Most protein testing in the food industry still relies on measuring nitrogen, not protein itself. The two dominant workhorses, the Kjeldahl method and the Dumas combustion method, both estimate total protein by detecting how much nitrogen is in a sample and multiplying by a conversion factor. Beyond these, a range of dye-binding assays, spectroscopic tools, and antibody-based tests serve different purposes depending on whether the goal is a bulk protein number on a nutrition label, a quality check on a production line, or the detection of a specific allergenic protein hiding in a processed product.
Kjeldahl and Dumas, the Two Pillars
If you see a protein value on a food label, it almost certainly traces back to one of two nitrogen-based methods. The Kjeldahl method, developed in the 1880s, digests a food sample in concentrated sulfuric acid to convert organic nitrogen into ammonium, which is then measured. It has been the official reference method for protein in food for over a century, and regulatory agencies worldwide still treat it as the benchmark.1Europe PMC. Measuring Protein Content in Food: An Overview of Methods
The Dumas method works differently: it burns the sample at high temperature in the presence of oxygen, then traps and measures the nitrogen gas released.2Journal of Cereal Science. Cereal protein analysis via Dumas method: Standardization of a micro-method using the EuroVector Elemental Analyser It is faster than Kjeldahl, often finishing a single sample in a few minutes rather than the hour or more Kjeldahl can take. It also avoids the corrosive chemicals. For these reasons, Dumas has gained ground in grain and cereal testing, though Kjeldahl remains the legal standard in many regulatory frameworks.
Both methods share a fundamental design choice: they measure nitrogen, not protein. That distinction matters more than it might seem.
The 6.25 Problem
Once you know how much nitrogen is in a food, you need to convert that number into a protein estimate. The traditional conversion factor is 6.25, based on the assumption that proteins contain about 16 percent nitrogen. Multiply the nitrogen content by 6.25, and you get your protein value. Simple enough, except that this assumption is only roughly correct and often overshoots.
Different proteins have different amino acid compositions, which means different nitrogen contents. Dairy proteins, for instance, contain proportionally more nitrogen per gram of protein than soy proteins do. Amino acid analyses have shown that the blanket factor of 6.25 overestimates protein content for most foods.3Journal of the American Oil Chemists’ Society. Calculation of Nitrogen‐to‐Protein Conversion Factors: A Review with a Focus on Soy Protein Food-specific factors do exist. Wheat uses 5.70, milk uses 6.38, and so on. But the lack of standardized methods for deriving these food-specific factors means that many labs and regulations still default to 6.25.4PubMed. Converting nitrogen into protein–beyond 6.25 and Jones’ factors
The practical result is that many protein values on labels are slight overestimates. For the average consumer, the difference is small. For researchers studying protein quality, or for companies formulating sports nutrition products where protein content is a selling point, the gap can be more meaningful.
When Nitrogen Is Not Protein at All
The deeper weakness of nitrogen-based methods is that not all nitrogen in food comes from protein. Compounds like urea, ammonia, free amino acids, and nucleic acids all contain nitrogen but are not dietary protein. In honest food, this non-protein nitrogen is usually a small fraction and does not meaningfully distort the result. The problem arises when someone deliberately adds a nitrogen-rich compound to inflate apparent protein content.
The most notorious example is the 2008 melamine scandal, in which producers added melamine to milk and infant formula to inflate protein readings on standard nitrogen tests. Conventional nitrogen-based methods could not distinguish melamine nitrogen from real protein nitrogen.5Food Control. Detection of non-protein nitrogen (NPN) milk adulterants using Raman spectroscopy with machine learning The episode sickened hundreds of thousands of children and killed several. It remains one of the starkest demonstrations that measuring nitrogen alone has a built-in blind spot.
Beyond melamine, other nitrogen-rich adulterants like ammonium sulfate and urea have been documented in the dairy supply chain.5Food Control. Detection of non-protein nitrogen (NPN) milk adulterants using Raman spectroscopy with machine learning In the aftermath of the melamine crisis, labs worldwide developed and implemented targeted screening methods using techniques like chromatography and mass spectrometry to detect these specific compounds in products.6PubMed. Methods for the analysis of melamine and related compounds in foods: a review Testing for melamine in milk-containing products has since become routine in many countries. Studies have found detectable melamine in a wide range of commercial products that contain powdered milk, though typically at levels well below regulatory limits.7PubMed Central. Determination of melamine contamination in chocolates containing powdered milk by high-performance liquid chromatography (HPLC)
This vulnerability has prompted a broader conversation about whether nitrogen-based methods are sufficient for protecting public health, or whether the industry needs to move toward methods that measure actual protein rather than a proxy for it.8PubMed. Total Protein Methods and Their Potential Utility to Reduce the Risk of Food Protein Adulteration
Dye-Binding and Colorimetric Assays
For labs that need to measure protein in a dissolved sample rather than a bulk commodity, several colorimetric methods offer a different approach. These work by reacting a chemical with the protein in solution and measuring the color change, which correlates with protein concentration.
The Bradford assay is one of the most widely used. It works by mixing a dye called Coomassie Brilliant Blue G-250 with a protein solution. The dye binds to protein molecules and shifts its peak light absorption from 465 to 595 nanometers, a color change that can be read by a spectrophotometer.9Analytical Biochemistry. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding The Bradford assay is fast, cheap, and sensitive to very small amounts of protein, which makes it popular in research labs. Its main drawback is that it responds differently to different proteins, so you need to compare your results against a known standard, and the choice of standard affects your answer.
The Lowry method and the bicinchoninic acid (BCA) assay both rely on copper-based chemistry. They react copper ions with the peptide bonds in proteins, then use a secondary color reaction to amplify and detect the signal. These methods tend to be more consistent across different protein types than Bradford, but they are also more sensitive to interference from sugars, lipids, and other common food components. The biuret assay is a simpler, less sensitive copper-based method that has been adapted for certain food matrices like corn-based products.10PubMed. A practical method for extending the biuret assay to protein determination of corn-based products
An important advantage of these colorimetric assays is that they respond to actual protein, not just nitrogen. That means they are not fooled by melamine or urea the way Kjeldahl and Dumas can be. The trade-off is that they generally require the sample to be in liquid form, need careful calibration, and are less suited to high-throughput testing of whole food commodities. They work best in research and quality-control settings where individual samples are being assessed rather than truckloads of grain.
UV Absorbance at 280 Nanometers
A straightforward way to estimate protein in a clear solution is to shine ultraviolet light through it and measure how much is absorbed at 280 nanometers. Certain amino acids, particularly tryptophan and tyrosine, absorb UV light strongly at that wavelength. By comparing the measured absorbance against a standard curve or known values for the specific protein, you can calculate the concentration.11PubMed. Spectrophotometric determination of protein concentration
This method is fast and requires no reagents, which makes it attractive for routine monitoring in settings like dairy processing or beverage production. Its main limitation is that it works only on clear solutions, since anything that scatters light or absorbs at the same wavelength will throw off the reading. It also depends on the protein having a predictable amino acid composition. For mixed or unknown protein sources, it is less reliable than dye-binding methods.
Near-Infrared Spectroscopy for Rapid Screening
Near-infrared (NIR) spectroscopy has become one of the most practical tools for protein testing in industrial settings. It works by bouncing near-infrared light off a sample and analyzing which wavelengths are absorbed. Different chemical bonds absorb at different wavelengths, so a trained model can pick out the protein signal from the overall spectrum and estimate concentration.12PubMed Central. The Role of Near-Infrared Spectroscopy in Food Quality Assurance: A Review of the Past Two Decades
The selling point of NIR is speed and non-destructiveness. You can scan a pile of grain, a conveyor belt of cereal, or a batch of protein powder without grinding, dissolving, or destroying the sample. Readings take seconds. Many grain elevators, dairy plants, and feed mills use NIR instruments as their primary screening tool, calibrating them against Kjeldahl or Dumas reference measurements periodically to maintain accuracy.
NIR has been explored for newer applications too, like estimating protein and oil content in plant-based product mixtures for emerging food categories.13Journal of Food Composition and Analysis. Near-infrared spectroscopy-based quantification of sunflower oil and pea protein isolate in dense mixtures for novel plant-based products The main limitation is that NIR needs a robust calibration model built from many reference samples, and it can struggle with unfamiliar sample types or complex matrices it was not trained on. It is a screening tool, not a standalone analytical method, and results that look unusual still get sent to a wet chemistry lab for confirmation.
ELISA and Immunological Methods for Specific Proteins
Sometimes the question is not “how much total protein is in this food” but rather “is a specific protein present, and how much of it is there?” This matters enormously for allergen testing. A consumer with a soy allergy does not care about total protein; they need to know whether soybean proteins are lurking in a product labeled soy-free.
Enzyme-linked immunosorbent assays, known as ELISAs, use antibodies that bind specifically to target proteins. A well-designed sandwich ELISA can detect soybean protein residues at levels as low as a fraction of a milligram per kilogram of food, with recovery rates that confirm the method is measuring real contamination rather than noise.14Process Biochemistry. Development and validation of a specific sandwich ELISA for determination of soybean allergens and its application in processed foods Similar approaches have been developed for cashew allergens, with detection limits in the low nanogram-per-milliliter range, sensitive enough to flag trace contamination in commercial products.15PubMed. Developing a dual-antibody Sandwich ELISA and LFIA for detecting the cashew allergen Ana o 3 in foods
ELISA kits are available commercially for all of the major food allergens, and food manufacturers use them routinely to verify cleaning procedures and check finished products. Lateral flow strips, which work on the same antibody principle but in a simpler format, provide rapid yes-or-no answers on the factory floor without lab equipment. These methods do not tell you total protein content. They tell you whether a specific protein is present and, in quantitative versions, roughly how much. That makes them complementary to nitrogen-based and colorimetric methods rather than replacements.
Mass Spectrometry and Electrophoresis for Authentication
At the high end of protein analysis, mass spectrometry (MS) can identify individual proteins and even individual peptide fragments with extreme precision. In food authentication, liquid chromatography coupled with tandem mass spectrometry has been used to detect species-specific peptide markers in processed meat products. One comprehensive effort identified over a hundred peptides, with dozens verified as reliable markers that could pinpoint which animal species were present in a highly processed sample.16PubMed. Liquid chromatography-mass spectrometry bottom-up proteomic methods in animal species analysis of processed meat for food authentication and the detection of adulterations This kind of analysis is what regulators turn to when they suspect a product labeled as beef actually contains horse or pork.
Mass spectrometry has also been applied to detecting undeclared milk proteins in processed foods, where even trace amounts can trigger allergic reactions in sensitive individuals.17Journal of AOAC INTERNATIONAL. Development and Validation of a Method for the Quantification of Milk Proteins in Food Products Based on Liquid Chromatography with Mass Spectrometric Detection The advantage over ELISA is that MS does not depend on antibodies, which can degrade or cross-react with non-target proteins. The disadvantage is cost and complexity; mass spectrometry requires expensive instruments and trained operators, making it impractical for routine screening.
Gel electrophoresis, particularly SDS-PAGE, occupies a middle ground between simple and advanced methods. It separates proteins by molecular weight, producing a banding pattern that serves as a kind of fingerprint. Researchers have used SDS-PAGE protein profiles to verify the botanical origin of honey by looking for specific protein bands associated with particular floral sources.18Food Analytical Methods. The Application of SDS-PAGE Protein and HPTLC Amino Acid Profiling for Verification of Declared Variety and Geographical Origin of Honey Similar work has applied protein profiling to berry seeds, identifying specific polypeptide bands that can distinguish between species and detect adulteration in berry-based products.19PubMed. Comprehensive electrophoretic profiling of proteins as a powerful tool for authenticity assessment of seeds of cultivated berry fruits Electrophoresis does not give you a total protein number, but it tells you what kinds of protein are present, which is invaluable when the question is authenticity rather than quantity.
Choosing the Right Method
The method you need depends entirely on what question you are asking. For total protein content in a bulk commodity like flour, milk powder, or animal feed, Kjeldahl or Dumas remains the standard, and NIR provides a fast screening complement. For research on protein in solution, the Bradford, Lowry, or BCA assay is more appropriate. For allergen testing, ELISA kits and lateral flow strips are the workhorses. For authentication and fraud detection, mass spectrometry and electrophoresis offer molecular-level detail that no other approach matches.
Cost and throughput matter too. A Kjeldahl analysis might cost a few dollars per sample in a commercial lab, but each sample takes significant hands-on time and uses corrosive reagents. Dumas is faster and cleaner but requires an expensive combustion analyzer. NIR instruments have a high upfront cost but are essentially free per sample after calibration. ELISA kits cost a moderate amount per test but need no major instrumentation. Mass spectrometry sits at the top in both analytical power and price.
No single method answers every protein question. The food industry has settled on a layered approach: fast screening tools like NIR for incoming raw materials, nitrogen-based reference methods for regulatory compliance, colorimetric assays and ELISA for targeted quality control, and chromatographic or mass spectrometric methods reserved for investigations into fraud, contamination, or product identity disputes.
Emerging Challenges with Novel Protein Sources
As the food industry expands into alternative proteins, including plant-based formulations, insect-derived ingredients, and cultured products, existing testing methods face new complications. Many of these novel sources have amino acid profiles and nitrogen contents that differ from the traditional commodities the testing infrastructure was built around. Applying the old conversion factors to a cricket flour or a pea-rice protein blend can yield numbers that do not accurately reflect true protein content.
Insect proteins pose particular challenges. The expansion of edible insects into Western diets has outpaced the development of safety and authenticity testing methods, with substantial information gaps around processing, contamination, and fraud risk in these supply chains.20PubMed Central. An analysis of emerging food safety and fraud risks of novel insect proteins within complex supply chains Insect exoskeletons contain chitin, a nitrogen-rich polymer that is not protein but registers as nitrogen on Kjeldahl and Dumas tests. Without chitin-specific corrections, protein values for insect-based foods can be substantially inflated.
For plant-based products that blend multiple protein sources, NIR models trained on single-ingredient commodities may not transfer well. Researchers are actively building new calibration sets for mixtures of ingredients like pea protein isolate and sunflower oil that reflect real formulations on the market. The testing landscape is adapting, but it lags behind the pace of product innovation. If you are evaluating protein claims on a novel food product, the safest assumption is that the number on the label was derived from nitrogen measurement with a generic or food-specific conversion factor, and that the true protein content could be slightly lower than reported.