P ProteinBenchmark

Complete Protein Foods: What Qualifies, and Why It Matters Less Than You Think

Most guides to complete protein answer a yes/no question and then tell you to pair rice with beans. The yes/no answer is easy and the pairing advice is obsolete. The number that actually separates protein sources — and that almost nobody publishes — is DIAAS.

What "complete" actually means

Your body can build eleven of the twenty amino acids it needs. The other nine — histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan and valine — are essential: they have to come from food. A protein is called complete when it supplies all nine in proportions adequate to meet human requirements, judged against a reference amino acid pattern.

That definition does two things badly. It is a binary applied to something continuous — a protein one percentage point short of the lysine reference is filed alongside one missing an amino acid entirely. And it says nothing about digestibility, which is at least as important: an amino acid you cannot absorb is not available regardless of what the label says.

DIAAS — the Digestible Indispensable Amino Acid Score — fixes both. It scores the limiting essential amino acid after accounting for how much of it is actually digested and absorbed, and it reports a number rather than a verdict (Mathai et al. 2017; Nutrients 2020 12(9):2670). The FAO recommended it in 2013 as the replacement for the older PDCAAS. It is the score we use across this site, and it is the column the complete-protein listicles leave out.

Every protein source, scored

Completeness verdict, DIAAS, leucine content, and what a 25 g serving of protein from each source is actually worth after the quality adjustment. Every figure is computed at build time from the same modules that drive our calculators — nothing here is typed by hand.

Source Complete? DIAAS Leucine Usable from 25 g
Whey Yes 1.15 10.5% 25.0 g
Dairy / casein Yes 1.18 9.5% 25.0 g
Egg Yes 1.13 8.5% 25.0 g
Meat, poultry, fish Yes 1.10 8.0% 25.0 g
Soy Yes 0.90 8.0% 22.5 g
Pea Yes 0.82 8.2% 20.5 g
Blended plant Yes 0.80 7.8% 20.0 g
Grain / wheat No 0.59 7.0% 14.8 g
Collagen No 0.00 2.5% 0.0 g

"Usable from 25 g" is our Bioavailability-Adjusted Yield: label protein multiplied by DIAAS, clamped at the label figure. A DIAAS greater than one means the limiting amino acid is in surplus relative to the reference pattern — it signals quality, not extra grams, so whey and dairy do not get credited above 25.0 g. Leucine is a percentage of protein by weight. Full formulas on the methodology page.

The protein-combining rule is obsolete

If you have read that plant proteins must be paired at the same meal — rice with beans, hummus with pita — to "make" a complete protein, that advice traces to Diet for a Small Planet (1971). Its author retracted it in a later edition, and the nutrition literature has been unambiguous since: the body maintains a free amino acid pool and draws on it continuously, so complementary proteins eaten hours apart still complement each other (Young & Pellett 1994). The Academy of Nutrition and Dietetics states the position directly in its vegetarian-diets paper — varied plant protein across the day is sufficient, and deliberate combining at each meal is not required (Melina et al. 2016).

This matters because the combining myth misdirects effort. It turns plant protein into a timing and pairing problem, which it is not, and distracts from the two things that genuinely differ.

What actually differs: quantity and leucine

The real plant-protein constraint is a quality discount you fix with volume. Of the four plant sources we model, 3 are complete (soy, pea, blended plant) and one is not (grain / wheat). But look at the DIAAS column rather than the completeness column: soy scores 0.90 and pea 0.82, against 1.15 for whey. On a 25 g serving that is 20.5 g of usable protein from pea versus 25.0 g from whey — a real gap, and one closed by eating more, not by pairing.

The second difference is leucine, the amino acid that actually triggers muscle protein synthesis (Norton & Layman 2006). Plant sources run lower — 8.2% for pea and 7.0% for grain against 10.5% for whey — so a plant-based serving has to be larger to clear the same per-meal leucine threshold. That is a dose question, and it is why our calculator applies a higher gram-per-kilogram target on its plant-based setting. The threshold itself is explained on our leucine page.

Put plainly: a plant-based eater does not have a completeness problem. They have a total-grams problem and a per-meal-leucine problem, and both are solved by eating more protein rather than by assembling it in a particular order.

The one completeness verdict that changes what you do

Across the nine sources on this page, only one incompleteness verdict has practical consequences — and it is not a plant. Collagen contains zero tryptophan and is low in several other essential amino acids, so it cannot be scored against the FAO reference pattern at all: DIAAS 0.00, and 0.0 g of usable protein from a 25 g serving. Leucine-matched randomised trials confirm the implication — hydrolysed collagen does not raise myofibrillar protein synthesis at rest or after exercise (Oikawa et al. 2020; Aussieker & van Loon 2023).

This is why collagen is the only source our model excludes from the muscle-protein target entirely. Its calories still count; its grams do not. Every other source on the table — including grain, which is limiting in lysine — contributes meaningfully and is counted, at its own DIAAS. Our collagen page covers what it is and is not good for.

Put it to work

Frequently asked questions

What is a complete protein?

A complete protein contains all nine essential amino acids — the ones your body cannot synthesise — in proportions adequate to meet human requirements. Animal proteins (whey, dairy, egg, meat, fish) are complete, and so are soy and pea. Grain protein is limiting in lysine, so it is classed as incomplete, though it still contributes usefully to your daily total. Collagen is the genuine outlier: it contains no tryptophan at all, so it cannot be scored as a complete protein under the FAO reference pattern, and a 25 g serving yields 0.0 g of protein usable for muscle.

Do I need to combine plant proteins at the same meal?

No. The "protein combining" rule — eat rice with beans at the same sitting or the protein does not count — was popularised in the 1970s and has been withdrawn by the researchers and organisations that spread it. The body maintains a free amino acid pool and draws on it across the day, so complementary proteins eaten at different meals still combine (Young & Pellett 1994; Academy of Nutrition and Dietetics position paper, Melina et al. 2016). Eating varied plant protein across a day is sufficient.

Is pea protein a complete protein?

Yes, effectively. Pea protein is low in methionine relative to the reference pattern but still adequate, so it counts as complete and contributes to a muscle-protein target. Its DIAAS is 0.82, against 1.15 for whey — so a 25 g serving of pea protein delivers about 20.5 g of effective protein where the same grams of whey deliver 25.0 g. Completeness is not the useful question; the quality discount is.

Is soy protein complete?

Yes. Soy is the highest-quality common plant protein and is unambiguously complete, with a DIAAS of 0.90 — the closest any widely eaten plant source gets to animal protein. Its leucine content (8.0% by weight) is also close to that of meat, which matters because leucine, not completeness, is what triggers muscle protein synthesis.

If completeness barely matters, what should I look at instead?

Two things: total protein quantity, and leucine per meal. Plant proteins carry a real quality discount — DIAAS 0.59 for grain and 0.82 for pea versus 1.15 for whey — but the fix is eating more of it, not pairing it correctly at a single meal. Randomised trials find plant and animal protein produce comparable muscle gains when total protein is matched at a sufficiently high intake.

Is collagen a complete protein?

No, and this is the one completeness verdict that changes what you should do. Collagen contains zero tryptophan and is low in several other essential amino acids, giving it a DIAAS of 0.00. Leucine-matched randomised trials show hydrolysed collagen does not raise muscle protein synthesis (Oikawa et al. 2020; Aussieker & van Loon 2023). Count its calories; do not count its grams toward a muscle-protein target.

Sources

  • Mathai et al. 2017 — DIAAS values for animal and plant proteins; primary source for the per-source DIAAS table used on this page.
  • Nutrients 2020 12(9):2670 — review of animal, plant and collagen amino acid profiles and their DIAAS scores.
  • Phillips 2017 — protein quality in the context of resistance training and muscle mass.
  • Norton & Layman 2006 — leucine as the trigger for muscle protein synthesis.
  • Oikawa et al. 2020 and Aussieker & van Loon 2023 — leucine-matched trials showing collagen does not raise muscle protein synthesis.
  • Young VR & Pellett PL 1994, American Journal of Clinical Nutrition — plant proteins in human amino acid nutrition; the standard reference for the obsolescence of meal-by-meal protein combining. (Cited from the literature; we have not independently verified the identifier, so it is listed without a link.)
  • Melina V, Craig W & Levin S 2016, Journal of the Academy of Nutrition and Dietetics — Academy position paper on vegetarian diets, stating that varied plant protein across the day meets requirements without deliberate combining. (Cited from the literature; identifier not independently verified, so listed without a link.)

Full formulas, the complete DIAAS table, and the whole source list are on the methodology page.

Disclaimer

ProteinBenchmark is educational. Nothing here is medical advice. If you're on medication (including GLP-1 agonists), pregnant, breastfeeding, have kidney disease, or any chronic condition, talk to a registered dietitian or your doctor before changing your protein intake.

Powered by ProteinBenchmark — free protein calculators & nutrition database