VNFlow: Integrating Normalizing Flows with Variational Autoencoders for Molecular Design

The exploration of the vast, largely unexplored chemical space is being transformed by generative Artificial Intelligence. However, state-of-the-art methods, including normalizing flows, struggle to balance the optimization of complex objectives and sampling speed, particularly when generating specific compound classes and more intricate scaffolds, such as aromatic rings. Here, we developed a generative model that efficiently samples novel molecules while optimizing their drug-likeness, ease of synthesis or chemical reactivity. By employing normalizing flows combined with variational autoencoders to generate samples, our framework efficiently generated a diverse range of organofluorine-phosphates, demonstrating that combining normalizing flows directly with SELFIES or group-SELFIES can address key limitations in inverse molecular design, particularly when variational autoencoders cannot be applied due to a lack of available training data. Normalizing flows capture the chemical structures in a holistic way which paves the way towards targeted therapies that enable the optimization of complex molecular objectives.