Application of generative autoencoder in de novo molecular design

TL;DR

This paper explores the use of generative autoencoders to create novel molecular structures with desired properties, demonstrating the preservation of chemical similarity and successful generation of active compounds.

Contribution

It introduces the application of generative autoencoders for de novo molecular design, highlighting their ability to generate chemically similar and active novel compounds.

Findings

Latent space preserves chemical similarity principles.

Autoencoders can generate novel compounds with predicted activity.

Systematic search in latent space identifies compounds similar to known actives.

Abstract

A major challenge in computational chemistry is the generation of novel molecular structures with desirable pharmacological and physiochemical properties. In this work, we investigate the potential use of autoencoder, a deep learning methodology, for de novo molecular design. Various generative autoencoders were used to map molecule structures into a continuous latent space and vice versa and their performance as structure generator was assessed. Our results show that the latent space preserves chemical similarity principle and thus can be used for the generation of analogue structures. Furthermore, the latent space created by autoencoders were searched systematically to generate novel compounds with predicted activity against dopamine receptor type 2 and compounds similar to known active compounds not included in the training set were identified.