Binding Site Prediction

• AF2BIND: AF2BIND utilizes AlphaFold2 for predicting protein-ligand binding sites.
  
• BindWeb: a web server for ligand binding residue and pocket prediction from protein structures (online)
  
• BiteNet: large-scale detection of protein binding sites (online)
  
• BiteNet: Spatiotemporal identification of druggable binding sites using deep learning (standalone)
  
• Caviar: Protein cavity identification and automatic subpocket decomposition (standalone)
  
• CavityPlus: A web server for protein cavity detection with pharmacophore modeling, allosteric site identification, and covalent ligand binding ability prediction.
  
• COACH-D: COACH-D: improved protein–ligand binding sites prediction with refined ligand-binding poses through molecular docking
  
• ConCavity: prediction from protein sequence and structure (geometry-based) (online)
  
• CRAFT: cavity prediction tool based on flow transfer algorithm (standalone), 2024.
  
• CSM-Potential: mapping protein interactions and biological ligands in 3D space using geometric deep learning (PPI and regular binding sites) (online)
  
• DeepPocket: Ligand Binding Site Detection and Segmentation using 3D Convolutional Neural Networks (standalone)
  
• DeepSite:
  
• DeepSurf: a surface-based deep learning approach for the prediction of ligand binding sites on proteins (standalone)
  
• DogSiteScorer: This proteins.plus platform offers DogSiteScorer for predicting and scoring protein-ligand binding sites.
  
• EquiPocket: E(3)-Equivariant Geometric Graph Neural Network for Ligand Binding Site Prediction
  
• ESSA: Essential site scanning analysis: A new approach for detecting sites that modulate the dispersion of protein global motions (Essential sites are defined as residues that would significantly alter the protein’s global dynamics if bound to a ligand) (standalone)
  
• fpocket: mainly geometry-based (standalone)
  
• FPocketWeb: FpocketWeb is a browser app for identifying pockets on protein surfaces where small-molecule ligands might bind, running calculations locally on the user's computer.
  
• FrustraPocket: A protein–ligand binding site predictor using energetic local frustration (standalone)
  
• Get_Cleft:
  
• GetCleft: The detection of cavities, both internal or surface-exposed (standalone)
  
• Graphsite-classifier: deep graph neural network to classify ligand-binding sites on proteins (standalone)
  
• GrASP: GrASP (Graph Attention Site Prediction) identifies druggable binding sites using graph neural networks with attention.
  
• GRaSP-py: a graph-based residue neighborhood strategy to predict binding sites (standalone)
  
• IF-SitePred: IF-SitePred is a method for predicting ligand-binding sites on protein structures. It first generates an embedding for each residue of the protein using the ESM-IF1 model, then performs point cloud clustering to identify binding site centres.
  
• Kalasanty: Improving detection of protein-ligand binding sites with 3D segmentation (standalone)
  
• LVPocket: 3D global-local information to protein binding pockets prediction
  
• P2rank: P2Rank is a machine learning-based tool for predicting ligand-binding sites from protein structures, capable of handling various structure formats including AlphaFold models.
  
• POCASA: geometry-based, pockets and cavities, volume... (online)
  
• PocketAnalyzerPCA: Identify cavities and crevices in proteins (standalone)
  
• PocketDruggability: A model that predicts the “attainable binding affinity” for a given binding pocket on a protein; this model relies on 13 physiochemical and structural features calculated using the protein structure (standalone)
  
• PocketQuery: energy-based (online)
  
• PocketVec: Comprehensive detection and characterization of human druggable pockets through binding site descriptors
  
• PointSite: a point cloud segmentation tool for identification of protein ligand binding atoms (standalone)
  
• PrankWeb 3: accelerated ligand-binding site predictions for experimental and modelled protein structures (online)
  
• ProteinPocketDetection: Protein pocket detection via convex hull surface evolution and associated Reeb graph (standalone)
  
• PUResNet: Predicting protein-ligand binding sites using deep convolutional neural network (binding pocket) (standalone)
  
• PUResNetV2.0: state-of-the-art deep learning model designed to predict ligand binding sites in protein structures.
  
• pyKVFinder: Python package for biomolecular cavity detection and characterization in data science (standalone)
  
• PyVOL: python library packaged into a PyMOL GUI for identifying protein binding pockets, find volume, also via python command line (standalone)
  
• SGPocket: A Graph Convolutional Neural Network to predict protein binding site.
  
• SiteFerret: beyond simple pocket identification in proteins (standalone)
  
• SiteHound-web: energy-based (send chemical probes) (online)
  
• SplitPocket: geometry-based (online)
  
• GENEOnet: XAI approach to protein pocket detection
  

Allosteric Site Prediction

• Allo: a tool for dicriminating and prioritizing allosteric pockets (standalone)
  
• AlloReverse: AlloReverse predicts multi-scale allosteric regulation information based on reversed allosteric communication theory, aiding in drug design and biological mechanism understanding.
  
• AlloSite: allosteric site (online)
  
• APOP: Predicting Allosteric Pockets in Protein Biological Assemblages (standalone)
  
• DeepAllo: DeepAllo: Allosteric Site Prediction using Protein Language Model (pLM) with Multitask Learning
  
• MEF-AlloSite: MEF-AlloSite: an accurate and robust Multimodel Ensemble Feature selection for the Allosteric Site identification model
  
• PASSer: Designed for accurate allosteric site prediction, PASSer offers three machine learning models for quick and extensive allosteric analysis.
  
• PASSerRank: Prediction of allosteric sites with learning to rank (standalone)
  

Cryptic Pockets

• PocketMiner: predicting the locations of cryptic pockets from single protein structures (standalone)
  

Benchmark

• CryptoBnech: Cryptic protein-ligand binding sites dataset and benchmark
  

Fragment Site Prediction

• FTMap: FTMap maps unbound protein surfaces to identify druggable hot spots where small molecules may bind.
  

From Molecular Dynamics simulations

• CHARMM-GUI LBS Finder & Refiner: Ligand Binding Site Prediction and Refinement (standalone)
  
• ColaBind: Cloud-Based Approach for Prediction of Binding Sites Using Coarse-Grained Simulations with Molecular Probes
  ;)
• POVME2: POVME2 identifies druggable protein pockets and their unique conformations within molecular dynamics simulations, facilitating the discovery of novel pharmacologically active molecules.
  
• TACTICS: machine-learning algorithm (trajectory-based analysis of conformations to identify cryptic sites), which uses an ensemble of molecular structures (such as molecular dynamics simulation data) as input (standalone)
  

Growth Site Prediction

• SINCHO: SINCHO protocol is a method for the prediction/suggestion of the desirable anchor atom and growth site pair for the modification of the hit compound in hit-to-lead process.
  

Metal-binding site

• PinMyMetal: Specific details about PinMyMetal, presumably a tool for predicting metal-binding sites in proteins, were not provided in the data fetched.
  

Reviews

• Barton et al.: Comparative evaluation of methods for the prediction of protein-ligand binding sites