Without the means to stabilize metabolites at ambient temperature after at-home collection, researchers typically resort to transporting raw fecal samples on dry ice or, alternatively, freezing them in a home freezer followed by shipping on wet ice.

These approaches have major disadvantages, including the safety risks associated with study participants handling dry ice, as well as risks to metabolite stability due to temperature fluctuations during transport¹³ DNA Genotek already offers a unique and comprehensive solution for the collection and stabilization of microbial DNA from feces in the form of its OMNIgene®•GUT collection device.

Using the physical design of the OMNIgene•GUT kit, which ensures ease-of-use, reproducibility and homogeneity of samples^{14, 15}, we have developed a novel complement of physical form and preservation solution for fecal metabolites, while preserving compatibility with metabolomics processing platforms, including Metabolon’s global metabolomics platform (Figure 1).

Figure 1

Integration with Metabolon’s global metabolomics platform.
DNA Genotek’s OMNImet•GUT device and Metabolon’s global metabolomics platform provide an integrated solution for collection, storage, transport and analysis of human fecal sample in metabolomics studies (figure provided by Metabolon).

As the gold standard of fecal sample collection, immediate freezing at -80˚C was used as the reference condition in this study. Sample collection and storage in OMNImet•GUT was compared to immediate freezing at -80˚C for sensitivity as reflected in the number of metabolites detected by Metabolon’s global metabolomics platform. Overall, the number of metabolites detected in samples from 7 donors was similar between the two collection methods (Figure 2), demonstrating that collection with OMNImet•GUT does not introduce a systematic bias or experimental artefacts and maintains a gold standard level of detection sensitivity for metabolomics analysis.

Figure 2

Preservation of metabolomic coverage.
Technical replicate samples (n=3) from each of 7 donors were collected into either a flash-frozen or an OMNImet•GUT device (ME-200), followed by immediate freezing at -80°C. The number of metabolites detected in all technical replicates (n=3) of each donor’s samples is plotted.

To be suitable for at-home collection and storage of fecal samples, a device must not distort the metabolomic profile of the raw fecal sample and must adequately stabilize metabolites over time. To assess the fidelity and stability of metabolomic profiles from OMNImet•GUT samples, Spearman correlations were calculated for all detected metabolites between flash-frozen samples and matched OMNImet•GUT samples held for 0, 1, 4 and 7 days at room temperature (table 1).

Condition	T0	T1	T4	T7	Freeze-thaw
Median R	0.89	0.88	0.86	0.86	0.89

Table 1: Correlation analysis across n=7 biological replicates (after averaging each donor’s n=3 technical replicates) in each OMNImet•GUT storage condition (baseline (T0), room temperature for 1, 4, or 7 days (T1, T4, T7), or freeze-thaw) vs. flash-frozen. The median Spearman coefficient across all detected metabolites is shown.

The median Spearman coefficient between flash-frozen and OMNImet•GUT-collected samples held at room temperature for 0 days (i.e., samples immediately frozen in OMNImet•GUT devices) was 0.89, indicating that OMNImet•GUT preserves the metabolomic profile of the raw sample with a high degree of fidelity. Furthermore, median Spearman correlations to flash-frozen remained in the 0.86-0.88 range at 1, 4 and 7 days of room temperature storage in OMNImet•GUT, indicating sustained metabolite stabilization within the device.

Further analysis of Spearman coefficients after grouping metabolites by their biochemical classification (superpathway) revealed that peptides were likely a contributing factor of discrepancy between flash-frozen and OMNImet•GUT samples at all timepoints (Figure 3). The 37 peptide superpathway biochemicals detected in this study were primarily dipeptides, which may originate from food sources, gut tissue and/or microbial activity and whose importance in fecal metabolomics is currently unclear. By contrast, the amino acid and xenobiotic superpathways, which contain most of the known bioactive microbial metabolites, displayed median Spearman coefficients in the 0.9 range at all timepoints (T0 to T7) in OMNImet•GUT relative to matched flash-frozen samples (Figure 3). These results indicate that storage in OMNImet•GUT preserves the vast majority of metabolites in raw fecal sample with both fidelity and stability.

Figure 3:

Accurate measurement of individual metabolites and metabolic pathways.
Correlation analysis across n=7 biological replicates (after averaging each donor’s n=3 technical replicates) in each OMNImet•GUT storage condition (baseline (T0), room temperature for 1, 4, or 7 days (T1, T4, T7), or freeze-thaw) vs. flash-frozen (FF). The median Spearman coefficient across each major class of biochemicals (metabolic superpathway) is plotted and the number of biochemicals in each superpathway (detected in flash-frozen samples) is labeled.

Because metabolite stabilization is such a critical performance metric for an at-home fecal collection and storage device, we sought to obtain another quantitative measure of metabolite stability over time in OMNImet•GUT. In this case, T0 samples (i.e., samples immediately frozen in OMNImet•GUT devices) were used as the reference condition in order to isolate the time effect. Intraclass Correlation Coefficients (ICC) between T0 and subsequent timepoints were calculated as described¹², using samples from 7 individual donors. The median ICC across all detected metabolites after 4 and 7 days of room temperature storage in OMNImet•GUT was 0.93 and 0.90, respectively (maximum = 1.0). ICC values above 0.9 reflect excellent agreement between datasets^16,17 , again indicating successful stabilization of metabolites over 7 days of storage at room temperature.

Given the relatively minor discrepancies between flashfrozen samples and samples stored in OMNImet•GUT for up to 7 days, we sought to assess the impact of storing samples in OMNImet•GUT on the ability to derive biological insights from human subject data. To do this, we performed Hierarchical Clustering Analysis (HCA) and Principal Component Analysis (PCA) on matched samples from 7 donors that had been flashfrozen or stored in OMNImet•GUT for varying lengths of time. In both analyses, samples clustered clearly by donor, while differences among storage conditions and timepoints were minor relative to inter-donor differences (Figures 4 and 5).

In particular, no timedependent trends were consistently observed in either the PCA or HCA. These results indicate that storage in OMNImet•GUT for up to 7 days at room temperature preserves the unique metabolomic fingerprint of human fecal samples. Taken together, the Spearman, ICC, PCA and HCA analyses demonstrate that OMNImet•GUT samples continue to accurately reflect the metabolomic profile of raw fecal samples after storage for up to 7 days at room temperature.

Figure 4: PCA. Principal component analysis of metabolomic data collected using fecal samples from 7 donors (n=3 technical replicates per donor) that were flash-frozen or stored in OMNImet•GUT as indicated: baseline (T0), room temperature for 1, 4, or 7 days (T1, T4, T7), or freeze-thaw.

Figure 5: HCA. Hierarchical clustering analysis of metabolomics data collected using fecal samples from 7 donors (n=3 technical replicates per donor) that were flash-frozen or stored in OMNImet.GUT as indicated: baseline (T0), room temperature for 1, 4, or 7 days (T1, T4, T7), or freeze-thaw. (The same dataset as Figure 4).

In addition to stability, precision is a critical consideration in metabolomic analysis. Therefore, we sought to benchmark the precision of OMNImet•GUT samples against that of flash-frozen samples. PCA (Figure 4) showed that within-donor variability (n=3 technical replicates per condition from each of 7 donors) was comparable for samples in OMNImet•GUT and flash-frozen samples. In a separate analysis of n=6 technical replicates per condition collected from each of three donors, no statistically significant differences were observed between the median relative standard deviation (RSD) across all biochemicals in the OMNImet•GUT samples relative to matched flash-frozen replicates (data not shown). Together, these data indicate that collection into OMNImet•GUT device does not introduce additional technical variability into metabolomic analyses relative to the gold standard collection method.

Simulating transport of samples through different temperature conditions was performed by subjecting samples collected into OMNImet•GUT to freezethaw at -20˚C followed by 30˚C. This treatment simulates a scenario of air transport followed by ground transport without temperature control, or, alternatively, samples that have been frozen (e.g., in a household freezer) and then delivered to a study site without temperature control.

As seen through PCA, HCA and Spearman correlation (Figures 4, 5 and table 1), freeze-thaw cycling had minimal effect on the metabolomic profile of samples from all donors, indicating that OMNImet•GUT-collected samples are not sensitive to freeze-thawing and can be shipped without temperature-controlled conditions.

Short-chain fatty acids (SCFAs) are key mediators of gut microbial activity and play important roles in regulating the immune system, maintaining the gut epithelial barrier and preventing disorders such as metabolic syndrome, inflammatory bowel disease and certain types of cancer¹⁸. Therefore, we assessed the performance of OMNImet•GUT in stabilizing SCFAs at room temperature. To this end, a pooled fecal sample derived from 7 donors was stored at room temperature for varying lengths of time in OMNImet•GUT and in an unstabilized condition (cryovial), followed by targeted, quantitative analysis of eight SCFAs. While concentrations of SCFAs in raw samples increased approximately 2- to 10-fold after 96h (4 days) at room temperature, concentrations in OMNImet•GUT samples remained unchanged between 0 and 13 days at room temperature (Figure 6). These results indicate that OMNImet•GUT quenches bacterial metabolism and that SCFAs are stable in the device for at least 7 days at room temperature. In a similar experiment with samples from 7 individual donors, all SCFAs from all donors remained stable in OMNImet•GUT for 7 days at room temperature (data not shown). These results demonstrate that OMNImet•GUT is compatible with Metabolon’s short-chain fatty acid assay and stabilizes SCFAs for at least 7 days at room temperature.

Figure 6

Stability of SCFA in OMNImet•GUT
Aliquots of a 7-donor pooled raw fecal sample were either left unstabilized (n=4 replicates) or stored in OMNImet•GUT (ME-200) (n=4 replicates, A/B/C/D) at room temperature for the indicated number of days (T0 to T13), followed by quantification of eight short-chain fatty acids (SCFAs) using targeted LC-MS/MS.

In the context of most fecal metabolomics studies, at-home collection is necessary to recruit a representative donor population and achieve adequate statistical power, which are prerequisites for identification of meaningful and translatable metabolite-disease associations. To date, no standardized, validated at-home collection kit for fecal metabolomics has been available.

Measurement of >800 individual biochemicals per human donor on Metabolon’s global metabolomics platform, as well as quantitative analysis of 8 SCFAs, showed that samples collected in OMNImet•GUT devices remained stable at room temperature for up to 7 days and recapitulated the metabolomic profiles of matched flash-frozen samples while maintaining similar technical precision. These results show that sample collection and storage in OMNImet•GUT compares favorably with the gold standard flashfreezing methodology. OMNImet•GUT will enable the collection of high-quality fecal metabolomics data at a scale previously not possible and drive a new wave of discoveries in the field.

¹ Schrimpe-Rutledge, A.C., Codreanu, S.G., Sherrod, S.D., and McLean, J.A. (2016). Untargeted Metabolomics Strategies-Challenges and Emerging
Directions. J Am Soc Mass Spectrom 27, 1897-1905. http://doi.org/10.1007/s13361-016-1469-y
² Lamichhane, S., Sen, P., Dickens, A.M., Oresic, M., and Bertram, H.C. (2018). Gut metabolome meets microbiome: A methodological perspective
to understand the relationship between host and microbe. Methods 149, 3-12. http://doi.org/10.1016/j.ymeth.2018.04.029
³ Smirnov, K.S., Maier, T.V., Walker, A., Heinzmann, S.S., Forcisi, S., Martinez, I., Walter, J., and Schmitt-Kopplin, P. (2016). Challenges of metabolomics
in human gut microbiota research. Int J Med Microbiol 306, 266-279. http://doi.org/10.1016/j.ijmm.2016.03.006
⁴ Lee-Sarwar, K.A., Lasky-Su, J., Kelly, R.S., Litonjua, A.A., and Weiss, S.T. (2020). Metabolome-Microbiome Crosstalk and Human Disease. Metabolites 10.
http://doi.org/10.3390/metabo10050181
⁵ Liu, R., Hong, J., Xu, X., Feng, Q., Zhang, D., Gu, Y., Shi, J., Zhao, S., Liu, W., Wang, X., et al. (2017). Gut microbiome and serum metabolome alterations
in obesity and after weight-loss intervention. Nat Med 23, 859-868. http://doi.org/10.1038/nm.4358
⁶ Pedersen, H.K., Gudmundsdottir, V., Nielsen, H.B., Hyotylainen, T., Nielsen, T., Jensen, B.A., Forslund, K., Hildebrand, F., Prifti, E., Falony, G., et al. (2016).
Human gut microbes impact host serum metabolome and insulin sensitivity. Nature 535, 376-381. http://doi.org/10.1038/nature18646
⁷ Wang, Z., Klipfell, E., Bennett, B.J., Koeth, R., Levison, B.S., Dugar, B., Feldstein, A.E., Britt, E.B., Fu, X., Chung, Y.M., et al. (2011). Gut flora metabolism
of phosphatidylcholine promotes cardiovascular disease. Nature 472, 57-63. http://doi.org/10.1038/nature09922
⁸ Sharon, G., Cruz, N.J., Kang, D.W., Gandal, M.J., Wang, B., Kim, Y.M., Zink, E.M., Casey, C.P., Taylor, B.C., Lane, C.J., et al. (2019). Human Gut Microbiota
from Autism Spectrum Disorder Promote Behavioral Symptoms in Mice. Cell 177, 1600-1618 e1617. http://doi.org/10.1016/j.cell.2019.05.004
⁹ Hertel, J., Harms, A.C., Heinken, A., Baldini, F., Thinnes, C.C., Glaab, E., Vasco, D.A., Pietzner, M., Stewart, I.D., Wareham, N.J., et al. (2019). Integrated
Analyses of Microbiome and Longitudinal Metabolome Data Reveal Microbial-Host Interactions on Sulfur Metabolism in Parkinson’s Disease. Cell Rep 29,
1767-1777 e1768. http://doi.org/10.1016/j.celrep.2019.10.035
¹⁰ Stevens, V.L., Hoover, E., Wang, Y., and Zanetti, K.A. (2019). Pre-Analytical Factors that Affect Metabolite Stability in Human Urine, Plasma, and Serum:
A Review. Metabolites 9. http://doi.org/10.3390/metabo9080156
¹¹ Lim, M.Y., Hong, S., Kim, B.M., Ahn, Y., Kim, H.J., and Nam, Y.D. (2020). Changes in microbiome and metabolomic profiles of fecal samples stored with
stabilizing solution at room temperature: a pilot study. Sci Rep 10, 1789. http://doi.org/10.1038/s41598-020-58719-8
¹² Loftfield, E., Vogtmann, E., Sampson, J.N., Moore, S.C., Nelson, H., Knight, R., Chia, N., and Sinha, R. (2016). Comparison of Collection Methods for Fecal
Samples for Discovery Metabolomics in Epidemiologic Studies. Cancer Epidemiol Biomarkers Prev 25, 1483-1490. http://doi.org/10.1158/1055-9965.EPI-16-0409
¹³ Gratton, J., Phetcharaburanin, J., Mullish, B.H., Williams, H.R., Thursz, M., Nicholson, J.K., Holmes, E., Marchesi, J.R., and Li, J.V. (2016). Optimized Sample
Handling Strategy for Metabolic Profiling of Human Feces. Analytical Chemistry 2016 88 (9), 4661-4668. http://doi.org/10.1021/acs.analchem.5b04159
¹⁴ Doukhanine, E., Bouevitch, A., Pozza, L., and Merino, C. (2014). OMNIgene®•GUT enables reliable collection of high quality fecal samples for gut
microbiome studies. PD-WP-00040, DNA Genotek, Inc. Available at https://dnagenotek.com/US/pdf/PD-WP-00040.pdf.
¹⁵ Doukhanine, E., Egloff, C., Bouevitch A., Lynch D., Smith C., and Iwasiow R. (2016). Choice of microbiome extraction methodology significantly affects
biomarker discovery sensitivity and requirements for cohort size. MK-00670, DNA Genotek, Inc.
¹⁶ Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology.
Psychological Assessment, 6(4), 284–290. https://doi.org/10.1037/1040-3590.6.4.284
¹⁷ Koo, T.K., and Li, M.Y. (2016). A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med 15,
155-163. http://doi.org/10.1016/j.jcm.2016.02.012
¹⁸ Tan, J., McKenzie, C., Potamitis, M., Thorburn A.N., Mackay C.R., and Macia, L. (2014). The role of short-chain fatty acids in health and disease. Adv
Immunol, 121:91‐119. doi:10.1016/B978-0-12-800100-4.00003-9