline or selectively run specific analytical components depending on their needs.
• Normalization Strategy Evaluation: Systematically evaluates multiple normalization strategies for microbiome data.
• Taxonomic Resolution Assessment: Evaluates multiple taxonomic resolutions for optimal classification performance.
• Class Imbalance Handling: Incorporates methods to address class imbalance in machine learning classification.
• Interpretable Models: Establishes a foundation for developing interpretable microbiome-based screening tools.

Scientific Applications:

• Colorectal Cancer Screening: Supports early detection of colorectal cancer through microbiome-based classification models.
• Microbiome Biomarker Discovery: Identifies microbial features associated with colorectal cancer for potential diagnostic use.

Methodology:

The framework systematically evaluates normalization strategies, taxonomic resolutions, and class imbalance handling within a machine learning pipeline for microbiome-based colorectal cancer screening.