BAGO: A Self-Optimizing Tool for LC-MS Gradient Design in Metabolomics.

Area of Science:

• Analytical Chemistry
• Metabolomics
• Chromatography

Background:

• Automated analytical method design is crucial for advancing metabolomics.
• Designing optimal liquid chromatography (LC) gradients is complex, hindering untargeted metabolomics.
• Current methods struggle to achieve comprehensive separation of all metabolites, known and unknown.

Purpose of the Study:

• To develop a self-optimizing framework, BAGO, for automated LC gradient design in mass spectrometry-based untargeted metabolomics.
• To enhance global metabolite detection by improving the separation of all compounds.
• To enable robust and structure-agnostic optimization across diverse sample types.

Main Methods:

• Implemented a data-driven Bayesian optimization process for iterative gradient improvement.
• Proposed a global separation index to quantify coelution for annotated and unannotated features.
• Benchmarked BAGO across four metabolomics assays with diverse sample matrices and conditions.

Main Results:

• BAGO achieved substantial improvements in LC gradient design within 10 optimization iterations.
• Optimized gradients increased Gaussian-shaped peaks, MS/MS acquisition rates, and annotated metabolites.
• Application to *Drosophila* metabolomics revealed a 41.9% increase in Gaussian peaks and 18 additional significant metabolites.

Conclusions:

• BAGO represents a significant step toward fully automated, self-optimizing experimental workflows in untargeted metabolomics.
• The framework enhances metabolite discovery and identification through improved chromatographic separation.
• BAGO is available as an open-source tool, promoting broader adoption in the field.