Enabling the bioeconomy revolution with
‍the fastest enzymes

We crack the fundamental bottlenecks of enzyme engineering with unprecedented screening and data collection capacity

See the breakthroughs Let us solve your enzyme problems

Limited-time offer
Money-back guarantee

Come back early November to check out our beta testing program—limited-time offer with money-back guarantee.

Three disruptive technologies

EvolutionaryDE

UnleashedDrop

MillionFull

7 orders of magnitude greater screening capacity

- The “numbers“ advantage (10⁷ x)

4 orders of magnitude greater data collection capacity

- The data advantage (10⁴ x)

- compared with mainstream enzyme engineering platforms.

Enzyme is the core hardware of biomanufacturing

Enzyme efficiency remains a foundational bottleneck for market-competitive biosolutions

The green transition of the world economy needs efficient biosolutions.

In labs around the world, there are now numerous proof-of-concept biosolutions, going from 0 to 1, offering endless promises for a sustainable economy.

Looking at the market, there are still too few biosolutions being economically competitive against fossil fuel-based ways, the “1 to 100”.

The green transition of the world economy needs efficient biosolutions.

A bioeconomy revolution needs an enzyme engineering revolution

In essence, synthetic biopathways are linked enzymatic steps catalyzing conversions one step after another, eventually from feedstocks to products.

It’s impossible to have a fast bioprocess when the enzymes are slow. It’s hard to have a slow bioprocess when the enzymes are fast.

A bioeconomy revolution needs an enzyme engineering revolution

the dawn of AN enzyme engineering revolution

The ENZIDIA^® Evolutionary Wheel

Cracking the fundamental bottlenecks of enzyme engineering

EvolutionaryDE x UnleashedDrop x MillionFull

Enzyme engineering is fundamentally constrained by two bottlenecks: screening throughput (the numbers game), and data (the information game).

1. The numbers game

The more you screen, the more likely you can find a great enzyme.

Combining EvolutionaryDE & UnleashedDrop (proprietary), the ENZIDIA platform has a library creation and screening capacity of 10⁹ - 10¹¹ variants per run. This is 3 orders of magnitude greater than state of the art, and 7 orders of magnitude greater than robotics-based platforms.

2. The information game

To have a great river of information flowing toward better sequences, we need more than just a river channel (AI). We need lots of water (data).

Our technology MillionFull (patent-pending) allows the collection of 10⁵ -10⁷ data points per run, an unprecedented throughput. This data capacity gives our use of AI a unique advantage.

The numbers game and the information game graphic image.

The ENZIDIA Evolutionary Wheel - an "AI-second" platform:

1. Build and screen a massive library (~10¹⁰ per run)
2. Collect big data (~10⁶ per run)
3. Learn from the big data through AI
4. Design the next massive & intelligent library, guided by AI + big data.
Loop to step 1.

This loop multiplies the “numbers“ advantage (10⁷x) and “data“ advantage (10⁴ x), poised to revolutionise the way enzymes are engineered.

The ENZIDIA platform has been breaking enzyme bottlenecks on a roll

Two enzymes, from two distinct classes, have been engineered by beta versions of the ENZIDIA platform. Large improvements were obtained in both cases, in just one to two rounds of engineering. Iterations can be easily done for further improvements.

Great surprises await us as the full platform is deployed for more enzymes.

Dataset with approximately 100000 data points.

A dataset with ~100,000 data points
‍(also plotted are ~400k sequences with premature stop codons, which clustered around 0)

Case 1. A rate-limiting enzyme in the bioproduction of a food/feed supplement: 94% performance improvement after one round of engineering.
The benchmark sequence is commonly used in patents and papers.
Enzyme performance is defined as product formation in a given timeframe.

Case 2. A rate-limiting enzyme for the synthesis of a high-value flavor compound: 140% performance improvement in two rounds.

During Case 2: >100,000 data points of enzyme sequence-fitness relationship in a demo run of MillionFull (see graph).
The data revealed >10k unique sequences with significantly improved performance, and ~100k positive & negative mutations/epistatic interactions.

A screening capacity of >10⁹ has been reached in an ongoing project for our third enzyme engineering target.

Should our next project be the super-charged engineering of an enzyme/pathway in your bioprocess?

Employ the ENZIDIA platform Find answers for FAQ

whERE DOES enzidia STand?

A new paradigm of enzyme engineering

The ENZIDIA platform is >3 orders-of-magnitude more powerful than state-of-the-art

Platform category

Selling points

Limitations

AI-centric platforms

Limited differentiation with open-source algorithms

No solution for (labeled) data

Metagenomic data-based platforms

Natural sequence data + annotation for enzyme discovery

No labeled data

No experimental optimisation

Automation-based platforms

Automation for high-throughput screening (10^{^2} - 10^{^4})

High cost: $20 - $50 / variant

Limited screening power & data

Microfluidics-based platforms

Microfluidics-based high-throughput screening (10^{^6} - 10^{^7})

Limited compatibility with big data extraction.

Limited applicability

Unprecedented screening power: 10^{^9} - 10^{^11}

Unprecedented data generation: 10^{^5} - 10^{^7}

Wide applicability

Low cost

New in town

Selling point:

Unprecedented screening power: 10⁹ - 10¹¹

Unprecedented data generation: 10⁵ - 10⁷

Wide applicability

Limitations:

New in town

Microfluidics-based platforms

Selling point:

Microfluidics-based high-throughput screening (10⁶ - 10⁷)

Limitations:

Limited generalizability

Limited compatibility with big data extraction

Automation-based platforms

Selling point:

Automation for high-throughput screening (10² - 10⁴)

Limitations:

High cost

Limited screening throughput & data generation

AI-centric platforms

Selling point:

Limitations:

Limited differentiation with open-source algorithms

No solution for the need of big (labeled) data

Metagenomic data-based platforms

Selling point:

Natural sequence data + annotation for enzyme discovery

Limitations:

No labeled data

No experimental optimisation

The faces of enzidia

The Enzidia Team

A group of relentless problem solvers and disciplinary boundary breakers

Jinbei Li, Ph.D.

CEO & CSO

Platform designer, big picture thinker, broad learner

Simon Krarup

Senior Bioengineer

"A most dedicated researcher who is too humble to call himself so" - anonymous

Bjarke Erichsen

AI Engineer

Machine learning engineer
and self-taught bioinformatician

Prof. Alex Toftgaard Nielsen

Founding Advisor

Co-founder of Cysbio, Again, Mycropt
Professor at DTU Biosustain

Lene Lange, Ph.D.

Scientific Advisor

Former director of Molecular Biotechnology at Novozymes (among many things)

Steffen Klamt, Ph.D.

Scientific Advisor

Group leader at Max Planck Institute for Dynamics of Complex Technical Systems

Prof. Paul Jensen

Scientific Advisor

Pioneer in integrating deep reinforcement learning and automated science (U. Michigan)

Eg Nicolajsen, MBA

Business Advisor

Former COO in Corporates, 2x founder, experienced mentor to 250+ startups

Morten Birkeland

IP Advisor

MSc, M.Law, EPA, D.E.S.U, CEO of IPTector, serial IP entrepreneur. Former Novo Nordisk, Novozymes, Dupont

WORK WITH enzidia

Let Enzidia accelerate your enzymes & cell factories

Accelerated enzyme engineering, for augmented synthetic biology

Model 1 — Enzyme & cell factory optimisation service partnership

Fast feasibility evaluation for problem-platform fit (hours to days). Most enzymes in industrial bioprocesses and biosynthetic pathways can fit the ENZIDIA platform.

Small upfront fee to cover part of the project execution and establish mutual commitment.

5-12 months for most projects to complete. No more fee is asked regardless of if we do more work. ENZIDIA will aim for the best possible result that can be delivered.

Success-contingent payment based on the improvement achieved by ENZIDIA’s results. The metric is pre-determined based on relevance for the success of the overall bioprocess.

In addition: no restriction on the partner to disclose the success or failure of partnership projects. This is not a common practice. While we may not remain successful in 100% of the cases we work on (so far we are), we believe that having such transparency is essential to establish ENZIDIA as the best and most trusted choice in the industry.

Model 2 — Bioprocess co-development partnership

In co-development partnerships, Enzidia provides continuous and exclusive enzyme and strain improvement based on co-ownership. This is most suitable for a process that is per-commercialization at your company. The project can be initiated by you, but we do have a long list of candidate chemicals to choose from.

Model 3 — Enzidia high-impact process development

Enzidia is founded for the ambition to realize our collective green transition. Not all projects have the same potential impact size. We are/will be incubating several long-term, high-impact projects internally.

The list includes: 1) Pathways for converting electrochemical feedstocks (formate, methanol, acetate) to chemicals or fuels, such as for long-duration energy storage. 2) Chitosan production from chitin, 3) 3-HP, 4) Isoprene, 5) PFAS degrading enzymes, 6) Plastic (non-PET)-hydrolysing enzymes.
‍
We will not take on service requests for these projects, but welcome partnerships for their financing (e.g., joint ventures), scale-up, & production.

Our uniqueness doesn't come from business models. No need to fit into the models above. - We fit them to your needs.
If you are looking to make your bioprocess the most competitive on the market, get in touch today.

Partner with Enzidia Answers to FAQ

FOR THE INTRIGUED

FAQ

Would our enzyme bottleneck fit the ENZIDIA platform?

Most likely yes.
‍
Most enzymes used in industrial biotech can fit the ENZIDIA platform. The current version of the platform is especially suited for enzymes involved in cell factories.
‍
We will let you know whether a problem can fit our platform, and if so, the difficulty level on a scale of 1-10. Just book a call or write a message with the buttons right below this section.

What conditions is the ENZIDIA platform most suited for?

Aqueous conditions, with close to physiological pH. We can suit a wide temperature range from 30 C to 80 C.

What aspects of enzyme performance is the ENZIDIA platform best suited for?

1. Higher Kcat

2. Lower Km

3. Higher stability / thermotolerance

The former two (Kcat and Km) are known to be among the hardest aspects in enzyme engineering, given the complexity of these parameters which makes rational design immensely challenging, compared with selectivity, pH, and stability. - That's where the ENZIDIA platform shines given its unique data output and screening throughput advantage.

What aspects of enzyme performance would ENZIDIA have a hard time for?

1. Product/substrate selectivity, and product inhibition. These aspects will be evaluated on a case-by-case basis.

2. Enzymes for the degradation of large polymers. We are developing our platform for such cases, but we will not be eager to take on polymer degradation enzymes before this aspect of the platform is more mature.

3. Performance of enzymes in non-aqueous environments. Our screening system is an aqueous environment. The screening results don’t directly translate to different environments, though they can still serve as promising leads, just as all industrial enzymes are sourced from nature.

What info do we need to send to you?

Essential: the reaction that needs to be catalysed.

Great to have: essential info about the enzyme's necessary operation environment (pH, temperature, solvent).

Good to have if available: the current version of the enzyme; if substrate/product inhibition is present.

Would a ceiling, or local optimum, be hit in an enzyme engineering campaign? Can this be alleviated?

Optimization ceiling can more likely be hit when deriving variants from one template (specifically, negative epistasis between an existing residue and new mutations).

When more diversity is included in the library, such as building libraries through diverse natural homologs, especially aided with the massive data that we can collect using MillionFull, a great diversity can be included in the library so that there are many paths to avoid a local optimum.

Here is one theoretical reference (Fig.3) on this topic.

You can screen so many. Is it necessary?

We can never screen too many, based on empirical evidence and theoretical analysis.

a) Most beneficial mutations can not be predicted rationally or computationally.

In our massive data collection demo project, the amazing finding is that, while the most damaging mutations happened around the active site, as one would expect, the most beneficial mutations were spread across the entire sequence.

In extensive literature where the enzyme engineering efforts weren't limited to the neighborhood of the active site, this observation was consistent (one exception is when the focus is the catalytic selectivity rather than catalytic rate).

b) With extensive enzyme engineering, the trend is clear: greater screening efforts lead to the identification of sequences with greater improvements, and these improved sequences have more mutations (>20).

See Fig. 2 in this paper to get a meta view.

c) Even considering just a small number of random mutations, the size of the possibility space already exceeds any existing screening capacity. Yes, including Enzidia's, not to mention other platforms:

(for a 400-AA enzyme)

1 mutation: ~8*10^{^3} possibilities

2 mutations: >3*10^{^7}‍

3 mutations: >8*10^{^10}‍

4 mutations: 1.6*10^{^14}

...

d) Most importantly, the fundamental problem: the system of an enzyme has all the features of a "complex system".

Present in complex systems are different components interacting with each other in a nonlinear way. Small perturbations (e.g., one mutation of an enzyme) can lead to unpredictable changes in the whole system. This is why enzyme behaviours are so unpredictable, even when we are getting better with static structure predictions. This is also why there are so many amazing wonders, or "emergent behaviours", from the complex systems of enzymes.

Can’t find the answer to your question?

Ask us directly

We are here for you

Connect directly with our team today

Meet

Explore problem-platform fit

Book a call

Get the collaboration journey started!

Jinbei@enzidia.com

what we are guided by

Enzidia’s Core Principles: Truth, Mission, Learning

Truth is the ground we operate on, for everything, including how we set our mission and design our path. Truth, not preconceptions or ego.

Mission is what guides us. Enzidia was created and built as a tool, for the mission of the global sustainability transition.

Learning is the path. Path for how we find truth, and how we achieve our mission. We learn from everyone. We follow no one (including our past selves). We decide based on facts and first principles.

Initial team of Enzidia discussing in a meeting

funding SOURCES

Enabling the bioeconomy revolution with ‍the fastest enzymes

Enzyme is the core hardware of biomanufacturing

The green transition of the world economy needs efficient biosolutions.

A bioeconomy revolution needs an enzyme engineering revolution

The ENZIDIA® Evolutionary Wheel

EvolutionaryDE x UnleashedDrop x MillionFull

1. The numbers game

2. The information game

The ENZIDIA platform has been breaking enzyme bottlenecks on a roll

Should our next project be the super-charged engineering of an enzyme/pathway in your bioprocess?

A new paradigm of enzyme engineering

The Enzidia Team

Let Enzidia accelerate your enzymes & cell factories

FAQ

We are here for you

Enzidia’s Core Principles: Truth, Mission, Learning

Enzidia’s journey is catalyzed by

Enabling the bioeconomy revolution with
‍the fastest enzymes

The ENZIDIA^® Evolutionary Wheel