MetPC

MetPC performs preprocessing, metabolite identification, and biomarker discovery from metabolomics datasets using a hierarchical statistical model, an Expectation-Maximization (EM) algorithm, and two-dimensional false discovery rate (fdr2d) control.

Key Features:

• Preprocessing Capabilities: Incorporates preprocessing steps to transform raw metabolomics data into formats suitable for downstream analysis.
• Metabolite Identification: Implements a hierarchical statistical model fitted with an Expectation-Maximization (EM) algorithm to manage latent variables and iteratively estimate parameters for accurate metabolite detection.
• Biomarker Discovery: Applies control of the two-dimensional false discovery rate (fdr2d) to mitigate type I errors during multiple hypothesis testing in biomarker selection.

Scientific Applications:

• Metabolic pathway analysis: Identifies and characterizes metabolites to support analysis and reconstruction of metabolic pathways.
• Disease mechanism elucidation: Detects metabolite changes relevant to understanding disease mechanisms.
• Therapeutic target identification: Enables biomarker-driven identification of potential therapeutic targets.

Methodology:

Computational methods explicitly include data preprocessing, a hierarchical statistical model fitted with an Expectation-Maximization (EM) algorithm to handle latent variables, and control of the two-dimensional false discovery rate (fdr2d) for biomarker discovery.