ABCstats

ABCstats applies an Adaptive Box-Cox (ABC) transformation to metabolomic datasets to improve data normality for accurate statistical identification of significantly different metabolic features.

Key Features:

• Adaptive Box-Cox Transformation: Dynamically adjusts a power parameter based on normality test results to improve distributional normality compared with conventional log and square root transformations.
• Distribution Classification: Fits metabolic features into nine distinct beta distributions and identifies the prevalence of two low-normality types within metabolomic data.
• Monte Carlo and Empirical Validation: Demonstrates superior normalization performance for positively and negatively skewed distributions using Monte Carlo simulations and real-world metabolomic studies.
• Increased Discovery of Significant Metabolites: In one real study with three pairwise comparisons, ABC transformation yielded additional 84, 44, and 57 significant metabolites, corresponding to increases of 70.6%, 13.4%, and 22.9%, respectively.
• Biological Relevance: Newly identified metabolites contribute to potential insights into metabolic pathways and biomarker identification.

Scientific Applications:

• Differential Metabolite Analysis: Improves data distribution characteristics to support accurate statistical detection of significantly different metabolic features.
• Biomarker and Pathway Discovery: Increases identification of candidate metabolites for downstream biological interpretation and pathway analysis.

Methodology:

Adaptive Box-Cox transformation that dynamically adjusts a power parameter based on normality test results; classification of features into nine beta distributions; validation using Monte Carlo simulations and application to real-world metabolomic studies including pairwise comparisons.