Integrating untargeted metabolomics, genetically informed causal inference, and pathway enrichment to define the obesity metabolome

bioinformatics
causal inference
mendelian randomization
metabolomics
multi-omics
  • Approach: A multi-omics framework was developed, combining untargeted metabolomics (measuring both known and unknown metabolites) with two-sample Mendelian Randomization (MR) using metabolite-QTLs (mQTLs).
  • Causal Findings: 23 metabolites (15 known and 8 unknown) were identified as causally associated with BMI, with specific pathways like amino acid catabolism and lipid metabolism being implicated in the obesity metabolome.
  • Innovation: A novel pathway enrichment method was used to infer the metabolic function of the causally associated unknown metabolites based on their shared genetic links with identified metabolites.
Published

23 January 2026

PubMed: 32467615 DOI: 10.1038/s41366-020-0603-x Overview generated by: Gemini 2.5 Flash, 27/11/2025

Background and Objective

Obesity is associated with significant metabolic disruption. Understanding the causal connections between obesity and changes in metabolite levels is crucial for identifying new intervention targets. Previous studies using Mendelian Randomization (MR) to infer causality between metabolites and obesity often ignored the majority of data generated by untargeted metabolomics—specifically, the signals from unknown or unidentified metabolites.

This study aimed to develop a comprehensive framework that integrates untargeted metabolomics, genetics, and pathway enrichment analysis to: 1. Identify a broad spectrum of metabolites causally related to Body Mass Index (BMI). 2. Characterize the biological pathways involved in the obesity metabolome, including those represented by unidentified metabolites.

Study Methods

The authors utilized a multi-stage approach across three large cohorts:

  1. Metabolomic Profiling: Untargeted metabolomic profiling was conducted on samples from the Framingham Heart Study (FHS), generating quantitative signals for both known (identified) and unknown (unidentified) metabolites.
  2. Genome-Wide Association Study (GWAS): A GWAS was performed to identify metabolite-QTLs (mQTLs)—genetic variants associated with the levels of both known and unknown metabolites. These mQTLs served as the instrumental variables (IVs) for the subsequent MR analysis.
  3. Two-Sample Mendelian Randomization (MR):
    • The mQTLs identified from the FHS metabolomics GWAS were used as IVs.
    • The exposure was each metabolite (known and unknown).
    • The outcome was BMI, using summary statistics from a large published BMI GWAS.
    • Multivariable MR was applied to test for independence among groups of putatively co-regulated metabolites.
  4. Pathway Enrichment Analysis: To interpret the role of the unknown metabolites, a novel Pathway Enrichment Analysis method was developed. This method uses the shared genetic architecture (i.e., common mQTLs) between unknown and known metabolites to infer the metabolic pathway of the unknown metabolites.

Key Results and Findings

Causal Metabolites

  • The study identified 23 metabolites (15 known and 8 unknown) that were causally associated with BMI.
  • Pro-Obesity Metabolites: Metabolites whose higher levels were causally linked to higher BMI included certain amino acid catabolites, medium-chain acylcarnitines, and metabolites related to the carnitine cycle and fatty acid oxidation.
  • Anti-Obesity Metabolites: Metabolites whose higher levels were causally linked to lower BMI included glycine, acetylated/formylated amino acids, and several glycerophospholipids.

Defining the Obesity Metabolome

The integrated analysis defined the obesity metabolome as being characterized by a metabolic shift involving: 1. Impaired Amino Acid Catabolism: Specifically, branched-chain amino acid (BCAA) and aromatic amino acid catabolites showed a strong causal link with higher BMI. 2. Dysfunctional Lipid Metabolism: The enrichment analysis linked BMI to pathways of lipid metabolism, particularly those involving glycerophospholipids and the carnitine shuttle.

Importance of Unknown Metabolites

The inclusion of 8 causally-associated unknown metabolites expanded the definition of the obesity metabolome. The novel pathway enrichment approach successfully mapped these unknown metabolites to relevant metabolic pathways (e.g., lipid and amino acid metabolism) based on their shared genetic control with known metabolites.

Conclusions and Significance

The comprehensive framework successfully leveraged the power of untargeted metabolomics in combination with genetically informed causal inference (MR) and a novel pathway enrichment tool. This approach provides a more complete, systems-level view of the metabolic disturbances associated with obesity.

The identified causal metabolites and pathways offer promising avenues for targeted therapeutic and diagnostic interventions aimed at treating obesity and its related metabolic diseases.