Ms-mint: Targeted metabolomics with Python

ABSTRACT:

Metabolomics has emerged as a mainstream approach for investigating the complex metabolic underpinnings of living systems. Over recent years, metabolomics has increasingly been applied to large cohort studies that tax the limits of existing computational tools. Though effective when applied to small datasets, many existing metabolomics software tools have three common shortcomings that limit their utility when applied to large studies: they store entire datasets in memory, they use batch-dependent fitting algorithms, and they use ambiguous representation of regions of interests leading to inconsistent peak-picking results across samples. To address this, we developed the mass-spectrometry metabolomics integrator (ms-mint), a Python library for processing, analyzing, and visualizing large LC-MS datasets. To enable reproducible large scale data processing tasks, ms-mint uses a region of interest (ROI)-based approach to extract data. We illustrate the function of this new tool using a large dataset (3,334 files) of standards and Staphylococcus aureus metabolite extracts. Via regression analyses, we show that ms-mint faithfully reproduces data generated from other software tools (R2 > 0.9999). Moreover, once ms-mint processing parameters have been set, analyses are fully automated and efficient (3334 files in less than 10 minutes) while maintaining a minimal memory footprint (~200MB). We demonstrate how the extendable open-source library can be used to efficiently process very large datasets using a deterministic and batch-independent protocol enabling targeted metabolomics at scale.

INSTRUMENT(S): Liquid Chromatography MS - negative - hilic

PROVIDER: MTBLS6696 | MetaboLights | 2025-05-21

REPOSITORIES: MetaboLights