Inspired by biology

Protera technology

We combine protein and food science expertise with advanced AI technology to develop healthier, sustainable, and clean-label ingredients.

Clean-label ingredients by our Natural Intelligence™

From avocados to olive oil to that delicious tofu dish, food and its ingredients are produced naturally by biology.

We have tried to understand nature's production process since we developed our first agricultural methods. However, our good intentions to make enough food for the world's population has led to no longer sustainable practices.

We believe that by aligning protein understanding, AI technology, and consumer-driven demands, we can create a more sustainable food system.

natural intelligence

Proteins

Biology evolved proteins to enable life to thrive and work in harmony with the environment. At Protera, we're amazed at how proteins work as efficient nano-scale machines that transform our world. They can form structures such as skin or convert carbon dioxide into sugars and oxygen in plants.

Artificial Intelligence

In our search to understand proteins to their fullest, we have developed MADI™, a deep learning algorithm trained to learn how biology build proteins. MADI™ foundations relay on data generated by many generations of brilliant scientists over the past 70 years.

Fermentation

MADI™ draws inspiration from biology itself to carefully craft functional protein-based ingredients. We then produce these ingredients at scale using fermentation, a natural process used in making wine, beer, yogurt, and cheese.
Clean-label
We produce the proteins proposed by our Natural Intelligence™ using fermentation. Protera also promotes transparent labels for consumers like “yeast extract” or “non-animal” protein alternatives.
Functional Proteins
The proteins we craft focus on challenging ingredient components where no alternatives exist. Additives, for example, are key to decreasing food waste, yet the ones used today are unhealthy and produced with massive environmental impact.
Healthy, Sustainable, and Next-generation
We eat proteins every day, of many kinds, from plants, from fungus (yeast in bread) and from animals (not necessary!). Our bodies naturally process proteins and use them as energy to grow.

How MADI™ works

MADI™ is Protera’s Natural Intelligence engine. We trained it to understand complex molecular patterns connecting the amino acid sequence of a protein, its three-dimensional structure, and its function.

To learn these patterns, our team used billions of data points encoded and generated through a proprietary process.
Internally, in this training process, amino acid sequences and protein structures are represented in a high-dimensional space or "latent space". This new algorithm can make correlations that humans would find impossible to grasp. MADI™ then uses these correlations to predict protein folding and performance.

MADI™ takes advantage of GPU computing. This process has an unprecedented speed, allowing our model to analyze billions of proteins that exist in nature that we have not yet been able to characterize.

Leveraging The Power of GPU Computing

MADI™ represents a breakthrough in protein design taking advantage of GPU parallelization and Deep Learning.

Protera technology

Millisecond timeframe calculations

GPU
Thousands of Cores

Other algorithms

Several weeks of calculations

CPU
Multiple Cores