Proteomics is the large-scale study of all the proteins in a cell, tissue, or organism, along with what those proteins do, how they change, and how they interact. Where genomics reads the instruction manual (DNA), proteomics looks at the actual molecular machinery doing the work. The distinction matters because genes alone cannot tell you which proteins are present at a given moment, how abundant they are, or what chemical modifications they carry. The field has grown from cataloging a handful of proteins at a time to profiling thousands in a single experiment, and its applications now reach from cancer diagnostics to gut microbiome research.
How Mass Spectrometry Reads the Proteome
The workhorse technology behind most proteomics research is mass spectrometry, which identifies proteins by measuring the mass of their molecular fragments with extreme precision. The most common approach, called bottom-up proteomics, works by first breaking proteins into smaller peptide pieces using digestive enzymes, then separating those peptides by their chemical properties using liquid chromatography, and finally feeding them into a mass spectrometer that measures each peptide’s mass and fragmentation pattern. Matching those patterns against databases of known protein sequences lets researchers figure out which proteins were in the original sample.
A typical bottom-up workflow involves extracting proteins from a tissue sample, digesting them, purifying the resulting peptides, and then running them through nanoscale liquid chromatography coupled to a high-resolution mass spectrometer. Each step introduces potential sources of error and bias, from incomplete digestion to peptide losses during cleanup, so careful sample preparation is critical to getting reliable results.1PubMed Central. A Critical Review of Bottom-Up Proteomics: The Good, the Bad, and the Future of this Field2PubMed. Mass Spectrometry-Based Bottom-Up Proteomics: Sample Preparation, LC-MS/MS Analysis, and Database Query Strategies
An alternative called top-down proteomics skips the digestion step entirely and feeds intact proteins into the mass spectrometer. This preserves information that bottom-up methods destroy: the specific combination of chemical modifications, sequence variants, and splice forms that make each protein molecule unique. Researchers call these distinct molecular forms “proteoforms,” and a single gene can give rise to dozens of them. Understanding which proteoforms are present in a sample offers deeper insight into how proteins actually function in their biological context.3PubMed Central. Progress in Top-Down Proteomics and the Analysis of Proteoforms4PubMed Central. Top-Down Proteomics and the Challenges of True Proteoform Characterization
A third strategy, data-independent acquisition (DIA), has gained popularity because it collects fragmentation data on all detectable peptides in a sample rather than selecting only the most abundant ones. This produces permanent digital maps that researchers can revisit and reanalyze later as databases improve or new questions arise, making experiments more reproducible and comprehensive over time.5PubMed. Advances in data-independent acquisition mass spectrometry towards comprehensive digital proteome landscape
Measuring Protein Abundance
Knowing which proteins are in a sample is only part of the picture. Researchers also need to know how much of each protein is present and how those levels change between, say, healthy and diseased tissue. Quantitative proteomics tackles this using two broad strategies: labeling and label-free methods.
Labeling approaches attach chemical or metabolic tags to proteins or peptides so that samples from different conditions can be mixed together and analyzed simultaneously. Several well-known labeling strategies exist, including isotope-coded tags for relative quantification and metabolic labeling methods where cells incorporate heavy amino acids as they grow. Label-free approaches, by contrast, skip the tagging step and instead compare the signal intensity or spectral counts of peptides across separate runs. Both families of methods have trade-offs in throughput, accuracy, and cost, and the choice depends on the experimental question.6PubMed. Stable isotope labelling methods in mass spectrometry-based quantitative proteomics
Why Post-Translational Modifications Matter
A cell does not rely solely on turning genes on and off to control what happens inside it. One of the most powerful regulatory mechanisms involves chemically modifying proteins after they have been built, a process called post-translational modification. Attaching a small phosphate group to a signaling protein, for example, can switch it on or off in milliseconds, much faster than making a new protein from scratch. Cells use hundreds of different types of modifications to fine-tune protein activity, localization, stability, and interactions.
Mass spectrometry-based proteomics has revealed just how extensive this modification landscape is. For several major modification types, including phosphorylation, ubiquitination, glycosylation, and acetylation, researchers can now confidently identify and pinpoint tens of thousands of individual modification sites across the proteome.7PubMed Central. Status of large-scale analysis of post-translational modifications by mass spectrometry Specialized enrichment workflows allow researchers to profile multiple types of modifications from the same biological sample. One such method enabled the quantitative analysis of roughly 8,000 proteins along with more than 20,000 phosphorylation sites, 15,000 ubiquitination sites, and 3,000 acetylation sites in a single experiment, providing a sweeping view of how signaling pathways respond to a drug treatment.8Nature Methods. Integrated proteomic analysis of post-translational modifications by serial enrichment
This is one of the reasons proteomics delivers information that genomics cannot. Two cells with identical DNA can have radically different protein modification profiles, and those differences can mean the difference between a cell that divides normally and one that becomes cancerous.
Affinity-Based Platforms Beyond Mass Spectrometry
Mass spectrometry is not the only way to profile proteins at scale. Affinity-based platforms use engineered binding molecules, either short DNA-like probes called aptamers or antibody-based reagents, to capture and measure specific proteins in blood or other fluids. These platforms are especially attractive for large clinical studies because they can process thousands of samples quickly and measure thousands of proteins from a small volume of plasma.
The two leading commercial platforms are SomaScan, which uses aptamer probes and can currently quantify roughly 1,100 to 11,000 proteins in a plasma sample, and Olink, which uses antibody-based proximity extension assays to measure between about 48 and 5,400 proteins.9Nature Communications. Cross-ancestry comparison of aptamer and antibody protein measures Head-to-head comparisons of these platforms have found trade-offs: Olink tends to show more reliable specificity for its protein targets and produces more associations with clinical traits, while SomaScan provides broader proteome coverage and higher measurement precision.10PubMed Central. Proteomic profiling platforms head to head: Leveraging genetics and clinical traits to compare aptamer- and antibody-based methods
These technologies have opened the door to large-scale population studies. Instead of analyzing a few hundred research participants by mass spectrometry, investigators can now profile the plasma proteome across tens of thousands of people, looking for protein signatures tied to cardiovascular disease, diabetes, and other conditions.11PubMed Central. Emerging Affinity-Based Proteomic Technologies for Large-Scale Plasma Profiling in Cardiovascular Disease
Mapping the Entire Human Proteome
Since the early 2010s, the Human Proteome Project has been working to catalog every protein encoded by the human genome and build reliable assays to detect each one. The project is framed around roughly 19,400 protein-coding genes, and the goal is to confirm experimental evidence of protein expression for each of them. As of the 2024 report, about 18,138 of those genes, roughly 93%, have confirmed protein-level evidence. That leaves around 1,273 “missing proteins” that have not yet been reliably detected.12PubMed Central. The 2024 Report on the Human Proteome from the HUPO Human Proteome Project
These missing proteins are not hypothetical. Their genes exist, and bioinformatics predicts they should produce proteins. But many are expressed only in trace amounts, in rare cell types, or during very narrow windows of development, making them extremely difficult to detect even with the most sensitive instruments. Researchers have pursued creative strategies like enriching low-abundance proteins before analysis to coax these hidden molecules into view.13PubMed. Digging More Missing Proteins Using an Enrichment Approach with ProteoMiner Completing the human proteome map remains one of the field’s defining challenges.14PubMed. Progress and pitfalls in finding the ‘missing proteins’ from the human proteome map
Single-Cell and Spatial Proteomics
Until recently, proteomics required pooling millions of cells to get enough protein material for a measurement. That meant every result was an average, hiding the diversity that exists between individual cells in the same tissue. Single-cell proteomics is changing that. Over the past two to three years, improvements in sample preparation, multiplexing strategies, and specialized mass spectrometry hardware have made it possible to profile proteins from tiny, picogram-level inputs.15PubMed Central. Single-cell proteomics using mass spectrometry Recent advances allow quantification of roughly half the expressed proteome from an individual cell.16PubMed Central. Entering the era of deep single-cell proteomics
Spatial proteomics adds another dimension by asking not just what proteins are present, but where they are. Inside a cell, the same protein can behave very differently depending on whether it sits at the cell membrane, in the nucleus, or in a recycling compartment. Spatial methods systematically map the locations of thousands of proteins in parallel, revealing how organelle composition shifts when a cell is stressed, dividing, or responding to a drug.17PubMed Central. Advances and Applications of Spatial Proteomics: From Organellar Maps to Clinical Translation
Cancer Biomarkers and Precision Oncology
One of the most active application areas for proteomics is cancer. Blood plasma is a rich source of potential biomarkers because it collects secreted proteins from tissues throughout the body. Mass spectrometry and affinity-based platforms are being used to search for protein signatures that could detect cancer early or distinguish aggressive tumors from indolent ones.18PubMed Central. Proteomics and mass spectrometry for cancer biomarker discovery19PubMed. Proteomic profiling of human plasma for cancer biomarker discovery
The ambition goes beyond finding diagnostic markers. Proteogenomics, which combines genomic sequencing with deep protein profiling of the same tumor, is being positioned as a way to guide treatment decisions in oncology. By layering protein data on top of mutation data, researchers can get a more functional picture of what is actually driving a particular patient’s tumor. Large consortia have applied this approach to multiple cancer types, and the case has been made that proteogenomics needs to be integrated into clinical trials for precision oncology to reach its potential.20PubMed Central. The next horizon in precision oncology: Proteogenomics to inform cancer diagnosis and treatment In liver cancer, for instance, researchers used patient-derived organoids combined with proteogenomic profiling to identify molecular features linked to drug responses, predicting which drug combinations might work best for individual patients.21PubMed Central. Pharmaco-proteogenomic characterization of liver cancer organoids for precision oncology
This work is still largely in the research phase. Few proteomic biomarker panels have crossed into routine clinical use yet, but the pipeline is growing and the underlying technology continues to improve.
Finding Drug Targets
Proteomics also plays a growing role in drug discovery, particularly in figuring out what a drug actually binds to inside a cell. One powerful method, thermal proteome profiling, exploits the fact that when a drug binds a protein, it typically changes how that protein responds to heat. By gradually heating cells that have been treated with a drug versus untreated cells and then measuring which proteins become more or less stable, researchers can identify drug targets across the entire proteome without needing to chemically modify the drug itself.22PubMed. Thermal Proteome Profiling for Drug Target Identification and Probing of Protein States
Thermal proteome profiling remains the only approach that can examine protein states in live cells under native conditions at a proteome-wide scale.23PubMed. Drug Target Identification in Tissues by Thermal Proteome Profiling It has been used to confirm intended drug targets, reveal unexpected off-target interactions, and explore how a protein’s binding partners shift under different cellular conditions. For drug development, these off-target hits can be just as informative as the intended target, sometimes explaining side effects or suggesting new therapeutic uses.
Protein Turnover and Dynamics
A static snapshot of which proteins are present at one moment does not capture the full picture of cellular life. Proteins are constantly being made and degraded, and the rate at which this happens, known as protein turnover, varies enormously from one protein to the next. Some proteins persist for weeks; others are replaced within hours. Disruptions in turnover are linked to diseases like neurodegeneration, where misfolded proteins accumulate because the cell’s disposal machinery fails.
Recent mass spectrometry methods allow researchers to measure the turnover rates of thousands of proteins simultaneously. By feeding cells or animals amino acids labeled with heavy isotopes and tracking how quickly those labels appear in newly made proteins, researchers can calculate degradation and synthesis rates across the proteome in a single experiment.24PubMed Central. JUMPt: Comprehensive Protein Turnover Modeling of In Vivo Pulse SILAC Data by Ordinary Differential Equations This kind of dynamic proteomics is especially useful for understanding how aging, disease, and drug treatments alter the balance between protein production and disposal.
Proteomics of the Gut Microbiome
Proteomics is not limited to human cells. Metaproteomics applies the same mass spectrometry tools to the complex microbial communities living inside the human gut and elsewhere in the environment. Unlike genomic methods that tell you which microbes are present and what genes they carry, metaproteomics reveals which proteins those microbes are actually producing, offering a direct readout of community-level metabolic activity.25PubMed Central. Microbial metaproteomics for characterizing the range of metabolic functions and activities of human gut microbiota
Ultra-deep metaproteomic analyses have revealed a striking degree of functional redundancy in the gut microbiome: many different bacterial species produce overlapping sets of proteins with similar functions. This redundancy may help explain why the gut ecosystem remains stable even as specific microbial species come and go, since the loss of one species’ contribution can be covered by another.26Nature Communications. Revealing proteome-level functional redundancy in the human gut microbiome using ultra-deep metaproteomics Metaproteomics also captures host-microbe interactions: alongside bacterial proteins, researchers detect human proteins secreted into the gut, providing clues about how the immune system and microbial inhabitants communicate.
Protein Aggregation in Neurological Disease
Proteomics has proved valuable for studying diseases in which proteins misfold and clump together. In neurodegenerative conditions, the buildup of insoluble protein aggregates is a hallmark of disease. Proteomic profiling of brain tissue after ischemic stroke in mice revealed that ischemia triggers the aggregation of many proteins already linked to neurodegeneration, including TDP-43, FUS, and several other RNA-binding proteins associated with amyotrophic lateral sclerosis and frontotemporal dementia. RNA-binding proteins turned out to constitute the largest group of aggregating proteins after ischemia.27Scientific Reports. Cerebral ischemia induces the aggregation of proteins linked to neurodegenerative diseases
Findings like these point to molecular overlaps between conditions traditionally studied in isolation. The idea that stroke and long-term neurodegeneration share aggregation-prone protein networks was not obvious from genetics alone. It emerged because proteomics could directly survey which proteins became insoluble in damaged brain tissue, then compare that list against the catalog of proteins implicated in other diseases. That kind of unbiased discovery, where you look at everything and let the data reveal unexpected connections, is one of proteomics’ greatest strengths.
The Computational Side of Proteomics
Generating proteomic data is only half the battle. A single mass spectrometry run can produce hundreds of thousands of spectra, each representing a candidate peptide. Matching those spectra to protein sequences requires sophisticated search algorithms, and controlling the rate of false identifications is a persistent challenge. The standard approach uses a “decoy” database of fake protein sequences: if the search engine matches spectra to fake sequences at a known rate, that provides an estimate of how many of the real matches might also be wrong. Refining this process to keep the false discovery rate accurate remains an active area of method development.28PubMed Central. False discovery rate estimation using candidate peptides for each spectrum
These statistical concerns are not just academic bookkeeping. In clinical biomarker studies, a falsely identified protein could lead to years of wasted follow-up research. In the Human Proteome Project, a single incorrect identification could cause a “missing protein” to be prematurely checked off the list. The field has invested heavily in standardizing both the computational pipelines and the reporting criteria so that results published by one laboratory can be trusted and reproduced by another.29PubMed. Gentle Introduction to the Statistical Foundations of False Discovery Rate in Quantitative Proteomics
Structural Proteomics and Protein Interaction Networks
Beyond identifying and quantifying proteins, researchers want to understand how proteins physically interact with each other. Structural proteomics uses techniques like cross-linking mass spectrometry, in which a chemical “staple” is used to freeze two proteins (or two regions of the same protein) in close contact. After digestion and mass spectrometry, the cross-linked peptides reveal which parts of the proteins were near each other, providing distance constraints that feed into computational modeling of protein complex structures.30PubMed Central. Cross-linking mass spectrometry for mapping protein complex topologies in situ
When combined with other structural information, these distance measurements can generate plausible models of how multi-protein complexes are assembled, even inside intact cells. The technique has been applied to large complexes like RNA polymerase, helping researchers map how accessory proteins dock onto the transcription machinery.31Nature Communications. Reliable identification of protein-protein interactions by crosslinking mass spectrometry As AI-based structure prediction tools continue to improve, cross-linking data provides an experimental check on computationally generated models, grounding predictions in physical reality.