BQuant

BQuant quantifies and identifies metabolites in 1H NMR spectroscopy data using a Bayesian probabilistic framework to resolve overlapping peaks and estimate metabolite concentrations.

Key Features:

• Probabilistic Bayesian model selection: Uses Bayesian model selection within a probabilistic framework to infer metabolite identities and abundances from spectral data.
• Reference profile database: Represents NMR spectra as mixtures of reference profiles drawn from a comprehensive database of metabolite spectra.
• Linear mixed model: Incorporates a linear mixed model to account for technological characteristics and experimental variability in NMR spectra.
• Automation and accuracy: Performs automated identification and quantification and has been demonstrated to outperform existing automated alternatives in accuracy.

Scientific Applications:

• High-throughput metabolomics: Applied to rapid analysis of complex biological samples in large-scale metabolomic studies.
• Plasma metabolomics: Used for identification and quantification of metabolites in plasma samples.
• Simulated datasets: Validated on simulated NMR datasets to assess performance under controlled conditions.
• Spike-in experiments: Applied in spike-in experiments to validate quantification accuracy using known metabolite additions.

Methodology:

Represent NMR spectra as mixtures of reference profiles from a database and apply Bayesian model selection to infer identities and concentrations, with a linear mixed model used to model technological variability.