metabomxtr

metabomxtr performs mixture-model analysis to model missingness in non-targeted metabolomics data, distinguishing undetectability from true absence and estimating parameters and likelihoods for truncated normal or lognormal distributions.

Key Features:

• Mixture-Model Analysis: Implements mixture models that distinguish true absence from undetectability and model truncated measurements under normal or lognormal distributions.
• Optimized Parameter Estimation: Returns optimized parameter estimates and log-likelihoods for fitted mixture models.
• High-Throughput Compatibility: Processes large-scale non-targeted metabolomics datasets typical of high-throughput studies.

Scientific Applications:

• Handling missingness in non-targeted metabolomics: Accounts for missing values arising from detection limits or true absence in non-targeted metabolomics datasets.
• Metabolic pathway and biomarker inference: Improves inference of metabolite presence and abundance to support metabolic pathway analysis and biomarker discovery.

Methodology:

Fits mixture models to truncated metabolomics data assuming normal or lognormal components and returns optimized parameter estimates and log-likelihoods.