MetaboLINK Is a Novel Algorithm for Unveiling Cell-Specific Metabolic Pathways in Longitudinal Datasets

Authors

Jared Lichtarge
Gerarda Cappuccio
Soumya Pati
Alfred Kwabena Dei-Ampeh
Senghong Sing
LiHua Ma
Zhandong Liu
Mirjana Maletic-Savatic

Publication Date

1-1-2024

Journal

Frontiers in Neuroscience

DOI

10.3389/fnins.2024.1520982

PMID

39872998

PMCID

PMC11769959

PubMedCentral® Posted Date

1-13-2025

PubMedCentral® Full Text Version

Post-print

Published Open-Access

yes

Keywords

metabolome, PCA, Glasso, MetaboLINK, hESC, embryonic bodies, rosettes, neuroprogenitors

Abstract

INTRODUCTION: In the rapidly advancing field of 'omics research, there is an increasing demand for sophisticated bioinformatic tools to enable efficient and consistent data analysis. As biological datasets, particularly metabolomics, become larger and more complex, innovative strategies are essential for deciphering the intricate molecular and cellular networks.

METHODS: We introduce a pioneering analytical approach that combines Principal Component Analysis (PCA) with Graphical Lasso (GLASSO). This method is designed to reduce the dimensionality of large datasets while preserving significant variance. For the first time, we applied the PCA-GLASSO algorithm (i.e., MetaboLINK) to metabolomics data derived from Nuclear Magnetic Resonance (NMR) spectroscopy performed on neural cells at various developmental stages, from human embryonic stem cells to neurons.

RESULTS: The MetaboLINK analysis of longitudinal metabolomics data has revealed distinct pathways related to amino acids, lipids, and energy metabolism, uniquely associated with specific cell progenies. These findings suggest that different metabolic pathways play a critical role at different stages of cellular development, each contributing to diverse cellular functions.

DISCUSSION: Our study demonstrates the efficacy of the MetaboLINK approach in analyzing NMR-based longitudinal metabolomic datasets, highlighting key metabolic shifts during cellular transitions. We share the methodology and the code to advance general 'omics research, providing a powerful tool for dissecting large datasets in neurobiology and other fields.