DynamicBind: A Deep Learning Approach for Dynamic Protein-Ligand Docking and Drug Discovery

While significant progress has been made in predicting static protein structures, understanding protein dynamics, influenced by ligands, is essential to understanding protein function and advancing drug discovery. Traditional docking methods typically treat proteins as rigid, which limits their accuracy. Although molecular dynamics simulations can suggest relevant protein conformations, they are computationally intensive. Recent advances, such as AlphaFold, predict structures from sequences but generate only a few conformations, overlooking the dynamic nature of proteins. This limitation affects the docking accuracy, as the structures predicted by AlphaFold may not reflect the optimal configurations for ligand binding, leading to inaccurate predictions.

Researchers from Galixir Technologies, the School of Pharmaceutical Sciences at Sun Yat-sen University, the Center for Theoretical Biological Physics and the Department of Chemistry at Rice University, and the Global Future technology Institute at Shanghai Jiao Tong University have developed dynamic link, a deep learning method that uses geometric equivariant diffusion networks to create a fluid energy landscape, allowing efficient transitions between different equilibrium states. DynamicBind accurately predicts ligand-specific conformations from unbound protein structures without holostructures or extensive sampling. It excels at virtual docking and screening benchmarks, accommodating large protein conformational changes and identifying hidden pockets in new protein targets. This approach holds promise for accelerating the development of small molecules for previously non-drug targets, promoting computational drug discovery.

DynamicBind, a geometric deep generative model for dynamic docking, efficiently tunes protein conformations from initial AlphaFold predictions to holographic states. It handles significant conformational changes, such as the DFG in-out transition in kinases, better than traditional molecular dynamics simulations. DynamicBind achieves this by learning a channeled energy landscape that minimizes frustration during transitions between biologically relevant states. Unlike traditional Boltzmann generators, DynamicBind can be generalized to new proteins and ligands.

The DynamicBind model is a diffusion-based graph neural network equivalent to E(3) that uses a coarse-grained representation to predict protein-ligand binding conformations. It efficiently transforms input structures to account for 3D parity and transrotational changes, outperforming traditional methods with less data. The model employs a transformation-like transformation for training, interpolating between crystal structures and AlphaFold. Using a graphical representation, each protein residue and ligand atom is a node with several characteristics. DynamicBind updates these nodes through tensor products and diffusion processes to predict side chain dihedrals, torsion angles, translations, and rotations, improving binding affinity predictions.

DynamicBind is a versatile tool for predicting complex protein-ligand structures, capable of adapting to significant protein conformational changes. During inference, it gradually adjusts ligand positions and internal angles over 20 iterations and at the same time adapts protein conformations, particularly side chain angles, improving the structures predicted by AlphaFold. Unlike traditional models, it employs a morph-like transformation instead of Gaussian noise perturbations, which improves the model's ability to capture biologically relevant conformational changes. DynamicBind excels at predicting ligand pose, reducing clashes and revealing cryptic zones, as demonstrated in several benchmarks and case studies, demonstrating its potential for drug discovery applications.

In conclusion, DynamicBind integrates protein conformation generation and ligand pose prediction into a single end-to-end deep learning framework, significantly faster than traditional MD simulations. Unlike conventional docking methods that require predefined binding pockets, DynamicBind performs global docking, which is ideal for identifying cryptic pockets. This reduces potential side effects by targeting specific proteins and aids drug discovery by predicting unwanted protein targets or identifying targets in phenotype screening. Although it shows excellent performance, improvements are needed for better generalization to proteins with low sequence homology. Advances in Cryo-EM and the incorporation of binding affinity data may enhance DynamicBind's capabilities.

Review the Paper and GitHub. All credit for this research goes to the researchers of this project. Also, don't forget to follow us on twitter.com/Marktechpost”>twitter. Join our Telegram channel, Discord channeland LinkedIn Grabove.

If you like our work, you will love our Newsletter..

Don't forget to join our 42k+ ML SubReddit