Anyone weighing MS versus antibody spatial proteomics is choosing between roughly two orders of magnitude in breadth and roughly two orders of magnitude in spatial continuity, in opposite directions. That is the honest shape of the trade. What tips it in practice is often neither of those, but how many samples you intend to run.
Key Takeaways
|
Two Philosophies, or One Distinction?
The framing usually offered is targeted against untargeted, and it is not quite right. As set out in Spatial Proteomics by Mass Spectrometry: LCM, Single-Cell, and Imaging Approaches, both approaches involve prior choices; they differ in where the choice is made. Antibody methods target the analyte, since you detect the proteins whose antibodies you applied. MS-based spatial proteomics frequently targets the cell or region, selecting what to measure by morphology or marker staining and then measuring its proteome without specifying which proteins.
That distinction has a practical consequence worth stating before any comparison of numbers. An antibody experiment cannot discover a protein, because a protein absent from the panel produces no signal and leaves no trace that it was missed. An MS experiment cannot discover a cell type it did not select, but it can discover any protein present in what it did select. The two blind spots are in different places, which is why the approaches are genuinely complementary rather than ranked.
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
Discovery Breadth Against Targeted Plex
This is the axis where the difference is largest, and it is worth putting numbers to rather than describing qualitatively.
| Antibody-Based Imaging | MS-Based Spatial Proteomics |
Breadth per experiment | Documented plex limits of around 66 markers for cyclic immunofluorescence and around 40 for mass-based imaging platforms | Thousands of proteins. Published single-cell work reports over 5,000 proteins from individual cells |
Can it discover? | No. A protein without an antibody in the panel is invisible and its absence is undetectable | Yes. Proteins are identified without prior specification |
Post-translational modifications | Only where a modification-specific antibody exists and has been validated | Accessible in principle, since modified peptides differ in mass |
Proteoforms and isoforms | Generally not distinguished | Distinguishable where the peptides differ |
Sensitivity for one chosen target | High. Specific binding and amplification favour the antibody | Lower for any individual protein at equivalent material |
Reagent development | Substantial. Antibody validation and conjugation per marker | None. No affinity reagents required |
Table 1. Breadth and specificity compared. Plex figures are as documented in the review literature and will rise as platforms develop; the order-of-magnitude gap is the durable point rather than any specific ceiling.
A review of multiplexed tissue imaging platforms documents plex limits in this range, and the Modern Pathology review of high-plex biomarker assessment sets out what achieving them costs: time-consuming antibody conjugation and assay optimisation alongside challenging data processing and extensive operator training. On the MS side, a chip-and-trap workflow published in Nature Methods reports over 5,000 proteins from single cells, and region-level microdissection reaches comparable or greater depth as covered in Laser Capture Microdissection + LC-MS Proteomics.
Two qualifications keep this fair. Antibody plex ceilings are rising and any specific figure will date, so the durable claim is the order of magnitude rather than the number. And breadth is not automatically an advantage: if you know which twenty proteins matter, measuring five thousand is not five thousand times better; it is a different and in some respects harder experiment, with more multiple-testing burden and more scope for spurious findings.
Which Gives Better Resolution and Cell Assignment?
Antibody imaging, clearly and by a wide margin, and this is the axis where the MS community should concede rather than argue.
Three separate advantages combine.
- Resolution is subcellular and instrument-determined. Mass-based antibody imaging platforms reach sub-micron and low-micron resolution, sufficient to assign signal to compartments within cells.
- Coverage is continuous. Every position in the imaged area is measured, so there is no sampling decision and nothing falls between measurements. MS approaches measure the regions or cells they selected and nothing else.
- Cell assignment is direct. Because the image is continuous and resolution is subcellular, signal can be assigned to segmented cells natively, which supports single-cell phenotyping across a whole section.
By contrast, MS spatial resolution is set by dissection precision, not by the mass spectrometer. That is a genuine limitation and also the reason it is adjustable: you choose where to draw the boundary, and depth follows from how much material that boundary encloses. Imaging-guided approaches narrow the gap considerably by selecting individual cells, as described in Deep Visual Proteomics: Imaging-Guided Mass Spectrometry, but even there the measurement is of discrete selected objects rather than a continuous field.
Throughput on whole sections belongs in the same discussion, and it does not favour either side unambiguously. High-plex antibody imaging of a cohort is not fast: work on a mass-based platform reported a tissue microarray of biopsies from 41 patients at 36-plex requiring 12 consecutive days of continuous acquisition. That figure predates current hardware and should be read as an order-of-magnitude anchor, but it is a useful corrective to the assumption that imaging is inherently the faster route.
What Do They Actually Cost?
Comparisons rarely address this, because it needs published rates rather than estimates. Where rates are published, the structural difference is larger than the headline numbers suggest.
The Boston University Spatial Biology Core rate card publishes antibody conjugation at $139 per marker and panel staining at $289 per sample for internal users. A 40-marker panel therefore costs around $5,560 in conjugation before a single sample has been stained, and around $9,880 at the commercial rate. Mass spectrometry has no equivalent line at all, because it requires no affinity reagents.
Study Size | Panel Development | Staining at $289 Each | Effective Cost per Sample |
5 samples | $5,560, one time | $1,445 | Approximately $1,401 |
20 samples | $5,560, one time | $5,780 | Approximately $567 |
50 samples | $5,560, one time | $14,450 | Approximately $400 |
200 samples | $5,560, one time | $57,800 | Approximately $317 |
Table 2. How antibody panel development amortises, using published internal rates for a 40-marker panel. Reagent and staining costs only; instrument time, analysis, and personnel are excluded, and rates are institution-specific.
The Cost Curves Have Different Shapes Cost per sample for the antibody route falls 4.4-fold between a five-sample study and a two-hundred-sample study, purely because a fixed development cost spreads. Mass spectrometry has no such fixed cost and therefore no such amortisation — its cost is instrument time and low-input handling per sample, which is roughly constant regardless of study size. Which reframes the cost question usefully. It is not which technique is cheaper but how many samples you have. For a small exploratory study, antibody panel development is a substantial cost against very few measurements, and MS avoids it entirely. For a large cohort, the same development costs a few hundred dollars per sample and antibody imaging becomes the economical option. The crossover depends on your local rates for both, which is why the honest advice is to obtain them rather than to accept a general claim. |
One further cost that belongs in the comparison but resists tabulation: the reagent development burden is not only financial. Antibody conjugation and assay optimisation take time and expertise, and a panel that does not work is a delay as well as an expense. MS-based approaches shift that burden to low-input sample handling and instrument access, which is a different skill set rather than a smaller one.
Sample and Tissue Requirements
Both approaches work on archival material, which is worth stating because it is often assumed only antibody methods do.
- Fixation. Both are compatible with formalin-fixed paraffin-embedded tissue. MS-based work on FFPE has reported up to around 2,000 proteins from microdissected archival material, so biobank collections are accessible to either.
- Tissue consumed. Antibody imaging generally consumes one section and preserves it as an image. MS approaches consume the dissected material, though the remainder of the section survives.
- Staining interaction. MS-based microdissection is affected by staining, since haematoxylin reduces protein detection, whereas antibody imaging depends on staining by design. This is an asymmetry worth knowing when the same block must serve both.
- Serial sections. Running both approaches on adjacent sections from the same block is entirely feasible and is how many programmes actually operate, which sidesteps the choice for the cost of one additional section.
That last point deserves more attention than it usually gets. The two approaches are not mutually exclusive at the specimen level. Adjacent sections differ by a few microns of tissue, which for most questions is negligible, so a study can have antibody imaging for spatial context across the whole section and MS for untargeted depth in selected regions, from the same block.
Which Should You Choose?
Work down this list. The first condition that applies is usually decisive.
If This Is True | Choose |
You do not know which proteins matter | MS. An antibody panel cannot detect what it does not contain, and cannot signal that it missed anything |
You need post-translational modifications or proteoform resolution | MS. Antibody methods generally cannot distinguish these at all |
You need subcellular localisation across a whole section | Antibody imaging. Continuous coverage and subcellular resolution are not available from dissection-based approaches |
You know your targets and have many samples | Antibody imaging. Panel development amortises, and sensitivity per target is higher |
You know your targets and have few samples | Consider MS. Panel development is a large fixed cost against few measurements |
You need single-cell phenotyping across a large population | Antibody imaging. Every cell in the field is measured rather than selected |
You need untargeted depth from a defined region | MS with microdissection. Deepest coverage of any spatial proteomics approach |
You can afford adjacent sections | Both. This is what most mature programmes do, and it is cheaper than choosing wrongly |
Table 3. A selection framework. The final row is the honest answer for most well-resourced programmes and is worth considering before treating this as a binary decision.
The pattern across that table is that the two approaches fail in opposite directions, which is precisely why combining them works. Antibody imaging gives you every cell and a limited set of proteins. Mass spectrometry gives you selected cells and effectively all their proteins. A discovery experiment by MS followed by targeted validation across a cohort by antibody imaging plays to both, and is a more defensible study design than either alone.
For the individual MS approaches and how they differ from one another, see Single-Cell Proteomics by Mass Spectrometry. For where spatial proteomics sits alongside metabolite and lipid imaging, and why those classes have no antibody route at all, see Spatial Analysis in Analytical Science: Mass Spectrometry Imaging and Spatial Omics.
This article was produced under Separation Science's AI Editorial Guidelines.
