Disparate Metabolomics Data Reassembler: A Novel Algorithm for Agglomerating Incongruent LC-MS Metabolomics Datasets

Area of Science:

• Biochemistry
• Analytical Chemistry
• Bioinformatics

Background:

• Liquid chromatography-mass spectrometry (LC-MS) metabolomics has emerged as a powerful tool for studying metabolic processes.
• The field is currently fragmented due to incompatible datasets arising from diverse instrumentation and data formats.
• This lack of data coherency hinders the accumulation of large-scale datasets necessary for significant discoveries.

Purpose of the Study:

• To develop a novel algorithm, Disparate Metabolomics Data Reassembler (DIMEDR), for integrating incongruent LC-MS metabolomics datasets.
• To bridge the inconsistencies between datasets from different studies and platforms.
• To enable meaningful cross-study analysis of metabolomics data.

Main Methods:

• DIMEDR utilizes a primary dataset as a template for reassembling and integrating subsequent disparate datasets.
• A novel procedure for universal retention time correction and comparison is employed.
• Ubiquitous features in the primary dataset are identified and used as endogenous internal standards for data integration.

Main Results:

• DIMEDR successfully unified two human and two mouse urine metabolomics datasets from four unrelated studies.
• The algorithm maximized spectral feature similarity across all integrated samples.
• Meaningful analysis was enabled across datasets that were previously incomparable.

Conclusions:

• DIMEDR provides a robust framework for overcoming data fragmentation in LC-MS metabolomics.
• The algorithm facilitates the creation of unified metabolomics datasets, enabling larger and more impactful biological discoveries.
• This approach has the potential to accelerate advancements in the field of metabolomics.