Tofu tops Cambridge researcher’s new protein ranking, with cheddar last

Tofu has ranked first in a new scoring system developed by University of Cambridge researcher Dr. Chris Macdonald to compare protein sources across environmental impact, nutritional characteristics, affordability and availability. Cheddar cheese finished last among the 20 foods assessed.

• Tofu scored 36 out of 40, while cheddar scored 13. All four highest-scoring foods were plant-based.
• The BPI Score combines four environmental measures with six measures covering nutrition and accessibility.
• The first version uses UK supermarket prices and availability, with Macdonald acknowledging limits to its wider application.

Published in July in Frontiers in Sustainable Food Systems, the perspective paper introduces the Better Protein Institute’s BPI Score. Macdonald, who founded the institute, developed the framework to examine questions that individual nutrition labels or carbon footprint measurements cannot answer alone.

A food may perform well on greenhouse gas emissions but poorly on freshwater use, for example. A nutritional comparison may overlook cost. The BPI Score brings those measurements together, although its results depend on which factors are included and how points are allocated.

“Without multidimensional analysis, we could be missing important trade-offs; a product may be a clear winner in a given category but a clear loser in another,” Macdonald wrote.

The first version divides its assessment into two categories: people and planet. Environmental performance accounts for up to 16 points, while nutritional and accessibility measures contribute up to 24, giving a maximum overall score of 40.

Tofu scored 36 points, followed in the highest-scoring group by chia seeds, beans and quinoa. At the other end were mozzarella, lamb, beef and cheddar, which scored 13 points.

The environmental assessment draws on a dataset published by researcher Michael Clark and colleagues in 2022. It compares greenhouse gas emissions, water pollution, land use and freshwater consumption, with each measured per 100g of protein.

Using protein content as the common denominator allows foods with different protein concentrations to be compared on the same basis. The figures therefore describe the impact of supplying 100g of protein, rather than producing a 100g serving of food.

Some differences were substantial. The dataset put tofu’s greenhouse gas emissions at 0.61kg of carbon dioxide equivalent per 100g of protein, compared with 64.19kg for beef. Land use was 2.16m² for tofu and 211.38m² for beef.

Freshwater consumption produced a less uniform result. Nuts had the highest figure cited in the paper, at 2,239.1 liters per 100g of protein, while several plant-based foods recorded much lower values.

Tofu received the maximum 16 environmental points. Beef received one.

For the people category, Macdonald assessed sodium, saturated fat and cholesterol using the Nutritionix database. He also awarded points according to whether each food was classified as a complete protein source, alongside its affordability and availability.

The nutritional measures did not produce a consistent plant-versus-animal split. Cod had the lowest saturated fat figure cited, while cheddar had the highest. Brazil nuts also fell within the poorest-performing group on that measure.

Protein completeness was scored as a binary measure, with four points for foods classified as complete and none for those classified as incomplete. Macdonald acknowledged that this choice could be contested because consumers can obtain the essential amino acids they need by eating varied protein sources.

Affordability was based on prices from Aldi’s UK website, adjusted to the cost of providing 100g of protein. Lentils were the cheapest option and quinoa the most expensive, putting plant-based foods at both ends of the price comparison.

Availability contributed no distinction between the foods assessed: all were available in a UK supermarket and received four points.

That measure was included partly to accommodate future comparisons involving proteins made through precision fermentation or cell cultivation as products reach retail markets.

The paper describes several limitations. Prices from one supermarket and a UK-based availability test cannot establish how accessible the same foods are elsewhere. Agricultural subsidies, trade tariffs and local purchasing power can all change the comparison.

The ranking method also compresses large differences. Foods receive between zero and four points according to their position within five groups, so exceptionally high environmental impacts are not fully reflected in the final score.

“Another limitation of the current score is that the simplified five-point grading system doesn't sufficiently capture the diverse spread of values, nor does it sufficiently reflect outliers,” Macdonald wrote.

A second version is already in development, with a wider grading system among the planned changes.

If you liked this, check these out...‍

• Price remains the biggest hurdle for alternative proteins, major Lancet review finds

• Global review finds whole-food plant-based diets generally cost less than meat-based diets

• UK retail plant-based protein sales could double by 2040, Systemiq report finds