Complete Proteins

Dear Jack,

I recently got into a debate with a friend who insisted that quinoa, hemp, amaranth, buckwheat, and spirulina are all complete proteins. When I told him soy is the only complete vegetarian protein, he told me he avoids all soy due to its negative effect on men’s health. I wondered where this guy learned this and a quick google search on vegetarian complete proteins got me here. I checked out your page on protein. That chart listing the essential amino acids in so many plant foods made me realize that in fact many plant foods do contain all the essential amino acids but at such low levels it’s hard to list them as ‘complete proteins’ in the true meaning intended. So my take is that the only vegetarian protein that comes close to really being a complete protein (meaning it has essential amino acid levels similar to those of animal proteins) is soy. Would you agree?


[As an aside, the research to date indicates that moderate amounts of soy do not harm men’s health.]

Because this question about complete proteins comes up on a regular basis, I decided to do some number crunching to get to the bottom of it. Mind you, there are many more numbers to crunch if someone wanted to take the time, but I think I’ve covered enough to draw some conclusions.

I see two ways of looking at the question of whether a food has complete protein:

– How much of the food is needed to meet the RDA for all the essential amino acids (EAA)?
– How much variation is there in a food’s ability to meet the RDA for each individual EAA?

If you look at Table 3 of Where Do You Get Your Protein?, you will see what it takes to meet the RDA for each EAA for a variety of foods.

I have taken some of the data from Table 3 and constructed the table below to look more closely at what foods might contain a “complete protein”. In the table below, the numbers in the right-hand column represent the following equation:

A – B = R


A = servings required to meet the RDA for least plentiful EAA
B = servings required to meet the RDA for the most plentiful EAA
R = range
Food Range of Servings to
meet RDA for all EAA
Tuna 1.7 – 1.3 =   .4
Chicken leg 2.9 – 1.9 = 1.0
Ground beef 3.2 – 1.4 = 1.8
Edamame 3.3 – 1.8 = 1.5
Lentils 3.4 – 1.7 = 1.7
Pinto beans – refried 3.7 – 2.0 = 1.7
Tofu 3.8 – 1.1 = 2.7
Milk 5.6 – 2.9 = 2.7
Quinoa 6.1 – 3.6 = 2.5
Soy milk 6.2 – 3.3 = 2.9
Egg 6.6 – 3.9 = 2.7
Almonds 9.6 – 4.7 = 5.2
Corn 11.5 – 5.0 = 6.5
Spirulina 12.9 – 5.4 = 7.5

From the chart above, it appears that tuna is the most complete protein with a range of only .4 servings and only 1.7 servings required to meet the RDA for all the EAA. Chicken, beef, edamame (whole, cooked soybeans), lentils, and pinto beans all do quite well. I think it’s fair to consider all of them a “complete protein”. Tofu, milk, quinoa, soy milk, and eggs do significantly better than most grains and nuts which have a much wider range.

As for spirulina…not so much.

One thing to note about this is that all the numbers depend on the serving size. I tried to pick what I thought were reasonable (or common) serving sizes for each food (you can see what they are in Table 3).

The USDA database has a lot of specific entries for some of the food categories above (like ground beef). I chose what looked like a common version of the food, but I did not average the data across more than one version.

As for the remaining, supposed complete proteins mentioned in the original question above, here is what I found:

The USDA lists 9 g of protein per cup of cooked amaranth. That’s a good amount when compared to other grains, but I’m not sure if it’s as easy to eat a cup of amaranth as a cup of, say, rice or corn. The USDA had no amino acid info.

There was no info in the USDA database. I found many sources saying that it is a complete protein but none that I know to be reliable.

Nutritional yeast
No info in the USDA database. According to this site, it has 9 g of protein per 3 tbsp serving. That’s a decent amount, but I’m not sure if we can trust that info.

In conclusion, edamame, lentils, and pinto beans fared pretty well with chicken and beef for being a complete protein. Tofu, soymilk, and quinoa were on par with eggs and milk.

31 Responses to “Complete Proteins”

  1. Redwood Says:

    With all due respect, I don’t think you engaged this question well.

    In my amateur opinion, it’s a big fat “Who cares?”

    Which vegetarians or vegans are falling over from protein deficiency? So long as they are eating regularly and reasonably, don’t have some chronic illness that interferes with food absorption and are not using a plant-based diet to cover up some eating disorder — things that effect people on any diet strategy – then all is well.

    Most plant foods have all the necessary amino acids, in degrees that vary but they are not deficient. People don’t just eat tomatoes three or more times a day or for a whole week straight, we eat meals. Between eating mixed foods during meals and eating varied meals throughout a day and still further through the week and with the body’s ability to store amino acids and combine them as needed, this ideological quest for the singular foods with complete proteins is and antiquated venture.

    Hit up, plug in some foods and you’ll see a little purple graph indicating amino acid score and even low scoring vegetable items have a decent amount of the graph filled in. Prepare reasonable varied meals that makes you feel satisfied and which amino acids come from where are taken care of. Plant-based eating meets and often exceeds the recommended daily allowances, and no, we don’t have to be particularly “well-planned” or “careful” as is often stated, certainly not with protein anyway.

    “How about Mexican? We’ll grab some burritos!”
    “Nah, I had a burrito for lunch yesterday. Let’s do the Indian buffet.”
    “I was going to make Indian tonight for dinner. How about the Korean joint?”
    “Yeah, that’s cool. I could go for some japchae and kimchi.”

    That’s normal. How much protein is in each grain or vegetable is irrelevant.

    If we’re talking about diets that fuel athletics that may have higher protein demands, it’s a slightly more interesting inquiry. There are vegan athletes out there in a variety of demanding activities along with forums and websites to support and inform the vegan sports community. It’s not like if an athlete isn’t vegan they don’t think about protein sources or use supplements of all sorts (or pharmaceuticals) because even eating whole food diet including animal products may not deliver protein as quickly and conveniently.

  2. Jack Norris RD Says:


    I agree that, in regards to protein, what matters for someone’s health is the overall diet, not just a few foods. But I, and many others, have answered this question as you suggest many times over the years and in this one case, I decided to look more closely at what foods actually are complete proteins.

    That site is pretty handy. I was unaware of it previously. It gives foods an amino acid score and they consider 100 or above to be a complete protein. Quinoa and refried beans both score 106, while tuna scores 148 (for the version I chose) so that’s pretty much in agreement with what I came up with. But they give spirulina a 103, which is not in agreement with my findings. I wonder if this is where some people are getting the idea that spirulina is a complete protein. When I get some time I will try to figure out why the discrepancy. They had no amino acid info on hemp or aramanth.

  3. jeff gibson Says:

    Good post presenting another way to look at the protein picture Jack. I do agree with Redwood though that in general, with the caveats he states, this is way overrated as a critical issue in vegan health. However, the general public’s perceptions have been shaped by decades of poor information on this subject, so like you said such questions will repeatedly come up and we must patiently answer them. It has always struck me as a poor choice of words to characterize proteins as “complete” or “incomplete”. My dictionary defines incomplete as “lacking a part”, and as applied to protein (in the vast majority of plant foods) this is not the case. With very few exceptions, every one of the essential amino acids is there, just in varying proportions. If we only ate one food item all day every day that proportion would be hugely important. Fortunately, such is not the case.
    Just as an aside: I’m very grateful for the work you do Jack and appreciate your insight and perceptions. Thanks!

  4. Lisa A. Says:

    Thank you for the information. So what is the official definition of a complete protein if there is one? I understood it as presence of all nine essential amino acids in similar proportion. Is that correct?

  5. Jack Norris RD Says:


    I don’t think there is an exact definition. It would have to be some sort of combination of the two questions I posed above, in my original post:

    – How much of the food is needed to meet the RDA for all the essential amino acids (EAA)?
    – How much variation is there in a food’s ability to meet the RDA for each individual EAA?

  6. beforewisdom Says:

    I found this post to be interesting and worthwhile.

    IMHO too many people take the idea that they only have to complement once a day ( the ADA paper ) too liberally and are very sloppy about their protein. Yet, as others have mentioned that is a non-worry compared to other things like EFAs, calcium and the various pollutants out there.

    I have new respect for the humble pinto bean. I had no idea it was so protein nutritious.

  7. Redwood Says:

    I originally posted first thing in the morning and probably should have held off until I settled into a cup of coffee; my intention wasn’t to shoot the nutritional messenger. It’s fair that your post is intended to be a specific analysis and I really do appreciate your input on the subject. I hope I didn’t come off as combative or flippant.

    I confess that I hold a certain annoyance regarding the association of protein quality and plant-based diets and my spidey-sense began tingling moment I read the title Complete Proteins. I can’t stand it when I hear people exclaim that they included “a protein” in their meal, really the don’t-I-sound-nutritionally-savvy code for animal stuffs.

    There isn’t any compelling research associating complete protein foods or diets as a factor concerning longevity. Tuna scores a 148 amino acid profile on NutritionData, but so what, there doesn’t seem to be a meaninful advantage for such a high score. Seems more like there is a trend of caution against too much dietary protein even in conservative nutritional guidelines and it’s easier to overdo it than otherwise, so long as we’re not talking about food scarcity environments.

    I further appreciate your interpretation on the NutritionData profiles. I gather that it’s looking at amino acid balance alone, not the amount of total protein per calories as your equations are working out. If you have any additional insight of information as to how accurate or meaningful their amino acid profiling is, that would be great, because when I see number like:

    —–Cucumber, with peel, raw

    —–Beef, ground, 75% lean meat / 25% fat, crumbles, cooked, pan-browned [hamburger]

    —–Milk, whole, 3.25% milkfat

    —–Spinach, raw

    —–Potatoes, baked, flesh and skin, with salt

    —–Rice, white, long-grain, regular, cooked

    —–Blackbeans, boiled

    Seems like the lowly cucumber fairs quite well against the mighty hamburger. I have no idea why cows’ milk is venerated, the only reason it gets a high score is that the naturally occurring fat is artificially separated out and higher scores for low fat and fat free is furthered by fat reduction concentrating the protein. Note how when this is done with soybeans people call it “unnatural” and” toxic.”

    Popeye has been right all along about spinach. Potatoes are complete which is interesting considering that most people regard them as “just carbs.” Black beans (and you mentioned refried beans) are complete all by themselves so no need to ever fuss over combining them with rice (though rice and beans are yummy so I’ll continue anyway), and rice, which is supposedly “all carbs,” makes a pretty decent showing.

    This data confirms my attitude that debates on complete proteins are far more academic than practical and I wish there was a way to eradicate the prevalent unwarranted association with plant-based diets.

    I acknowledge that such scoring isn’t the whole nutritional picture (calories, absorption, etc.), but if there is something way-off-goofy with that amino acid profile data, it would be good to know.

    Thanks for your continued work.

  8. Jack Norris RD Says:


    Now that you bring up these other foods, I’m starting to wonder how useful the NutritionData’s assessment of a complete protein actually is. If you look at my numbers for a baked potato (, the range to meet the RDA for all EAA is very wide.

    There is one benefit for eating complete proteins, particularly for people with kidney disease, and that is that complete proteins allow you to limit your total protein consumption and thus limit your nitrogenous waste (ammonia and urea). For most people, this probably isn’t very important.

  9. Lisa A. Says:


    Thank you for your response. I greatly appreciate your informative posts about different aspects of nutrition. And the topic of complete protein is definitely an interesting one.

    One more question. I noticed that different sources state that there is a different number of essential amino acids. Some say 8, some say 9. One person was trying to tell me that there are 10. So what is the correct number?

  10. Jack Norris RD Says:


    That National Academy of Sciences’ report, Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids, lists nine amino acids as essential, but puts an asterisk next to histidine that says:

    Although histidine is considered indispensable, unlike the other eight indispensable amino acids, it does not fulfill the criteria used in this report of reducing protein deposition and inducing negative nitrogen balance promptly upon removal from the diet.

    They also list six of the non-essential amino acids to be conditionally indispensible which they describe as “limited under special pathophysiological conditions.”

  11. Roger Says:

    Re:’s protein completeness score–

    I think the reason cucumber gets a very high protein completeness score might be that it contains very similar levels of each amino acid–but very low levels, as well. They also say a 301g serving of cucumber has only 2g of protein.

    On the other hand, the formula of this post would probably give a very low completeness score to something like cucumber, because with such low levels of each amino acid, you’re going to require a *whole lot* of servings to get your RDAs of each–and at such low levels, a very small variation in the amount of each AA can lead to a big change in how many servings are required to meet your RDA.

    Also, note that you can get amino acid numbers from the pages by clicking on “more details” on the “protein & amino acids” box.

    So here are some numbers. Using the 52g (1/2 cup slices) serving size for cucumber, with peel, raw, and using RDAs from the table linked to in the post for a 140lb vegan person, here’s what I get for # of servings needed to reach the RDA, with amounts per serving in parentheses:

    HIS: 188.5 (5.2mg)
    ISO: 122.0 (10.9mg)
    LEU: 194.7 (15.1mg)
    LYS: 176.2 (15.1mg)
    MET CYS: 255.8 (5.2mg)
    PHE TYR: 146.1 (15.6mg)
    THR: 141.4 (9.9mg)
    TRY: 134.6 (2.6mg)
    VAL: 147.4 (11.4mg)

    PRO: 186.7 (0.3g)

    So using the measure of this post, we get 255.8 – 122.0 = 133.8. Massive!

    But if you use a measure like, say, what percentage difference there is between the values per serving of the highest & lowest AA (even normalized for RDA), you’d get a small difference–because there isn’t much of any of them! I suspect that’s how are getting their completeness scores.

    So in short, to get value out of those completeness scores, you’ve also got to look at the total amount of protein per serving. Doesn’t matter if something’s a complete protein if there isn’t much of it anyway.

  12. 1VeganMom Says:

    Hi there folks…. here is JMHO:
    I do appreciate all the info posted. The numbers, for me are something worth investigating in relation to my current choices in my vegan diet. Though I haven’t had any iron or protein issues within my diet (my doctor is aware of my lifestyle and we check this about twice a year), I have seen – without reference back to the sites listed yet, that my choices may need to go up the ladder a few rungs. Then again, since I do not have nutritional issues, other than weight in my climbing age, I may not change much.
    I do think that when choosing a protein within ANY diet, one factor, which hasn’t been discussed here is that the accompanying foods consumed with that protein do have an affect of protein absorbed into one’s body. Like calcium is with V-D or K3. If other things are not chosen appropriately, the amount of protein contained in that source may not matter – as what one’s body is able to absorbed is not only dependent on other food items, but greatly varies from person to person.
    In summary, information is always better than none. But to optimize one’s diet for maximal health, this information needs to be properly divided and interpreted with a nutritionist/doctor and the person seeking to eat for health.

    As stated in the beginning of my post: this is Just My Humble Opinion. I respect all messengers and the info provided. I am in no way refuting any of it. I am simply implying that of all this information, it should be carefully applied to each person’s individual health and not be a “total blanket” for everyone.


  13. Jack Norris RD Says:


    I’m not aware of other foods eaten at the same time as higher protein foods inhibiting protein absorption. I would doubt any normal foods in typical amounts have much of an effect on protein absorption.

    Also, the vitamin D consumed with calcium has not direct affect on the calcium it is consumed with. It has be to be converted by the kidneys and the liver into calcitriol before it can increase calcium absorption.

  14. Michael Maltzer Says:

    I read that soy inherently in its natural form does not have a remarkable level of protein (and no, not from a blog, but from a website that detailed the processing of soy in America according the FDA standards). And, that the only way it becomes a complete protein is through infusion of amino acids during its processing.

    So, if this is the case, then *any* food item could be considered a complete protein. And, as has been shown in Jack’s research, quinoa is a more reliable source of or protein than soy. So, then why is soy touted as the best source of protein? (my speculation…special interest groups- such as the soy industry). Soy blocks iodine absorption, which is a compound that makes up 1/2 of the hormone created and used by the thyroid (TSH)- and the thyroid absorbs 95% of the iodine used by the body. If soy blocks iodine absorption, people with a hypothyroid issue (which is ever-increasing in this country) are not being served very well by ingesting soy.

    “Everything in moderation”, right? So then, why is soy components included in everything packaged? (soy protein, soybean oil, soy protein, etc etc). We, as a country, are getting too much when everything is infused with some component of soy.

  15. Jack Norris RD Says:


    > And, that the only way it becomes a complete protein is through infusion of amino acids during its processing.

    I seriously doubt that’s true.

    > And, as has been shown in Jack’s research, quinoa is a more reliable source of or protein than soy.

    My research showed that edamame and tofu are better sources than quinoa, and that soy milk is about the same.

    > Soy blocks iodine absorption,

    My understanding is that if people eat enough iodine, their thyroid will not be affected by soy. But if someone does have hypothyroidism, they should be careful not too eat too much soy just to be safe.

  16. CA Says:

    I’ve been trying to track down an amino profile for gluten out of curiosity without any success. Your pages on protein at veganhealth are incredibly informative for everything else that I eat. Any info on vital wheat gluten/seitan? I’m guessing it is low in Lysine and THR like white flour, but I’d still be interested in seeing how it fares. Any info?

    Thanks for all your work!

  17. Jack Norris RD Says:


    The USDA does not list the amino acid levels in wheat gluten. I would assume it’s similar to bread.

  18. Dave Says:

    This is an informative and interesting discussion. I think it’s particularly useful because it shows that one can get their RDA of each amino acid from foods typically considered to be “incomplete” proteins and as such, maybe a better working definition of “complete” is called for. (btw, if you guys haven’t played with Jack’s excel sheet on protein, check it out.)

    The term incomplete protein has always irked me because it implies that amino acids are completely missing when it’s just that they are present in lower amounts (as others have already pointed out), and simply eating more of that food will compensate for that. (or in practice, eat a varied diet)

    It appears that more than one “official” body has set amino acid thresholds for meeting the definition of a complete protein. The measure I came across when I was reading up on this awhile back is the one set by the IOM, referenced in the National Academies Press report that Jack mentioned above. It is more conservative (i.e, recommends more of the essential acids) than others I have seen. Check out the bottom of this free pdf summary:

    Basically, they define a complete protein as a protein that meets or exceeds these amino acid thresholds per gram protein:

    (mg of esential amino acid required per gram protein to be called “complete”)
    Tryptophan 7
    Threonine 27
    Isoleucine 25
    Leucine 55
    Lysine 51
    Methionine+Cystine 25
    Phenylalanine+Tyrosine 47
    Valine 32
    Histidine 18

    These numbers have and will change as more research is done on human nutrition. Ultimately, though, this is an idealized protein formula. It’s probably close to tuna, I would imagine. That is, a food that matches or exceeds these numbers will require the smallest overall protein consumption to hit the RDA.

    Realistically, though, we don’t try to eat the least protein possible in obtaining our RDA. This is more like a marker for one end of the protein consumption spectrum. On the other end would probably be fruit, where you could actually get all your essential AAs but would probably need to eat hundreds (or more!) of pieces to get there.

  19. Jack Norris RD Says:

    I finally got the time to check into the difference in the numbers between what used for their “complete protein” ratings and what I use. uses the protein digestibility corrected amino acid scoring pattern (PDCAAS) which Dave lists above, while I use the RDA. The ratios are very similar.

    Also, does not take into account how much protein is in the food when determining what is a complete protein.

    Mystery solved.

  20. Ycuk Says:

    Hi, Jack! Now we have some debates on different corrections to ‘ideal pattern’, including PDCAAS. As I understand, do not use PDCAAS (from their Help materials).
    You also do not use PDCAAS. (Is it right?) But protein digestibility differs depending on food, food cooking and state of body (even if source products have similar AA patterns).
    I do not like PDCAAS corrections due to some of their specifics, but use of plain AA scores seems not to be very accurate, since actual digested amino acids will differ.
    What you think about it?

  21. Jack Norris RD Says:

    Ycuk, explains how they figure their protein quality score here:

    The numbers they use are listed on p. 589 of the National Academy of Sciences’ document “Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients), which says:

    The recommended protein digestibility corrected amino acid scoring pattern (PDCAAS) for proteins for children 1 year of age and older and all other age groups is as follows (in mg/g of protein): isoleucine, 25; leucine, 55; lysine, 51, methionine + cysteine (SAA), 25; phenylalanine + tyrosine, 47; threonine, 27; tryptophan, 7; valine, 32; and histidine, 18.

    I would therefore conclude that, yes, is using the PDCAAS. I, however, used the amino acid RDA rather than the PDCAAS to figure out if a food is a complete protein. When it comes to the amino acid ratios, both ways are fairly similar, the big difference is that I considered the amount of protein in a serving, not just the ratios of amino acids.

  22. Ycuk Says:


    Thank You very much for the link to “Reference”.

    As I understand it, there are many ‘ideal patterns’ to calculate scores depending on target group and a source of patterns. Differences between these patterns are not very big. use one of such scoring patterns (but they do not apply corrections to them to calculate PDCAAS itself).

    PDCAAS corrections (i.e. protein digestibility, see p.690, Table 10-27) are also not very big (usually about 10-20%, depending on product).

    In some cases, however, they are important. For instance, protein digestibility for wheat is low (0.85, see p.690), and lysine score is also low (0.49). It leads to very low PDCAAS (42).

    But I do not found yet complete list of protein digestibility values (which are essential to calculate PDCAAS), and I do not know, how they depend on cooking.

  23. Konstantinos Says:

    Hello from Greece!

    Hello from Greece. Great work!!

    I would like to ask a question about this table.

    The servings that you calculated for your equations at this new table, for which person’s weight is required?

    Because when we change the weight at the previous table 3, automatically it changes the servings required to meet the RDA EAA too.

    Thank you

  24. Jack Norris RD Says:


    That’s true and it would change the number of servings needed to meet the RDA (it would go down for lighter people and up for heavier people), but the relative value of the different foods would stay the same and would not change my analysis.

  25. Konstantinos Says:

    Thank you for your answer!

    It’s impressive to see lentils so high, even higher than soya and milk.

    Is it possible to tell me for which exactly weight have been calculated your servings for your equations at this new table.

    Because I calculate the servings required to meet the RDA for least plentiful EAA, for exable Lentils (I see methionine is the least plentiful) for a person’s weight 140lb (63 Kg) at table 3, to be 6.3, quite higher than the table in this article.

    If we want to be 3.4 the least plentifl EAA for the Lentils (methionine) as this table, the weight should be 75 lb (34 kgr), which is too low.

    Am I calculating something wrong?

    Thank you.

  26. Jack Norris RD Says:

    Something seems really messed up as I am not getting the same numbers I did back when I wrote this post. I think when I wrote Vegan For Life, I might have changed a bunch of the serving sizes in that chart to more standard serving sizes, such as changing the legumes from 1 cup to 1/2 cup cooked, which would double the amount needed to meet the RDA. But I need to take a closer look at this tomorrow and possibly redo the chart. That’s really going to mess up the results, I think.

  27. Leo Says:

    Hey Jack,

    I have a question about protein combining and would be very grateful if you’d be willing to answer it.

    Should 1 gram of protein from grains/nuts/seeds be combined with 1 gram of protein from legumes to make a “complete” protein, or are the proportions different?

    Thank you!

  28. Jack Norris RD Says:


    You only need to make sure you’re meeting the RDA for each of the amino acids on a daily basis. Perhaps this chart will help you see what’s need to meet the RDA for various amino acids or how foods contribute to various amount:

  29. Leo Says:

    Okay, I see. Thank you.

    Do you think that vegan athletes should aim for a higher protein intake than vegetarian athletes, since plant proteins have a lower digestibility than proteins from egg and milk? If one chooses a protein intake of 2 g/kg (2.5 times higher than the RDA of 0.8 g/kg for people 18-59 years of age), is the lysine RDA then 38*2.5 = 95 mg/kg? For my body weight of 58 kg, that would mean more than 10 servings of black beans, which I find quite a lot, but perhaps I’ve not understood it correctly?

    Thank you once again!

  30. Jack Norris RD Says:


    Yes, you would theoretically want to meet the RDA for amino acids. In my opinion, since your body would have no way to use that much protein for building muscle, it won’t matter if you meet it or not *and* I would assume you’re eating foods other than black beans. If you really want that much protein and are finding it hard to eat enough to get it across your entire diet, I’d try a protein supplement.

  31. Jess Says:

    All are complete proteins that your friend mentioned except spirulina, which must be combined with grains or nuts for example.

Leave a Reply