First, we compile thousands of enzyme sequences (from both public and private databases) with probabilities to make the reaction of interest. Then, we generate the structural models using crystallographic data. If not available, we generate the 3D structures with Alphafold2 or homology modeling.

First, we compile thousands of enzyme sequences (from both public and private databases) with probabilities to make the reaction of interest.  Then, we generate the structural models using crystallographic data. If not available, we generate the 3D structures with Alphafold2 or homology modeling.

Back

Here begins the structural filtering process. At this stage, it’s crucial to identify the catalytic site and to properly prepare the elements involved in the catalysis (i.e. cofactors, ions). Highest quality structural models are preferred.

Here begins the structural filtering process. At this stage, it’s crucial to identify the catalytic site and to properly prepare the elements involved in the catalysis (i.e. cofactors, ions). Highest quality structural models are preferred.

Here begins the structural filtering process. At this stage, it’s crucial to identify the catalytic site and to properly prepare the elements involved in the catalysis (i.e. cofactors, ions). Highest quality structural models are preferred.

Then we conduct the molecular filtering: each structure is modeled with the ligand, performing physics-based simulations to review the chances of success and filter the most promising candidates.

Then we conduct the molecular filtering: each structure is modeled with the ligand, performing physics-based simulations to review the chances of success and filter the most promising candidates.

Then we conduct the molecular filtering: each structure is modeled with the ligand, performing physics-based simulations to review the chances of success and filter the most promising candidates.

Finally, we test for different parameters: e.g. “can the reactant enter the active site?” or “does the enzyme have a complementary shape to the reactants?”. We select the candidates with the highest chances to work out in lab testing.

Finally, we test for different parameters: e.g. “can the reactant enter the active site?” or “does the enzyme have a complementary shape to the reactants?”. We select the candidates with the highest chances to work out in lab testing.

In the end, less than 20 enzymes are selected. These are the ones that are sent to the lab for testing and finding out the “best” enzyme for the reaction of interest. Plus, the selected enzymes can be further improved with ZYMEVOLVER technology!