Chapter 3 Pretreatment
Pretreatment strongly affects metabolomics results and covers the workflow from crude samples to injection vials for instrumental analysis. The goal is to preserve biologically meaningful compounds while removing as much irrelevant matrix background as possible. In metabolomics, however, the compounds of interest are often not known in advance, so pretreatment must be chosen according to study purpose, sample matrix, and analytical platform. For example, Gel Permeation Chromatography (GPC), Florisil, alumina, and silica gel can remove lipids, while alcohols and strong acids or bases can denature proteins and release additional compounds. If the study focuses on small lipids or peptides, these choices may need to change substantially. In practice, sample collection, quenching, extraction, derivatization, and storage should all be optimized as part of pretreatment.
3.1 Collection
Those papers investigated different fecal collection methods(Loftfield et al. 2016; Deda et al. 2017).
This paper discusses the influence of sample normalization(Wu and Li 2016).
3.2 Quenching
Quenching solvent is always used to stop enzymatic activity.
In this review(Lu et al. 2017), authors said:
A classical approach, which works well for many analytes, is boiling ethanol. Although the boiling solvent raises concerns about thermal degradation, it reliably denatures enzymes. In contrast, cold organic solvent may not fully denature enzymes or may do so too slowly such that some metabolic reactions continue, interconverting metabolites during the quenching process.
This review(Kim et al. 2020) summarized the urease-dependent metabolome sample preparation and found:
activities of urease and endogenous urinary enzymes and metabolite contaminants from the urease preparations introduce artefacts into metabolite profiles, thus leading to misinterpretation.
3.3 Extraction
According to this research(Bennett et al. 2009):
The total metabolome concentration is approximately 300 mM, whereas the protein concentration is approximately 7 mM., which implies that most cellular metabolites are in free form.
Dmitri et al.(Sitnikov et al. 2016) thought the most orthogonal methods to methanol-based precipitation were ion-exchange solid-phase extraction and liquid-liquid extraction using methyl-tertbutyl ether.
Another study used stable isotope labeled sample and found the use of a water-methanol-acetonitrile mixture for global metabolite extraction instead of aqueous methanol or aqueous acetonitrile alone (Doppler et al. 2016).
Metabolic information was highly influenced by the extraction solvent(Ibáñez et al. 2017).
Tissue samples need to first be pulverized into fine powders.
Feces collected with 95% ethanol or FOBT would be more reproducible and stable.
In this review(Lu et al. 2017), authors said:
In our experience, for both cell and tissue specimens, 40:40:20 acetonitrile:methanol:water with 0.1 M formic acid (and subsequent neutralization with ammonium bicarbonate) is generally an effective solvent system for both quenching and extraction, including for ATP and other high-energy phosphorylated compounds. We typically use approximately 1 mL of solvent mix to extract 25 mg of biological specimen. …Thus, although drying is acceptable for most metabolites, care must be taken with redox-active species.
nano LC-MS could be used to analyze small numbers of cells(Luo and Li 2017).
For plant like soybeans(Mahmud et al. 2017), ammonium acetate/methanol could be selected as extraction strategies compared with water/methanol and sodium phosphate/methanol. For general plant samples, check this comprehensive investigation(Bijttebier et al. 2016).
For blood plasma and serum sample, a comprehensive evaluation of 12 sample preparation methods (SPM) using phospholipid and protein removal plates (PLR), solid phase extraction plates (SPE), supported liquid extraction cartridge (SLE), and conventionally used protein precipitation (PPT) were performed. Results show PPT and PLR on the same samples by implementing a simple analytical workflow as their complementarity would allow the broadening of the visible chemical space (Chaker et al. 2022).
3.3.1 Extraction selection flowchart
In practice, extraction should be selected according to the study purpose, sample matrix, and metabolite class of interest rather than by habit alone. A simple decision logic is:
Is the study untargeted or targeted?
Untargeted studies usually prefer broader and more orthogonal extraction conditions, while targeted studies may prioritize recovery for a defined chemical class.Is the major interest polar metabolites, lipids, or both?
Polar metabolites often favor aqueous organic mixtures such as methanol/acetonitrile/water, while lipids often require biphasic extraction or non-polar solvents.Does the matrix contain high protein or phospholipid background?
Plasma, serum, and tissue extracts may need additional cleanup such as protein precipitation, phospholipid removal, SPE, or liquid-liquid extraction.Is sample amount limited?
Small-volume studies may need simpler and more reproducible extraction strategies with fewer transfers.Is absolute quantification or high-throughput screening required?
High-throughput studies usually favor simpler workflows, while quantitative studies may justify more selective cleanup or isotope-labeled standards.
As a rough guide:
Broad untargeted polar metabolomics: water/methanol/acetonitrile mixtures are often a good starting point.
Lipid-focused or broad lipid coverage: LLE or biphasic extraction is often more suitable.
Clinical plasma/serum screening: protein precipitation is often the simplest first-pass method, followed by SPE or phospholipid removal if matrix effects are severe.
Very dirty matrices: SPE or multi-step cleanup may improve selectivity at the cost of coverage.
3.3.2 Protein precipitation, SPE, and LLE
The three common strategies have different goals and trade-offs:
| Method | Main purpose | Strengths | Limitations | Common use |
|---|---|---|---|---|
| Protein precipitation (PPT) | Remove proteins quickly with organic solvent | Fast, cheap, high-throughput, simple to automate | Limited selectivity, matrix effects may remain, lipid/phospholipid background can still be high | Plasma, serum, routine untargeted screening |
| Solid-phase extraction (SPE) | Selective cleanup or enrichment on sorbent | Better cleanup, can enrich target classes, reduces matrix complexity | More steps, more cost, analyte loss possible, method development required | Targeted assays, dirty matrices, class-selective enrichment |
| Liquid-liquid extraction (LLE) | Partition compounds by polarity into different phases | Good orthogonality, useful for lipids and broad chemical space separation | More labor-intensive, emulsion risk, reproducibility can suffer if handling is inconsistent | Lipidomics, biphasic extraction, broad coverage workflows |
PPT is often the most practical starting point for biofluids because it is simple and scalable. SPE is useful when background suppression is severe or when the study targets a narrower compound class. LLE is often preferred when the aim is to separate polar and non-polar metabolite fractions or to improve lipid coverage.
3.4 Derivatization
Derivatization is always used in GC-based metabolomics study because many metabolites are too polar, thermally unstable, or non-volatile to be directly analyzed by GC-MS. This paper(Miyagawa and Bamba 2019) compared sequential derivatization methods and found different compounds would show different fluctuations during oximation or silylation process. This paper summarized derivatization methods for LC-MS (Zhao and Li 2020).
For GC-MS metabolomics, derivatization is not a minor technical detail. It is a major source of variation and should be planned as part of pretreatment. In many workflows, oximation is performed first to stabilize carbonyl compounds and reduce multiple derivative forms, followed by silylation to increase volatility and improve chromatographic behavior. If the timing, reagent freshness, residual water, or temperature is not controlled, the same metabolite may show unstable peak areas or multiple derivative peaks across samples.
The main practical points are:
Control water content because moisture can strongly interfere with silylation efficiency.
Keep derivatization timing consistent across all samples because reaction time can change peak intensity.
Use the same reagent lot and temperature conditions within a study when possible.
Treat batch processing carefully because derivatized samples may not remain stable for long waiting times.
Expect compound-dependent behavior since sugars, amino acids, organic acids, and carbonyl compounds may respond differently.
Therefore, derivatization should be optimized and documented with the same care as extraction solvent, especially in GC-MS studies or whenever semi-quantitative comparison is important.
3.5 Isotope label
You might try heavy water to exchange oxygen atom with samples to track certain metabolites(Osipenko et al. 2022) or MS-IDF(Wang et al. 2022).
3.6 Microsampling techniques
Traditional blood sampling requires large volumes (several mL) drawn by trained personnel, which limits the temporal resolution and ecological validity of metabolomics studies. Microsampling techniques collect much smaller volumes (typically 10–30 \(\mu\)L) and can often be performed by participants themselves at home, making them especially attractive for longitudinal and remote studies.
3.6.1 Solid-phase microextraction (SPME)
Solid-phase microextraction (SPME) uses a coated fiber or thin film to extract analytes directly from a biological matrix without the need for solvent extraction(Pawliszyn 2012). The coating selectively absorbs or adsorbs analytes based on their affinity to the coating material. SPME has been coupled with both GC-MS and LC-MS for metabolomics applications, and in vivo SPME probes have been developed for direct tissue sampling with minimal invasiveness(Reyes-Garcés et al. 2018). A key advantage is that SPME provides a balanced coverage of metabolites across a wide polarity range, since the extraction is equilibrium-based rather than exhaustive(Li et al. 2022).
3.6.2 Dried blood spots (DBS)
Dried blood spots involve collecting a small volume of capillary blood (typically from a finger prick) onto filter paper cards. After drying, the cards are stable at room temperature and can be shipped by mail, making DBS ideal for large-scale population studies and remote sampling(Déglon et al. 2012). DBS have been successfully used in newborn screening for decades and are now being applied to metabolomics studies. Challenges include hematocrit effects on spot size and analyte recovery, and the relatively small amount of material available for analysis(Manier et al. 2021).
3.6.3 Volumetric absorptive microsampling (VAMS)
Volumetric absorptive microsampling (VAMS, commercially available as Mitra devices) collects a fixed volume of blood (typically 10 or 20 \(\mu\)L) onto an absorbent polymer tip, overcoming the hematocrit-dependent volume variability of DBS(Denniff and Spooner 2014). VAMS devices are straightforward for self-collection and the fixed volume simplifies quantitative analysis. They have been validated for a range of metabolomics and pharmacokinetic applications.
3.7 Automation
Automation is increasingly important in metabolomics pretreatment, especially for large cohorts, longitudinal studies, and multi-batch projects. Automated liquid handling can improve reproducibility by reducing manual pipetting error, controlling timing more tightly, and standardizing solvent addition, mixing, and aliquoting across many samples.
Automation is especially useful for:
protein precipitation in plates
SPE plate workflows
addition of internal standards
batch-wise derivatization with controlled timing
aliquoting pooled QC and blanks
However, automation does not automatically solve pretreatment problems. Plate layout, dead volume, edge effects, carryover, tip selection, solvent compatibility, and evaporation still need to be checked carefully. In some cases, a simple manual workflow may be more robust than a poorly optimized automated one. Therefore, automation should be treated as a reproducibility tool that still requires validation.
3.8 Storage
Samples should be stored after sample collection or sample pretreatment. -80°C or -20°C is always preferred to store samples. Dry ice should be used during sample pretreatment. However, comprehensive investigation of storage influences found the metabolites profile will change after one day storage at -80°C(Yu et al. 2020) . Rapid analysis of samples should be considered to capture more accurate information in the samples.
Storage conditions such as temperature and time can affect the metabolite composition of various samples. Laparre et al.(Laparre et al. 2017) noted that the metabolite profiles of urine samples were significantly changed after 5 days of storage at 4°C , while Wandro and colleagues(Wandro et al. 2017) observed that the metabolomic profiles of cystic fibrosis sputum samples underwent notable changes after only 1 day of storage at 4°C . Likewise, Roszkowska et al. demonstrated that various signaling molecules were lost from the lipidome profile of tissue after storing the samples for one year at 80°C (Roszkowska et al. 2018). To date, most metabolomics studies involving storage of samples prior to the analysis have used a storage temperature of 80°C , as previous investigations have shown that low temperatures or freeze-thaw cycles do not significantly change the metabolite profile of certain samples(Lin et al. 2007) .
For gut microbiota, this paper could be checked for storage issue(Zubeldia-Varela et al. 2020).
For blood sample storage, you could check this paper(Hernandes et al. 2017).
For urine sample storage, check this(Laparre et al. 2017).
This piece reviewed the stability of energy metabolites(Gil et al. 2015).