How Much Stool Do I Need for Metabolomics?
About 50 mg of stool per sample. That is the amount our lab requests for untargeted metabolomics, and it is exactly what Zhao and colleagues extracted in Frontiers in Endocrinology from 44 adults with cardiovascular-kidney-metabolic (CKM) syndrome, pairing 50 mg of feces in 4:1 methanol:water with 100 µL of fasting serum. Mass turned out to be the easy part. The harder question is whether the stool layer adds anything a blood draw would miss, and this paper gives an unusually candid answer.
Key Takeaways
Untargeted stool metabolomics needs about 50 mg of feces per sample; a 2026 study of cardiovascular-kidney-metabolic syndrome extracted exactly that amount.
Homogenizing the specimen before weighing and normalizing to total signal matter more than adding mass, because stool water content and diet vary between donors.
In 44 adults with CKM syndrome, serum showed 470 corrected metabolite differences by stage and stool showed six, so pair stool with a blood draw.
Cross-matrix correlations are fragile in small cohorts: none of 816 fecal-serum metabolite pairs survived multiple-testing correction, though pathway-level pyrimidine scores did.
How the Paired Stool and Serum Metabolomics Study Was Run
Forty-four patients at one Beijing hospital were staged by the American Heart Association's CKM framework, 19 at stages 1-2 and 25 at stages 3-4, with no healthy control arm. Fecal aliquots stored at -80 °C were weighed to 50 mg and extracted in methanol:water (4:1) spiked with L-2-chlorophenylalanine; serum went into four volumes of acetonitrile:methanol. Both ran on a Q Exactive HF-X over an HSS T3 column in positive and negative mode, with intensities normalized to total signal and log10-transformed. The same stool went to 16S rRNA sequencing, and 35 serum samples received DIA proteomics on an Orbitrap Astral.
Key Findings
Six fecal features survived correction: Of 11,837 fecal features after QC, 498 met exploratory cutoffs and six held after Benjamini-Hochberg adjustment. Turicine, a hydroxyproline betaine, was the only annotated one (Cliff's delta 0.768, FDR 0.038).
Serum told a much longer story: 470 of 8,294 serum features stayed significant, with lower tryptophan, carnitine and uridine in stages 3-4 and a chenodeoxycholic acid isomer falling furthest (Cliff's delta -0.815, FDR 0.002).
Pyrimidine metabolism bridged the two matrices: no individual fecal-serum pair passed correction across 816 tests, yet pathway scores correlated after CKM adjustment (partial rho 0.429, FDR 0.0246), with serum uridine recurring among candidates.
The microbiome scattered rather than shifted: overall composition did not differ (PERMANOVA P = 0.506), but between-person dispersion was higher in stages 3-4 (mean Bray-Curtis 0.513 vs 0.436, P = 0.011).
Figure 1. Study design and multi-omics profiling. Forty-four participants were sorted into CKM stages 1-2 (n = 19) and 3-4 (n = 25). Serum from all 44 went to untargeted LC-MS/MS metabolomics, with 35 also run by proteomics, while fecal samples went to untargeted metabolomics and 16S rRNA sequencing. Adapted from Zhao et al. (2026), Frontiers in Endocrinology.
Why 50 mg of Stool Is Enough
Stool is roughly three-quarters water, and that fraction moves with transit time, so a fixed wet mass never means a fixed amount of dry matter. Ting and colleagues (2025) tested this directly: 50 mg of lyophilized feces was enough to capture between-person variation, and a single extraction in 1 mL of methanol yielded more features than a second pass. The limiting factor is almost never mass. It is where in the specimen the spatula went, which is why homogenizing the whole collection before weighing, and normalizing to total ion signal as Zhao's group did, beats doubling the input.
What the Stool Layer Added, and What It Did Not
Six corrected hits against 470 is a lopsided result, and the authors do not oversell it. Fecal metabolite profiling mostly confirmed that advanced CKM reshapes the blood more than the gut lumen, echoing the plasma-first findings in Multi-Omics: Plasma Score Flags Cardiovascular Risk in PAD. Where stool earned its place was at pathway level: the pyrimidine link to serum uridine held after adjusting for stage, and serum L-carnitine tracked eGFR at rho 0.752, which points to the kidney as the main driver. Exploratory DIA proteomics on 35 participants added 221 differential proteins, enriched for complement and coagulation cascades. My reservation is design rather than chemistry: candidates were chosen and tested in the same 44 people, identities rest on database matches rather than authentic standards, and a cross-sectional cohort cannot separate cause from declining kidney function.
In Practice: Pairing Stool With Serum in One Project
A stool aliquot that has thawed twice has usually lost part of its short-chain fatty acid and volatile content, so each tube's freezer history matters as much as its mass. For stool projects in our lab at Dalton we request 50 mg per sample, the same input used here, and we'd rather receive a second 50 mg aliquot frozen in reserve than one 200 mg tube opened repeatedly. That is also how we scope Metabolomics Services projects pairing stool with a serum draw, since blood is where this cohort found most of its stage signal.
Frequently Asked Questions
How much stool is needed for untargeted metabolomics?
About 50 mg per sample for LC-MS metabolite profiling. Zhao and colleagues extracted 50 mg of feces per participant, and Ting and colleagues found 50 mg of lyophilized stool enough to capture inter-individual variation. Homogenize before weighing, freeze at -80 °C, and keep a second aliquot in reserve.
What is stool metabolomics used for?
Stool metabolomics measures small molecules in feces, many of them produced or modified by gut bacteria, usually by LC-MS. It links diet and microbial activity to disease, screens for candidate biomarkers in gut, metabolic and kidney conditions, and gives functional meaning to 16S or metagenomic sequencing.
Do fecal metabolites reflect serum changes in cardiovascular-kidney-metabolic syndrome?
Only partly. In 44 participants, none of 816 fecal-serum metabolite pairs survived multiple-testing correction, yet pathway-level pyrimidine metabolism scores did correlate across the two matrices (partial rho 0.429). Serum uridine was the most recurrent candidate and needs targeted validation.
Conclusion
It's now reasonable to plan stool metabolomics around 50 mg per sample and to spend the saved effort on homogenization and a matched serum draw. What isn't settled is whether turicine or the uridine-centered pyrimidine link means anything beyond this one Beijing cohort, where candidates were picked and tested in the same 44 people. Anyone scoping a CKM study should budget for authentic standards and a second cohort before reading biology into a fecal hit.
Related Reading
See how we run these analyses in one lab: One-Shot Multi-Omics CRO for Drug Development.
Citation
Zhao, Z., Dai, J., Zou, Y., An, X., Zhang, Y., Zhu, C., Hao, L., & Jia, L. (2026). Paired metabolomics identify associations between fecal and serum metabolites in cardiovascular-kidney-metabolic syndrome. Frontiers in Endocrinology, 17, 1982623. https://doi.org/10.3389/fendo.2026.1982623
Note
This blog post summarizes findings from the above-cited research. Figures are adapted from the original publication. For full details, please refer to the source article.
By Seungjun Yeo, CEO at Dalton Bioanalytics. Specializing in multi-omics mass spectrometry for drug discovery and biomarker research.