Molecular Graph Convolutions: Moving Beyond Fingerprints

Abstract

Molecular "fingerprints" encoding structural information are the workhorse of cheminformatics and machine learning in drug discovery applications. However, fingerprint representations necessarily emphasize particular aspects of the molecular structure while ignoring others, rather than allowing the model to make data-driven decisions. We describe molecular "graph convolutions", a machine learning architecture for learning from undirected graphs, specifically small molecules. Graph convolutions use a simple encoding of the molecular graph---atoms, bonds, distances, etc.---which allows the model to take greater advantage of information in the graph structure. Although graph convolutions do not outperform all fingerprint-based methods, they (along with other graph-based methods) represent a new paradigm in ligand-based virtual screening with exciting opportunities for future improvement.

See "Version information" section

References in corpus (4)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Spectral Networks and Locally Connected Networks on Graphs
• AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery
• Massively Multitask Networks for Drug Discovery

Cited by in corpus (243)

• A Comprehensive Survey on Graph Neural Networks
• Automatic chemical design using a data-driven continuous representation of molecules
• Axiomatic Attribution for Deep Networks
• Graph Convolutional Neural Networks for Web-Scale Recommender Systems
• SchNet - a deep learning architecture for molecules and materials
• Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals
• Machine Learning Force Fields
• Graph Neural Networks: A Review of Methods and Applications
• Recent Advances and Applications of Deep Learning Methods in Materials Science
• Atomistic Line Graph Neural Network for Improved Materials Property Predictions
• Machine learning prediction errors better than DFT accuracy
• Big-Data Science in Porous Materials: Materials Genomics and Machine Learning
• ChemBERTa: Large-Scale Self-Supervised Pretraining for Molecular Property Prediction
• Self-Supervised Graph Transformer on Large-Scale Molecular Data
• How Powerful are Graph Neural Networks?
• Foundations and modelling of dynamic networks using Dynamic Graph Neural Networks: A survey
• Hierarchical modeling of molecular energies using a deep neural network
• Graph Representation Learning in Biomedicine
• Developing an improved Crystal Graph Convolutional Neural Network framework for accelerated materials discovery
• Autonomous discovery in the chemical sciences part II: Outlook
• Polymer Informatics: Current Status and Critical Next Steps
• Most Ligand-Based Classification Benchmarks Reward Memorization Rather than Generalization
• Applying machine learning techniques to predict the properties of energetic materials
• Neural Relational Inference for Interacting Systems
• OrbNet: Deep Learning for Quantum Chemistry Using Symmetry-Adapted Atomic-Orbital Features
• An Overview of Deep Semi-Supervised Learning
• Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
• Geometric deep learning for computational mechanics Part I: Anisotropic Hyperelasticity
• Transferability in Machine Learning for Electronic Structure via the Molecular Orbital Basis
• DGL-LifeSci: An Open-Source Toolkit for Deep Learning on Graphs in Life Science
• Conditional molecular design with deep generative models
• GraphAF: a Flow-based Autoregressive Model for Molecular Graph Generation
• Rapid Prediction of Electron-Ionization Mass Spectrometry using Neural Networks
• Atom2Vec: learning atoms for materials discovery
• Improving Chemical Autoencoder Latent Space and Molecular De novo Generation Diversity with Heteroencoders
• Coarse Graining Molecular Dynamics with Graph Neural Networks
• Multi-view Graph Contrastive Representation Learning for Drug-Drug Interaction Prediction
• Deep learning and the Schrödinger equation
• Graph Neural Networks in Particle Physics
• SMILES Transformer: Pre-trained Molecular Fingerprint for Low Data Drug Discovery
• Machine learning on DNA-encoded libraries: A new paradigm for hit-finding
• Deep Learning on Graphs: A Survey
• Graph Neural Networks: Taxonomy, Advances and Trends
• Graph Neural Networks for an Accurate and Interpretable Prediction of the Properties of Polycrystalline Materials
• A graph representation of molecular ensembles for polymer property prediction
• Learning Multimodal Graph-to-Graph Translation for Molecular Optimization
• Graph Dynamical Networks for Unsupervised Learning of Atomic Scale Dynamics in Materials
• Relational Graph Attention Networks
• Using Attribution to Decode Dataset Bias in Neural Network Models for Chemistry
• Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels
• MR-GNN: Multi-Resolution and Dual Graph Neural Network for Predicting Structured Entity Interactions
• Unified Graph Neural Network Force-field for the Periodic Table
• Supervised Community Detection with Line Graph Neural Networks
• Deep Learning and Knowledge-Based Methods for Computer Aided Molecular Design -- Toward a Unified Approach: State-of-the-Art and Future Directions
• A Universal Density Matrix Functional from Molecular Orbital-Based Machine Learning: Transferability across Organic Molecules
• Hierarchical Generation of Molecular Graphs using Structural Motifs
• Molecule Attention Transformer
• Generative Models for Automatic Chemical Design
• Graph convolutional neural networks as "general-purpose" property predictors: the universality and limits of applicability
• Strategies for the Construction of Machine-Learning Potentials for Accurate and Efficient Atomic-Scale Simulations
• Image-based deep learning for classification of noise transients in gravitational wave detectors
• Combining docking pose rank and structure with deep learning improves protein-ligand binding mode prediction
• Graph Neural Network and Spatiotemporal Transformer Attention for 3D Video Object Detection from Point Clouds
• Graph Kernels: A Survey
• Quantum deep field: data-driven wave function, electron density generation, and atomization energy prediction and extrapolation with machine learning
• Covariant Compositional Networks For Learning Graphs
• Graph Neural Network Predictions of Metal Organic Framework CO2 Adsorption Properties
• Graph Neural Networks for Temperature-Dependent Activity Coefficient Prediction of Solutes in Ionic Liquids
• Graph Neural Networks for Molecules
• Representation Learning for Natural Language Processing
• TransGCN:Coupling Transformation Assumptions with Graph Convolutional Networks for Link Prediction
• Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules
• Deeply learning molecular structure-property relationships using attention- and gate-augmented graph convolutional network
• PiNN: A Python Library for Building Atomic Neural Networks of Molecules and Materials
• Machine learning-assisted design of material properties
• ChemGrapher: Optical Graph Recognition of Chemical Compounds by Deep Learning
• A Graph to Graphs Framework for Retrosynthesis Prediction
• Three-Dimensionally Embedded Graph Convolutional Network (3DGCN) for Molecule Interpretation
• Hierarchical Visualization of Materials Space with Graph Convolutional Neural Networks
• Regression-clustering for Improved Accuracy and Training Cost with Molecular-Orbital-Based Machine Learning
• Generating Focussed Molecule Libraries for Drug Discovery with Recurrent Neural Networks
• Materials Representation and Transfer Learning for Multi-Property Prediction
• Molecular Design in Synthetically Accessible Chemical Space via Deep Reinforcement Learning
• Graph Neural Networks Accelerated Molecular Dynamics
• ChemBO: Bayesian Optimization of Small Organic Molecules with Synthesizable Recommendations
• SMILES2Vec: An Interpretable General-Purpose Deep Neural Network for Predicting Chemical Properties
• Quantum Graph Neural Networks
• Physical Pooling Functions in Graph Neural Networks for Molecular Property Prediction
• KekuleScope: prediction of cancer cell line sensitivity and compound potency using convolutional neural networks trained on compound images
• Explainability Techniques for Graph Convolutional Networks
• A Survey of Deep Learning for Scientific Discovery
• Graph Neural Network Based Coarse-Grained Mapping Prediction
• Machine Learning Prediction of Accurate Atomization Energies of Organic Molecules from Low-Fidelity Quantum Chemical Calculations
• GraphVAMPNet, using graph neural networks and variational approach to markov processes for dynamical modeling of biomolecules
• MoleculeNet: A Benchmark for Molecular Machine Learning
• HNHN: Hypergraph Networks with Hyperedge Neurons
• LatentGNN: Learning Efficient Non-local Relations for Visual Recognition
• Graph Convolutional Neural Networks with Node Transition Probability-based Message Passing and DropNode Regularization
• Learning Gradient Fields for Molecular Conformation Generation
• Message-Passing Neural Networks Learn Little's Law
• Machine learning framework to predict the performance of lipid nanoparticles for nucleic acid delivery
• All SMILES Variational Autoencoder
• Chemi-net: a graph convolutional network for accurate drug property prediction
• Robust Mid-Pass Filtering Graph Convolutional Networks
• Hashing-Accelerated Graph Neural Networks for Link Prediction
• Generating valid Euclidean distance matrices
• Relational Pooling for Graph Representations
• Graphite: Iterative Generative Modeling of Graphs
• Chemical-Reaction-Aware Molecule Representation Learning
• Benchmarking Accuracy and Generalizability of Four Graph Neural Networks Using Large In Vitro ADME Datasets from Different Chemical Spaces
• Multi-View Self-Attention for Interpretable Drug-Target Interaction Prediction
• A Multiple Filter Based Neural Network Approach to the Extrapolation of Adsorption Energies on Metal Surfaces for Catalysis Applications
• Symmetrical Graph Neural Network for Quantum Chemistry, with Dual R/K Space
• Graph neural networks for the prediction of molecular structure-property relationships
• AI-guided inverse design and discovery of recyclable vitrimeric polymers
• SpreadGNN: Serverless Multi-task Federated Learning for Graph Neural Networks
• Hierarchical Graph-to-Graph Translation for Molecules
• TF3P: Three-dimensional Force Fields Fingerprint Learned by Deep Capsular Network
• Assessing Graph-based Deep Learning Models for Predicting Flash Point
• Probing triple Higgs coupling with machine learning at the LHC
• Accurate Molecular-Orbital-Based Machine Learning Energies via Unsupervised Clustering of Chemical Space
• A Universal Framework for Accurate and Efficient Geometric Deep Learning of Molecular Systems
• A Physically-informed Graph-based Order Parameter for the Universal Characterization of Atomic Structures
• Learn molecular representations from large-scale unlabeled molecules for drug discovery
• Deep Learning Based Regression and Multi-class Models for Acute Oral Toxicity Prediction with Automatic Chemical Feature Extraction
• What graph neural networks cannot learn: depth vs width
• Describe, Transform, Machine Learning: Feature Engineering for Grain Boundaries and Other Variable-Sized Atom Clusters
• Formulation Graphs for Mapping Structure-Composition of Battery Electrolytes to Device Performance
• Optimal Transport Graph Neural Networks
• DeepHealth: Review and challenges of artificial intelligence in health informatics
• ASNets: Deep Learning for Generalised Planning
• An End-to-End Framework for Molecular Conformation Generation via Bilevel Programming
• Meta-Learning GNN Initializations for Low-Resource Molecular Property Prediction
• Active Importance Sampling for Variational Objectives Dominated by Rare Events: Consequences for Optimization and Generalization
• Learning 3D Representations of Molecular Chirality with Invariance to Bond Rotations
• Application and Assessment of Deep Learning for the Generation of Potential NMDA Receptor Antagonists
• How hard is to distinguish graphs with graph neural networks?
• Deep Learning Development Environment in Virtual Reality
• Graph Nets for Partial Charge Prediction
• Message Passing Networks for Molecules with Tetrahedral Chirality
• SimGNN: A Neural Network Approach to Fast Graph Similarity Computation
• Energy-based View of Retrosynthesis
• Graph Warp Module: an Auxiliary Module for Boosting the Power of Graph Neural Networks in Molecular Graph Analysis
• DASH: Dynamic Attention-Based Substructure Hierarchy for Partial Charge Assignment
• BatmanNet: Bi-branch Masked Graph Transformer Autoencoder for Molecular Representation
• Reliable Graph Neural Networks for Drug Discovery Under Distributional Shift
• Predicting drug properties with parameter-free machine learning: Pareto-Optimal Embedded Modeling (POEM)
• A Systematic Comparison Study on Hyperparameter Optimisation of Graph Neural Networks for Molecular Property Prediction
• LiDAR-based Online 3D Video Object Detection with Graph-based Message Passing and Spatiotemporal Transformer Attention
• FunQG: Molecular Representation Learning Via Quotient Graphs
• Making Graph Neural Networks Worth It for Low-Data Molecular Machine Learning
• A Quantum-Inspired Method for Three-Dimensional Ligand-Based Virtual Screening
• Graph Neural Networks Inspired by Classical Iterative Algorithms
• Rotation-equivariant Graph Neural Networks for Learning Glassy Liquids Representations
• Factor Graph Neural Network
• Graph Neural Networks for the Prediction of Substrate-Specific Organic Reaction Conditions
• Learning a Continuous Representation of 3D Molecular Structures with Deep Generative Models
• Towards Interpretable Sparse Graph Representation Learning with Laplacian Pooling
• A Tailored Convolutional Neural Network for Nonlinear Manifold Learning of Computational Physics Data using Unstructured Spatial Discretizations
• Discriminative structural graph classification
• Dynamically Pruned Message Passing Networks for Large-Scale Knowledge Graph Reasoning
• Discovering Molecular Functional Groups Using Graph Convolutional Neural Networks
• Bi-Level Graph Neural Networks for Drug-Drug Interaction Prediction
• Graph Neural Network Based Surrogate Model of Physics Simulations for Geometry Design
• Learning Attributed Graph Representations with Communicative Message Passing Transformer
• Improving Compound Activity Classification via Deep Transfer and Representation Learning
• A Generative Model for Molecular Distance Geometry
• Contrastive Dual-Interaction Graph Neural Network for Molecular Property Prediction
• Zoo Guide to Network Embedding
• Jointly Learning Entity and Relation Representations for Entity Alignment
• Gated Graph Recursive Neural Networks for Molecular Property Prediction
• Molecule Property Prediction and Classification with Graph Hypernetworks
• Artificial Intelligence in Drug Discovery: Applications and Techniques
• PEMP: Leveraging Physics Properties to Enhance Molecular Property Prediction
• Molecular Dipole Moment Learning via Rotationally Equivariant Gaussian Process Regression with Derivatives in Molecular-orbital-based Machine Learning
• HGCH: A Hyperbolic Graph Convolution Network Model for Heterogeneous Collaborative Graph Recommendation
• Tiered Latent Representations and Latent Spaces for Molecular Graphs
• Molecule3D: A Benchmark for Predicting 3D Geometries from Molecular Graphs
• Online Graph Learning from Social Interactions
• Probabilistic Generative Deep Learning for Molecular Design
• Adversarial Graph Disentanglement
• Drug-Target Indication Prediction by Integrating End-to-End Learning and Fingerprints
• Neural Message Passing for Multi-Label Classification
• On the equivalence of molecular graph convolution and molecular wave function with poor basis set
• Demand Forecasting from Spatiotemporal Data with Graph Networks and Temporal-Guided Embedding
• Attention-Based Learning on Molecular Ensembles
• Molecular distance matrix prediction based on graph convolutional networks
• PushNet: Efficient and Adaptive Neural Message Passing
• SB-MTL: Score-based Meta Transfer-Learning for Cross-Domain Few-Shot Learning
• Optimizing Diffusion Rate and Label Reliability in a Graph-Based Semi-supervised Classifier
• A comprehensive study on the prediction reliability of graph neural networks for virtual screening
• Leveraging binding-site structure for drug discovery with point-cloud methods
• Equivariant Entity-Relationship Networks
• Lipophilicity Prediction with Multitask Learning and Molecular Substructures Representation
• Enhancing the Association in Multi-Object Tracking via Neighbor Graph
• Accurate Prediction of Free Solvation Energy of Organic Molecules via Graph Attention Network and Message Passing Neural Network from Pairwise Atomistic Interactions
• Artificial Intelligence based Autonomous Molecular Design for Medical Therapeutic: A Perspective
• Prediction of Small Molecule Kinase Inhibitors for Chemotherapy Using Deep Learning
• MLSolv-A: A Novel Machine Learning-Based Prediction of Solvation Free Energies from Pairwise Atomistic Interactions
• Unveiling Molecular Moieties through Hierarchical Grad-CAM Graph Explainability
• Scaffold Embeddings: Learning the Structure Spanned by Chemical Fragments, Scaffolds and Compounds
• Prediction of Carbon Nanostructure Mechanical Properties and Role of Defects Using Machine Learning
• Pretraining a Foundation Model for Small-Molecule Natural Products
• Transformer Based Molecule Encoding for Property Prediction
• ENT-DESC: Entity Description Generation by Exploring Knowledge Graph
• Structure-Aware Label Smoothing for Graph Neural Networks
• Designing compact training sets for data-driven molecular property prediction
• Graph-to-Sequence Learning using Gated Graph Neural Networks
• Graph Attribute Aggregation Network with Progressive Margin Folding
• InteractionNet: Modeling and Explaining of Noncovalent Protein-Ligand Interactions with Noncovalent Graph Neural Network and Layer-Wise Relevance Propagation
• Fea2Fea: Exploring Structural Feature Correlations via Graph Neural Networks
• Unified theory of atom-centered representations and message-passing machine-learning schemes
• A Generative Framework for Predictive Modeling of Multiple Chronic Conditions Using Graph Variational Autoencoder and Bandit-Optimized Graph Neural Network
• MolMiner: Toward Controllable, 3D-Aware, Fragment-Based Molecular Design
• Hier-EgoPack: Hierarchical Egocentric Video Understanding with Diverse Task Perspectives
• End-to-End Learning from Complex Multigraphs with Latent-Graph Convolutional Networks
• Geometric learning of the conformational dynamics of molecules using dynamic graph neural networks
• LiteGEM: Lite Geometry Enhanced Molecular Representation Learning for Quantum Property Prediction
• Graph-Convolutional Deep Learning to Identify Optimized Molecular Configurations
• Graph Filtration Learning
• Towards explainable message passing networks for predicting carbon dioxide adsorption in metal-organic frameworks
• Particle Track Reconstruction using Geometric Deep Learning
• Molecular modeling with machine-learned universal potential functions
• Augmenting Molecular Images with Vector Representations as a Featurization Technique for Drug Classification
• Framework for Designing Filters of Spectral Graph Convolutional Neural Networks in the Context of Regularization Theory
• Deep Learning on Attributed Graphs: A Journey from Graphs to Their Embeddings and Back
• Neural Consciousness Flow
• Improving Graph Property Prediction with Generalized Readout Functions
• Drug cell line interaction prediction
• A deep learning approach to the structural analysis of proteins
• Implicit vs Unfolded Graph Neural Networks
• IL-Net: Using Expert Knowledge to Guide the Design of Furcated Neural Networks
• SeaPearl: A Constraint Programming Solver guided by Reinforcement Learning
• Batched Sparse Matrix Multiplication for Accelerating Graph Convolutional Networks
• SPECTRe: Substructure Processing, Enumeration, and Comparison Tool Resource: An efficient tool to encode all substructures of molecules represented in SMILES
• Transportation Scenario Planning with Graph Neural Networks
• Detect the Interactions that Matter in Matter: Geometric Attention for Many-Body Systems
• Generate Novel Molecules With Target Properties Using Conditional Generative Models
• Classification of vertices on social networks by multiple approaches
• Probabilistic hypergraph grammars for efficient molecular optimization
• Scalar Coupling Constant Prediction Using Graph Embedding Local Attention Encoder
• Weisfeiler and Leman go Hyperbolic: Learning Distance Preserving Node Representations
• libmolgrid: GPU Accelerated Molecular Gridding for Deep Learning Applications