Food · Grains & Legumes · 豆类
Soybeans & Tofu
少见的植物完全蛋白 · 异黄酮是植物雌激素≠ 雌激素 · 拆穿豆制品让男性女性化、致乳腺癌· 钙凝固豆腐是好钙源 · 整豆 ≫ 大豆分离蛋白 · 甲状腺药需错峰
Story path
- 1What is soy · one bean, a whole familyWhat is soy · one bean, a whole family
- 2Complete protein · a rarity among plantsComplete protein · a rarity among plants
- 3Isoflavones · what 'phytoestrogen' really meansIsoflavones · what 'phytoestrogen' really means
- 4Debunked · 'soy feminizes men'Debunked · 'soy feminizes men'
- 5Debunked · 'breast cancer means no soy'Debunked · 'breast cancer means no soy'
- 6Tofu & calcium · a hidden benefit set by the coagulantTofu & calcium · a hidden benefit set by the coagulant
- 7Whole soy ≫ isolate · don't let processing misleadWhole soy ≫ isolate · don't let processing mislead
- 8How to eat · how much · who should careHow to eat · how much · who should care
Chapter 1
What is soy · one bean, a whole family
What is soy · one bean, a whole family
The soybean (Glycine max) is the seed of a legume, cultivated and eaten in East Asia for over two thousand years — one of the few crops to travel from 'the poor person's meat' all the way into the spotlight of modern nutrition science. What makes it special: in a plant-staple world that is mostly 'starch-heavy, protein-light,' soy is an outlier — high in protein, and high-quality protein at that (next scene).
More importantly, soy is not a single food but a whole family:
Edamame: immature green soybeans, eaten wholeSoy milk: ground soybean slurryTofu: soy milk set with a coagulant (calcium or magnesium salt) into a gel — silken, firm, and dried/pressed grades by water contentFermented soy: natto, miso, tempeh, soy sauce — fermentation lowers anti-nutrients and creates vitamin K2 and probiotics (that thread belongs to fermented foods)Soy protein isolate: an industrially purified protein powder found in protein bars, plant-based meats, and processed foods — NOT the same thing as a whole soybean (covered later)
Soy earns its own island for two reasons. Half is that it is so central to Chinese diets and so underrated nutritionally. The other half is that it is the most myth-entangled food online — 'soy is full of estrogen,' 'men who eat soy get feminized,' 'breast cancer patients must avoid soy.' This island's job is to take those apart one mechanism and one citation at a time.
More importantly, soy is not a single food but a whole family:
Edamame: immature green soybeans, eaten wholeSoy milk: ground soybean slurryTofu: soy milk set with a coagulant (calcium or magnesium salt) into a gel — silken, firm, and dried/pressed grades by water contentFermented soy: natto, miso, tempeh, soy sauce — fermentation lowers anti-nutrients and creates vitamin K2 and probiotics (that thread belongs to fermented foods)Soy protein isolate: an industrially purified protein powder found in protein bars, plant-based meats, and processed foods — NOT the same thing as a whole soybean (covered later)
Soy earns its own island for two reasons. Half is that it is so central to Chinese diets and so underrated nutritionally. The other half is that it is the most myth-entangled food online — 'soy is full of estrogen,' 'men who eat soy get feminized,' 'breast cancer patients must avoid soy.' This island's job is to take those apart one mechanism and one citation at a time.
Chapter 2
Complete protein · a rarity among plants
Complete protein · a rarity among plants
Soy's hardest credential: it is one of the few complete plant proteins.
A 'complete protein' contains all 9 essential amino acids in proportions close to human need. Most plant proteins are 'incomplete' — lentils and rice each lack a piece and must be paired for complementarity (dive to protein for the amino-acid story). Soy needs no pairing: on its own it covers all 9 essential amino acids, with a Protein Digestibility-Corrected Amino Acid Score (PDCAAS) near 1.0 — the same tier as milk and egg. That is rare among plants (quinoa is another exception).
By content (USDA, per 100 g):
Firm tofu: ~12-17 g protein, ~110-145 kcal, mostly unsaturated fatSilken tofu: ~5-8 g protein, softer and wetterEdamame: ~11 g protein, plus fiberDry soybeans: ~36 g protein
For comparison: a palm-sized block of firm tofu carries protein in the range of a small piece of chicken, yet with almost no cholesterol and low saturated fat. For anyone cutting red or processed meat (dive to processed-meat) while worried that plant protein is 'not good enough,' soy is a scientifically solid anchor protein — not a compromise.
A 'complete protein' contains all 9 essential amino acids in proportions close to human need. Most plant proteins are 'incomplete' — lentils and rice each lack a piece and must be paired for complementarity (dive to protein for the amino-acid story). Soy needs no pairing: on its own it covers all 9 essential amino acids, with a Protein Digestibility-Corrected Amino Acid Score (PDCAAS) near 1.0 — the same tier as milk and egg. That is rare among plants (quinoa is another exception).
By content (USDA, per 100 g):
Firm tofu: ~12-17 g protein, ~110-145 kcal, mostly unsaturated fatSilken tofu: ~5-8 g protein, softer and wetterEdamame: ~11 g protein, plus fiberDry soybeans: ~36 g protein
For comparison: a palm-sized block of firm tofu carries protein in the range of a small piece of chicken, yet with almost no cholesterol and low saturated fat. For anyone cutting red or processed meat (dive to processed-meat) while worried that plant protein is 'not good enough,' soy is a scientifically solid anchor protein — not a compromise.
Chapter 3
Isoflavones · what 'phytoestrogen' really means
Isoflavones · what 'phytoestrogen' really means
Almost every soy myth points at one word: isoflavones — mainly genistein and daidzein. They are called 'phytoestrogens,' and that name has misled countless people. Let's open the mechanism.
The key: 'phytoestrogen' ≠ estrogen
Isoflavones are structurally somewhat similar to the body's estrogen (estradiol), so they can bind estrogen receptors. But 'can bind' is not 'is equivalent.' Two crucial differences:
1. Extremely weak: isoflavones activate the receptor orders of magnitude more weakly than the body's own estrogen — they sit in the same chair but barely push.
2. Different receptor preference: humans have two estrogen receptors (ERα and ERβ). Estradiol hits both; isoflavones prefer ERβ. Yet ERα is the main driver of proliferative signaling in breast and uterus, while ERβ often acts as a regulatory 'brake' in many tissues.
So isoflavones are more accurately a selective estrogen receptor modulator (SERM)-like plant molecule — in different tissues they may act slightly 'estrogen-like' or 'anti-estrogenic,' depending on the local receptor type and the body's own estrogen level. That is why human studies show them to be neither as dangerous as 'taking estrogen' nor as miraculously 'hormone-balancing' as supplements claim.
Individual variation: gut bacteria can convert daidzein into a more active metabolite (equol), but only some people (a higher share in Asian populations) carry the bacteria for it — which partly explains why results differ between individuals and studies.
The one line from this scene: isoflavones are weak molecules that 'sit in the chair, barely push, and are picky about which chair' — not the estrogen you imagined. The next two scenes test this mechanism against human evidence.
The key: 'phytoestrogen' ≠ estrogen
Isoflavones are structurally somewhat similar to the body's estrogen (estradiol), so they can bind estrogen receptors. But 'can bind' is not 'is equivalent.' Two crucial differences:
1. Extremely weak: isoflavones activate the receptor orders of magnitude more weakly than the body's own estrogen — they sit in the same chair but barely push.
2. Different receptor preference: humans have two estrogen receptors (ERα and ERβ). Estradiol hits both; isoflavones prefer ERβ. Yet ERα is the main driver of proliferative signaling in breast and uterus, while ERβ often acts as a regulatory 'brake' in many tissues.
So isoflavones are more accurately a selective estrogen receptor modulator (SERM)-like plant molecule — in different tissues they may act slightly 'estrogen-like' or 'anti-estrogenic,' depending on the local receptor type and the body's own estrogen level. That is why human studies show them to be neither as dangerous as 'taking estrogen' nor as miraculously 'hormone-balancing' as supplements claim.
Individual variation: gut bacteria can convert daidzein into a more active metabolite (equol), but only some people (a higher share in Asian populations) carry the bacteria for it — which partly explains why results differ between individuals and studies.
The one line from this scene: isoflavones are weak molecules that 'sit in the chair, barely push, and are picky about which chair' — not the estrogen you imagined. The next two scenes test this mechanism against human evidence.
Chapter 4
Debunked · 'soy feminizes men'
Debunked · 'soy feminizes men'
'Men who eat soy will lower their testosterone / grow breasts / get feminized' — the most widespread soy myth, usually paired with an anecdotal photo of 'a guy who drank soy milk daily and grew breasts.' Let's unpack it.
The myth's 'source': almost entirely extreme single cases plus a literal misreading of the word 'phytoestrogen' from the previous scene. There were 1-2 case reports of a man who drank 3 liters of soy milk a day (an isoflavone dose far beyond any normal diet) developing breast tissue that resolved after stopping. That extreme-dose anecdote got inflated into 'men can't eat soy.'
What the evidence says (Hamilton-Reeves 2010, *Fertility & Sterility*): a meta-analysis pooling 32 studies tested the core question directly — whether soy protein or isoflavones lower men's testosterone, free testosterone, or SHBG. The conclusion is clear: no significant effect. Normal, even high, soy/isoflavone intake does not change men's androgen levels.
Why it is counterintuitive: exactly the mechanism from the previous scene — isoflavones bind the receptor very weakly and prefer ERβ, and at dietary doses they never reach 'endogenous estrogen' strength. The isoflavones in a cup of soy milk or a block of tofu are worlds away from the '3 liters a day' extreme case.
The verdict: within a normal diet, men eating soy will not be feminized and will not have lowered testosterone. This is confirmed repeatedly by meta-analysis — not a 'maybe.' Just eat tofu and soy milk as the high-quality protein they are.
The myth's 'source': almost entirely extreme single cases plus a literal misreading of the word 'phytoestrogen' from the previous scene. There were 1-2 case reports of a man who drank 3 liters of soy milk a day (an isoflavone dose far beyond any normal diet) developing breast tissue that resolved after stopping. That extreme-dose anecdote got inflated into 'men can't eat soy.'
What the evidence says (Hamilton-Reeves 2010, *Fertility & Sterility*): a meta-analysis pooling 32 studies tested the core question directly — whether soy protein or isoflavones lower men's testosterone, free testosterone, or SHBG. The conclusion is clear: no significant effect. Normal, even high, soy/isoflavone intake does not change men's androgen levels.
Why it is counterintuitive: exactly the mechanism from the previous scene — isoflavones bind the receptor very weakly and prefer ERβ, and at dietary doses they never reach 'endogenous estrogen' strength. The isoflavones in a cup of soy milk or a block of tofu are worlds away from the '3 liters a day' extreme case.
The verdict: within a normal diet, men eating soy will not be feminized and will not have lowered testosterone. This is confirmed repeatedly by meta-analysis — not a 'maybe.' Just eat tofu and soy milk as the high-quality protein they are.
Chapter 5
Debunked · 'breast cancer means no soy'
Debunked · 'breast cancer means no soy'
'Soy has estrogen, it feeds or triggers breast cancer, and patients especially must avoid it' — the most anxiety-inducing myth, and the one most worth correcting, because it drives many people who would benefit to avoid a good food.
Where the worry comes from: in some cell-culture and mouse experiments, high concentrations of genistein stimulated the proliferation of estrogen-receptor-positive breast cancer cells. The doses and metabolism in a petri dish or a rodent are nothing like a real person eating tofu, but those early experiments spread as 'proof soy causes cancer.'
Human evidence says the opposite:
**Shu 2009 (*JAMA*): the Shanghai Breast Cancer Survival Study followed 5,042 breast cancer patients. Result — higher soy intake after diagnosis was associated with lower** mortality and recurrence (an inverse association), not higher.**Nechuta 2012 (*Am J Clin Nutr*): a pooled analysis of three US and Chinese cohorts totaling 9,514 breast cancer survivors. Those consuming ≥10 mg isoflavones per day (below the average Asian diet) had about a 25% lower recurrence risk** and lower all-cause mortality.
In other words, in real populations, breast cancer patients who eat soy do better, not worse. The mechanism fits: isoflavones prefer ERβ and activate weakly, behaving more like an 'occupier / anti-estrogen' in breast tissue rather than fuel on the fire.
Thyroid (Messina 2006, *Thyroid*): a related worry, in passing — in iodine-replete, euthyroid people, soy/isoflavones do not harm thyroid function. (The one genuine caveat is thyroid-medication absorption, covered in the last scene.)
The verdict: major authorities (e.g. the American Institute for Cancer Research) now state clearly that moderate soy is safe, and possibly beneficial, for breast cancer patients. Treating cell/mouse experiments as a human conclusion is the root error of this myth.
This scene is general education; a cancer patient's specific diet should still be discussed with their treating physician.
Where the worry comes from: in some cell-culture and mouse experiments, high concentrations of genistein stimulated the proliferation of estrogen-receptor-positive breast cancer cells. The doses and metabolism in a petri dish or a rodent are nothing like a real person eating tofu, but those early experiments spread as 'proof soy causes cancer.'
Human evidence says the opposite:
**Shu 2009 (*JAMA*): the Shanghai Breast Cancer Survival Study followed 5,042 breast cancer patients. Result — higher soy intake after diagnosis was associated with lower** mortality and recurrence (an inverse association), not higher.**Nechuta 2012 (*Am J Clin Nutr*): a pooled analysis of three US and Chinese cohorts totaling 9,514 breast cancer survivors. Those consuming ≥10 mg isoflavones per day (below the average Asian diet) had about a 25% lower recurrence risk** and lower all-cause mortality.
In other words, in real populations, breast cancer patients who eat soy do better, not worse. The mechanism fits: isoflavones prefer ERβ and activate weakly, behaving more like an 'occupier / anti-estrogen' in breast tissue rather than fuel on the fire.
Thyroid (Messina 2006, *Thyroid*): a related worry, in passing — in iodine-replete, euthyroid people, soy/isoflavones do not harm thyroid function. (The one genuine caveat is thyroid-medication absorption, covered in the last scene.)
The verdict: major authorities (e.g. the American Institute for Cancer Research) now state clearly that moderate soy is safe, and possibly beneficial, for breast cancer patients. Treating cell/mouse experiments as a human conclusion is the root error of this myth.
This scene is general education; a cancer patient's specific diet should still be discussed with their treating physician.
Chapter 6
Tofu & calcium · a hidden benefit set by the coagulant
Tofu & calcium · a hidden benefit set by the coagulant
Many people don't know that how much calcium a block of tofu contains depends on which coagulant was used to make it. This is a practical mechanism worth remembering.
Soy milk itself isn't especially high in calcium. To set it into tofu, you add a coagulant:
Calcium salt (calcium sulfate / gypsum): common in traditional firm tofu. The calcium enters the tofu matrix directly, so the finished product is markedly higher in calcium — this tofu can reach 150-350 mg calcium per 100 g, near or above the same weight of milk.Magnesium salt (magnesium chloride / nigari): common in Japanese-style tofu; cleaner flavor but much lower calcium.Glucono-delta-lactone (GDL): common in silken/lactone tofu; also low calcium.
So whether 'tofu builds bone' depends on the first coagulant on the ingredient list. To use tofu for calcium, choose calcium-sulfate (gypsum)-set firm tofu. This calcium is reasonably well absorbed, making it an excellent plant calcium source for the lactose-intolerant (dive to milk) or those who don't drink milk. For the full mechanism of how calcium gets into bone, dive to calcium.
Incidentally, soy also supplies magnesium, potassium, iron (non-heme — better absorbed with vitamin C, dive to iron), and cardiovascular-neutral-to-friendly unsaturated fat. Overall it is a nutrient-dense food that can be eaten minimally or heavily processed.
Soy milk itself isn't especially high in calcium. To set it into tofu, you add a coagulant:
Calcium salt (calcium sulfate / gypsum): common in traditional firm tofu. The calcium enters the tofu matrix directly, so the finished product is markedly higher in calcium — this tofu can reach 150-350 mg calcium per 100 g, near or above the same weight of milk.Magnesium salt (magnesium chloride / nigari): common in Japanese-style tofu; cleaner flavor but much lower calcium.Glucono-delta-lactone (GDL): common in silken/lactone tofu; also low calcium.
So whether 'tofu builds bone' depends on the first coagulant on the ingredient list. To use tofu for calcium, choose calcium-sulfate (gypsum)-set firm tofu. This calcium is reasonably well absorbed, making it an excellent plant calcium source for the lactose-intolerant (dive to milk) or those who don't drink milk. For the full mechanism of how calcium gets into bone, dive to calcium.
Incidentally, soy also supplies magnesium, potassium, iron (non-heme — better absorbed with vitamin C, dive to iron), and cardiovascular-neutral-to-friendly unsaturated fat. Overall it is a nutrient-dense food that can be eaten minimally or heavily processed.
Chapter 7
Whole soy ≫ isolate · don't let processing mislead
Whole soy ≫ isolate · don't let processing mislead
When praising soy's benefits, you must separate 'whole soybeans / traditional soy foods' from 'soy protein isolate' — their health meaning differs.
This end (recommended): whole-soy foods — edamame, tofu, soy milk, natto, miso, tempeh. They keep soy's natural matrix: protein + fiber + minerals + intact plant structure. The fermented ones (natto / miso / tempeh) add a bonus: fermentation lowers anti-nutrients like phytate, creates vitamin K2 (dive to vitamin-k2), and brings probiotics — that thread is developed under 'fermented foods.'
That end (watch out): soy protein isolate (SPI) is an industrially purified protein powder with the fiber and fat stripped out, appearing widely in protein bars, meal replacements, plant-based meats, and processed foods. SPI itself is not 'poison,' and as a protein supplement it has its uses; the issue is that it usually rides inside highly processed products (added sugar, salt, additives). In other words: what makes some 'plant-based meats' unhealthy is not the soy — it is their ultra-processed nature (dive to superfoods for how a 'health halo' gets exploited by marketing).
How to apply it: prioritize whole soy and traditional soy foods, treating soy as real food rather than 'raw material for protein powder.' A serving of edamame, a block of tofu, a bowl of miso soup beats a 'high-protein bar' wrapped around soy protein isolate and sugar.
This end (recommended): whole-soy foods — edamame, tofu, soy milk, natto, miso, tempeh. They keep soy's natural matrix: protein + fiber + minerals + intact plant structure. The fermented ones (natto / miso / tempeh) add a bonus: fermentation lowers anti-nutrients like phytate, creates vitamin K2 (dive to vitamin-k2), and brings probiotics — that thread is developed under 'fermented foods.'
That end (watch out): soy protein isolate (SPI) is an industrially purified protein powder with the fiber and fat stripped out, appearing widely in protein bars, meal replacements, plant-based meats, and processed foods. SPI itself is not 'poison,' and as a protein supplement it has its uses; the issue is that it usually rides inside highly processed products (added sugar, salt, additives). In other words: what makes some 'plant-based meats' unhealthy is not the soy — it is their ultra-processed nature (dive to superfoods for how a 'health halo' gets exploited by marketing).
How to apply it: prioritize whole soy and traditional soy foods, treating soy as real food rather than 'raw material for protein powder.' A serving of edamame, a block of tofu, a bowl of miso soup beats a 'high-protein bar' wrapped around soy protein isolate and sugar.
Chapter 8
How to eat · how much · who should care
How to eat · how much · who should care
How to choose / cook:
For calcium: choose firm tofu whose ingredient list says 'calcium sulfate'High-protein, low-processing: edamame, tofu, soy milk (unsweetened), nattoMake soy milk at home or buy the no-added-sugar kind; flavored soy drinks are often sugary — read the labelWhole dry soybeans need thorough soaking + full cooking to inactivate lectins (same mechanism as the anti-nutrient scene — dive to lentils)
How much: large population evidence supports 1-2 servings of soy foods per day (1 serving ≈ 100 g tofu / 1 cup soy milk / half a cup edamame) as a safe and beneficial range. Traditional Asian diets have done this for generations without 'hormone problems.' There is no need to pile on to extreme amounts (the source of the anecdotal myths), nor any need to fear it to the point of abstaining entirely.
Who should pay attention:
People on thyroid medication (levothyroxine): this is the one genuine interaction — soy does not harm thyroid function, but it reduces levothyroxine absorption. The fix is simple: separate the medication from soy (and other high-calcium / high-fiber foods) by a few hours (commonly ≥4 hours after taking the drug on an empty stomach before eating soy). Follow your physician's instructions. Dive to hashimoto.People with soy allergy: a true soy allergy (especially in children) means avoidance — this is an immune issue, unrelated to the mythsGout: soy was once thought to trigger gout, but modern evidence shows plant purines (including legumes) are weakly associated with flares; moderate intake is fine. Still consult a doctor during acute flares.
This scene provides general information only and does not replace a physician's judgment of your individual situation.
For calcium: choose firm tofu whose ingredient list says 'calcium sulfate'High-protein, low-processing: edamame, tofu, soy milk (unsweetened), nattoMake soy milk at home or buy the no-added-sugar kind; flavored soy drinks are often sugary — read the labelWhole dry soybeans need thorough soaking + full cooking to inactivate lectins (same mechanism as the anti-nutrient scene — dive to lentils)
How much: large population evidence supports 1-2 servings of soy foods per day (1 serving ≈ 100 g tofu / 1 cup soy milk / half a cup edamame) as a safe and beneficial range. Traditional Asian diets have done this for generations without 'hormone problems.' There is no need to pile on to extreme amounts (the source of the anecdotal myths), nor any need to fear it to the point of abstaining entirely.
Who should pay attention:
People on thyroid medication (levothyroxine): this is the one genuine interaction — soy does not harm thyroid function, but it reduces levothyroxine absorption. The fix is simple: separate the medication from soy (and other high-calcium / high-fiber foods) by a few hours (commonly ≥4 hours after taking the drug on an empty stomach before eating soy). Follow your physician's instructions. Dive to hashimoto.People with soy allergy: a true soy allergy (especially in children) means avoidance — this is an immune issue, unrelated to the mythsGout: soy was once thought to trigger gout, but modern evidence shows plant purines (including legumes) are weakly associated with flares; moderate intake is fine. Still consult a doctor during acute flares.
This scene provides general information only and does not replace a physician's judgment of your individual situation.