TopologyNet: Topology based deep convolutional neural networks for biomolecular property predictions

Abstract

Although deep learning approaches have had tremendous success in image, video and audio processing, computer vision, and speech recognition, their applications to three-dimensional (3D) biomolecular structural data sets have been hindered by the entangled geometric complexity and biological complexity. We introduce topology, i.e., element specific persistent homology (ESPH), to untangle geometric complexity and biological complexity. ESPH represents 3D complex geometry by one-dimensional (1D) topological invariants and retains crucial biological information via a multichannel image representation. It is able to reveal hidden structure-function relationships in biomolecules. We further integrate ESPH and convolutional neural networks to construct a multichannel topological neural network (TopologyNet) for the predictions of protein-ligand binding affinities and protein stability changes upon mutation. To overcome the limitations to deep learning arising from small and noisy training sets, we present a multitask topological convolutional neural network (MT-TCNN). We demonstrate that the present TopologyNet architectures outperform other state-of-the-art methods in the predictions of protein-ligand binding affinities, globular protein mutation impacts, and membrane protein mutation impacts.

20 pages, 8 figures, 5 tables

References in corpus (13)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Deep Learning in Neural Networks: An Overview
• The NumPy array: a structure for efficient numerical computation
• Multi-Task Feature Learning Via Efficient l2,1-Norm Minimization
• AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery
• Massively Multitask Networks for Drug Discovery
• Persistent homology analysis of protein structure, flexibility and folding
• Multi-task Neural Networks for QSAR Predictions
• Persistence weighted Gaussian kernel for topological data analysis
• Toxicity Prediction using Deep Learning
• Automatic parametrization of implicit solvent models for the blind prediction of solvation free energies
• Persistent Homology for The Quantitative Prediction of Fullerene Stability
• Topological fingerprints reveal protein-ligand binding mechanism

Cited by in corpus (44)

• DeepAffinity: Interpretable Deep Learning of Compound-Protein Affinity through Unified Recurrent and Convolutional Neural Networks
• OnionNet: a multiple-layer inter-molecular contact based convolutional neural network for protein-ligand binding affinity prediction
• Representability of algebraic topology for biomolecules in machine learning based scoring and virtual screening
• A review of mathematical representations of biomolecules
• HAC-Net: A Hybrid Attention-Based Convolutional Neural Network for Highly Accurate Protein-Ligand Binding Affinity Prediction
• Topological Deep Learning: A Review of an Emerging Paradigm
• Drug-Target Interaction/Affinity Prediction: Deep Learning Models and Advances Review
• Persistent Dirac for molecular representation
• Persistent homology analysis of ion aggregation and hydrogen-bonding network
• Persistent homology analysis of osmolyte molecular aggregation and their hydrogen-bonding networks
• Evolutionary homology on coupled dynamical systems
• Quantitative toxicity prediction using topology based multi-task deep neural networks
• Weighted (Co)homology and Weighted Laplacian
• Benchmarking AlphaFold3's protein-protein complex accuracy and machine learning prediction reliability for binding free energy changes upon mutation
• Review of the mechanisms of SARS-CoV-2 evolution and transmission
• Persistent spectral based machine learning (PerSpect ML) for drug design
• Evolutionary Khovanov homology
• Mathematical deep learning for pose and binding affinity prediction and ranking in D3R Grand Challenges
• The stability of persistent homology of hypergraphs
• Persistent cohomology for data with multicomponent heterogeneous information
• Sequence-based Multiscale Model (SeqMM) for High-throughput chromosome conformation capture (Hi-C) data analysis
• Weighted persistent homology for biomolecular data analysis
• MathDL: Mathematical deep learning for D3R Grand Challenge 4
• Integrating Physics and Topology in Neural Networks for Learning Rigid Body Dynamics
• Dirac-Equation Signal Processing: Physics Boosts Topological Machine Learning
• HERMES: Persistent spectral graph software
• Persistent spectral graph
• Evolutionary de Rham-Hodge method
• Algebraic graph learning of protein-ligand binding affinity
• Weighted persistent homology for osmolyte molecular aggregation and hydrogen-bonding network analysis
• InteractionNet: Modeling and Explaining of Noncovalent Protein-Ligand Interactions with Noncovalent Graph Neural Network and Layer-Wise Relevance Propagation
• The de Rham-Hodge analysis and modeling of biomolecules
• Topology Estimation of Simulated 4D Image Data by Combining Downscaling and Convolutional Neural Networks
• DG-GL: Differential geometry based geometric learning of molecular datasets
• A Unified Topological Approach to Data Science
• Multiscale Laplacian Learning
• Generative network complex (GNC) for drug discovery
• Ollivier persistent Ricci curvature (OPRC) based molecular representation for drug design
• Inferring Mesoscale Models of Neural Computation
• libmolgrid: GPU Accelerated Molecular Gridding for Deep Learning Applications
• Pull-back Geometry of Persistent Homology Encodings
• PotentialNet for Molecular Property Prediction
• Mathematics at the eve of a historic transition in biology
• Geometric Bounds for Persistence