HPLC testing laboratory · Davis, CaliforniaPublished pricesBook onlineinfo@openchemworks.com
Insights

Nitrogen to protein conversion factors: why 6.25 is the wrong number for most foods

Published November 5, 2026Updated November 19, 2026OpenChemWorks Laboratory

Updated November 19, 2026

At a glance: what the numbers say

  • The default factor is wrong in a known direction for most honest ingredients. 6.25 assumes protein is 16.00% nitrogen. Published Jones factors imply nitrogen contents from 15.67% (milk) to 19.31% (almonds). Using 6.25 where the Jones factor applies overstates protein for 12 of the 16 ingredient classes in our table, understates it for 2 (milk at 6.38 and wheat bran at 6.31), and is exact for the 2 classes whose Jones factor is 6.25 itself.
  • Almonds are the largest honest error: +20.7%. Jones factor 5.18 vs 6.25. On FoodData Central's almond profile (21.15 g protein per 100 g, Jones basis), declaring at 6.25 would print 25.52 g, an overstatement of 4.37 g per 100 g, with no spiking, no fraud, and no bad lab work. Just the wrong constant.
  • Gelatin overstates by 12.6%, peanuts by 14.5%, soy by 9.5%, wheat flour by 9.6%. Computed per class in Table 1; anchored to FoodData Central profiles in Table 2 (soybeans: 36.49 g becomes 39.94 g; peanuts: 25.80 g becomes 29.53 g).
  • Milk runs the other way. Jones factor 6.38: 6.25 understates milk protein by 2.0%. On whole milk's 3.15 g per 100 g, the 6.25 declaration would be 3.09 g. The default is not a conservative choice; it is simply a different error per ingredient.
  • Per gram of nitrogen, the spread is 5.18 to 6.38 g of "protein." One gram of measured nitrogen becomes 6.25 g of declared protein by default, but 5.18 g in almonds and 6.38 g in milk, a 23.2% spread (6.38 ÷ 5.18 = 1.232) created entirely by the choice of constant.
  • So what: If your label, your specification, or your feed guarantee was computed at 6.25 for an ingredient with a Jones factor well below 6.25, the number is overstated by a computable amount, and an amino acid sum exposes it without any argument about methods. Check your factor before someone else does.
  • Why it matters: Protein is the headline number on protein foods. A systematic 10 to 21% overstatement on nuts, seeds, soy, and wheat ingredients is large enough to fail a retailer verification, misprice a formulation, and mislead a consumer, even when every measurement underneath it was done correctly.

Introduction

How much protein does a nitrogen measurement really support? The measurement itself gives nitrogen; the protein figure on a label comes from multiplying that nitrogen by a conversion factor, and the default factor, 6.25, assumes every protein carries the same share of nitrogen. This article asks how large the error from that assumption is, ingredient by ingredient, and in which direction it runs.

The computation uses the published Jones factors for each ingredient class, from which the implied nitrogen share and the signed error of using 6.25 are derived. It then anchors that arithmetic to real FoodData Central profiles, backing out the nitrogen each profile implies and re-declaring it at the default factor, so the abstract error becomes grams per 100 g on foods a reader can recognise.

Read the first table as the factor map: where the default overstates, where it understates, and where it is exact. Read the second for what that error is worth on real profiles. The perspectives that follow cover the label, analytical, and historical angles, including the amino acid sum as a cross-check that never passes through a nitrogen constant at all.

Signed error of using 6.25 instead of the Jones factor

Positive bars overstate protein; the two negative bars understate it. Orange is almonds, the largest error.

Almonds+20.7%
Other tree nuts and oil seeds+17.9%
Peanut+14.5%
Brazil nut+14.5%
Gelatin+12.6%
Wheat flour (endosperm)+9.6%
Soybean+9.5%
Barley+7.2%
Oats+7.2%
Rye+7.2%
Wheat, whole kernel+7.2%
Rice+5.0%
Corn (maize)0.0%
Other animal products0.0%
Wheat bran-1.0%
Milk and milk products-2.0%
-5%0+25%
Figure 1: Error introduced by the 6.25 default for each ingredient class, relative to its Jones factor. Most classes are overstated; milk and wheat bran run the other way.

Table 1. Jones factors and the signed error of the 6.25 default

Jones factors from Jones (1941), USDA Circular 183, as tabulated in FAO food composition tables and in Mariotti, Tomé & Mirand (2008), Critical Reviews in Food Science and Nutrition 48:177–184. Implied nitrogen % = 100 ÷ factor (computed). Error of 6.25 = (6.25 − Jones) ÷ Jones × 100 (computed; positive = 6.25 overstates).

Ingredient class Jones factor Implied N % of protein Error of using 6.25
Almonds 5.18 19.31 +20.7%
Other tree nuts and oil seeds 5.30 18.87 +17.9%
Peanut 5.46 18.32 +14.5%
Brazil nut 5.46 18.32 +14.5%
Gelatin 5.55 18.02 +12.6%
Wheat flour (endosperm) 5.70 17.54 +9.6%
Soybean 5.71 17.51 +9.5%
Barley 5.83 17.15 +7.2%
Oats 5.83 17.15 +7.2%
Rye 5.83 17.15 +7.2%
Wheat, whole kernel 5.83 17.15 +7.2%
Rice 5.95 16.81 +5.0%
Corn (maize) 6.25 16.00 0.0%
Other animal products 6.25 16.00 0.0%
Wheat bran 6.31 15.85 −1.0%
Milk and milk products 6.38 15.67 −2.0%

Worked example (Table 1, almonds). Implied nitrogen % = 100 ÷ 5.18 = 19.305 → 19.31%. Error = (6.25 − 5.18) ÷ 5.18 × 100 = 1.07 ÷ 5.18 × 100 = 20.66 → +20.7%. Interpretation: almond protein really carries about 19.31 g of nitrogen per 100 g of protein (the amandin figure traces to Osborne & Campbell's 19.3% nitrogen for the almond globulin amandin, cited by Jones). A nitrogen method that assumes 16% therefore invents roughly a fifth more protein than the nitrogen can represent.

What the same nitrogen declares at 6.25, on real profiles

Protein per 100 g after re-declaring the nitrogen in each FoodData Central profile at the 6.25 default.

Gelatin, dry powder96.40 g
Soybeans, mature seeds, raw39.94 g
Peanuts, all types, raw29.53 g
Almonds25.52 g
Egg, whole, raw12.56 g
Wheat flour, white, all-purpose11.33 g
Rice, white, long-grain, raw7.49 g
Milk, whole3.09 g
048.20 g96.40 g
Figure 2: Declared protein at 6.25 for the anchor profiles in Table 2. Gelatin shows the largest absolute gap between the Jones-basis value and the 6.25 declaration.

Table 2. What the error is worth on real profiles (g per 100 g)

Method: treat the FoodData Central protein value as the Jones-basis value (FDC protein for these foods is conventionally derived with specific factors; see Limitations), back out the nitrogen the measurement must have seen (N = protein ÷ Jones factor), then re-declare that same nitrogen at 6.25. The nitrogen does not change; only the constant does.

Anchor food (FDC ID) Class factor FDC protein Implied N (computed) Protein at 6.25 (computed) Over/understatement
Nuts, almonds (170567) 5.18 21.15 4.083 25.52 +4.37
Peanuts, all types, raw (172430) 5.46 25.80 4.725 29.53 +3.73
Soybeans, mature seeds, raw (174270) 5.71 36.49 6.391 39.94 +3.45
Gelatins, dry powder, unsweetened (169599) 5.55 85.60 15.423 96.40 +10.80
Wheat flour, white, all-purpose (168894) 5.70 10.33 1.812 11.33 +1.00
Rice, white, long-grain, raw (168877) 5.95 7.13 1.198 7.49 +0.36
Egg, whole, raw (171287) 6.25 12.56 2.010 12.56 0.00
Milk, whole (171265) 6.38 3.15 0.494 3.09 −0.06

Worked example (Table 2, soybeans, one full chain). FDC 174270 protein = 36.49 g/100 g. Implied nitrogen = 36.49 ÷ 5.71 = 6.391 g N per 100 g. Re-declared at 6.25: 6.391 × 6.25 = 39.94 g/100 g. Overstatement = 39.94 − 36.49 = 3.45 g per 100 g, matching the +9.5% from Table 1 (3.45 ÷ 36.49 × 100 = 9.46 → 9.5%). A 30 g protein claim built on soy at 6.25 carries about 2.8 g of constant, not protein (30 × 0.095 ≈ 2.85).

Perspectives

Label: where 6.25 overstates and where it understates. Overstatement concentrates in plant storage proteins and gelatin: nuts and seeds (+14.5 to +20.7%), soy (+9.5%), wheat (+7.2 to +9.6%), rice (+5.0%). Understatement is small and dairy-specific (−2.0% milk, −1.0% wheat bran). US regulation (21 CFR 101.9(c)(7)) permits protein by nitrogen × 6.25, so a 6.25 label can be legal and still be 20% high on an almond product. "Legal" and "accurate" are separate properties of the same number, and retailers' verification programs increasingly test the second one.

Analytical: the amino acid sum is the factor-free reference. Summing the amino acids from a total amino acid analysis never passes through a nitrogen constant at all. Our total profile (AA-T) reports 16 amino acids: it does not report tryptophan, and cysteine needs the AA-3 workup, so the sum is a close cross-check, not a complete one. Where the sum and the nitrogen declaration disagree by the amounts in Table 2, the factor is the first suspect, before spiking (the subject of our companion nitrogen-math analysis) and before lab error. That ordering matters: honest factor error is common; fraud is the later hypothesis.

Historical: why a 19th-century constant still runs modern labels. 6.25 is 100 ÷ 16: an average nitrogen content for the mixed proteins early analysts had at hand. Jones showed in 1941 that specific proteins deviate systematically (amandin at 19.3% nitrogen is the extreme in this table), and FAO tables have carried his specific factors ever since. The default survived because it is simple, comparable across foods, and written into regulation, not because it is right per ingredient. Mariotti and colleagues (2008) argue the Jones factors themselves are imperfect (they were derived from purified protein fractions and ignore non-protein nitrogen) and propose lower class averages; we use Jones here because it is the published, citable standard a label can actually be checked against, and we flag the critique rather than hiding it.

Feed and food crossover: the same error, two industries. Feed guarantees and food labels both ride on crude protein (nitrogen × factor). A +14.5% factor error on peanut meal or a +9.5% error on soy moves formulation least-cost solutions, amino acid supplementation decisions, and price-per-unit-protein comparisons in both directions across the two industries. The amino acid profile, not crude protein, is what animals and label claims actually use; pricing protein by a mis-factored nitrogen number pays for constant, not nutrition.

Cross-check your declaration against the amino acid sum

If your product's protein declaration was built at 6.25 on a nut, seed, soy, wheat or gelatin base, a total amino acid analysis gives you a factor-free number to compare it against, before a retailer or a plaintiff's lab does.

What we can and can't measure for this question

AA-T is $230 per sample, hydrolysis included, and reports 16 amino acids in g/100 g and mol%. Tryptophan is destroyed in acid hydrolysis and cysteine is not reported (AA-3 adds it on request), so the sum falls short of total protein by those amounts. Asparagine and glutamine are reported with aspartate and glutamate. Solids and powders that need extraction add PREP-1 at +$30. We do not offer protein by nitrogen.

Order total amino acid analysis See an example report

Overnight turnaround available, with expedited service on request. Standard results in 3 to 5 business days from receipt.

Methods & Sources

  • Jones factors (accessed November 5, 2026): Jones, D.B. (1941), "Factors for converting percentages of nitrogen in foods and feeds into percentages of protein," USDA Circular 183, as tabulated in FAO, Food Composition Table for Use in East Asia (1972), Part I (milk 6.38, wheat flour refined 5.70, wheat whole-kernel 5.83, almonds 5.18, soybeans 5.71) and in Mariotti, Tomé & Mirand (2008), Table 3 (adding gelatin 5.55, barley/oats/rye 5.83, wheat bran 6.31, rice 5.95, corn 6.25, peanut 5.46, Brazil nut 5.46, other oil seeds and nuts 5.30, other animal products 6.25).
  • Food profiles (accessed November 5, 2026): USDA FoodData Central, SR Legacy, protein (g/100 g) retrieved per FDC ID from the FDC API: almonds 170567 (21.15), peanuts 172430 (25.80), soybeans 174270 (36.49), gelatin 169599 (85.60), wheat flour 168894 (10.33), rice 168877 (7.13), egg 171287 (12.56), whole milk 171265 (3.15).
  • Formulas: implied N % = 100 ÷ factor; error % = 100 × (6.25 − factor) ÷ factor; implied N (g/100 g) = FDC protein ÷ factor; protein at 6.25 = implied N × 6.25.
  • Assumptions and limitations: Table 2 treats the FDC protein value as the Jones-basis value. FDC protein derivation varies by entry and era (many SR Legacy proteins were computed with specific factors; some were not), so Table 2 is a controlled illustration of factor arithmetic on real profiles, not a re-audit of FDC. The factor error percentages in Table 1 do not depend on that assumption. Jones factors themselves are contested (Mariotti et al. 2008). Non-protein nitrogen is excluded from both tables by construction; a food with substantial non-protein nitrogen carries an additional, separate overstatement that only direct amino acid measurement resolves.
Order builder

Have samples in mind?

Pick the tests, the sample type and any preparation. The total updates as you go, and you can book in the same place. Every report includes the chromatogram and QC behind the numbers.

For the short version, see how much does amino acid testing cost: the per-sample numbers, and why quotes vary.

  • Free amino acids$150
  • Total amino acids$230
  • Creatine$120