Investigating the gut microbiome using metabolomics

Jessica Pandohee1 , Mary Boyce1, Scott Campbel2, John Moncur2, and David Broadhurst1

1Centre for Integrative Metabolomics and Computational Biology (CIMCB), School of Science, Edith Cowan University, WA 6027, Australia; 2SpectralWorks Limited, Cheschire, WA7 4EB, United Kingdom.

First Published: Scottish Metabolomics Network Syposium 2018


Metabolomics has a great potential in addressing the metabolic gap in microbe-host systems and expanding knowledge on the human gut microbial metabolism. Metabolomics as a field is the analysis and unbiased relative quantification of all metabolites in a biological sample. However in order to study the entire metabolome of a biological system, it is essential that most, if not all, of the metabolites present in the system in question are extracted and identified through analytical platforms. Human stool is a complex solid sample that contains indigestible food matter, bacterial biomass, as well as unwanted/excess metabolites in the body and metabolites produced from gut bacteria. Performing metabolomics on faecal samples for studying gut microbial metabolism is an idea still in its infancy. As a result, methods adapted from targeted short chain fatty acids or bile acids protocols have been utilised for faecal analyses. The aim of this presentation is to give an overview of optimised extraction methodologies to maximise the metabolite coverage.


Sample Preparation

Pooled samples were prepared by homogenising equal amounts of faecal material from five healthy individuals. 100 mg of material was then weighed in centrifuge tubes and stored at -80 ºC.

Table 1. Characteristics of solvents tested for metabolite extractions.

Figure 1. Extraction protocol for faecal material.

Table 2. GC-MS Conditions.

Data Processing

RAW files were analysed using AnalyzerPro (SpectralWorks Limited, UK). A targeted approach was performed using CIMCB’s in-house high resolution accurate mass metabolomics library developed by running pure metabolite standards on the QE GC-MS.


Figure 2. Chromatographic profiles of faecal extracts in four solvents.


A target library of 288 metabolites was used to do a preliminary screen for a mixture of amino acids, sugars, organic acids, peptides, fatty acids and bile acids.

Figure 3. Hierarchical Cluster Analysis of peak areas.

A total of 161 metabolites were identified across the 12 extraction solvents  (Level 1 annotation, Metabolomics Standards Initiative).

Three major clusters were observed:

Cluster A shows that polar solvents such as MeOH, IPA and water at 4 different pH extracts mainly polar metabolites and specifically the organic acids, sugars, amino acids and amines;

Cluster B shows that non-polar solvents extract the fatty acids, steroids, but not the more polar metabolites; and

Cluster C shows a group of metabolites that are extracted in solvents with intermediate polarities. More interestingly, some of these compounds are not extracted in neither highly polar nor highly non-polar solvents.

Acetone is a promising solvent that extracts the most metabolites (90) within the 161 compounds studied, followed by MeOH (84).


  • Polarity of solvents used for metabolite extractions affects the coverage of compounds.
  • To perform truly untargeted metabolomics, it is essential to use solvents of different polarities to maximise the metabolite coverage in a study
  • Based on the 161 metabolites studied, acetone is the best solvent for a targeted analysis of aqueous soluble metabolites and ethyl acetate for a targeted analysis of non-aqueous soluble metabolites.
  • The target library is being expanded to screen for a wider range of compounds.

Content retrieved from: