BRENNEN

Industries / Food, farming and forestry / Food and beverage processing / Plant-based and alternative protein

Taste and nutrition

This work makes alternative proteins taste and smell like the foods they replace, and measures what they do for the people who eat them.

7

Milestones achieved, from 2013 to 2026.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 2013

    A better protein score

    Since a 1991 FAO/WHO report, protein quality had been rated with PDCAAS, a score that corrects a food's mix of essential amino acids for how well its protein is digested, and caps the result at 1.0. An expert meeting of the UN Food and Agriculture Organization (FAO) held in Auckland in 2011, published in 2013, recommended a new score, DIAAS, which counts each essential amino acid digested by the end of the small intestine and is not capped.

    Lessonthe cap hid real differences, since milk and soya both got the top score although milk holds more of some essential amino acids, the FAO report says.

    fao.org

    Credit

    FAO Expert Consultation on Protein Quality Evaluation in Human Nutrition

  2. 2020

    Only soy milk counts

    The Dietary Guidelines for Americans 2020–2025 count soy drinks and soy yogurt fortified with calcium, vitamin A and vitamin D in the dairy group, because their nutrients and uses match milk and yogurt. Almond, rice, coconut, oat and hemp drinks are left out, because their overall nutrient content is not similar to dairy milk.

    Lessonin official advice, a plant drink replaces milk only when its nutrients match, not because it looks and pours like milk.

    dietaryguidelines.gov

    Credit

    US Department of Agriculture; US Department of Health and Human Services

  3. 2020

    Swapping meat in a trial

    In the SWAP-MEAT trial at Stanford, 36 adults ate at least two servings a day of Beyond Meat products for eight weeks and animal meat for eight weeks, keeping the rest of their diet the same. In the plant phase their LDL cholesterol was lower (109.9 against 120.7 mg/dL) and they weighed 0.9 kg less. Beyond Meat supplied the plant products and funded the study with an unrestricted gift to Stanford.

    Lessonthe trial's main measure, blood TMAO, was lower overall but significantly so only in people who ate the plant products second, so the order of a short swap can shape its result.

    pmc.ncbi.nlm.nih.gov

    Credit

    Gardner and colleagues, Stanford University

  4. 2023

    Blind tasters rank burgers

    A study gave 175 untrained consumers four burgers: two plant-based, one 75% beef and 25% mushroom, and one all-beef. Half tasted them blind and half were told what each was. Both groups ranked the plant-based burger built to resemble meat first and beef second; blind tasters found no significant difference between beef and the other two burgers, and they could pick out the beef yet still preferred the meat-like plant burger.

    ebi.ac.uk

    Credit

    Sogari, Caputo, Petterson, Mora and Boukid, in Food Research International

  5. 2024

    No LDL change in Singapore

    In an eight-week trial in Singapore, 89 adults at raised risk of diabetes replaced their usual protein foods with 2.5 servings a day of either plant-based meats or the matching animal meats. LDL cholesterol, the main measure, did not differ between the groups. The plant group ate more fibre and more sodium, and its diastolic blood pressure was lower.

    Lessonthe cholesterol drop seen in the US SWAP-MEAT trial did not repeat with another population and other products.

    ebi.ac.uk

    Credit

    Toh, Henry and colleagues, in the American Journal of Clinical Nutrition

  6. 2024

    Processed plant foods and hearts

    A study of 126,842 adults in the UK Biobank, followed for a median of nine years, found that each extra 10 percentage points of daily energy from plant foods that were not ultra-processed went with a 7% lower risk of heart and circulatory disease. The same amount from ultra-processed plant foods went with a 5% higher risk. Ultra-processed foods are a class in the NOVA system for industrial products made largely from refined ingredients and additives; the group studied here covers every plant-sourced one, not only meat and milk substitutes.

    Lessona food made from plants is not a healthier choice by that fact alone; how far it is processed also counts.

    ebi.ac.uk

    Credit

    Rauber, Levy and colleagues; funded by World Cancer Research Fund

  7. 2026

    Plant chicken scores lower

    Using INFOGEST, a standard laboratory model of human digestion, researchers compared chicken breast and thigh with five commercial plant-based chicken products. Chicken scored a DIAAS of 100% or more and the plant products 52–65%; 81–93% of the plant products' amino acids were digested, against about 100% for chicken. A 2025 study of soy-based meat analogues made by high-moisture extrusion found in vitro DIAAS of 81 to 102, so the score depends on the protein and the process.

    ebi.ac.ukebi.ac.uk

    Credit

    Alotaibi, Muleya, Salter, Hoad and Eldeghaidy; Hammer, Moretti and colleagues

Watch for

  1. Preferred by meat eaters

    Impossible Foods calls itself "the plant-based protein brand preferred by meat eaters in taste tests" in announcing its first UK supermarket range, at Tesco. The company has said it runs blind paired tests with 100 to 200 regular meat eaters at independent university facilities, but it has not published the data.

    Achieved whenan independent sensory study publishes blind-test data in which meat eaters prefer Impossible's current recipe over beef.

    impossiblefoods.comimpossiblefoods.com

    Claimed by

    Impossible Foods, September 2026

Open

  1. 01

    Aroma scores from lab data

    A tool that takes the list of aroma compounds measured in a plant-based meat by gas chromatography–mass spectrometry (GC-MS, an instrument that separates and identifies the volatile chemicals in a sample) and predicts whether a trained panel would judge its smell meaty or off.

    The brief
    Credit · unclaimed
  2. 02

    Protein scores for best-sellers

    A public table of in vitro DIAAS for the 30 best-selling plant-based meats and milks in at least two countries, all measured with one method and listed with brand, batch and purchase date.

    The brief
    Credit · unclaimed

Open 01 · the brief

Aroma scores from lab data

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A tool that takes the list of aroma compounds measured in a plant-based meat by gas chromatography–mass spectrometry (GC-MS, an instrument that separates and identifies the volatile chemicals in a sample) and predicts whether a trained panel would judge its smell meaty or off.

Why it is still open

To tie compounds to meaty and off flavours in commercial plant-based meats, a 2024 study had to pair a trained tasting panel with gas chromatography–olfactometry, in which people sniff each compound as it leaves the instrument. It named compounds behind meaty notes, such as 2-methyl-3-furanthiol, and behind off notes, such as (E,E)-2,4-decadienal. A search found no published tool that predicts such panel results from instrument data alone.

ebi.ac.uk

Built with what exists

OpenPOM, an open-source reproduction of the 2023 Principal Odor Map, a model that predicts a molecule's smell from its structure (github.com/ARY2260/openpom); RDKit for chemical structures; published GC-MS and olfactometry tables; a large language model to pull compound tables out of papers.

  1. First step

    Copy the odourant table from the 2024 study into a spreadsheet with each compound's structure written as a SMILES string, and run OpenPOM on it.

  2. Done looks like

    From instrument data alone, the tool places at least 80% of the products in the 2024 study in the same meaty or off-flavour group as the panel did, and does as well on the products of at least one other published study. Code and data are public.

Open 02 · the brief

Protein scores for best-sellers

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public table of in vitro DIAAS for the 30 best-selling plant-based meats and milks in at least two countries, all measured with one method and listed with brand, batch and purchase date.

Why it is still open

Published measurements cover a few products at a time, such as five plant-based chicken products in 2026. A new Protein Quality Hub collects 7,775 published protein values by food type and method, so measured branded products would have a public home.

ebi.ac.ukebi.ac.uk

Built with what exists

The INFOGEST static in vitro digestion protocol; amino acid analysis by HPLC at a university or contract food laboratory; Open Food Facts for product and label data.

  1. First step

    Get quotes from two laboratories for INFOGEST digestion plus amino acid analysis of three products, and draw up the 30-product list from published retail sales rankings.

  2. Done looks like

    30 products, each measured in duplicate with a reference protein such as casein in every batch, duplicates within 5 DIAAS points of each other; results and methods published under an open licence and submitted to the Protein Quality Hub.