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

Protein Quality Scores, Computed: DIAAS-Style Ratings for Commercial Ingredients

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

Updated November 19, 2026

ESTIMATES, NOT MEASUREMENTS, read this first. Every score on this page labeled "DIAAS-style estimate" or "chemical score" is an estimate assembled from published components: amino acid profiles from USDA FoodData Central, reference patterns from FAO (2013), and digestibility assumptions stated in the tables. True DIAAS is a measured value from ileal digestibility studies in humans or pigs. Nobody at OpenChemWorks ran those studies. Where a published measured DIAAS exists (Mathai, Liu & Stein, 2017), we print it separately and label it "measured (published)" so the two kinds of number can never be confused.

At a glance: what the numbers say

  • Gelatin (collagen) scores 0.00, at every age pattern and every digestibility assumption. Its tryptophan content in FoodData Central is 0 mg per gram of protein, and a score built on the lowest ratio is zero when any indispensable amino acid is zero.
  • Among the eight ingredients we computed, only wheat flour (apart from gelatin at 0.00, above) falls below the FAO "good quality" line on chemistry alone. Wheat's chemical score (older child/adult pattern) is 0.46, limited by lysine at 22.1 mg/g protein against a 48 mg/g pattern. Split peas score 0.88 (limited by sulfur amino acids). Everything dairy, egg, and soy scores above 1.00 before any digestibility correction.
  • Digestibility is where plant scores are decided. Soy isolate's chemical score is 1.07 (sulfur amino acids limiting). Apply a 0.85 digestibility factor to the limiting amino acids and the DIAAS-style estimate is 0.91. Apply 0.75 and it is 0.80. Same chemistry, three different claims, decided entirely by an assumption, which is why this page prints the assumption next to every estimate.
  • A 50:50 wheat + pea protein blend (by protein) scores 1.03, higher than either ingredient alone (wheat 0.46, pea 0.88). Wheat brings the sulfur amino acids peas lack (38.9 vs 20.2 mg/g); peas bring the lysine wheat lacks (76.6 vs 22.1 mg/g). The minimum pea share that lifts the blend's lysine to the adult pattern is 47.5% of protein. Blend math, not marketing, is how plant proteins become complete.
  • Published measured DIAAS and our computed chemical scores rank ingredients the same way. Mathai et al. (2017), measuring ileal digestibility in pigs, found dairy proteins at or above 100 on the young-child pattern (milk protein concentrate 120, whey concentrate 107, whey isolate 100), soy isolate at 84, and pea concentrate at 62. Our undigested chemical scores are higher, as they must be, because digestion only removes amino acids. The gap between the two columns is the digestibility cost, made visible.
  • So what: If you are formulating or substantiating a protein-quality claim, the claim rests on three separable numbers: the amino acid profile (measurable today, on your actual ingredient, for $230), the reference pattern (published, free), and digestibility (borrowed from literature, and the weakest link unless your ingredient is the one that was measured). Know which of the three your claim is actually standing on.
  • Why it matters: A total amino acid analysis (AA-T) supplies most of the input measurement every score needs, and the amino acid profile is the only one of the three inputs you can own outright for your specific lot. Tryptophan and the sulfur amino acids need extra work. Literature profiles describe a generic ingredient. Your claim describes your product.

Introduction

Protein quality claims are usually quoted as a single score, as if the score were a property of the ingredient. This article takes the score apart. A quality score stands on three separable inputs: an amino acid profile, a reference pattern that says how much of each indispensable amino acid a gram of protein should carry, and a digestibility figure that says how much of what is present is actually absorbed. The question the article answers is which of those three inputs each published score is really standing on, and how far the score moves when an assumption changes.

That matters to anyone formulating a product or substantiating a claim, because only one of the three inputs can be measured on your own ingredient: its amino acid profile. The reference pattern is published and free. Digestibility is almost always borrowed from literature on a related material, and for mid-scoring plant proteins that borrowed assumption, not the chemistry, decides whether the score clears 1.00.

The profiles are computed from USDA FoodData Central entries, the patterns are those published by FAO, and the measured comparison values are the published DIAAS results of Mathai and colleagues, kept in a separate column throughout so that estimates and measurements are never confused. Read the chemical scores first: they are the ceiling every estimate starts from. Then read the scenario estimates beside the measured values, with the digestibility assumption printed next to each one, and finish with the worked blend, which shows how two incomplete proteins can complete each other.

Chemical scores (undigested), older child / adult pattern

The ceiling every DIAAS-style estimate starts from: lowest amino-acid ratio against the FAO reference pattern, before any digestibility correction.

Milk, dry, nonfat1.49
Egg, whole, dried1.42
Egg, white, dried1.38
Whey, sweet, dried1.14
Soy protein isolate1.07
Peas, split, raw0.88
Wheat flour, white0.46
Gelatins, dry powder0.00
00.751.50

1.00 = meets the FAO reference pattern

Figure 1: Chemical scores from Table 2. Dairy, egg, and soy sit above the reference line on chemistry alone; wheat and gelatin fall below it.

Reference patterns used (FAO 2013)

Digestible indispensable amino acid scoring patterns, mg per gram of protein (FAO, Dietary protein quality evaluation in human nutrition, Report of an FAO Expert Consultation, 2013):

Amino acid Infant, birth–6 mo Child, 6 mo–3 y Older child, adolescent, adult (>3 y)
Histidine 21 20 16
Isoleucine 55 32 30
Leucine 96 66 61
Lysine 69 57 48
Sulfur amino acids (Met + Cys) 33 27 23
Aromatic amino acids (Phe + Tyr) 94 52 41
Threonine 44 31 25
Tryptophan 17 8.5 6.6
Valine 55 43 40

Table 1: Amino acid content, mg per gram of protein (computed from FoodData Central)

Each value = (amino acid g per 100 g food / protein g per 100 g food) × 1000. SAA = methionine + cystine (FoodData Central's term); AAA = phenylalanine + tyrosine.

Ingredient (FDC ID) His Ile Leu Lys SAA AAA Thr Trp Val EAA total
Milk, dry, nonfat (172195) 27.1 60.5 98.0 79.3 34.3 96.6 45.1 14.1 66.9 464
Egg, white, dried (172204) 22.6 56.5 84.3 68.0 60.3 97.3 45.4 12.3 63.7 446
Egg, whole, dried (172188) 25.0 50.7 86.4 69.5 53.4 93.9 44.3 16.1 62.2 438
Whey, sweet, dried (171283) 18.3 55.6 91.7 79.7 38.2 59.6 63.2 15.9 53.9 428
Soy protein isolate (174276) 26.1 48.2 76.8 60.3 24.6 88.5 35.5 12.6 46.4 371
Peas, split, raw (172428) 25.3 42.5 72.7 76.6 20.2 72.2 35.2 6.9 44.8 362
Wheat flour, white (169761) 22.3 34.6 68.7 22.1 38.9 80.5 27.2 12.3 40.2 295
Gelatins, dry powder (169599) 7.7 13.5 28.7 40.4 7.1 23.8 17.2 0.0 24.3 159

Worked example. Soy isolate lysine: FoodData Central lists lysine 5.327 g and protein 88.32 g per 100 g. (5.327 / 88.32) × 1000 = 60.3 mg lysine per gram of protein. Soy isolate SAA: (1.130 + 1.046) / 88.32 × 1000 = 24.6 mg/g. EAA total counts the nine indispensable amino acids individually (methionine and phenylalanine counted once each, without their sparing partners): for soy isolate, 2.303 + 4.253 + 6.783 + 5.327 + 1.130 + 4.593 + 3.137 + 1.116 + 4.098 = 32.74 g per 100 g food; 32.74 / 88.32 × 1000 = 371 mg/g protein. Note: dried sweet whey (FDC 171283) is a lactose-rich dairy ingredient (12.93% protein), not a whey protein concentrate; its per-gram-of-protein profile is what is scored here, and it should not be read as the profile of a commercial WPC/WPI.

Table 2: Chemical scores (undigested), the ceiling every DIAAS-style estimate starts from

Chemical score = lowest (content / pattern) ratio across the nine scoring groups. No digestibility applied. This is an estimate component, not a DIAAS.

Ingredient Score, >3 y pattern Limiting group (>3 y) Score, 6 mo–3 y pattern Limiting group (child)
Milk, dry, nonfat 1.49 SAA (34.3/23) 1.27 SAA (34.3/27)
Egg, whole, dried 1.42 Leucine (86.4/61) 1.22 Lysine (69.5/57)
Egg, white, dried 1.38 Leucine (84.3/61) 1.13 Histidine (22.6/20)
Whey, sweet, dried 1.14 Histidine (18.3/16) 0.92 Histidine (18.3/20)
Soy protein isolate 1.07 SAA (24.6/23) 0.91 SAA (24.6/27)
Peas, split, raw 0.88 SAA (20.2/23) 0.75 SAA (20.2/27)
Wheat flour, white 0.46 Lysine (22.1/48) 0.39 Lysine (22.1/57)
Gelatins, dry powder 0.00 Tryptophan (0.0/6.6) 0.00 Tryptophan (0.0/8.5)

Worked example. Soy isolate, >3 y: ratios are His 26.1/16 = 1.63, Ile 48.2/30 = 1.61, Leu 76.8/61 = 1.26, Lys 60.3/48 = 1.26, SAA 24.6/23 = 1.07, AAA 88.5/41 = 2.16, Thr 35.5/25 = 1.42, Trp 12.6/6.6 = 1.91, Val 46.4/40 = 1.16. The lowest is SAA at 1.07, so the chemical score is 1.07, sulfur-amino-acid limited. The same ingredient on the child pattern scores 0.91 because the child pattern asks for more SAA per gram (27 vs 23). Pattern choice alone moves this ingredient across the 1.00 line.

Published measured DIAAS (Mathai et al. 2017)

Measured ileal digestibility scores on the young-child pattern, printed separately from the computed chemical scores so the two kinds of number are never confused.

Milk protein concentrate120
Whey protein concentrate107
Whey protein isolate100
Soy protein isolate84
Pea protein concentrate62
060120
Figure 2: Measured DIAAS ranks the ingredients the same way the computed chemical scores do; the gap between the two charts is the digestibility cost.

Table 3: DIAAS-style ESTIMATES under stated digestibility assumptions, with published measured DIAAS for comparison

Estimate formula: chemical score (limiting group, >3 y pattern) × assumed standardized ileal digestibility of the limiting amino acids. The 0.85 and 0.75 columns are scenarios, printed to show sensitivity. They are not measurements and must not be quoted as DIAAS.

Ingredient Chemical score (>3 y) Estimate at SID 0.85 Estimate at SID 0.75 Measured DIAAS (published, Mathai et al. 2017, young-child pattern unless noted)
Milk proteins (concentrate/powders) 1.49 (nonfat dry milk, this page) 1.27 1.12 120 (milk protein concentrate)
Whey proteins 1.14 (dried sweet whey, this page) 0.97 0.86 107 (concentrate), 100 (isolate)
Soy protein isolate 1.07 0.91 0.80 84
Pea protein concentrate 0.88 (whole split peas, this page) 0.75 0.66 62
Wheat 0.46 (white flour, this page) 0.39 0.35 Below 75; no quality claim (Stein laboratory summaries)
Gelatin / collagen 0.00 0.00 0.00 0 by construction (no tryptophan)

Worked example. Soy isolate estimate at SID 0.85: 1.07 × 0.85 = 0.91. At SID 0.75: 1.07 × 0.75 = 0.80. Sensitivity: a 0.10 change in the digestibility assumption moves the estimate by 0.11 (1.07 × 0.10), which is larger than the entire gap between "good" (75–99) band boundaries for several ingredients in this table. That is the finding: for mid-scoring plant proteins, the digestibility assumption is the score.

Caveats printed with the table, not under it: (1) our ingredient proxies differ from Mathai's test materials (whole split peas are not pea protein concentrate; dried sweet whey is not whey isolate), so the estimate and measured columns are expected to differ in both directions; (2) Mathai's values use the young-child pattern, ours use the >3 y pattern, which is less demanding; (3) FAO bands: DIAAS ≥ 100 "excellent/high," 75–99 "good," below 75 no protein-quality claim. PDCAAS, the score FDA uses for %DV, truncates at 1.00 and uses fecal rather than ileal digestibility and the older preschool pattern, so a PDCAAS of 1.00 and a DIAAS-style estimate of 0.91 for the same ingredient are not a contradiction. They are different instruments.

Worked blend: wheat + pea (computed)

Blend contents are protein-weighted means of Table 1. At 50% wheat protein + 50% pea protein: lysine = (22.1 + 76.6)/2 = 49.4 mg/g (score 49.4/48 = 1.03); SAA = (38.9 + 20.2)/2 = 29.6 mg/g (score 1.29); valine = (40.2 + 44.8)/2 = 42.5 (score 1.06); all other groups score higher. Blend chemical score = 1.03, lysine-limited, against 0.46 and 0.88 for the components. Minimum pea share to reach lysine 48 mg/g: solve 22.1 + x(76.6 − 22.1) = 48, so x = 25.9/54.5 = 0.475. At 47.5% pea protein the blend's next-lowest group is valine at 1.06, so lysine remains limiting exactly at the pattern. One blend per cycle will be added to this section.

Essential amino acid density by ingredient

Which protein ingredient delivers the most essential amino acids depends on the unit you rank by. Ranked per gram of protein, whey protein isolate leads (516.3 mg/g) and the full spread across the six ingredients is only 19% (516.3 / 434.1 = 1.19). Ranked per 100 kcal the order changes: fat carries calories that carry no amino acids, so whole dry milk, third per gram of protein, falls to last. Pick the axis your label or your customer's panel runs on. Density is not quality: this table does not adjust for digestibility or limiting amino acids, so quote it only with that caveat. The whey profiles are the USDEC reference profiles for isolate and concentrate grades (the leucine density of the isolate, 143.3 mg/g protein, is computed from that source and is not comparable with the commercial-whey average used in the nitrogen article); the egg and milk profiles are USDA FoodData Central. For plant proteins, Gorissen et al. (2018) measured essential amino acids at about 30% of protein for pea and 28% for brown rice, against 43% for commercial whey in the same study.

Essential amino acids per gram of protein

Milligrams of total essential amino acids per gram of protein. Orange is the leader on this axis.

Whey protein isolate (WPI)516.3 mg/g
Egg white powder468.8 mg/g
Whole dry milk464.9 mg/g
WPC80462.7 mg/g
WPC34450.8 mg/g
Whole dried egg434.1 mg/g
0516.3 mg/g protein
Figure 1: Per gram of protein, whey protein isolate leads, and the full spread across the six ingredients is narrow. The same profiles reorder on the two axes that follow.

Table 4. Essential amino acid density per gram of protein and per 100 kcal

Profiles: whey grades from the U.S. Dairy Export Council Reference Manual for U.S. Whey and Lactose Products (2005), p.51–52, which reproduces USDA nutrient database values; egg and milk profiles from USDA FoodData Central (SR Legacy). EAA total = His + Ile + Leu + Lys + Met + Phe + Thr + Trp + Val.

Ingredient Protein (g/100 g) kcal/100 g EAA (g/100 g) EAA (mg/g protein) EAA (mg/100 kcal) Leucine (mg/g protein)
Whey protein isolate (WPI) 90.75 371 46.85 516.3 12,628 143.3
Egg white, dried, powder 82.4 376 38.63 468.8 10,274 87.0
WPC80 80.00 412 37.02 462.7 8,985 101.0
Milk, dry, whole 26.32 496 12.24 464.9 2,467 97.9
WPC34 34.36 369 15.49 450.8 4,198 102.4
Egg, whole, dried 48.1 576 20.88 434.1 3,625 84.1

Worked example, WPI per gram of protein: EAA sum = 1.35 + 5.90 + 13.00 + 9.15 + 2.05 + 2.30 + 6.25 + 1.50 + 5.35 = 46.85 g per 100 g. 46.85 g ÷ 90.75 g protein = 0.5163 g/g = 516.3 mg/g protein. Per 100 kcal: 46,850 mg ÷ 371 kcal × 100 = 12,628 mg/100 kcal.

Read the milk row carefully: whole dry milk ranks third per gram of protein (464.9 mg/g) and last per calorie (2,467 mg/100 kcal). Both numbers are correct. They answer different questions.

Essential amino acids per 100 kcal

Fat carries calories that carry no amino acids, so fat-bearing ingredients fall on this axis.

Whey protein isolate (WPI)12,628 mg
Egg white powder10,274 mg
WPC808,985 mg
WPC344,198 mg
Whole dried egg3,625 mg
Whole dry milk2,467 mg
012,628 mg/100 kcal
Figure 2: Per calorie, whole dry milk drops from near the top of the per-gram ranking to last, and whole dried egg falls with it. Egg white powder holds second place.

Essential amino acids per 100 kcal

Fat carries calories that carry no amino acids, so fat-bearing ingredients fall on this axis.

Whey protein isolate (WPI)12,628 mg
Egg white powder10,274 mg
WPC808,985 mg
WPC344,198 mg
Whole dried egg3,625 mg
Whole dry milk2,467 mg
012,628 mg/100 kcal
Figure 2: Per calorie, whole dry milk drops from near the top of the per-gram ranking to last, and whole dried egg falls with it. Egg white powder holds second place.

Perspectives

Formulator: pair by limiting amino acid, not by marketing category

Table 1 reads as a pairing chart. Legumes are lysine-rich and SAA-poor; cereals are the reverse. The wheat + pea computation above is the template: identify each ingredient's limiting group, check the partner's surplus in exactly that group, solve for the share. Soy isolate needs no partner on chemistry alone at the adult pattern (1.07); its problem is digestibility evidence, not a partner.

Label: PDCAAS vs DIAAS-style, where they disagree

FDA protein %DV still runs on PDCAAS, truncated at 1.00. Truncation hides real differences at the top (milk at 1.49 chemical score and egg white at 1.38 both print as 1.00) and the fecal digestibility basis flatters some plant proteins relative to ileal scoring. If your label claim and your marketing claim use different scores, this table is where the discrepancy will be found, by you or by someone else.

Ingredient buyer: score per dollar, data gap, reported honestly

A score-per-dollar ranking requires reliable published prices for each ingredient grade. We verified public price reporting for dairy ingredients (USDA Dairy Market News covers whey grades) but did not verify equivalent public reporting for pea and soy ingredient grades on the access date. The per-dollar table is therefore not published in this version rather than built on spot marketplace listings. Gap recorded; it closes when a citable price series for plant ingredients is verified.

Skeptic: how much of this ranking survives a different assumption?

Change the pattern (adult to child) and whey drops from 1.14 to 0.92, soy from 1.07 to 0.91 (Table 2). Change the digestibility assumption by 0.10 and mid-table estimates move by roughly 0.10 (Table 3). Change nothing and gelatin is still 0.00 and wheat is still last. The top and bottom of the ranking are robust; the middle is assumption-driven, and this page's estimates should be cited with their assumptions attached or not cited at all.

Get the input measurement for your ingredient

Every score above starts from an amino acid profile. Literature profiles describe generic ingredients; a claim needs your lot's profile.

What we can and can't measure for this question

AA-T ($230 per sample, hydrolysis included) measures 16 amino acids on your lot and gives lysine, threonine, leucine, isoleucine, valine, histidine and phenylalanine. A DIAAS-style score also turns on tryptophan and the sulfur amino acids, and AA-T does not cover them: tryptophan is destroyed in acid hydrolysis and is not offered here, cysteine needs the AA-3 workup (on request), and methionine can partly oxidize. Digestibility is a separate input that no amino acid analysis supplies. Solids that need extraction add PREP-1 at +$30.

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

  • Whey profiles: USDEC Reference Manual for U.S. Whey and Lactose Products (2005), p.51 to 52. Egg white powder, whole dried egg and whole dry milk: USDA FoodData Central, SR Legacy (whole dry milk converted from per-cup values by dividing by 1.28). EAA = His + Ile + Leu + Lys + Met + Phe + Thr + Trp + Val; cysteine and tyrosine are excluded.
  • Gorissen, S.H.M. et al. (2018), Amino Acids 50:1685 to 1695 (PMC6245118).
  • FAO (2013), Dietary protein quality evaluation in human nutrition: Report of an FAO Expert Consultation, FAO Food and Nutrition Paper 92: scoring patterns for the three age groups (His 21/20/16, Ile 55/32/30, Leu 96/66/61, Lys 69/57/48, SAA 33/27/23, AAA 94/52/41, Thr 44/31/25, Trp 17/8.5/6.6, Val 55/43/40 mg/g protein), DIAAS definition, and quality bands. Patterns cross-verified against the IDF factsheet summary of FAO 2013 and the Stein laboratory's published pattern footnotes. Accessed November 5, 2026.
  • USDA FoodData Central, SR Legacy, via the FDC API: FDC 172195, 172204, 172188, 171283, 174276, 172428, 169761, 169599 (protein and amino acids per 100 g). https://fdc.nal.usda.gov/, accessed November 5, 2026.
  • Mathai, J.K., Liu, Y. & Stein, H.H. (2017), British Journal of Nutrition: measured DIAAS in growing pigs for dairy and plant proteins (young-child pattern: milk protein concentrate 120, whey protein concentrate 107, whey protein isolate 100, soy flour 89, soy protein isolate 84, pea protein concentrate 62; dairy proteins first-limited by histidine or sulfur amino acids, soy and pea by sulfur amino acids), via the University of Illinois Stein laboratory publications and published summaries of the paper. Accessed November 5, 2026.
  • Formulas: content (mg/g protein) = (g per 100 g food / protein g per 100 g food) × 1000; ratio = content / pattern; chemical score = minimum ratio; DIAAS-style estimate = chemical score × assumed SID of limiting group; blend content = protein-weighted mean.
  • Limitations: chemical scores ignore digestibility entirely and are ceilings, not quality claims. Scenario estimates apply one uniform SID to the limiting group only; true DIAAS applies a measured SID per amino acid. Ingredient proxies (whole foods for commercial isolates/concentrates) are named in every table. FoodData Central entries are SR Legacy values for generic foods, not specifications for any commercial ingredient.
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