Fecal 1H-NMR Metabolomics: A Comparison of Sample Preparation Methods for NMR and Novel in Silico Baseline Correction

Abstract

Analysis of enteric microbiota function indirectly through the fecal metabolome has the potential to be an informative diagnostic tool. However, metabolomic analysis of feces is hampered by high concentrations of macromolecules such as proteins, fats, and fiber in samples. Three methods—ultrafiltration (UF), Bligh–Dyer (BD), and no extraction (samples added directly to buffer, vortexed, and centrifuged)—were tested on multiple rat (n = 10) and chicken (n = 8) fecal samples to ascertain whether the methods worked equally well across species and individuals. An in silico baseline correction method was evaluated to determine if an algorithm could produce spectra similar to those obtained via UF. For both rat and chicken feces, UF removed all macromolecules and produced no baseline distortion among samples. By contrast, the BD and no extraction methods did not remove all the macromolecules and produced baseline distortions. The application of in silico baseline correction produced spectra comparable to UF spectra. In the case of no extraction, more intense peaks were produced. This suggests that baseline correction may be a cost-effective method for metabolomic analyses of fecal samples and an alternative to UF. UF was the most versatile and efficient extraction method; however, BD and no extraction followed by baseline correction can produce comparable results.

Feather pulp: a novel substrate useful for proton nuclear magnetic resonance spectroscopy metabolomics and biomarker discovery

“…One potential limitation of the BD extraction method is the occurrence of baseline distortions due to the incomplete removal of macromolecules in samples that are high in proteins, fats, and/or fiber such as feces ( Brown et al., 2022 ). We did not observe substantive baseline distortions for feather pulp samples extracted with the BD method.…”

Comparative Analysis of the Temporal Impacts of Corticosterone and Simulated Production Stressors on the Metabolome of Broiler Chickens

“…Metabolite spectra were obtained using the 1-D NOESY gradient pulse pre-saturation water suppression pulse sequence ‘noesygppr1d’ with 10 ms mixing time as previously described [ 30 ]. Briefly, each sample was run for 512 scans to a total acquisition size of 128 k, a spectral window of 20.5 ppm, a total data acquisition time of 4.56 s, a transmitter offset of ≈4.7 ppm, and a recycle delay of 1 s (total T1 relaxation recovery time of 5.56 s).…”

“…Although there are many potential means of assessing stress, metabolomics has shown promise as a highly sensitive method of determining a stress condition based on relative concentrations of metabolites in biological samples [ 24 ]. In fact, recent nuclear magnetic resonance (NMR) spectroscopy-based metabolomics studies have shown that the metabolomes of eggs, breast muscle, liver, kidney, feces, and feathers can be used in a variety of ways to better understand domestic poultry physiology, including stress [ 6 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 ]. However, the use of metabolomics to identify biomarkers of health in poultry remains in its infancy.…”

1H NMR Signals from Urine Excreted Protein Are a Source of Bias in Probabilistic Quotient Normalization

“…This can be performed experimentally 21 or with computational methods. 22 , 23 However, we advocate the use of SMolESY, because of its ease of application to 1 H NMR spectra (including retrospective application where proteinuria is observed after the fact), being fast and highly effective method for removing broad baseline signals from protein and improving the estimation of normalization coefficients. Our observations and proposed methodology are of high importance for the accurate normalization of urine biofluid 1 H NMR spectra, especially in the context of studies on diseases and phenotypes where proteinuria is likely to be present (e.g., diabetes, chronic kidney disease, pregnancy, infection, or protein rich diet).…”

“…We recommend the removal of protein baseline signals prior to estimation of normalization coefficients. This can be performed experimentally or with computational methods. , However, we advocate the use of SMolESY, because of its ease of application to 1 H NMR spectra (including retrospective application where proteinuria is observed after the fact), being fast and highly effective method for removing broad baseline signals from protein and improving the estimation of normalization coefficients. Our observations and proposed methodology are of high importance for the accurate normalization of urine biofluid 1 H NMR spectra, especially in the context of studies on diseases and phenotypes where proteinuria is likely to be present (e.g., diabetes, chronic kidney disease, pregnancy, infection, or protein rich diet)…”