Case Study — Machine Learning-Guided Enzyme Engineering: Improving CO2-Fixation rate of Glycolyl-CoA Carboxylase

Guiding Leading Synthetic Biologists 2.8x Beyond Nobel Prize Winning Method

The Challenge

Maximizing enzyme optimization, minimizing resource use

The research group has spent years attempting to improve the speed at which this complex enzyme can catalyse CO2. They used rational design and directed evolution, a method that earned Frances Arnold the Nobel prize just four years earlier. These industry-standard approaches can take years of hands-on research and trial and error to deliver results. From the 8,000 variants, the group tested in their final round of directed evolution, only 20% turned out to be functional proteins. Of those, just one variant displayed enhanced properties.

Our client explains,

“After several rounds of directed evolution we gave up on further improvement.”
That’s when they brought Exazyme on board. Our goal was to help them understand and improve the complex enzyme.

The Solution

Discovering a 2.8x more effective variant in under 5 minutes

minutes to identify protein variants
wet lab tests
catalytic rate

Our AI-based algorithm created a list of novel protein variants to take to the wet lab for the client’s research team. The calculation took less than five minutes but replaced weeks of manual modeling.

The team experimentally tested the 10 variants our AI-tool produced. 90% of those tested were active. Two of the 10 variants tested displayed enhanced properties:

  • One showed a 2.8-fold increase in catalytic rate
  • Another consumed less ATP, making it 50% more energy-efficient

Before Exazyme
Rational design and directed evolution
Wet lab testing of 15,000+ variants
~20% active variants
Enhanced properties in ~1 in 8,000
After Exazyme
Algorithmic design and wet lab validation
Wet lab testing of 10 variants
90% active variants
Enhanced properties in 1 in 5

The Results

“We were amazed by the results”

Before partnering with us, the client had nearly given up on further optimizing the carboxylase. With Exazyme, they achieved a rapid and major breakthrough. The 2.8x improvement in protein catalysis speed surpassed even the best results achieved during years of using traditional methods.

The head of this research group says,

“We were amazed by the results. We were surprised to see such a strong performance jump with the variant Exazyme suggested."
Ingmar Schuster, Exazyme CEO agrees. “We are extremely proud of this breakthrough and the impact it will have on the field of protein engineering. Our software is a game changer for researchers, making it easy to achieve better results with fewer experiments.”

The Future

Pushing the boundaries of protein engineering

Today, our partnership with this client continues on their upcoming research projects. As the team continues to improve the carboxylase catalysis rate and more, we offer customized tools that make designing proteins as easy as using an app.

Ingmar says, “We look forward to continuing to push the boundaries of what's possible with AI for protein engineering." We can’t wait to see what we’ll discover.

Contact our team to see how we can work together.

This field is for validation purposes and should be left unchanged.