Draft:Spatial Proteomics
Spatial Proteomics
From Wikipedia, the free encyclopedia
Spatial proteomics refers to a broad collection of methods focused on analyzing proteins in a spatial context within cells or tissues. This field is an intersection of computational advancement, such as machine learning and artificial intelligence (AI), with traditional proteomics approaches. Data is high dimensional, multiplexed, and often multimodal. By conserving spatial architecture, these techniques enable investigation of protein localization, interactions and cellular organization within biological systems. Spatial proteomics has applications in developmental biology, cancer research, neuroscience, and precision medicine, when understanding tissue architecture and cellular heterogeneity is essential.
 | Review waiting, please be patient.
This may take 8 weeks or more, since drafts are reviewed in no specific order. There are 2,922 pending submissions waiting for review.
Where to get help
How to improve a draft
You can also browse Wikipedia:Featured articles and Wikipedia:Good articles to find examples of Wikipedia's best writing on topics similar to your proposed article. Improving your odds of a speedy review To improve your odds of a faster review, tag your draft with relevant WikiProject tags using the button below. This will let reviewers know a new draft has been submitted in their area of interest. For instance, if you wrote about a female astronomer, you would want to add the Biography, Astronomy, and Women scientists tags. Editor resources
Reviewer tools
|
Comment: There are some issues with your citations. Please fix them. GreenRedFlag (talk) 18:42, 9 March 2026 (UTC)
Background
Proteins are an essential macromolecule and the functional result of the central dogma of biology, DNA → RNA → Protein.[1] They are composed of amino acids, which are the smallest units of protein.[2] These are joined together by peptide bonds, making small protein units, called peptides.[2] There are 20 common amino acids that exist, and proteins can use any combination of these to create a meaningful structure.[2][3] Proteins are divided into 4 levels: primary, secondary, tertiary, and quaternary.[2][4] Functionally, proteins are only active at the tertiary or quaternary level, serving diverse and extensive roles in cells and homeostasis.[3][5] For example, proteins can be structural, regulatory, or enzymatic, all of which can be excreted.[2][4]
Following translation, proteins can be subjected to a number of post translational modifications.[2][5][6] For many proteins, these modifications are crucial for functionality.[2][5][6] Each kind, or a combination of these modifications can have a distinct effect on a protein's function, role, or on/off state. Some examples include phosphorylation, glycosylation, ubiquitination, acetylation and more.[5][6] Proteins are made as inactive precursors (zymogen), and must be cleaved to become active (enzyme).
Given the incredibly versatile and comprehensive integration of proteins in biological systems scientists are interested in characterizing them, learning about their structure, function, and interactions within biological systems, earning the term proteomics. The investigation of proteomics in a spatial context (subcellular or within tissues) is spatial proteomics.
Spatial proteomics technologies
Spatial proteomics is a broad field that includes many different methods for studying proteins in tissues. These techniques allow researchers to detect many proteins and location in the cell at the same time while keeping the tissue structure intact.[7][8] This means scientists can identify different types of cells, see how they are functioning, and analyze how neighbouring cells interact.[9][10]
Keeping the tissue architecture intact also makes it possible to study natural cellular environments, such as immune cell niches, interactions between tumours and the immune system (tumour-immune interactions),[9][10] and cell neighbourhoods (microenvironments) linked to disease progression.[9][11] Many spatial proteomics methods can measure multiple proteins at once (multiplexing), though the way this is done depends on the specific technology used.
Emerging technologies
Emerging technologies in spatial and subcellular (within a cell) proteomics are increasingly combined with other omics approaches, such as spatial transcriptomics, metabolomics, and epigenomics. This interdisciplinary strategy provides more comprehensive insights into how gene expression, protein localization, and metabolic activity interact.
For example, advanced image segmentation and machine learning–based analysis is improving the identification of cells and subcellular structures. These tools enhance spatial resolution, improve workflow automation, and increase both the accuracy and scalability of high-content imaging and proteomics workflows.[12]
Characteristics and Properties
Spatial proteomics technologies vary widely but share several core technical characteristics, including spatial resolution, multiplexing capacity, sample requirements, and compatibility with other spatial omics approaches.
- Spatial resolution: Spatial proteomics methods preserve physical location of proteins within cells or tissue. Depending on the technology used, spatial resolution can range from tissue-level to single-cell or subcellular regions. Imaging-based techniques such as imaging mass cytrometry (IMC) and multiplexed ion beam imaging (MIBI) can achieve single-cell resolution, and in some cases approach subcellular resolution through pixel-level ion or laser scanning of antibody-labeled proteins.[13] Resolution in other approaches, such as barcoding or region-of-interest sampling, depend on the size of capture regions or pixels instead of individual cells.[14]
- Multiplexing capacity: An important feature of spatial proteomics is the ability to detect multiple proteins simultaneously while preserving spatial context. Standard immunofluorescence microscopy is often limited to a few markers due to spectral overlap. Modern multiplexed spatial proteomics technologies overcome this through metal-tagged antibodies, DNA-barcoded probes, or iterative staining cycles, detecting dozens to over 50 protein targets in a single experiment.[15] These multiplexing capacities allow for mapping of cell types, signaling pathways, and immune interactions within tissues.
- Sample requirements: Intact tissue sections or fixed cells are required to preserve spatial information. Most spatial proteomics tools are therefore compatible with widely used tissue preservation practices, including formalin-fixed paraffin-embedded (FFPE) tissue sections, fresh-frozen tissues, or cultured cells.[13]
- Integration with other spatial omics: Spatial proteomics is frequently combined with other molecular profiling approaches, such as spatial transcriptomics, spatial metabolomics, and epigenomic profiling. Multimodal spatial omics often integrate protein measurements alongside RNA or genomic analyses, generating comprehensive molecular maps of tissues and cellular microenvironments.[16] These integration approaches link protein localization with gene expression, cellular identity, and tissue architecture.
Methods overview
Methods for spatial proteomics are diverse and generally employ mass spectrometry and imaging (using fluorescence or antibodies) technologies. These can act as the basis of methods individually or in combination.
Mass spectrometry-based
Mass spectrometry-based (MS) methods often involve identification of proteins within cell lysate or organellar[17] fractions (fractionation), affinity purification, and proximity labelling. Approaches can be divided into labelled and label-free MS.
- Affinity purification (AP-MS): Protein(s) of interest is pulled down using antibody or affinity tags (such as enzymes that tag proteins very close to the protein of interest). This method enables detection of the interactome, in which stable interactions between groups of bound proteins can be isolated and profiled with subsequent MS analysis.[18]
- Proximity labelling: Protein(s) of interest is fused to an enzyme that chemically tags nearby proteins within living cells. Labeled proteins are then purified and identified by mass spectrometry. This method captures spatially close proteins, including weak or transient interactions that may not survive purification.[19]
- Fractionation: Cells are physically separated sub-cellularly (nucleus, cytoplasm, mitochondria, etc.) with centrifugation or biochemical fractionation. Lysates are solutions of broken down cells, which can be from the whole cell, or subcellular. Each fraction is then analyzed separately by mass spectrometry. This method helps determine protein localization by enriching specific cellular compartments.[20][21]
Imaging / in situ methods
These methods importantly preserve tissue architecture. These approaches generally rely on antibodies or fluorophores for signal detection. Antibodies may be conjugated to fluorescent dyes or oligonucleotides. To detect these signals, conventional fluorescence microscopes are used, providing a low barrier of entry to use these methods.
- Antibody-based: These methods use antibodies that specifically bind to target proteins within preserved tissue sections. Usually antibodies are joined to fluorescent dyes, metal tags, or DNA barcodes which allows for multiplexed detection and protein expression spatial mapping.
- Fluorescence-based: These methods use fluorescent molecules (fluorophores) that emit light upon excitation to visualize proteins or other biomolecules within the tissue (in situ). Signals are detected using fluorescence microscopy, enabling spatial localization and quantification within intact tissue architecture.
Commonly integrated techniques
LC-MS
Liquid chromatography-tandem mass spectrometry (LC-MS) is a widely used, untargeted method of studying protein abundance.[22] LC-MS is generally used to complement spatially proteomics methods such LCM-MS or MALDI-MS, but can also be used to detect metabolites or fats (lipids).[22] The ability of LC-MS to provide a more comprehensive analysis is due to its high sensitivity and depth of molecular data. [22] In LC-MS, samples are broken down into a liquid solution (solubilized) also known as the liquid/mobile phase. The molecules of interest (analytes) in the mobile phase are then measured by their interaction with a stationary phase (a chromatography column).[23] In the context of protein analysis, liquid chromatography provides a fast, highly sensitive measurement informing protein polarity and size. This can be integrated with mass spectrometry (one machine) to produce reliable, high quality, spectral data on proteins of interest. [22][23]
LCM
Laser capture microdissection (LCM) is a technique that combines imaging with microscopic laser cutting.[22] This method is often used alongside mass spectrometry methods for its specificity in selecting precise regions of interest, for example, on particularly heterogenous samples to isolate cell types of interest.[24] LCM uses infrared or ultraviolet lasers. Either approach works for FFPE or fresh frozen tissues.[24]
The resolution of LCM can go down to the single-cell level, meaning it can select precise regions of interest in a tissue or sample alongside preservation of spatial information. Regions of interest are identified first by histology, staining, or other immunohistochemical (IHC) methods, before LCM is performed.[22] Traditional mass spectrometry is often used on these selected regions afterwards for spatial proteomic profiling.[22]
*Table is not comprehensive, but provides an example of the diversity of spatial proteomics methods.
Mass spectrometry-based methods
Labelled mass spectrometry
Labelled MS uses metal-tagged antibodies, which enable clearer distinction of unique targets due to less spectral overlap compared to fluorophores as seen in imaging-based methods.[40]
LOPIT/hyperLOPIT
Localization of Organelle Proteins by Isotope Tagging (LOPIT) and (hyperLOPIT) use isobaric labeling (conjugated tags with the same total mass), subcellular fractionation, and quantitative mass spectrometry to assign proteins to cellular compartments based on their distribution across fractions. Comparing abundance patterns across gradients generates high-resolution organelle maps and spatial proteome atlases. HyperLOPIT increases multiplexing depth and resolution, improving organelle classification.[29][30]
Some challenges LOPIT encounters include requiring well-characterized marker proteins for accurate protein profiling. Proteins are classified using average abundance profiles, so cell-to-cell variability is lost. Proteins that localize to multiple organelles can also be difficult to assign confidently.[30]
Label-free mass spectrometry
Label-free mass spectrometry (IMS) allows detection of biomolecules directly from tissue sections without antibodies, fluorescent probes, or other labels. It generates ion intensity maps representing thousands of molecular species, such as lipids, metabolites, peptides, and small proteins, while preserving spatial context. Because prior target selection and antibody reagents are not required, IMS is considered discovery-driven.[25] Spatial resolution depends on the instrument and ionization method, typically ranging from about ~1–200 μm (eukaryotic cells .[41]
MALDI-MS
Matrix-assisted laser desorption/ionization imaging mass spectrometry (MALDI-MS) is a common method used to map molecules directly in tissue samples. In this technique, a special chemical coating (the matrix) is placed on the tissue. A laser then scans across the surface, and at each spot it measures the molecules present, creating a map that shows where different substances are located.[26] The method does not require antibodies and can detect many types of molecules at the same time, including lipids, peptides, and small proteins.
The level of detail of MALDI-MS is ~5–50 μm, although some advanced systems can reach ~1–5 μm.[25] It can also detect small metabolites, but very small molecules can be harder to see because signals from the chemical coating may interfere. MALDI-IMS is used in research to study differences within tumours, track how drugs spread through tissues, discover new biomarkers, and examine spatial proteomic profiles.[25][26]
DESI-MS
Desorption electrospray ionization mass spectrometry imaging (DESI-MSI) is an ambient ionization technique in which charged solvent droplets desorb and ionize molecules from tissue surfaces. Nano-DESI enhances spatial precision using a controlled liquid microjunction.[27] The method is antibody-free, with spatial resolution of ~20–200 μm (DESI) and ~10–20 μm (nano-DESI). It is best suited to small metabolites, lipids, and drugs, while peptide detection is more limited than with MALDI.[27]
SIMS
Secondary ion mass spectrometry (SIMS) uses a focused ion beam to sputter secondary ions from a sample surface. Spatial resolution is ~1–10 μm, with NanoSIMS achieving ~50–100 nm.[28] SIMS is effective for small metabolites, lipid fragments, and elemental or isotopic imaging, but intact protein detection is limited due to fragmentation.[28]
Imaging/in situ-based methods
CycIF
Cyclic immunofluorescence (CycIF) is one approach using highly multiplexed immunofluorescent imaging. This method is commonly used due to its accessibility– imaging can be conducted with a conventional fluorescence microscope, and the protocol calls for reagents and antibodies available in most labs. Four to six channels can be used for imaging simultaneously, after which the fluorophores are inactivated through bleaching, and a new set of fluorophores can be used, repeating the cycle. Therefore, images from each iteration displays different signals while still maintaining the original architecture of the sample. Using this approach, images from different iterations can be compiled, with up to 30 channels. A notable advantage of this technique is that cell morphology is preserved through multiple CycIF rounds, and the signal to noise ratio increases.[34]
There are 3 core protocols for protein detection outlined in previous studies: direct immunofluorescence with dye-conjugated antibodies, indirect immunofluorescence, and live cell imaging for expressed fluorescent proteins. Overall, CycIF is a method that provides a cost-effective means to acquire quantitative information on protein abundance, localization, microenvironment, and cell morphology. [34]
IBEX
Iterative bleaching extends multiplexity (IBEX) is a high content imaging method that can be done in 2-5 days at relatively low cost and basic laboratory skills.[35] IBEX is also iterative, with multiple cycles of antibody staining, fluorescence imaging, and chemical fluorophore inactivation. More than 65 protein targets can be visualized within a single tissue section.[36]
In each cycle, tissue sections are labeled with fluorophore-conjugated antibodies, imaged, then treated with a reducing agent (commonly lithium borohydride) which chemically inactivates most fluorophores without removing bound antibodies or damaging tissue architecture. This enables repeated rounds of staining and imaging on the same section, increasing the number of detectable markers beyond conventional immunofluorescence. IBEX can spatially resolve complex phenotypes in tissues from a variety of sample types such as animal models and human samples.[35] IBEX is designed as an open workflow compatible with conventional microscopes and commercially available antibodies.[35] The number of markers detected per cycle depends on microscope configuration and fluorophore compatibility and imaging may be limited by tissue thickness.[35][36]
CODEX
Co-detection by indexing (CODEX) is a multiplexed fluorescence microscopy platform for imaging proteins in intact tissue sections.[37] Early versions used oligonucleotide-conjugated antibodies detected through DNA polymerase-mediated incorporation of fluorescent nucleotides.[42] This approach was time-consuming and required large amounts of enzymes and specialized buffers.
Newer CODEX systems use automated microfluidics with a standard fluorescence microscope. Fluorescent DNA probes complementary to antibody-linked oligonucleotides are iteratively hybridized, imaged, and stripped.[37] This enables visualization and quantification of up to ~60 targets in a single tissue section after one staining step, with relatively short run times.[37] Importantly, the cyclic process preserves tissue morphology, allowing simultaneous identification of cell types and their spatial relationships.
However, CODEX has two main limitations: cost and lack of signal amplification. Experiments can be expensive due to the reliance on antibodies, and especially when custom antibodies are required. In addition, without signal amplification, low-abundance proteins (such as transcription factors or cytokines) may be difficult to detect.[37]
Other methods
Mixed methods
IMC
Imaging mass cytometry (IMC) integrates immunohistochemistry with high-resolution laser ablation (using a laser to heat or vaporize cell or tissue samples) and time-of-flight mass spectrometry.[31] Tissue sections are stained with metal isotope–conjugated antibodies, and regions of interest are ablated using a pulsed laser with ~1 μm resolution.[31] The ablated material is transported to a mass cytrometer, where metal tags are quantified to generate multiplexed, spatially resolved protein maps.[31]
IMC preserves tissue architecture and enables analysis of FFPE or frozen sections.[31] It can simultaneously detect 40+ biomarkers without spectral overlap or autofluorescence interference due to the use of distinct metal isotopes.[31] The technique generates quantitative single-cell data while maintaining spatial context, supporting studies of tumour microenvironment organization and cellular neighbourhoods in solid tissues.[31] However, acquisition is time-intensive because of serial laser ablation, and panel design requires extensive antibody validation.[31][43]
MIBI
Multiplexed ion beam imaging (MIBI) uses secondary ion mass spectrometry (SIMS) to image antibodies tagged with elemental metal reporters.[32][33] SIMS provides high resolution and sensitivity, supporting tissue imaging.[44] Tissue sections are stained with labeled antibodies and exposed to a primary ion beam, releasing secondary ions that are quantified by a mass spectrometer.[44] MIBI can measure up to 100 targets simultaneously with minimal interference between detection channels.[32]
MIBI commonly uses dynamic SIMS instruments, which reduce acquisition time but limit the number of masses detected in parallel, as each mass requires a separate ion detector. These instruments are typically limited to a maximum of seven ion detectors due to cost and complexity. Although SIMS can be performed in parallel to reduce acquisition time, cost and technical complexity remain limiting factors.[44]
A newer approach, MIBI-TOF (multiplexed ion beam imaging by time of flight), uses static SIMS instruments to simultaneously detect and quantify all naturally occurring elements (e.g. hydrogen to uranium).[44] MIBI-TOF enables highly multiplexed, subcellular imaging of target expression and localization.[44] However, detection remains limited by the availability of proteoform-specific antibodies.[45]
Artificial intelligence-integrated
Deep Visual proteomics (DVP)[38]
Deep visual proteomics (DVP) is a method that integrates AI for image processing at the stage of data collection. DVP combines high-sensitivity mass spectrometry and single cell/nucleus laser microdissection. Spatial information is collected, while linking specific locations to the amount of protein and their relationship to cellular or subcellular profiles.
This approach uses BIAS (Biology Image Analysis Software),[39] a software employing deep learning-based cell segmentation and machine learning-based identification of cell types from a reference database of FFPE tissue or cell culture images. Regions of interest can be cellular or subcellular, and are precisely selected for either manually or by AI. These regions are then isolated with laser microdissection, and processed to analyse their proteomic profile. [38][39]
Bioinformatics
Analysis of spatial proteomics data typically involves identifying proteins with similar distribution profiles, detecting changes in protein localization, and characterizing sets of proteins that can potentially coordinate for specific biological functions. Protein annotation (such as to organellar compartments) is also an important component.
These analyses use specialized computational frameworks, many of which are available in the R/Bioconductor ecosystem. Increasingly user-friendly tools, such as interactive web browsers, are becoming more common, in which data analysis, visualization, and even integration of other "omics" layers may be performed.
Following preprocessing, quality control, or data preparation, dimensionality reduction is commonly used for visualization of the overall structure of the data set. Subsequent clustering can then reveal underlying biological similarity between samples. This can be unsupervised– where organelle assignments are inferred after clustering, or supervised– where prior knowledge of organellar markers is incorporated into model training to guide classification.
Protein distribution profiles are commonly analyzed with approaches such as profile subtraction[46], particularly in time-course experiments, to detect dynamic changes in protein localization.
Bioinformatic analyses of spatial proteomics data requires data fluency, familiarity with quality control metrics and data structure, and appropriate preprocessing and normalization strategies to ensure robust interpretation. Ongoing challenges in the field include limited pipeline standardization, incomplete analytical workflow automation, and technical constraints regarding low-input samples, all of which affect accessibility and reproducibility.
| Tool[47] | Primary analytical role | Methodological approach | |
|---|---|---|---|
| Data processing / quantification | maxQuant[48] | Protein identification and quantification |
|
| Spatial localization/ classification | pRoloc (software suite) | Organelle classification | Machine learning (semi-supervised, Bayesian) |
| TRANSPIRE[49] | Prediction of intracellular protein translocation | Machine learning (probabilistic Gaussian process classifier) | |
| TransGCN[50] (extension of TRANSPIRE) | Prediction of intracellular protein translocation | Machine learning (semi-supervised graph convolutional network) | |
| Spatial imaging analysis | histoCAT[51] | High-dimensional multiplexed imaging analysis |
|
| DVP[38][39] | High-resolution image analysis | Machine learning (semi-supervised)
| |
| Network-based modeling | DOM-ABC[54] | Domain-informed functional inference | Machine learning (support-vector machines) |
*Table is not comprehensive, but provides an example of the diversity of spatial proteomics data analysis tools.
Limitations and Challenges
For spatial proteomics methods, the limitations and challenges are similar across methods. For imaging/in situ-based methods, spectral overlap is the largest concern, due to the usage of fluorescent antibodies.[55][56] This limits how many antibodies can be used and identified in each sample.[55][56] Additionally, sample/tissue thickness can limit detection for proper imaging (signal detection) for accurate identification and quantification.[56]
Mass spectrometry-based methods have two main limitations: cost and run time. MS instruments are expensive to purchase, and generally run on a per-sample basis.[55][56] These methods are also very time consuming, as tissue sections are scanned in small segments.[55][56] Regardless of the approach used, availability of antibodies targeting specific proteins of interest is always a challenge. If not commercially available, custom antibodies must be ordered, which may be costly, and validation of all the antibodies used (panel validation) adds another layer of cost, time, and complexity to protocols.[55][56]
Overall, the greatest challenge for any method is detection of low abundance proteins.[55][56] As proteins cannot be amplified, proteomic advances are driven by improving sensitivity of technologies instead.[55] Signal amplification of proteins remains a technical limitation and challenge; ultimately detection methods and technologies can still be improved in this sector to overcome this challenge.
Applications
Clinical
Understanding the dynamic complexity and organizational structure of tissues is a fundamental component in understanding disease biology, including initiation, and progression. For instance, immune cell infiltration and its implication in forming the tumor microenvironment is important for understanding pathomechanisms and disease pathogenesis. Spatial proteomics provides spatial and subcellular or subtissue contexts, which inform disease phenotyping, therapeutic targets, and molecular patient stratification impact the development of precision medicine.[47][38][57][58][59]
Current clinical practice generally combines spatial proteomics, or proteomics, with other omics techniques (genomics, transcriptomics, epigenomics, metabolomics).[59]
Research
In the research setting, the incorporation of spatial proteomics has been particularly insightful and explored in biological settings. Spatial proteomic analysis of clinical data has been crucial in understanding developmental biology, disease and disorder biology, infectious disease, cancer biology, and discovery of rare cell populations. Spatial context supplements understanding of ligand-receptor interactions and facilitates identification and evaluation of predictive and prognostic biomarkers, which may provide novel drug and therapeutic targets. [58]
In combination with clinical outcomes such as therapeutic response or survival outcomes, integration of spatial proteomics with other omic techniques is already used in a variety of clinico-research contexts. For example, to guide clinical trial design, such as identifying patient subgroups most likely to benefit from specific treatments, for more accurate patient risk/disease assessment, and in developing precision medicine. Oftentimes, spatial proteomics is implemented retrospectively, on archived or biobanked FFPE tissues.
In pharmaceutical and biotechnology settings, spatial proteomics is increasingly integrated into drug discovery and translational development workflows. Spatial resolution enables target identification, pharmacodynamic mapping, and assessment of drug-target engagement within intact tissues. Imaging and mass spectrometry–based platforms have been applied to evaluate spatial drug distribution, characterize mechanisms of therapeutic resistance, and supports better patient selection for clinical trials, where patient stratification is key in optimizing drug effectiveness.[60][59]