Molecular graph generation method based on variational auto-encoder and message passing neural network

Abstract

The invention discloses a molecular graph generation method based on a variational auto-encoder and a message passing neural network. The molecular graph generation method is used for molecular generation and molecular target characteristic optimization. According to the method, the message passing neural network is constructed into the encoder and the decoder of the variational auto-encoder, so that the running time and the occupied memory of the training process are further reduced; in addition, by constructing a potential space of the variational auto-encoder, molecular properties are allowed to be optimized; in a molecule generation experiment on a QM9 chemical database, the model can generate 100% effective compounds, and the novelty rate and the uniqueness rate are also very high; in a target optimization experiment on a QM9 chemical database, target characteristics can be further optimized.

Description

technical field [0001] The invention relates to the technical field of molecular graph generation, in particular to a molecular graph generation method based on a variational autoencoder and a message passing neural network. Background technique [0002] The discovery of new molecules in materials chemistry has become a hot topic in modern society, and materials innovation is a key driver of many recent technological advances. Materials innovation is a key driver of many recent technological advances. From clean energy to the aerospace industry or drug discovery, research in chemistry and materials science is constantly evolving to develop compounds with novel uses, lower cost and better performance. [0003] At the highest level of abstraction, the design of molecules is formulated as a combinatorial optimization problem to find optimal solutions in a vast chemical space. Many important problems in drug discovery and materials science are based on the principle of designi...

AI Technical Summary

Problems solved by technology

However, this remains a challenging task due to the huge size of the chemical space

A medicinal chemist, or a brand-new molecular design software, is faced with an almost infinite search space. This search space is huge. Due to its discrete nature, it is very difficult to find target compound molecules in this space.

[0004] Despite tremendous advances in high-throughput screening techniques, an exhaustive search in such a large space is impossible

Images