PubMed · 2024 · 26 citations · 29 references
Life SciencesBioinformaticsComputational BiologyComputational Medicinal ChemistryDrug DiscoveryComputational Drug DiscoveryMolecular DockingProtein DockingProtein-ligand DockingProtein-protein DockingDiffusion ModelsDocking BenchmarkingDocking ValidationGeneralizability ModelsLigand-protein DockingProtein-protein Docking Benchmark DatasetsScaling Laws
Accurate blind docking has the potential to lead to new biological breakthroughs, but for this promise to be realized, docking methods must generalize well across the proteome. Existing benchmarks, however, fail to rigorously assess generalizability. Therefore, we develop DockGen, a new benchmark based on the ligand-binding domains of proteins, and we show that existing machine learning-based docking models have very weak generalization abilities. We carefully analyze the scaling laws of ML-based docking and show that, by scaling data and model size, as well as integrating synthetic data strategies, we are able to significantly increase the generalization capacity and set new state-of-the-art performance across benchmarks. Further, we propose Confidence Bootstrapping, a new training paradigm that solely relies on the interaction between diffusion and confidence models and exploits the multi-resolution generation process of diffusion models. We demonstrate that Confidence Bootstrapping significantly improves the ability of ML-based docking methods to dock to unseen protein classes, edging closer to accurate and generalizable blind docking methods.
29
Highly accurate protein structure prediction with AlphaFold
John Jumper, Alexander Pritzel, Tim Green et al. · Nature · 2021 · 45.2K citations · Full text
Mastering the game of Go with deep neural networks and tree search
David Silver, Aja Huang, Chris J. Maddison et al. · Nature · 2016 · 15.8K citations
UniProt: a hub for protein information
Nucleic Acids Research · 2014 · 5.3K citations · Full text