Published on Sun Nov 15 2020

Molecular Mechanics-Driven Graph Neural Network with Multiplex Graph for Molecular Structures

Shuo Zhang, Yang Liu, Lei Xie

The prediction of physicochemical properties from molecular structures is acrucial task for artificial intelligence aided molecular design. A growing number of Graph Neural Networks (GNNs) have been proposed to address this challenge. We aim to design a GNN which is both powerful and efficient for molecule structures.

0
0
0
Abstract

The prediction of physicochemical properties from molecular structures is a crucial task for artificial intelligence aided molecular design. A growing number of Graph Neural Networks (GNNs) have been proposed to address this challenge. These models improve their expressive power by incorporating auxiliary information in molecules while inevitably increase their computational complexity. In this work, we aim to design a GNN which is both powerful and efficient for molecule structures. To achieve such goal, we propose a molecular mechanics-driven approach by first representing each molecule as a two-layer multiplex graph, where one layer contains only local connections that mainly capture the covalent interactions and another layer contains global connections that can simulate non-covalent interactions. Then for each layer, a corresponding message passing module is proposed to balance the trade-off of expression power and computational complexity. Based on these two modules, we build Multiplex Molecular Graph Neural Network (MXMNet). When validated by the QM9 dataset for small molecules and PDBBind dataset for large protein-ligand complexes, MXMNet achieves superior results to the existing state-of-the-art models under restricted resources.

Sat Sep 26 2020
Machine Learning
Heterogeneous Molecular Graph Neural Networks for Predicting Molecule Properties
Graph neural network (GNN)-based methods have been widely used to predict quantum mechanical properties of molecules. Only limited works explicitly take many-body interactions, i.e., interactions between three and more atoms, into consideration. We introduce a novel graph representation of molecules, heterogeneous molecular graph (
0
0
0
Thu Aug 27 2020
Machine Learning
Graph Neural Network Architecture Search for Molecular Property Prediction
Predicting the properties of a molecule from its structure is a challenging task. By treating molecule structure as graphs, graph neural networks (GNNs) have been widely used to predict molecular properties. However, the design and development of GNNs rely on labor-intensive design and tuning.
0
0
0
Sat Aug 31 2019
Machine Learning
Gated Graph Recursive Neural Networks for Molecular Property Prediction
Molecule property prediction is a fundamental problem for computer-aided drug discovering and materials science. Quantum-chemical simulations such as density functional theory have been widely used for calculating the molecule's properties. We model a molecular as a directed complete graph in which each atom has a spatial position.
0
0
0
Wed May 29 2019
Machine Learning
Path-Augmented Graph Transformer Network
Path-Augmented Graph Transformer Networks (PAGTN) are built on longer-range dependencies in graph-structured data. We use path features in molecular graphs to create global attention layers. We show that our model consistently outperforms GCNs on molecular property prediction.
0
0
0
Mon Dec 14 2020
Machine Learning
Molecular graph generation with Graph Neural Networks
Drug Discovery is a fundamental and ever-evolving field of research. The design of new candidate molecules requires large amounts of time and money. Machine learning methods are ideal for the design of potential candidate molecules, which are naturally represented as graphs.
0
0
0
Mon Jul 19 2021
Machine Learning
Learning Attributed Graph Representations with Communicative Message Passing Transformer
We propose a Communicative Message Passing Transformer (CoMPT) neural network to improve the molecular graph representation. CoMPT uses message interactions between nodes and edges to leverage the graph connectivity inductive bias and reduce the message enrichment explosion.
3
5
20