MGCVAE: Multi-objective Inverse Design via Molecular Graph Conditional Variational Autoencoder

TL;DR

This paper introduces MGCVAE, a molecular graph generative model that efficiently designs molecules with specific properties, outperforming previous models in generating drug-like molecules with targeted physical characteristics.

Contribution

The study presents a novel multi-objective conditional variational autoencoder for molecular design, demonstrating improved generation of molecules with desired properties compared to existing models.

Findings

MGCVAE generated 25.89% optimized molecules versus 0.66% by MGVAE.

MGCVAE effectively produces drug-like molecules with targeted properties.

Multi-objective optimization successfully satisfies two physical property targets.

Abstract

The ultimate goal of various fields is to directly generate molecules with desired properties, such as finding water-soluble molecules in drug development and finding molecules suitable for organic light-emitting diode (OLED) or photosensitizers in the field of development of new organic materials. In this respect, this study proposes a molecular graph generative model based on the autoencoder for de novo design. The performance of molecular graph conditional variational autoencoder (MGCVAE) for generating molecules having specific desired properties is investigated by comparing it to molecular graph variational autoencoder (MGVAE). Furthermore, multi-objective optimization for MGCVAE was applied to satisfy two selected properties simultaneously. In this study, two physical properties -- logP and molar refractivity -- were used as optimization targets for the purpose of designing de novo molecules, especially in drug discovery. As a result, it was confirmed that among generated molecules, 25.89% optimized molecules were generated in MGCVAE compared to 0.66% in MGVAE. Hence, it demonstrates that MGCVAE effectively produced drug-like molecules with two target properties. The results of this study suggest that these graph-based data-driven models are one of the effective methods of designing new molecules that fulfill various physical properties, such as drug discovery.