Enzymes are important for life. They act as catalysts, break down food so our bodies can use it, and help make strong medicines to fight diseases. Scientists want to make new enzymes that can do useful things.

Now, making new enzymes is much easier. A group of researchers, including bioengineers and synthetic biologists, have created a new way to design enzymes on a computer. They can predict how these enzymes will work and test them in many chemical reactions without ever leaving the computer. Their work is shared in a paper published in Nature Communications.

This new method is faster because it doesn't need living cells to make enzymes. Instead, the scientists have used machine learning to predict how good new enzymes will be. These predictions come from changes made to DNA sequences on a computer, not in a lab, saving time and effort.

AI predicts the best enzymes quickly

Before, scientists started with an enzyme found in nature and changed it bit by bit in living cells to make it do what they wanted. This took a lot of effort and time. Now, with computer models, scientists can skip all that and predict the best enzymes quickly.

Enzyme engineering has been around for a while, but using artificial intelligence, and in particular machine learning, speeds it up. Enzymes are proteins, and proteins are made of amino acids. Changing the DNA changes the amino acids, which changes the protein's shape and function. Scientists can now make and test these proteins outside of living organisms, which makes things even faster.

The researchers have already shown this can work and hope to work with pharmaceutical companies to develop it further. They're also looking at using this method for other reactions that could help the environment, make food healthier, or make industrial processes safer and cheaper.

However, they face challenges, especially with not having enough data. Machine learning needs a lot of good data to work well, and the need for data will only grow.