AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery

Abstract

Deep convolutional neural networks comprise a subclass of deep neural networks (DNN) with a constrained architecture that leverages the spatial and temporal structure of the domain they model. Convolutional networks achieve the best predictive performance in areas such as speech and image recognition by hierarchically composing simple local features into complex models. Although DNNs have been used in drug discovery for QSAR and ligand-based bioactivity predictions, none of these models have benefited from this powerful convolutional architecture. This paper introduces AtomNet, the first structure-based, deep convolutional neural network designed to predict the bioactivity of small molecules for drug discovery applications. We demonstrate how to apply the convolutional concepts of feature locality and hierarchical composition to the modeling of bioactivity and chemical interactions. In further contrast to existing DNN techniques, we show that AtomNet's application of local convolutional filters to structural target information successfully predicts new active molecules for targets with no previously known modulators. Finally, we show that AtomNet outperforms previous docking approaches on a diverse set of benchmarks by a large margin, achieving an AUC greater than 0.9 on 57.8% of the targets in the DUDE benchmark.

References in corpus (6)

• ADADELTA: An Adaptive Learning Rate Method
• Understanding Neural Networks Through Deep Visualization
• Going Deeper with Convolutions
• Massively Multitask Networks for Drug Discovery
• Multi-task Neural Networks for QSAR Predictions
• Toxicity Prediction using Deep Learning

Cited by in corpus (62)

• DeepDTA: Deep Drug-Target Binding Affinity Prediction
• Molecular Graph Convolutions: Moving Beyond Fingerprints
• Machine Learning Unifies the Modelling of Materials and Molecules
• DeepAffinity: Interpretable Deep Learning of Compound-Protein Affinity through Unified Recurrent and Convolutional Neural Networks
• Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point clouds
• Self-Supervised Graph Transformer on Large-Scale Molecular Data
• TopologyNet: Topology based deep convolutional neural networks for biomolecular property predictions
• Deep Neural Networks for Data-Driven Turbulence Models
• Representability of algebraic topology for biomolecules in machine learning based scoring and virtual screening
• WideDTA: prediction of drug-target binding affinity
• Deep convolutional networks for quality assessment of protein folds
• Chemception: A Deep Neural Network with Minimal Chemistry Knowledge Matches the Performance of Expert-developed QSAR/QSPR Models
• Deep Confidence: A Computationally Efficient Framework for Calculating Reliable Errors for Deep Neural Networks
• Learning Deep Architectures for Interaction Prediction in Structure-based Virtual Screening
• Molecule Attention Transformer
• Visualizing Convolutional Neural Network Protein-Ligand Scoring
• Atomic Convolutional Networks for Predicting Protein-Ligand Binding Affinity
• Adaptive Graph Convolutional Neural Networks
• Machine learning-assisted design of material properties
• Three-Dimensionally Embedded Graph Convolutional Network (3DGCN) for Molecule Interpretation
• Combining Machine Learning with Knowledge-Based Modeling for Scalable Forecasting and Subgrid-Scale Closure of Large, Complex, Spatiotemporal Systems
• SMILES2Vec: An Interpretable General-Purpose Deep Neural Network for Predicting Chemical Properties
• Hierarchical, rotation-equivariant neural networks to select structural models of protein complexes
• Simulations meet Machine Learning in Structural Biology
• MoleculeNet: A Benchmark for Molecular Machine Learning
• Deep Learning for Computational Chemistry
• Multi-Residual Networks: Improving the Speed and Accuracy of Residual Networks
• Drug-Target Interaction/Affinity Prediction: Deep Learning Models and Advances Review
• Latent Molecular Optimization for Targeted Therapeutic Design
• Multi-View Self-Attention for Interpretable Drug-Target Interaction Prediction
• Artificial Intelligence in Materials Science and Engineering: Current Landscape, Key Challenges, and Future Trajectorie
• TinySpeech: Attention Condensers for Deep Speech Recognition Neural Networks on Edge Devices
• Generating 3D Molecular Structures Conditional on a Receptor Binding Site with Deep Generative Models
• Independent SE(3)-Equivariant Models for End-to-End Rigid Protein Docking
• A Survey on Applications of Artificial Intelligence in Fighting Against COVID-19
• Convolutional Rectifier Networks as Generalized Tensor Decompositions
• How Much Can We Really Trust You? Towards Simple, Interpretable Trust Quantification Metrics for Deep Neural Networks
• A Novel Framework Integrating AI Model and Enzymological Experiments Promotes Identification of SARS-CoV-2 3CL Protease Inhibitors and Activity-based Probe
• DeepAtom: A Framework for Protein-Ligand Binding Affinity Prediction
• Multimodal Deep Neural Networks using Both Engineered and Learned Representations for Biodegradability Prediction
• Deep Neural Networks for Detecting Statistical Model Misspecifications. The Case of Measurement Invariance
• Learning a Continuous Representation of 3D Molecular Structures with Deep Generative Models
• An Oracle for Guiding Large-Scale Model/Hybrid Parallel Training of Convolutional Neural Networks
• Capsule Networks for Protein Structure Classification and Prediction
• Discovering Molecular Functional Groups Using Graph Convolutional Neural Networks
• Learning Graph While Training: An Evolving Graph Convolutional Neural Network
• Regression Enrichment Surfaces: a Simple Analysis Technique for Virtual Drug Screening Models
• DGCNN: Disordered Graph Convolutional Neural Network Based on the Gaussian Mixture Model
• Explainable Deep Relational Networks for Predicting Compound-Protein Affinities and Contacts
• Leveraging binding-site structure for drug discovery with point-cloud methods
• Accelerated Convolutions for Efficient Multi-Scale Time to Contact Computation in Julia
• Learning physics confers pose-sensitivity in structure-based virtual screening
• Structure-aware Interactive Graph Neural Networks for the Prediction of Protein-Ligand Binding Affinity
• Classificação de espécies de peixe utilizando redes neurais convolucional
• PotentialNet for Molecular Property Prediction
• libmolgrid: GPU Accelerated Molecular Gridding for Deep Learning Applications
• ParaVS: A Simple, Fast, Efficient and Flexible Graph Neural Network Framework for Structure-Based Virtual Screening
• How to Train your DNN: The Network Operator Edition
• Improved Drug-target Interaction Prediction with Intermolecular Graph Transformer
• Landscape of Big Medical Data: A Pragmatic Survey on Prioritized Tasks
• IL-Net: Using Expert Knowledge to Guide the Design of Furcated Neural Networks
• Drug cell line interaction prediction