(Credit: Olimpia Davies For Unsplash+)
In a Nutshell
- Tofu scored highest overall among 20 popular protein sources, while cheddar cheese scored lowest.
- Plant-based foods claimed every spot in the top 20% of rankings, while animal-based foods occupied the entire bottom 20%, when health and environmental factors were combined.
- A combined rating folds pollution, land and water use, sodium, saturated fat, affordability, and availability into one score.
Tofu just topped a new ranking of the world’s best protein sources, scoring 36 out of a possible 40 points. Cheddar cheese came in dead last.
Chris Macdonald, founder of the Better Protein Institute, built a tool called the BPI Score to answer a question that sounds simple but is anything but: What are the best protein sources? Writing in the journal Frontiers in Sustainable Food Systems, he scored 20 popular high-protein foods on health and environmental factors, and one pattern jumped out right away. Plant-based proteins dominated the top tier, while animal-based proteins filled the entire bottom tier.
That gap matters because popular labels like nutrition traffic lights or carbon footprint stickers only tell part of the story. They tend to measure one thing at a time. A food might look great on a carbon label but hide a health problem, or look healthy on a nutrition label while quietly damaging a water supply. The BPI Score combines those pieces into one rating, and that fuller picture reveals a divide between plant and animal protein sources that narrower labels miss.
How the New Protein Scoring System Works
Macdonald built the BPI Score around two categories: a Planet Score and a People Score. The goal was to avoid the trap of single-issue labels that look only at, say, greenhouse gas emissions or sodium content while ignoring everything else.
For the Planet Score, Macdonald pulled from an existing dataset covering climate-warming emissions, water pollution, land use, and water use, all calculated per 100 grams of protein so comparisons would be fair across very different foods. A bowl of beans and a steak don’t contain the same amount of protein per serving, so standardizing the measurement this way lets a nut be compared fairly against a cut of meat.
On emissions, the gap was especially wide. Tofu produced far lower emissions than beef when measured per 100 grams of protein, with the paper citing a gap of roughly two orders of magnitude, meaning tofu’s emissions were about 100 times lower. Land use told a similar story: tofu required a small fraction of the land that beef needed to produce the same amount of protein. Water pollution followed the same pattern, with plant foods like chia seeds producing far less pollution than foods like pork.
Water use was messier. Nuts used the most water overall, yet plant foods still ranked lowest, with tofu, sunflower seeds, and beans using very little water compared to most other options.
For the People Score, researchers focused on health markers linked to obesity and heart disease: sodium, saturated fat, and cholesterol, again measured per 100 grams of protein. Sodium levels showed the same plant-versus-animal split seen in the environmental numbers, with beans containing very little sodium compared to foods like mozzarella. Saturated fat results were more mixed, with some animal products landing at both ends of the scale: cod scored well, while cheddar scored poorly. Even a plant-based food, Brazil nuts, ranked near the bottom of this category, reminding us that the plant-versus-animal divide wasn’t absolute across every measure.
Cholesterol produced an odd result. Every plant food tested scored zero, while animal products varied widely, with tuna scoring relatively well and eggs scoring poorly.
Researchers also considered whether a food counted as a “complete” protein, meaning it contains all nine amino acids the body needs but can’t make on its own. Every animal-based food in the study qualified as complete, while most plant-based foods didn’t, with tofu among the exceptions.
Finally, the study looked at affordability by checking prices at a major UK grocery chain and whether each food was available in UK supermarkets. Lentils were the cheapest per 100 grams of protein, while quinoa was the most expensive.
The Full Ranking: 20 Protein Sources, Best to Worst
How each food placed on the BPI Score, which combines environmental impact (emissions, water pollution, land use, freshwater use) with health and accessibility (sodium, saturated fat, cholesterol, protein completeness, price, availability). Maximum score: 40 points.
- 1Tofu36 ptsPlant
- 2Chia seedsPlant
- 3BeansPlant
- 4QuinoaPlant
- 5Sunflower seedsPlant
- 6LentilsPlant
- 7Chicken breastAnimal
- 8TunaAnimal
- 9PeanutsPlant
- 10ChickpeasPlant
- 11WalnutsPlant
- 12Brazil nutsPlant
- 13SalmonAnimal
- 14EggsAnimal
- 15Pork chopsAnimal
- 16CodAnimal
- 17Beef steakAnimal
- 18Lamb chopsAnimal
- 19MozzarellaAnimal
- 20Cheddar cheese13 ptsAnimal
Measurements are per 100 grams of protein where applicable. Prices and availability are based on UK supermarket data. The published paper (Macdonald, Frontiers in Sustainable Food Systems, 2026) reports the overall winner and loser, their point totals, and the top and bottom four foods. The full 1–20 order comes from the Better Protein Institute’s results summary, which is based on an updated version (V1.1) of the score.
Why Tofu Is the Best Protein Source, and Cheddar Isn’t
When researchers added up all the scores, tofu took the top spot, and cheddar cheese finished at the bottom. The pattern held at a wider level too. Every food in the top fifth of the rankings was plant-based, including tofu, chia seeds, beans, and quinoa. Every food in the bottom fifth was animal-based, including mozzarella, lamb, beef, and cheddar.
This wasn’t a huge study. The study examined only 20 products, and the data came from existing databases rather than new lab testing. Pricing information came from a single UK supermarket chain, so the affordability results may not hold up the same way in other countries where food costs and availability differ.
What the Best Protein Source Score Leaves Out, for Now
Macdonald is upfront that this first version of the BPI Score, which he calls V1, doesn’t capture everything. This round left out factors like the treatment of farmed animals, the risk of antibiotic resistance from livestock farming, and broader health risks tied to certain meats, mostly because they’re hard to measure in a simple point system.
Macdonald argues that including these missing pieces, even in a limited way, would likely widen the gap between plant and animal proteins rather than close it. He also raises a bigger tradeoff: giving farmed animals more space to roam can mean less space for wild animals and natural habitats, while cramming farmed animals into less space spares wild land but worsens conditions for those animals. There’s no clean win here, which is part of why these ethical questions are so hard to reduce to a single number.
A next version of the score, already in development, sets out to tackle some of these harder questions. For now, the findings make one thing clear: under this scoring system, when health costs, pollution, land use, and water use are weighed together instead of looked at one at a time, popular animal proteins like beef and cheese come out looking far worse than their reputation as kitchen staples would imply.
Paper Notes
Limitations
Macdonald examined only 20 high-protein food products, a small sample he describes as a first version, or “V1,” meant to be improved over time. Environmental and health data were pulled from existing databases rather than newly collected measurements, and the simplified five-point grading scale used for scoring does not fully capture outliers or the wide spread of values seen in categories like pollution. Affordability and availability data came from a single UK supermarket chain, Aldi, which the paper notes limits how well those findings apply outside the UK, since food costs and access vary by region due to factors like subsidies, trade policies, and food deserts. This score also currently leaves out harder-to-quantify factors such as the ethical treatment of farmed animals, antibiotic resistance linked to livestock farming, and zoonotic disease risk.
Funding and Disclosures
Macdonald reported receiving financial support for this work, with Lucy Cavendish College, University of Cambridge, providing funding for publication costs. He declared no commercial or financial relationships that could be seen as a conflict of interest and stated that he did not use generative AI to write the manuscript.
Publication Details
Paper Title: “What are the best sources of protein? Introducing the BPI Score”
Author: Chris Macdonald
Journal: Frontiers in Sustainable Food Systems on July 29, 2026. It is listed as a Perspective article, with the citation: Macdonald C (2026) What are the best sources of protein? Introducing the BPI Score. Front. Sustain. Food Syst. 10:1677796.







