Single Cell Metabolomics Service — CE-MS and Microflow LC-MS

Single-Cell vs. Global Cell Metabolomics

A global cell metabolomics measurement returns the average of every cell in the sample, and that average is dominated by whichever cell type is most abundant or most metabolically active. A resistant subpopulation, a rare progenitor, or a few cells that respond to a drug while their neighbors do not will all disappear inside that mean. Single cell metabolomics exists to recover the differences that averaging destroys.

The measurement is technically harder than a population-scale analysis for three reasons: the amount of each metabolite in one cell is orders of magnitude below what a conventional assay consumes, the pool turns over faster than most preparation workflows take, and the act of separating cells can itself perturb the metabolism being measured. Each of those problems is addressed in the design of the workflow rather than in the analysis afterwards.

Because a single cell yields a small analyte pool, the information that comes back is different in kind from a population-level profile: fewer metabolites measured with certainty, but each one attributable to an identified cell and to a position in a cluster. Where a global profile of cultured cells is what a project needs, a cell metabolomics service is the more efficient route; where the question is how cells differ within a tissue, the spatial dimension is handled by spatial metabolomics instead.

Service Modules — Isolation, Microscale Extraction & Acquisition

Projects arrive at different stages. Some need the whole chain from live cells to a per-cell table, some already have sorted populations frozen and ready, and some need a targeted assay on cells that were isolated elsewhere. The service is organized so that a project can enter at whichever step matches its current state.

Single Cell Isolation and Sorting

Cells are separated by fluorescence-activated cell sorting, magnetic bead sorting, laser microdissection from tissue sections, or microfluidic capture, depending on the sample and on how much material the downstream assay can tolerate. The isolation method is chosen for the analyte class and for the amount of stress it imposes on the cell, and the gating strategy is recorded so that the population being sampled is defined rather than assumed.

Quenching and Microscale Extraction

Metabolite pools change within seconds, so quenching has to happen faster than the turnover it is meant to freeze. Cells are washed cold and flash-frozen immediately after isolation, then extracted in a minimal volume with isotopically labeled internal standards spiked before any analyte can be lost. Extraction volume is scaled to the sample so that the extract stays concentrated enough to inject.

CE-ESI-MS for Charged and Polar Metabolites

Capillary electrophoresis coupled to electrospray ionization mass spectrometry separates metabolites by charge and hydrodynamic radius in a capillary rather than on a column, which suits the small, polar, and often charged species that dominate central metabolism. The separation needs very little sample volume, which is the property that makes it work at the single cell level.

Microflow LC-MS/MS for Targeted and Untargeted Panels

Microflow liquid chromatography with tandem mass spectrometry covers the analytes that CE handles poorly, including many lipids and neutral compounds, and supports scheduled multiple reaction monitoring when a defined panel is being quantified. High-resolution acquisition is used where the question is which metabolites are present rather than how much of a known list is there.

Per-Cell Data Processing and Interpretation

Raw data are processed to a per-cell feature table, normalized against the internal standards and the pooled QC samples run in the same batch, and then analyzed by dimensionality reduction such as PCA, UMAP, or t-SNE, followed by clustering, differential metabolite testing between clusters, and mapping of the discriminating metabolites onto pathways. Cell-level metadata from the sorting gate is carried through so that a cluster can be described by the markers that defined it.

Key Challenges in Single-Cell Metabolomics and How We Solve Them

• Analyte amounts far below conventional assay needs — A single cell holds an amount of each metabolite that is orders of magnitude below the loading of a routine population-level measurement, and a method transferred directly from population-level workflows returns a chromatogram dominated by baseline. We scale extraction volume to the sample, select the separation technique that tolerates the smallest injection, and confirm detection with standards run at single-cell equivalent amounts rather than assuming that a method validated on larger samples still works.
• Metabolite turnover faster than the preparation — Glycolytic intermediates and energy cofactors can shift measurably while cells are being counted, gated, or transferred, so a slow workflow reports the stress of isolation rather than the state of the cell. Quenching is done cold and immediately after isolation, and the elapsed time between sorting and freezing is recorded for every batch so that a batch with an extended sort can be identified and interpreted rather than silently averaged in.
• Isolation stress that changes the metabolism being measured — Enzymatic dissociation, shear forces in a sorter, and extended time in a collection tube all alter the metabolite pool, and the effect differs between cell types. We compare sorted cells against cells recovered under gentler conditions, include a short recovery step where the biology allows it, and report the isolation method alongside the data so that the direction of any isolation effect is known.
• Ion suppression and matrix at very small volumes — At microliter extraction volumes, salts, culture medium carryover, and the small-molecule background of the separation system are proportionally larger and can suppress ionization for the analytes of interest. Salt and medium carryover are controlled during the cold wash, the separation is chosen to move analytes away from the salt region, and suppression is monitored with spiked standards rather than assumed to be uniform across cells.
• Normalization without a shared internal standard per cell — Every cell carries a slightly different amount of material, and normalizing by cell volume, protein, or total signal each introduce their own assumptions. Isotopically labeled standards are added to the extraction of every cell, pooled QC samples are run at intervals across the batch, and the normalization actually used is stated so that a between-cluster difference can be compared with a between-batch one.
• Separating biology from stochastic variation — Two cells of the same type never give identical numbers, and cell cycle phase, stochastic gene expression, and measurement noise all contribute to that spread. The analysis reports the technical spread measured from replicate injections and reference cells next to the biological spread between cells, so a genuine subpopulation is not declared on the basis of variation that the method itself produces.

Metabolites We Quantify at the Single-Cell Level

Coverage is organized by metabolite class, because the class determines both the separation technique and how confidently a single-cell measurement can be interpreted. Classes are combined within one acquisition where the chemistry allows it, and the measured list is reported per cell with its annotation level.

Analytical Platform & Method for Single-Cell Analysis

The platform choice follows from the analyte class and from how much material the isolation step can deliver, not from a fixed workflow. A charged, polar panel and a lipid panel from the same experiment are therefore acquired on different systems, and the results are merged at the per-cell data level.

Instrumentation

Method Controls

Every batch carries the controls that make a per-cell comparison interpretable. Isotopically labeled standards are added to each extraction before the analyte can be lost. Pooled QC samples prepared from the same cell population are injected at intervals, so that instrument drift can be separated from biological spread, and extraction and matrix blanks bound the background contributed by reagents and the separation system. Replicate injections of a reference extract then quantify the technical spread that the reporting subtracts from the spread between cells.

Where a defined panel is being quantified, calibration is built with matrix-matched standards at amounts typical of a single cell rather than extrapolated from a calibration curve built for larger sample amounts. Where that panel is the specific need, targeted metabolomics workflows provide the same quantitative framework for larger cell numbers. Discovery work is run in the same laboratory, so a metabolite first seen by untargeted metabolomics can be moved into a scheduled assay without transferring methods between sites.

Single Cell Metabolomics Workflow

Sample Types & Cell Preparation Requirements

The service accepts live or freshly isolated cells, frozen sorted populations, and tissue sections for dissection-based isolation. Requirements below are the general case; a project that works with fewer cells than listed can still be assessed, because the limiting factor is the analyte class rather than the cell number alone.

• Viability is checked before sorting, and a preparation below the agreed viability threshold is reported rather than carried forward, because dead and dying cells give a metabolite profile that reflects degradation
• For workflows that require sorting, the starting cell number is what matters rather than the number analyzed: isolation consumes most of the material, so the input has to account for the cells that will be lost at the gate
• Include the sample metadata sheet: origin, treatment arm, time point, harvest method, and any handling condition that differed between samples
• Do not submit fixed or permeabilized cells for metabolite measurement, because fixation and permeabilization remove or redistribute the small-molecule pool being measured
• If you are sending processed data for analysis only, also send the raw acquisition files where available, together with the internal standard used and the normalization that has already been applied

Single Cell Metabolomics Deliverables — What You Receive

• Per-cell feature table — Measured metabolites with intensity, migration time or retention time, mass-to-charge ratio, and annotation confidence level, one row per analyzed cell, delivered in Excel and CSV.
• Internal standard response table — Response of the labeled standards per cell extract, which is the evidence that the extraction and the injection behaved consistently across the batch.
• Clustering and dimensionality reduction — Cluster assignment and PCA, UMAP, or t-SNE projections with the cell-level metadata from the isolation gate carried through.
• Differential metabolite results — Per-cluster comparisons with fold changes, the statistical test applied, and the correction used, listed with the direction of change for every contributing metabolite.
• Pathway mapping — Discriminating metabolites mapped onto curator databases with the database name and release stated, so enrichment can be reproduced.
• Quality control report — Technical spread from replicate injections and pooled QC samples, batch drift, blank levels, and the number of cells that passed quality criteria relative to the number analyzed.
• Methods documentation — Isolation method and parameters, quenching and extraction volumes, instrument settings, internal standards, and the normalization applied, written so that the analysis subsection can be assembled from the report.

Representative Data

Applications of Single Cell Metabolomics

For projects that already hold a per-cell dataset, metabolomics data analysis can be carried out on the submitted matrix, and the wider service portfolio is described under metabolomics services.

Frequently Asked Questions

Selected Publications in Single-Cell Metabolomics