Вештачка интелигенција

New Method Uses Artificial Intelligence to Predict Optimal Drug Synthesis

Summary

An innovative method developed by a team of researchers from LMU University, ETH Zurich, and Roche Pharmaceutical Research and Development (pRED) has the potential to revolutionize drug synthesis. By utilizing artificial intelligence (AI), the method predicts the optimal way to […]

New Method Uses Artificial Intelligence to Predict Optimal Drug Synthesis

An innovative method developed by a team of researchers from LMU University, ETH Zurich, and Roche Pharmaceutical Research and Development (pRED) has the potential to revolutionize drug synthesis. By utilizing artificial intelligence (AI), the method predicts the optimal way to synthesize drug molecules, reducing the need for laborious and time-consuming laboratory experiments.

David Nippa, the lead author of the study published in the journal Nature Chemistry, states that this breakthrough has the potential to significantly increase the efficiency and sustainability of chemical synthesis. Nippa is a doctoral student with Dr. David Konrad’s research group at LMU University and also works at Roche.

Durable pharmaceutical ingredients often consist of frameworks with attached functional groups that enable specific biological functions. In order to achieve new or improved medical effects, these functional groups are modified and added to different positions within the framework. However, this process poses challenges in chemistry as the frameworks, primarily composed of carbon and hydrogen atoms, are inherently unreactive. One method to activate these frameworks is through a borrelation reaction, where a boron-containing group binds to a carbon atom within the framework. This borrelation group can then be replaced with various medicinally-effective groups. Despite its potential, controlling borrelation in a laboratory environment is difficult.

In collaboration with Kenneth Atzom, a doctoral student at ETH Zurich, David Nippa developed an AI model trained on data from reliable scientific papers and experiments conducted in Roche’s automated laboratory. This model successfully predicts the borrelation position for any given molecule and provides optimal conditions for the chemical transformation.

An interesting aspect is that the predictions are improved when considering three-dimensional information about the initial materials, rather than solely relying on their two-dimensional chemical formulas.

The method has already been successfully applied to identify positions in existing active ingredients where additional active groups can be introduced. This assists researchers in rapidly developing new and more efficient variations of known drug active ingredients.

Frequently Asked Questions (FAQ)

1. How does artificial intelligence aid in drug synthesis?

Artificial intelligence enables the development of innovative methods that predict the optimal way to synthesize drug molecules. By analyzing and providing recommendations for chemical transformations using data from scientific papers and experiments, artificial intelligence reduces the number of necessary laboratory experiments and increases process efficiency.

2. What is a borrelation reaction?

A borrelation reaction is a chemical reaction where a boron-containing group binds to a carbon atom within a molecule. This reaction allows for the substitution of the borrelation group with various medicinally-effective groups.

3. How does this new method contribute to the development of new variations of existing drugs?

This new method enables the identification of positions within existing drug active ingredients where additional active groups can be introduced. This aids researchers in rapidly developing new and more efficient variations of known drugs, thereby advancing the pharmaceutical industry.