Directional molecule generation method based on graph neural network

Abstract

The invention relates to a directed molecule generation method based on a graph neural network, and relates to the technical field of material molecules. Comprising the following steps: converting an organic molecular structure graph into a molecular graph in a topological mapping mode, and taking embedded representation of the molecular graph as input of a graph neural network model; through a graph neural network model, learning the molecular graphs based on a message propagation process, including representations of nodes and edges in the molecular graphs; the generated representations are learned through a graph neural network so that various decisions can be made in the graph generation process; in the decision-making process, the new structure is added to the existing graph in a form conforming to the organic molecule chemical rule, and the probability of the addition event depends on the historical graph derivation process of the graph. The finally generated novel molecules are confirmed through chemical valence constraint, and the chemical effectiveness of the generated molecules can be ensured. According to the method, effective novel molecular structures with chemical properties similar to those of original molecules can be generated aiming at an organic molecule database.

Description

technical field [0001] The invention relates to the technical field of material molecules, in particular to a method for generating directed molecules based on a graph neural network. Background technique [0002] In our daily life, graph neural networks can be seen everywhere. They are used to build complex systems, such as many topological structures composed of economics, nature, and social sciences, such as social networks, performance networks in the field of biomedicine, synthesis and performance prediction of material molecules in the field of chemical materials, etc., in many aspects of social life. It has practical meaning in the actual scene. For example, recommending content and users of interest to users in social networks, identifying protein functions in PPI biological networks, or predicting the ability of existing materials to target certain physical and chemical properties. In recent years, machine learning has become an efficient research method for graph...

AI Technical Summary

Problems solved by technology

The specific implementation process is to map the chemical molecular structure into a molecular graph, and learn the graph through the graph neural network. At the same time, for the order of the generated nodes, the breadth-first search algorithm is used to overcome the large number of invalid generation caused by the random order in the prior art. molecular problem

Images