Bioactive peptides for nutrition

01Overview and value chain

Markers: [EC: Food Supplements Directive 2002/46/EC | OECD: Food Systems | Regulator: FDA (USA), EFSA (EU), NMPA (China)]

Bioactive peptides are short chains of 2–20 amino acids, locked inside large plant or animal protein globulins in their native state but released as potent cell-signalling molecules once freed. Unlike bulk dietary protein that serves mainly as a plastic amino-acid source, these peptides act on receptors in the gut, immune system, vasculature and muscle, triggering specific metabolic cascades. The 2026 industry is defined by an AI-discovery revolution: generative platforms such as Nuritas Magnifier and Brightseed Forager scan trillions of candidate sequences in plant genomes, predicting bioactivity, gastrointestinal stability and target binding, and compressing ingredient discovery from a 5–7 year wet-lab campaign to a matter of months. The flagship plant peptide complex PeptiStrong, isolated by AI from faba bean (Vicia faba), down-regulates the muscle-atrophy genes Atrogin and MURF1 and has cleared FDA GRAS for use in foods, drinks and supplements.

The key directions of bioactive peptides for nutrition are:

1. AI-peptide discovery (AI-Peptide Discovery): deep-learning screening of plant genomes for latent bioactive motifs, cutting lead time from 5–7 years to months.
2. Targeted enzymatic hydrolysis (Targeted Enzymatic Hydrolysis): controlled proteolysis with bespoke protease cascades that cut the protein only at AI-specified sites, yielding a defined peptide pattern.
3. Membrane downstream fractionation (Membrane Downstream Fractionation): tangential-flow ultrafiltration with a sharp molecular-weight cut-off (<3 kDa) that isolates ultra-short signalling peptides from larger allergenic fragments.
4. Functional nutrition formats (Functional Nutrition Formats): GRAS-grade peptide ingredients formulated into sports recovery, sarcopenia/longevity, infant and enteral-clinical products.

Sectoral value chain

[AI Proteolysis Design] ──> [Protein Extraction] ──> [Targeted Hydrolysis]
           │                          │                         │
     (Magnifier / Forager)     (faba bean, rice,          (AI-specified proteases)
                               soy, oat)                          │
                                                                  ▼
[B2B Supplements / Nutrition] <── [Drying & Packing] <── [Membrane Downstream]

Value chain levels

Cross-cutting technologies of the sector:

• Deep-learning QSAR for peptide activity (Deep-Learning QSAR): neural networks trained on biological data to predict gastrointestinal stability and bioavailability.
• Molecular-weight-controlled hydrolysis (MW-Controlled Hydrolysis): precision arrest of proteolysis (thermal inactivation, rapid pH shift) to prevent over-degradation to free amino acids.
• Tangential-flow ultrafiltration (Tangential-Flow Ultrafiltration): industrial membrane cascades separating undigested protein and allergenic peptides from ultra-short signalling peptides.

02US

The US leads commercial deployment of AI-discovered peptides, building a large market in functional sports nutrition, geroprotective formulas against sarcopenia, and personalized nutraceuticals, anchored by a Silicon-Valley venture hub at the intersection of generative AI and nutrition.

GRAS clearance, sports & longevity nutrition, AI-peptide venture hub

• FDA GRAS notifications: Nuritas PeptiStrong cleared the FDA GRAS process, removing restrictions on its use in foods, beverages and supplements across the United States.
• Sports nutrition & longevity: PeptiStrong became a premium recovery ingredient and a sarcopenia-prevention nutraceutical for ageing adults, while Brightseed scaled Forager-discovered phytonutrients with healthy-food brand partners.
• AI-peptide venture hub: the majority of funds and accelerators financing generative-AI-meets-nutrition start-ups are concentrated in the US, making it the sector’s main technology driver.

03CN

China is scaling bioactive peptides as a key tool against an ageing demographic profile and to raise the value-added processing of its agricultural crops, with a focus on traditional Asian feedstock and large-volume clinical and infant nutrition.

traditional feedstock peptides, infant & clinical nutrition, MARA standardization

• Peptides from traditional feedstock: research centres, including the Chinese Academy of Agricultural Sciences, derive bioactive peptides from soy, rice, wheat, ginseng and marine organisms, with domestic AI screening platforms for medicinal-herb phytochemistry.
• Infant & clinical nutrition: deep-hydrolysis peptides are widely built into hypoallergenic infant formula and enteral clinical-nutrition products for bedridden and post-surgical patients.
• Standardization & state regulation: MARA and the National Health Commission are developing national standards for functional peptides to order a fast-growing market and protect consumers from adulteration.

04EU

The European Union prioritizes scientific evidence, safety and environmental sustainability of peptide ingredients, acting as the principal R&D base for fundamental peptidomics research and AI-discovery start-ups.

R&D base & peptidomics, EFSA health-claim control, Green Deal upcycling

• R&D cradle: Nuritas was founded by Nora Khaldi in Dublin with European grant support and the Irish National Institute for Food Research; Ireland and Denmark remain the key fundamental peptidomics centres, alongside University College Dublin research on plant hydrolysates and metabolic-syndrome markers.
• Strict EFSA control: health claims for peptides undergo the most rigorous EFSA verification, requiring randomized placebo-controlled clinical evidence that screens out pseudo-scientific supplements.
• Green Deal agenda: peptide production is framed as part of the transition to sustainable plant feedstock, with enzymatic upcycling of food-industry side streams from oat, pea and lupin subsidized under Horizon Europe.

05Leading companies and research institutes

06Tech stack and innovations

The AI-peptide stack fuses generative bioinformatics with precision bioprocess engineering, where a 1 °C temperature or 0.1 pH shift at the hydrolysis step can change the protease cleavage site and destroy a batch’s bioactivity.

1. AI platforms and predictive modelling (In-Silico Prediction):
   • deep convolutional networks analyze plant protein structure to flag fragments that, once released, interact with cell receptors (e.g. down-regulating NF-kB-driven inflammation).
   • digestion simulation filters for resistance to pepsin, trypsin and gut peptidases, keeping only motifs that reach the intestine intact and absorbable.
2. Precision enzymatic hydrolysis (Precision Enzymatic Hydrolysis):
   • given the AI cleavage map, technologists deploy commercial proteases (e.g. alcalase, papain, thermolysin) in reactors with tight parameter hold, because a 1 °C or 0.1 pH deviation shifts the cut site.
   • the reaction is arrested by rapid heating to 90 °C for 10 min, irreversibly denaturing the proteases before peptides degrade to inactive free amino acids.

07Value chains and production pipelines

Industrial pipeline of directed hydrolysis of bioactive peptides (ISO 22000)

┌───────────────────────────┐      ┌───────────────────────────┐
│ 1. AI modelling and       │ ───> │ 2. Extraction and         │
│    protease selection     │      │    isolate dissolution    │
└───────────────────────────┘      └───────────────────────────┘
                                                  │
                                                  ▼
┌───────────────────────────┐      ┌───────────────────────────┐
│ 4. Thermal inactivation   │ <─── │ 3. Enzymatic hydrolysis   │
│    (reaction arrest)      │      │    (pH, temperature)      │
└───────────────────────────┘      └───────────────────────────┘
              │
              ▼
┌───────────────────────────┐      ┌───────────────────────────┐
│ 5. Ultrafiltration        │ ───> │ 6. Spray-drying and       │
│    (>3 kDa cut-off)       │      │    bioactivity grading    │
└───────────────────────────┘      └───────────────────────────┘

Stage 1: AI design and protease selection

The Magnifier platform models the cleavage of faba bean protein and selects a specific cascade of two commercial bacterial proteases that excise the PeptiStrong peptide pattern with minimal formation of free amino acids.

Stage 2: Isolate preparation and dissolution

Plant faba-bean isolate with >85% protein is dissolved in deionized water to 10% dry solids, heated to 55 °C and set to pH 8.0 with sodium-hydroxide dosing for the first protease.

Stage 3: Directed controlled proteolysis

The first protease is dosed in and hydrolysis runs for 4 hours under continuous mixing, with a pH-stat auto-dosing alkali to neutralize freed carboxyl groups; a second protease then finishes cutting the peptide chains.

Stage 4: Thermal inactivation of enzymes

To stop the peptides degrading to inactive free amino acids, the mass is flash-heated to 90 °C and held for 10 minutes, irreversibly denaturing the proteases.

Stage 5: Microfiltration and cascaded ultrafiltration

The hydrolysate is cooled to 45 °C, centrifuged to remove insoluble fibre, then pumped through a tangential-flow cascade: a 10 kDa membrane removes undigested protein and denatured enzyme, and a 3 kDa membrane removes mid-size peptides, leaving a permeate of ultra-short signalling peptides under 3 kDa.

Stage 6: Concentration, drying and packing

The peptide solution is vacuum-concentrated to 35% solids and spray-dried (175 °C inlet, 80 °C outlet); the light-cream hygroscopic powder at <4% moisture is packed in nitrogen-flushed foil pouches for B2B delivery to sports-nutrition and supplement manufacturers.