Generation of novel Diels-Alder reactions using a generative adversarial network

Area of Science:

• Computational chemistry
• Organic chemistry
• Artificial intelligence in chemistry

Background:

• Deep learning, particularly generative adversarial networks (GANs), shows promise in molecular generation.
• The application of GANs in organic chemistry, especially for reaction generation, remains underexplored.

Purpose of the Study:

• To investigate the utility of a GAN as a generative model for creating Diels-Alder reactions.
• To assess the capability of a MaskGAN model in learning and generating organic reactions.

Main Methods:

• A MaskGAN model was trained using a dataset of 14,092 known Diels-Alder reactions.
• The trained model was then employed to generate novel Diels-Alder reaction examples.

Main Results:

• The MaskGAN model successfully generated 1,441 novel Diels-Alder reactions.
• Analysis of the generated reactions demonstrated that the model had learned fundamental organic reaction rules.

Conclusions:

• The MaskGAN model is effective for generating organic reactions, specifically Diels-Alder reactions.
• This deep learning approach can assist chemists in discovering and exploring new chemical reactions.