Biological proteins: the invisible cornerstone of hundreds of billions in biomanufacturing, now being rewritten by AI

When it comes to 'biological proteins,' many people's first thoughts might be the 'collagen' in skincare, 'whey protein' in the gym, or 'plant protein' on the dining table. But in the biotech and biopharmaceutical industries, the weight of the term 'biological protein' goes far beyond that.

It’s the 'active ingredient' in antibody drugs, the 'industrial engine' in enzyme catalysis, the 'core antigen' in vaccine development, and a key functional additive in cell culture—from basic research to clinical drugs, from the food industry to environmental management, biological proteins are the invisible 'heroes' underpinning almost all biotech sectors.

This article will systematically break down the definition of industrial biological proteins, their commercialization pathways, traditional R&D bottlenecks, and how AI-driven R&D platforms are reshaping industry research paradigms.

1. Industrial Perspective: What Are Biological Proteins?

In an industrial context, biological proteins refer to protein substances with clear functional activity that can be used as production raw materials or end products. Among these, recombinant proteins produced via genetic engineering in host cells are currently the most widely applied and the most technically challenging category. Unlike nutritional proteins for consumption, their value comes from their functional attributes rather than their nutritional content.

Currently, commercially available biological proteins are mainly divided into three categories:

1. Medical functional proteins: Including antibody proteins, cytokines, lactoferrin, recombinant collagen, recombinant epidermal growth factor, and other repair proteins. They are widely used in innovative drug development, in vitro diagnostics, skin repair, and precision medicine, making them core raw materials in the biopharmaceutical industry.

2. Industrial enzyme proteins: Acting as core catalysts in industrial biomanufacturing, they are used in food processing, textile washing, biodegradation, bioenergy, and more. They replace traditional chemical catalysts, significantly reducing industrial energy consumption and pollution.

3. New alternative proteins: Centered on microbial proteins, mycelium proteins, and microalgae proteins, they offer high nutrition, low land use, and short production cycles, making them excellent alternatives to traditional animal and plant proteins in food and feed.

2. Three Core Tracks: How the Commercial Value of Biological Proteins Is Realized

  Three major commercialisation sectors

With the maturation of biotechnology, the industrialization of biological proteins is accelerating, and the market size of various sub-sectors is steadily growing, forming three high-potential core application scenarios that support a biological protein industry market worth hundreds of billions of yuan.

2.1. Biomedicine: The Core Foundation of Precision Medicine

Medical biological proteins are the backbone of modern biomedicine and one of the fastest-growing segments of the global pharmaceutical industry. Leveraging proteomics technology, researchers and industry players can accurately explore disease mechanisms, screen drug targets, and develop biomarkers, providing essential support for innovative drug development, disease screening, and prognosis monitoring.

Currently, antibody drugs, recombinant protein drugs, and recombinant repair proteins are widely used in cancer treatment, immune modulation, skin barrier repair, and anti-infection therapies. These proteins are highly targeted and have low side effects, making them core biological materials that traditional chemical drugs cannot replace.

2.2. Industrial Biomanufacturing: The Core of Green Industry Upgrades

Industrial enzyme proteins are the most widely used category of biological proteins in the industrial field and a key driver of traditional industry's green transformation. According to the Biocatalysis Industry White Paper, the global industrial enzyme market is expected to reach $8-9 billion by 2026, with stable long-term growth.

Compared to chemical catalysis processes, enzyme protein-catalyzed reactions operate under mild conditions, have high conversion efficiency, and generate no secondary pollution. They are widely applied in food fermentation, grain and oil processing, textile treatment, detergents, and bioenergy production. Emerging areas like plastic biodegradation also show huge application potential, effectively addressing traditional industry's high energy consumption, pollution, and waste, and helping manufacturing achieve low-carbon upgrades.

2.3. Food and Wellness: A New Growth Point for Consumer Needs

Driven by consumption upgrades and health demands, new bio-based substitute proteins are emerging rapidly. New protein sources like microbial protein, mycelium protein, and microalgae protein can be mass-produced through microbial fermentation without relying on land cultivation or livestock, while being rich in high-quality amino acids and dietary fiber and effectively alleviating global protein resource shortages.

Currently, these biological proteins are gradually being used in alternative meat, functional health products, and high-end feed. Thanks to their safety, health, and sustainability advantages, they have become a new growth driver in the wellness industry.

3. Industry Pain Points: The Core Bottleneck of Traditional Biological Protein R&D

 The Three Major Bottlenecks in Traditional Research and Development

Biological proteins may seem like just a 'raw material,' but their research and production barriers are extremely high. For a functional protein, going from sequence design to industrial-scale production requires crossing three major industry hurdles, which is also the core reason why high-end biological proteins have long depended on overseas supply:

First hurdle: Functional design — the protein you want may not exist in nature

Natural proteins are products of evolution and often are only 'good enough,' let alone 'user-friendly.' For example:

Natural Protein A isn't alkali-resistant and needs the column replaced after a few uses;

Wild-type enzymes often can't meet the stability or substrate adaptability requirements for industrial production, making direct large-scale use difficult;

Natural antibodies' affinity and stability don't necessarily meet pharmaceutical needs.

To make a protein 'user-friendly,' it has to be engineered — site-directed mutagenesis, directed evolution, domain rearrangement... Traditional methods rely on 'expert experience + high-throughput screening,' which is time-consuming, costly, and has low success rates. A mutant with several times better performance often requires screening tens of thousands or even hundreds of thousands of clones, taking months or years.

Second hurdle: Expression and preparation — just because you designed it doesn’t mean you can make it

Even with a good sequence, you still need the host cell to 'cooperate' and express the protein. Different proteins have different quirks:

Some tend to form inclusion bodies — the expressed protein misfolds into insoluble aggregates, requiring complex refolding processes to obtain active protein, or re-optimizing expression conditions for soluble expression; some have complex post-translational modifications like glycosylation or disulfide bond formation, usually needing eukaryotic expression systems like yeast or mammalian cells, which increases cost and time; some are toxic to the host cells, killing them before they can grow; and some require extremely high purity, as pharmaceutical-grade proteins often need over 99% purity, with each extra purification step adding cost and loss.

Which expression system should be used? How to optimize culture conditions? How to design the purification process? Every step directly affects final yield, activity, and cost.

Third hurdle: Scaling up — what works in the lab may not work in the factory

Being able to produce milligrams doesn’t mean kilograms can be stably produced. During scale-up, common problems include decreased expression, loss of activity, increased aggregates, or endotoxin overages. Each scale-up, from shake flasks to tens of liters to thousands of liters for industrial production, tests engineering capabilities.

This is also why high-end biological proteins have long been dominated by a few international giants. High technical barriers, long R&D cycles, and large capital investment make it difficult for latecomer companies to catch up.

4. AI is Rewriting the Game of Biological Protein Development

If traditional biological protein development is like 'crossing a river by feeling for stones,' AI is turning it into 'building according to blueprints.'

In recent years, structural prediction models represented by AlphaFold have, for the first time, allowed researchers to obtain high-precision 3D structures of most proteins without experiments, reducing the cost of acquiring protein structures by several orders of magnitude. But 'being able to see it' doesn’t equal 'being able to make it.' The real industrial challenge is—given a functional requirement, can you directly design a protein sequence that meets it?

This is exactly the problem that the new generation of AI protein design platforms aims to solve. In this wave of technology, domestic companies have already taken the lead in turning AI protein design from papers into products. Take the MatwingsVenus™ (Xiaowu™) platform from Shanghai Matwings Technology, for example. It doesn’t just predict structures; it directly designs and modifies proteins from scratch based on functional requirements:

If you want a 'more alkali-resistant Protein A ligand,' AI can directly provide an optimized sequence; if you want an 'industrial enzyme with higher catalytic efficiency,' AI can quickly scan key sites and suggest mutation combinations; if you want an 'antibody with improved stability,' AI can enhance thermal stability and expression while maintaining affinity.

This direct mapping from 'requirement' to 'sequence' breaks the ceiling of traditional 'trial-and-error R&D.' Its core advantages are reflected in three areas:

1. Faster speed. Traditional rational design combined with directed evolution often takes months to years, while AI-assisted design can reduce the cycle to weeks or even days, improving development efficiency by several to tens of times. In drug development, where 'time is life,' the value of speed is self-evident.

2. Higher success rate. AI, based on massive protein data and physicochemical prior knowledge, can accurately identify 'valuable mutations' in the vast sequence space, avoiding resource waste from blind screening. Mutations that once required screening hundreds of thousands of clones to find can now be hit by designing just a few dozen candidates.

3. Broader exploration boundaries. Human experts’ experience is limited to known structures and mechanisms, whereas AI can discover some 'counterintuitive' but effective designs, expanding the possible boundaries of protein engineering. Some mutation combinations might seem 'unreasonable' from a traditional perspective, but experiments often show better results—this is exactly what makes AI design so exciting.

More importantly, AI design cannot stay confined to computers. The real value lies in the 'dry-lab to wet-lab' closed loop—AI designs candidate sequences, automated experimental platforms quickly verify them, and the data is fed back to iteratively optimize the design. The MatwingsVenus™ (Xiaowu™) platform integrates a protein label database with billions of entries, over 200 specialized protein design tools, and more than 30 domain-expert-tuned skills, building a complete 'design—verify—iterate' chain.

5. Closing Thoughts: The 'China Moment' of Biological Proteins

For a long time, the high-end biological protein market has been dominated by a few international brands, with domestic companies mostly in the role of 'followers.' But the situation is changing.

On one hand, the explosive growth of downstream industries in China—biopharmaceuticals, in vitro diagnostics, synthetic biology, and more—is providing upstream biological protein companies with a huge local market demand. Data shows that the recombinant protein market in China is expected to reach about $6.8 billion by 2025 and could hit $13.5 billion by 2030, making it the fastest-growing single-country market in the world. At the same time, the self-sufficiency rate of domestic enterprises in the local market has risen from 35% in 2020 to 52% in 2025, and domestic replacement is moving from being just a 'slogan' to real 'performance.'

On the other hand, the wave of AI-driven protein design gives later entrants a chance to 'overtake in another lane'—as the research paradigm shifts from 'experience-driven' to 'data-driven,' the barriers of first-mover advantage are being redefined.

More importantly, biological proteins are not just a 'business'; they are a key part of maintaining autonomy and control in the biopharmaceutical industry chain. From Protein A resins used in antibody drug production, to recombinant antigens for vaccine development, to various enzymes used in molecular biology reagents—every key protein that is domestically produced is strengthening the foundational base of China's biotech industry.

When AI meets biological proteins, when design capabilities meet manufacturing capabilities, and when local demand meets global competition—a new era of protein supply that is more efficient, more precise, and more sustainable is accelerating toward us.